Flexible flaps for separation control on a wing with low aspect ratio Dipl. Ing. G. Patone Dr. W. Müller Dr. R. Bannasch Prof. Dr. Ing. I. Rechenberg

bionics pilot project: In this project, we were trying to prove that flaps that are built according to covert feathers of birds, can indeed function as eddy breaks and therefore prevent the sudden drop in lift generation during stall. Aeroflexible surface flaps as "eddy-breaks" built after the covert feathers of birds Project partners: DLR Abt. Turbulenzforschung Firma Stemme GmbH Fachgebiet Bionik und Evolutionstechnik an der TU-Berlin

main objective: As all of you know, during stall an eddy starts to propagate from the trailing edge towards the leading edge of the wing. Thereby the flow on the upper surface is separated with the result of a sharp drop in lift generation. This is especially dangerous because it happens mainly in situations when high lift generation is needed (like during landing or narrow curves). eddy

Brown Skua with “eddy-flaps” How do birds deal with this problem? When looking at this photograph of a Brown Skua you will notice that on the left wing the covert feathers are lifted off the surface. Our working hypothesis was that these coverts were actually able to prevent the spreading of the eddy, thereby working as eddy-flaps. eddy lifted covert feathers eddy- flaps

aerofoil with eddy- flaps Here you can easily see how we think these flaps work. They are lifted off the surface by the eddy, thereby preventing it from spreading over the entire upper surface. eddy- flaps

cL cL cD cL cL cD goal: not like this but like this a a Our goal can be shown quite clearly with these polar diagrams: We want to transform the polar curve of an aerofoil from right shape by attaching flaps to left shape. not like this but like this cL cL cD a cL cL cD a

improving the stall behavior of an aerofoil research goal main objective is to improve the stall behaviour of a certain aerofoil by attaching flaps. Of course, these flaps should not have any negative effects on the quality of the aerofoil while they are not in use, meaning during regular flight when they are not lifted off the surface. And third, according to natures role model, these flaps should work without external controls, meaning they should be smart, acting as sensors and effector as well. improving the stall behavior of an aerofoil flaps should not have any negative effects while not active flaps should work without external controls

experimental setup at the wind tunnel The experimental setup at our wind tunnel: We conducted our experiments in an open circuit wind tunnel with a nose opening of 1.2 meters diameter. The aerofoils used are 70 cm long and 20 cm deep, (this will make for an aspect ratio of 3.5). The experiments I will tell you about today were those run with a NACA 2412 Reynolds Number of 130’000. The aerofoil was hung on three thin wires that connected to an electronic balance for measuring lift. A second electronic balance for measuring drag was connected via rods and wires to the trailing edge of the aerofoil. For measuring the pressure distribution a scanivalve and a baratron pressure sensor was used. lift-balance wind tunnel 10m/s NACA 2412 aerofoil with an aspect ratio of 3.5 Re= 130.000 drag-balance pressure sensor and scanivalve

porosity schematic pressure distribution px > py Working as a bionics project it is only natural to look at natures role model first. So we tested the porosities of the bird feathers and found that the covert feathers are slightly porous. This makes sense when we look at a schematic pressure distribution on an aerofoil. When using impervious materials like plastic sheets, the flaps were lifted off the surface before the critical angle of attack is reached. As a result, lift and drag are influenced negatively schematic pressure distribution px > py impervious materials are lifted off by the pressure differences porous materials remain on the surface

aerofoil with silk flaps This slide shows you diagrammatically the dimensions of the best flaps used. We tested about 300 configurations. The flaps that worked best were made out of silk, that was supported by thin steel wires. The size was 100 by 600mm and they were attached at 35% chord length. A representative polar diagram of these flaps is shown here silk steel wire

silk flaps The effect of these silk flaps is quite impressive. The diagram depicts the lift coefficient over the angle of attack. The red curve is the one for the aerofoil without flaps and you notice the characteristic drop in lift generation beyond the critical angle. The blue curve shows several things: First, while the flaps are not lifted off, they do not have a negative effect on the lift generation (about their effect on the drag I will tell you in a second). Then, when increasing the angle of attack beyond its critical value, the lift generation does NOT drop but even increases further a little bit.

polar diagram for ‘perforated plastic sheet’ flaps We can clearly see that again, while the flaps are not lifted off the surface, they do not have any negative effect, neither on the lift nor on the drag coefficient. But more importantly, beyond the critical angle of attack, the lift generation could be kept more or less constant for another 10 degrees increase in the angle of attack. After that the lift dropped, which was due to the flaps not being lifted off the surface anymore. Now, of course we were interested in the mechanism of these 'eddy-flaps' as we call them. In order to find out about it, we measured the pressure distribution on the aerofoil without and with flaps attached.

aerofoil for pressure distribution This sketch shows the aerofoil (again a NACA 2412 was used) with the arrangement of the holes. We drilled seven rows spread over one half of the aerofoil, each row consisting of 40 holes, spread over the upper and lower surface In the next three slides you can see the pressure distribution on the entire upper surface, of course, the right half of the aerofoil will be a mirror image of the left half (that we only measured).

pressure distribution at 16 degrees The pressure distribution on the upper surface at an angle of attack of 16 degrees. (that is shortly before flow separation). The leading edge, therefore the incoming flow, is on the bottom, and the sequence of the shading goes from black, for lowest pressure (meaning high lift generation) towards white, for areas of little lift generation. In the upper diagram the aerofoil without flaps is shown, in the lower diagram that with flaps attached. You can see in the upper diagram as well known, that close to the leading edge the area with the lowest pressure, is situated. Along the borders of the aerofoil, the effect of the 'trailing vortices' can be observed. This is basically the same for the aerofoil with flaps attached, only that there are some minor disturbances due to the flaps. without flaps with flaps incident airflow

pressure distribution at 19 degrees These two diagrams show the situation at an angle of attack of 19 degrees, which is right after stall. In the aerofoil without flaps you can see that flow separation has commenced almost to the leading edge. On both sides, however, an area of relatively high lift generation prevails due to the effect of the trailing vortices. If you look at the lower diagram with the lift distribution for the aerofoil with the flaps attached, it is obvious that the eddy, and therefore flow separation, is being prevented from spreading further towards the leading edge by the flaps. The eddy is restricted to the hind part of the aerofoil, where not to much lift is generated anyway. As a result, in front of the flaps the low pressure and therefore lift generation can remain at a much higher level than it were without flaps. without flaps with flaps incident airflow

flow visualisation trapped vortex trapped eddy When seeing this, of course we became interested in the actual flow pattern, so we used smoke for flow visualization and got this pattern. The flaps restrict the eddy to the hind part of the aerofoil. In front of the flaps, however, we found a trapped vortex, that fills in the angle between the aerofoil and the flaps. trapped vortex trapped eddy

pressure distribution at 36 degrees In the third slide of this series, the pressure distribution at an angle of attack of 36 degrees is shown. In the upper diagram for the aerofoil without flaps the flow separation has spread even further and the effect of the trailing vortices (which is basically a virtual decrease in the angle of attcack) cannot prevent this even in these areas. In the lower diagram, the further decrease of the high lift generation is visible as well, yet, in front of the flaps there is still the area with a fairly large lift generation. without flaps with flaps incident airflow

polar diagram (calculated from all measuring rows) In this picture the calculated lift coeffient is depicted over the angle of attack. The lift coefficient is calculated by integrating the pressure measurement of all measuring rows of upper and lower surface. The curve resembles that which I showed earlier which was obtained by measuring the lift and drag directly with a balance. Again, the curve for the aerofoil without flaps shows the sudden drop in lift generation, whereas the one with flaps shows nearly the desired behaviour. a

lift distribution without flaps at different angles of attack The diagram shows the lift distribution over the length of the aerofoil. The different curves represent different angles of attack. Let’s look at the curve for 6 degrees angle of attack. The points represents the lift coeffienct calculated for the seven measuring rows (please remember that we measured only one side and the other side is the mirror image). The lift distribution is what you would expect, namely that in the middle of the aerofoil most of the lift is generated. Let us take a look at the curve for 18 degrees, which is just below the critical angle of attack. The total lift generated has increased to a maximum, of course. A small increase of the angle of attack lead to the dramatical drop in lift generation. It is obvious, that in the middle of the aerofoil, the flow separation reduces the generated lift more drastically than in the lateral areas of the aerofoil. There, lift generation remains fairly high because of the effect of the trailing vortices.

lift distribution with flaps at different angles of attack the same diagram but with flaps attached. Again, for 6 degrees, we have the expected lift distribution like without flaps. Also for 18 degrees, the curve looks very similar to the one without flaps. Now, the first curve with separated flow (at 19 degrees) shows two things: First, again a drop of generated lift in the middle of the aerofoil is seen, but this drop is not as dramatical as without flaps. Second, in the lateral areas of the aerofoil, the lift generated dosent change.

eddy flaps prevent sudden drop in lift generation during stall summary: With flaps that were built according to the biomechanical properties of covert feathers of birds we were able to prevent the sudden drop in lift generation that usually accompanies flow separation. The pressure distribution on the aerofoil indicates that the ‘eddy-flaps’ are able to contain the eddy in the hind part of the aerofoil, so that it cannot spread up to the leading edge of the aerofoil, where it would cause flow separation and therefore a decrease in lift generation. eddy flaps prevent sudden drop in lift generation during stall pressure distribution indicates: eddy-flaps restrict eddy to hind part of aerofoil outlook: automatic contour adapting flaps dynamic stall behaviour