24TH INTERNATIONAL CONGRESS OF THE AERONAUTICAL SCIENCES

DELTA WITH LEADING EDGE EXTENSION AND PROPELLER PROPULSION FOR FIXED WING MAV

Cezary Galiński*, Nicholas J. Lawson**, Rafał Żbikowski** *Politechnika Warszawska, **Cranfield University

Keywords: MAV, , aerodynamic testing

Abstract required. On the other hand, the ability to hover or at least slow flight is desirable. Air turbulence is perceived as a major problem Air turbulence is perceived as a major for MAV outdoor applications. It seems problem in the case of such applications. reasonable to use MAVs in these situations According to recent work [1], short duration providing the flow is attached to the control vertical gusts may have velocity comparable to surfaces even at high angles of attack. This will MAV airspeed, so brief periods of flight at very ensure effective control during turbulence. large angles of attack have to be considered. In Delta are known to have excellent these circumstances it seems reasonable to large angle of attack qualities. However, apply a MAV design with as high angle of quantitative data about leading edge vortex attack as possible. In particular, the flow has to effectiveness at low Reynolds numbers were not be attached to control surfaces to perform available, so an experiment was undertaken to effective control during turbulent flight measure them. Generally, the effect appeared to conditions. be similar to that obtained for large aeroplanes Delta wings are known to have excellent including advantages of leading edge extensions large angle of attack qualities. Generation of a (LEX). There was no observed negative effect of leading edge vortex allows to the flow to LEX/propeller interference. In fact an increase reattach and improve stall qualities. Therefore a of stall angle was observed as a result of the delta wing was considered as a candidate for propulsion operation. MAV design. However, quantitative data about The following presents force and leading edge vortex effectiveness at low visualisation results from a series of MAV Reynolds numbers were not available. configurations. Therefore an experiment was undertaken in order measure them [2]. Generally, the effect 1 Introduction appeared to be similar to that obtained for large, manned aeroplanes including advantages of The Micro Aerial Vehicle is defined here as a additional application of Leading Edge small (hand launched, storable in portable Extensions (see Figure 1). container), light (weight 150-200g), simple and Design of the delta wing MAV’s with inexpensive unmanned flying vehicle for direct, LEX appears to be non-trivial because of the over the hill reconnaissance. The focus is on LEX propeller interference problem. Details of fixed wing, forward thrust since the the design, wind tunnel tests results, including ability to negotiate strong opposing winds is flow visualization of such configuration, are now presented in this paper.

1 C.GALIŃSKI, N.LAWSON, R.ŻBIKOWSKI

1.2 1.2 1.0 1.0

0.8 0.8

0.6 0.6 Cl Cl 0.4 configuration 0.4

0.2 clean 0.2 with LEX

0.0 0.0 -5 0 5 10 15 20 25 0.0 0.1 0.2 0.3 0.4 0.5 α [ o ] -0.2 -0.2 Cd

Fig.1 Leading Edge extension effect on the delta MAV wing performance

NACA 0004 were applied from the wing root to the wing tip. 2 MAV design The model was tested in clean Propeller propulsion seems the most suitable for configuration and with LEX of four different a fixed wing MAV. The propeller should be shapes: triangular, trapezoidal, circular and located in front of the CG if stable hovering is polynomial. The LEX area was the same in each desirable. Unfortunately, the propeller at the case. The span of LEX was always slightly front of the vehicle would strongly interfere longer than the slot length. with the leading edge vortex. Therefore an aircraft configuration was developed with propeller located in the slot inside the wing contour.

Fig.2 MAV general design Fig.3 Tested configurations The model built for wind tunnel tests had a wing area of 0.1m², a wing span of .45m, an There was a doubt about propulsive aspect ration of 2 and a leading edge sweep efficiency of such a configuration. So it was angle of 39º. Aerofoils NACA 23003 and decided to test whether ducted propeller concept could help to improve it. The ring would have 2 DELTA WING WITH LEADING EDGE EXTENSION AND PROPELLER PROPULSION FOR FIXED WING MAV

also the additional advantage of protecting the 3 Propulsion efficiency tests person hand-launching the vehicle as well as There were two experiments exploring isolating the propeller stream from the leading propulsion efficiency. First of them allowed to edge vortex. However, since this feature may measure the static thrust and second provided cause problems in storage, the ring design information about power needed to balance the should be foldable and deployed at launch. In drag for cruise with increasing wind tunnel such case, the ring may be divided into two airspeed. Results are presented in Figures 5 and parts and have the diameter slightly greater than 6. propeller to avoid collision with the propeller during the deployment. That is why in the case 0.8 of the test model the distance between propeller tip and the ring was as large as 1.5mm. It was also decided not to test the stream/vortex 0.6 isolation potential, since the test should verify

the concept in worst case scenario. Therefore a ]

simple, narrow ring was applied. [N 0.4

thrust without ring with ring 0.2

0.0 0 10203040 power [W]

Fig.5 Static thrust

80

Fig.4 MAV prepared for storage without ring 60 with ring The model was equipped with standard electric motor Speed 300 with standard propeller 6x3. 40 According to [3] wings equipped with power [W] membrane type covering provide more stable 3 C 2 20 coefficient and l ratio in an oscillating C d free stream. This feature was perceived as advantageous since turbulence resistance was 0 the goal of the experiment. Therefore a structure 4 6 8 10 12 14 with a carbon/epoxy torsion box near the airspeed [m/s] leading edge and ribs covered by membrane film at the rest of the wing was selected. Fig.6 Cruise power characteristics

Power required to provide static thrust appeared to be the same for both configurations

3 C.GALIŃSKI, N.LAWSON, R.ŻBIKOWSKI

with and without ring. Conversely, the cruise with airspeeds similar to those experienced in power was greater for the configuration flight, so a true Reynolds number was obtained. equipped with the ring. This result may in part Figure 7 shows the range of Reynolds numbers be due to the ring drag being larger than the applied during the measurements. increase in propulsive efficiency. Therefore both experiments did not reveal any advantages of 8E+5 the ducted configuration. Another reason for this discrepancy may be the large gap between the propeller and the ring which is not an 6E+5 optimized design. Further study should be undertaken to provide an answer about potential Reynolds number efficiency improvement due to the presence of in the free flight 4E+5 Re the ring. Therefore the ducted configuration was in the tunnel temporarily abandoned, since it was not critical for LEX verification. 2E+5

4 LEX/propeller cooperation

0E+0 0102030 4.1 Test procedure α [ o ]

The experiment was conducted in two parts. Fig.7 Wind tunnel Reynolds number applied to Firstly, steady flight lift/drag polars were measure the steady flight polar. For the measured for each LEX shape. Secondly, tests comparison Reynolds number experienced with running propeller were completed for in flight by a 170g MAV. selected steady flight conditions. The test sequence was as follows: Maximum wind tunnel airspeed was constrained because of the uncertainty • For certain deflections, angles concerning the shape of lift coefficient versus of attack and wind tunnel airspeeds, the angle of attack characteristic (Cl(α)). The motor was set to an rpm which provided maximum airspeed predicted for free flying a drag reading equal to zero. could damage it in the wind tunnel if • The angle of attack was gradually used with an incorrect angle of attack. The increased with all other parameters minimum wind tunnel airspeed was also constant. constrained. Thus, it was anticipated that flow instability at large angles of attack would be This sequence allowed the simulation of magnified if the wind tunnel airspeed were not entrance into a strong vertical gust. It was not an stabilised. Hence the airspeed was kept constant ideal simulation since measurement was static. after Cl=0.3 was achieved. Airspeeds that could Therefore dynamic effects were ignored. But the be achieved by 170 g airplane in the free flight measurements provided an estimate of LEX are presented for comparison. They were effect and LEX/propeller interaction. calculated given the lift coefficients measured in the experiment. 4.2 Test conditions Tests were conducted in the closed jet tunnel of 4.3 Measurement results the Cranfield University at RMCS Shrivenham. Figures 8–10 show the main result. In this case The facility allowed for tests of real size vehicle the elevator was set to the loitering position. It 4 DELTA WING WITH LEADING EDGE EXTENSION AND PROPELLER PROPULSION FOR FIXED WING MAV

is clear that all the LEX configurations provide could be increased by rotation of thrust vector increased maximum lift coefficient and stall only. Figures 11-15 show the results of this angle in the motor off mode as expected. operation. Motor off lift curves measured during Unexpectedly, both maximum lift coefficient first part of experiment are shown for and stall angle are even greater in the motor on comparison. Unfortunately, the Re number for mode. Thus, propeller operation appears not to the motor-off curves are slightly smaller than be problematic for the leading edge vortex in the steady flight Re number applied during the this configuration. The LEX effect is increased second phase of experiment. However, the rather than reduced. thrust vector rotation effect seems to be too The vertical thrust component was subtracted small to explain the total lift increase. from the measured lift curves to verify if lift

1.4 1.4 Re~138 000 Re~138 000 β=0o β=0 o 1.2 1.2 motor off motor off

1.0 1.0

0.8 0.8 Cl LEX Cl LEX 0.6 0 0.6 0 A A B 0.4 B C 0.4 C D D 0.2 0.2

0.0 0.0 0 5 10 15 20 25 30 0.00.10.20.30.40.50.60.7 Cd Fig.8 Characteristics measured with neutral and motor switched off

1.4 1.4 Re~138 000 Re~138 000 o o β=6 β=6 1.2 1.2 motor off motor off

1.0 1.0

0.8 0.8 Cl Cl

0.6 0.6 LEX LEX

0 0 0.4 A 0.4 A B B C C 0.2 D 0.2 D

0.0 0.0 0 5 10 15 20 25 30 0.00.10.20.30.40.50.60.7 Cd Fig.9 Characteristics measured with elevon angle of 6º and motor switched off

5 C.GALIŃSKI, N.LAWSON, R.ŻBIKOWSKI

1.4 1.4 Re~155 000 Re~155 000 o β=6 o β=6 1.2 motor on 1.2 motor on

1.0 1.0

0.8 0.8 Cl Cl LEX 0.6 LEX 0.6 0 0 A A B 0.4 B 0.4 C C D D 0.2 0.2

0.0 0.0

0 5 10 15 20 25 30 0.00.10.20.30.40.50.60.7 Cd

Fig.10 Characteristics measured with elevon angle of 6º, motor switched on and Re~60 000

Figures 9 and 10 allow another comparison of the LEX from vortex generation. In such cases where it can be seen that the difference between only outboard LEX segments would be LEX performance seems to be smaller if the responsible for lift increases. The difference motor is operating. This observation may suggest between LEX geometries was much smaller in that the angle of attack was effectively decreased outboard segments. in front of the propeller thus excluding this part

1.4 1.4

β=6 o β=6 o 1.2 1.2 LEX 0 LEX 0

1.0 1.0

0.8 0.8 Cl Cl 0.6 0.6

motor off, Re~138 000 0.4 motor on, Re~155 000 motor off, Re~138 000 0.4 motor on, Re~155 000 motor on, Re~155 000, thrust subtracted 0.2 0.2

0.0 0.0 0 5 10 15 20 25 30 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 Cd

Fig.11 Clean configuration characteristics measured with elevon angle of 6º

6 DELTA WING WITH LEADING EDGE EXTENSION AND PROPELLER PROPULSION FOR FIXED WING MAV

1.4 1.4

o β=6 β=6 o 1.2 LEX A 1.2 LEX A

1.0 1.0

0.8 0.8 Cl Cl

0.6 0.6

motof off, Re~138 000 motor on, Re~155 000 0.4 0.4 motor off, Re~ 138 000 motor on, Re~155 000 0.2 motor on, Re~155 000, thrust subtracted 0.2

0.0 0.0 0 5 10 15 20 25 30 0.00.10.20.30.40.50.60.7 Cd

Fig.12 Characteristics of the configuration with LEX A measured with elevon angle of 6º

1.4 1.4

o β=6 β=6 o 1.2 1.2 LEX B LEX B

1.0 1.0

0.8 0.8 Cl Cl 0.6 0.6

motor off, Re~138 000 0.4 motor on, Re~155 000 0.4 motor off, Re~ 138 000 motor on, Re~155 000 0.2 motor on, Re~155 000, thrust subtracted 0.2

0.0 0.0 0 5 10 15 20 25 30 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 Cd

Fig.13 Characteristics of the configuration with LEX B measured with elevon angle of 6º

7 C.GALIŃSKI, N.LAWSON, R.ŻBIKOWSKI

1.4 1.4

β=6 o β=6 o 1.2 1.2 LEX C LEX C

1.0 1.0

0.8 0.8 Cl Cl

0.6 0.6

motor off, Re~138 000 0.4 motor on, Re~155 000 0.4 motor off, Re~138 000 motor on, Re~155 000 0.2 motor on, Re~155 000, thrust subtracted 0.2

0.0 0.0 0 5 10 15 20 25 30 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 Cd

Fig.14 Characteristics of the configuration with LEX C measured with elevon angle of 6º

1.4 1.4

o β=6 o β=6 1.2 LEX D 1.2 LEX D

1.0 1.0

0.8 0.8 Cl Cl 0.6 0.6

motor off, Re~138 000 0.4 motor on, Re~155 000 0.4 motor off, Re~138 000 motor on, Re~155 000 0.2 motor on, Re~155 000, thrust subtracted 0.2

0.0 0.0 0 5 10 15 20 25 30 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 Cd

Fig.15 Characteristics of the configuration with LEX D measured with elevon angle of 6º

A flow visualisation experiment was also leading edge provide the clearest evidence. This undertaken to further explain the lift increases. shows that the lift increase is not caused by a vertical thrust component only, but also by more 4.4 Flow visualisation general flow improvement over the whole wing. The phenomenon observed here seems to be Figures 16-18 show the separated flow close to similar to sonic flow excitation effects described the wing tip in motor off mode and attached in in [4-7]. The major difference is the method of motor on mode. Three outboard tufts close the excitation. At this time propeller passes through 8 DELTA WING WITH LEADING EDGE EXTENSION AND PROPELLER PROPULSION FOR FIXED WING MAV

the slot with a frequency of about 250Hz thus the flow less likely to separate hence increasing generating pressure waves. This probably makes both the lift coefficient and stall angle.

Fig.16 Propeller effect for angle of attack of 20º.

Fig.17 Propeller effect for angle of attack of 25º.

9 C.GALIŃSKI, N.LAWSON, R.ŻBIKOWSKI

Fig.18 Propeller effect for angle of attack of 30º.

Number Airfoils”, Journal of Aircraft, Vol. 36, No.3, pp. 523-529, May-June 1999 5 Conclusions [4] K.K.Ahuja, R.H.Burin “Control of flow separation by sound”, AIAA Paper 84-2298, Oct.1984 • Leading edge extensions (LEX) can be [5] J.Wojciechowski, S.Skrzyński, M.Litwińczyk successfully integrated with the „Investigation of the Acoustic Wave Influence on the propeller propulsion, generating a Laminar Boundary Laser Separation Process on the leading edge vortex in the neighborhood Airfoil”, 332 EUROMECH Colloquium on Drag of the propeller stream. Reduction, Ravello, Italy, April 1995. [6] J.Wojciechowski “Active Control of the Boundary • The effect of the propeller rotating in the Layer on the Laminar Airfoil”, Proceedings of the slot seams to be similar to the effect of a Národni konference s mezinárodni účasti Inženýrská vibrating membrane. Mechanika’99, 17-20 květen, Svratka, Tch. • Flow asymmetry effects due to [7] U. Rist, K. Augustin “Control of Laminar separation LEX/propeller interference should be Bubbles”, Lectures on on Aircraft explored before described configuration Including Applications in Emerging UAV Technology, 24-28 November 2003, Von Karman is applied in a flying prototype. Institute for Fluid Dynamics.

6 References

[1] S. Watkins, W. Melbourne “Atmospheric Winds: Implications for MAVs”, proceedings of the XVIII International UAV Conference, 31 March – 2 April 2003, Bristol, UK [2] C.Galiński, M.Eyles, R. Żbikowski “Experimental Aerodynamics of Delta Wing MAVs and their Scaling”, Proceedings of the 18th International UAV Conference, 31 March – 2 April 2003, Bristol, UK [3] W. Shyy, F. Klevenbring, M. Nilsson, J. Sloan, B. Carrol, C. Fuentes “Rigid and flexible low Reynolds 10