Sailplane Aerodynamics at Delft University of Technology
Total Page:16
File Type:pdf, Size:1020Kb
NVvL 2006 1 Research on sailplane aerodynamics at Delft University of Technology. Recent and present developments. L.M.M. Boermans TU Delft Presented to the Netherlands Association of Aeronautical Engineers (NVvL) on 1 June 2006 Summary The paper consists of two parts. Part one describes the findings of recent aerodynamic developments at Delft University of Technology, Faculty of Aerospace Engineering, and applied in high-performance sailplanes as illustrated by the Advantage, Antares, Stemme S2/S6/S8/S9 family, Mü-31 and Concordia (Chapter 1). The same technology has been applied in the aerodynamic design of the Nuna solar cars as illustrated by the Nuna-3 (Chapter 2). Part two describes results of ongoing research on boundary layer control by suction, a beckoning perspective for significant drag reduction (Chapter 3). 1. Recent research on sailplane aerodynamics 1.1. Introduction The speed polar of a high-performance sailplane, subdivided in contributions to the sink rate of the wing, fuselage and tailplanes, Figure 1, illustrates that the largest contribution is due to the wing; at low speed due to induced drag and at high speed due to profile drag 1 . Figure 1 Contributions to the speed polar of the ASW-27 Hence, improvements have been focused on these contributions, followed by the fuselage and tailplane drag, in particular at high speed. These improvements will be illustrated by examples of various new sailplane designs. It should be pointed out, however, that each new sailplane design incorporates improvements similar to those presented here for the individual sailplanes. NVvL 2006 2 1.2. Induced drag Induced drag has been minimized by optimizing the wing planform with integrated winglets, as illustrated by the wing of the Advantage shown in Figure 2a. The Advantage is a new Standard Class high-performance sailplane with shoulder wing, Figure 3, being built by Sailplanes Inc. in New Zealand. Figure 2a Planform and lift distribution of the Advantage wing in comparison with Munk’s optimum lift distribution for minimum induced drag Winglets can be optimized at one lift coefficient only, as indicated in Figure 2a where the winglet has been bent down to show the actual lift distributions and the optimal lift distribution according to Munk’s third theorem 2 . But Figure 2b shows that the absolute minimum induced drag can be realized within 1% at all lift coefficients. Besides, the winglet airfoils have been designed for low profile drag in their operational region of lift coefficients, obtained by 100% laminar flow on the lower surface and 50% on the upper surface. In addition, ample reserve to separation in case of yaw has been applied, in particular at low speed when circling in thermals. NVvL 2006 3 1.04 2 k k C L C D === i πππ A 1.02 CL 1.00 0.2 0.4 0.6 0.8 1.0 1.2 1.4 0.98 0.96 0.94 Advantage 0.92 Munk 0.90 Figure 2b Induced drag factor of the Advantage wing in comparison with Munk’s minimum induced drag factor Figure 3 The Standard Class high-performance sailplane Advantage of Sailplane Inc., New Zealand 1.3. Profile drag Profile drag depends on the extent of the laminar boundary layer on the upper and lower surfaces and on airfoil thickness. As an example, the 12.7% thick airfoil with camber changing flap of 14% chord, applied as main airfoil in the Antares wing, is presented. The Antares is a new 20 m-span high-performance sailplane with retractable electric engine, Figure 4. The very efficient 42.5 kW DC/DC brushless outside rotating engine and the large slowly rotating propeller enable starting and climbing silently with 4.4 m/s to a maximum altitude of 3,000 m. Usually, at an altitude of about 500 m the engine is retracted and ample power is available to extend the engine and fly home when thermals are weakening. Lithium-ion batteries are placed in the inner wing and water ballast tanks – to increase the wing loading - in the outer wing NVvL 2006 4 Figure 4 The Antares 20 m span high-performance sailplane with retractable electric engine. The wing airfoil has laminar flow up to 95% of the chord on the lower surface at the high-speed zero-degree flap deflection, and up to 75% of the chord on the upper surface at the low-speed 20 degrees flap deflection, as indicated by the pressure distributions in Figure 5. The upper and lower surface flap gaps have been sealed by flexible mylar strips. This enables the boundary layer on the lower surface to remain laminar beyond the flap hinge position at the zero-degree flap deflection up to 95% chord where transition is artificially forced by zigzag tape. On the upper surface the sealing prevents low-pressure peaks and subsequent steep pressure gradients on the flap at 20 degrees deflection, thus postponing separation. Consequently, the profile drag is very low over a large range of lift coefficients, as shown by the measured characteristics in Figure 6. Full laminar flow on the lower surface cannot be realized at low lift coefficients and high Reynolds numbers due to the adverse pressure gradient needed to reach the overpressure at the trailing edge. And an increase of the laminar flow extent on the upper surface leads to a steeper pressure gradient in the rear part, which in turn enhances flow separation and a loss of lift beyond the upper boundary of the low drag bucket. This causes unfavorable flying qualities in thermals. This example shows that further drag reduction of airfoils by conventional means is hampered, this being the reason for the present research on boundary layer suction as described in the second part of the paper. NVvL 2006 5 Figure 5 Pressure distributions of airfoil DU97-127/15M at 0, 10 and 20 degrees flap deflection Figure 6 Measured aerodynamic characteristics of airfoil DU97-127/15M at various flap deflections and Reynolds number 5.1 *10 6 NVvL 2006 6 1.4. Fuselage drag Fuselage drag depends mainly on fuselage thickness, contraction behind the cockpit, and streamline shaping. Fuselage frontal area should be minimal. Contraction behind the cockpit – and the corresponding pressure gradient - is limited because of flow separation when the boundary layer on the fuselage is completely turbulent, for instance when flying in rain. Streamline shaping by fitting the fuselage shape to the streamlines produced by the wing minimizes cross-flow effects. For an undisturbed boundary layer development, continuity of curvature is required in flow direction. This is guaranteed by deriving the top, bottom, and line of largest width from airfoil shapes, and using Hügelschäffer curves (deformed ellipses) for the fuselage cross-sections. A remarkable finding 3 is that the cockpit length can be increased by 0.3 m without any drag increase, as illustrated by the accumulative development of the drag coefficient on a rotationally symmetrical body in the Figures 7a and 7b. Transition occurs at 33% of the original fuselage length, and the total drag is found at the tail. This result offers the possibility for improved crashworthiness measures as a longer crumbling nose cone and keeping the pilot’s feet out of this zone. These features have been implemented in the design of the fuselages of the Advantage and Antares. Figure 7a Pressure distribution on a fuselage with cockpit extension NVvL 2006 7 Figure 7b Development of the drag on a fuselage with cockpit extension 1.5. Wing-fuselage interference The aerodynamics of a wing-fuselage combination is complicated 3 . As sketched in Figure 8, main flow effects are, at an increasing angle of attack, an additional increase of the angle of attack in the wing root region due to cross-flow of the fuselage named alpha-flow, resulting in viscous flow effects as the forward shift of the transition position on the wing and, if the laminar airfoil is not modified, separated flow at the rear of the wing root. At a decreasing angle of attack, an additional decrease of the angle of attack occurs in the wing root region causing transition to move forward on the lower surface of the wing. Another viscous flow problem is the separation of the turbulent boundary layer on the fuselage in front of the wing root due to the steep adverse pressure gradient towards the stagnation pressure on the wing root leading edge, and the resulting vortical flow system on the fuselage around the wing root. Figure 8 Alpha-flow effect (left) and viscous flow effects (right) on a wing-fuselage combination NVvL 2006 8 Figure 9 Oil flow patters on the wing lower surface (upper picture), wing upper surface (middle picture) and fuselage of a high-performance sailplane windtunnel model NVvL 2006 9 Figure 9 shows flow patterns, visualized by fluorescent oil in UV light, on the wing and fuselage of a high-performance sailplane windtunnel model. The change of friction at transition and zero friction at separation are depicted in the oil flow. The curved transition lines on the wing due the alpha-flow effect, and the separated flow on the wing-root upper surface are clearly visible. Zigzag tape, intended to trip the laminar boundary layer, is located in the turbulent boundary layer now and produces small vortices as indicated by the oil traces. Transition on the fuselage and the separation line around the wing root are clearly visible as well. In addition to these friction and pressure drag producing viscous flow effects, the circulation distribution of the wing-fuselage combination produces more induced drag than the wing only. The following examples illustrate how these problems have been tackled.