We also wish to thank the FIRST lab and Jason Miller for sharing his expertise in building unique composite materials. We would like to acknowledge Dr. Tommy ...
Variable Camber Compliant Wing – Wind Tunnel Testing Christopher R. Marks 1 and Lauren Zientarski 2 University of Dayton Research Institute, Dayton, Ohio, 45469-0013 Adam Culler 3 and Benjamin Hagen 4 Sierra Lobo, Inc., Wright-Patterson Air Force Base, Ohio 45433 Brian Smyers 5 and James J. Joo 6 U.S. Air Force Research Laboratory, Wright-Patterson Air Force Base, Ohio 45433-7542
This paper describes initial wind tunnel testing of a Variable Camber Compliant Wing developed by the U.S. Air Force Research Laboratory. The current version of the Variable Camber Compliant wing has a two foot chord length and was designed to demonstrate the ability to actively change wing section camber at low flight speed. The intent of the design was to adjust the wing camber by six percent chord while holding maximum camber location, and maximum thickness constant (emulating a change from a NACA 2410 to NACA 8410 profile) using a single continuous outer skin. The design is unique; the entire skin is seamless, continuous, and made of a single piece of non-stretchable composite skin. Smooth elastic deformation of the wing is attained by the underlying compliant mechanism. 2D camber change is achieved by a single actuation direction to control both leading and trailing edge deflection. 3D shape change is also capable through variation of camber along the span wise direction using a distributed actuation system along the span. Relatively low speed design and testing conditions were chosen to support air vehicle noise modeling efforts, for which low speed flight results in lower noise. Wing shape change without aerodynamic load was measured using an optical measurement system, indicating nearly six percent camber change measured at the middle rib. The wing was tested in the U.S. Air Force Vertical Wind Tunnel Facility in order to demonstrate operation of the wing under aerodynamic load. Increasing the profile camber resulted in an increase in section lift coefficient and vice versa. The variable camber wing is an inherently flexible structure that deforms under aerodynamic load. Digital image correlation of the model with wind off and wind on was used to understand the flexibility of the structure and its effect on aerodynamic forces. Aerodynamic modeling of the wings actual windoff shape and wind on shape are used to provide insight into the wind tunnel results and discuss the complexity of aerodynamic measurements on a flexible structure.
Nomenclature αl=0 = angle of attack where Cl is equal to zero αspar = angle of attack measured relative to the spar horizontal surface Cl,ms = section lift coefficient at midspan
hc = maximum section camber L/D = Lift to drag ratio
Cp = static pressure coefficient %FS = % of full stroke
Uin = freestream velocity xact = actuator position
1
Re
= Reynolds number
Research Engineer, Aerospace Mechanics Division, 300 College Park, Senior Member AIAA. Research Engineer, Aerospace Mechanics Division, 300 College Park, Student Member AIAA. 3 Program Manager, 2145 Fifth St., Senior Member AIAA. 4 Test Engineer, 2145 Fifth St., Senior Member AIAA. 5 Research Mechanical Engineer, AFRL/RQVV, 2130 Eighth St., Senior Member AIAA. 6 Research Mechanical Engineer, AFRL/RQVC, 2130 Eighth St., Senior Member AIAA. 1 American Institute of Aeronautics and Astronautics 2
I. Introduction
T
he United States Air Force Research Laboratory has developed a novel adaptive Variable Camber Compliant Wing (VCCW) that can actively re-contour the airfoil camber using compliant mechanisms. Compliant mechanisms are flexible, often single piece structures that upon application of a load change shape through elastic deformation. The wing concept is shown in Figure 1. A wing with distributed camber control would be able to reconfigure itself through the continuous deformation of the structure to optimize its geometry to suit current altitude, airspeed, and lift-to-drag (L/D) ratio requirements. A VCCW also has the potential to avoid undesirable aerodynamic conditions such as separated flow and increased parasitic drag through the seamless skin construction without holes and gaps. This can increase overall range and endurance of an aircraft in addition to control surface effectiveness and power reductions. In addition to adapting to current flight conditions for efficiency gains, the VCCW would enable wing reconfiguration to decrease noise. Flap side edge noise can be a large contribution to overall airframe noise [1]. The VCCW with seamless skin would significantly reduce or eliminate flap side edge noise by eliminating control surface gaps and edges. Airfoil trailing edge noise scales with V5 [2], and in general, flying slower can reduce airfoil self-noise. A previous study investigating the relationship between airfoil self noise and airfoil camber has found that increasing camber, and thus lift, can decrease noise by allowing decreased airspeed, and decreased angle of attack [3]. A VCCW would allow reconfiguration of wing section shape during various mission/flight segments to decrease noise. Variation of section profile shape is certainly not a new concept. It is the objective of flap devices used on modern aircraft, and researchers have been actively developing methods that improve current designs. The Mission Adaptive Wing could adjust its camber by incorporating separate leading and trailing edge mechanisms made up of rigid mechanical linkages. This allowed variable camber, roll control, and gust load alleviation to optimize cruise and maneuver conditions. Although this structure provided a continuous contour to reduce drag, the structure was extremely complex, increasing the weight [4]. The Defense Advanced Research Projects Agency (DARPA) Smart Wing Project also used a smoothly contoured trailing edge that demonstrated variable wing twist. Shape Memory Alloy Material allowed a trailing edge maximum deflection of ±25 degrees. Use of the hingeless control surface provided approximately an 8 to 18% increase in rolling moment compared to the traditional conventional design. The associated difficulties with this program mentioned severe manufacturing complications, power consumption, and scalability [5]. More recently, FlexSys designed and tested a Mission Adaptive Compliant Wing that exercised an adaptive trailing edge. This wing offered a smooth, variable trailing edge designed for high altitude endurance aircraft using stretchable skin. The wing was mounted to the underside of the White Knight stub pylon’s forward fuselage attach point. Testing proved to lower drag over a large range of lift with minimal weight and power penalties. This design was optimized to resist deflection under significant external aerodynamic loading [6]. Multiple other studies have been able to demonstrate controlled twist of morphing wings. A warp controlled wing twist introduced variable torsion by incorporating a torsionally stiff wing that can be twisted without relying on extremely powerful actuators. This wing consists of ribs that rotate around a fixed aluminum rod. Continuous carbon fiber skin was applied, except for the trailing edge. The upper and lower surfaces of the skins were separated to enable the skins to slide over the ribs. Lift and drag characteristics were determined through wind tunnel experimentation, resulting in a higher lift-todrag ratio, mostly at lower angles of attack [7]. Once again this active wing twist was tested to demonstrate that an active wing twist can have sufficient gain to control the rolling motion of the aircraft and maximize the lift-to-drag ratio. Yokozeki et al. have been developing a variable camber morphing wing that uses corrugated Figure 1. The VCCW design concept enables profile camber structures to actuate a flexible seamless flap. change with a continuous outer mold line. The section view shows three different section profile configurations with This device was driven by a wire system in the different values of maximum camber. leading edge of the structure. Wind tunnel test 2 American Institute of Aeronautics and Astronautics
setup involved three-dimensional flow over a low aspect ratio wing and focused on comparing the experimental results with a hinged wing, suggesting that improvement in the morphing wing lift coefficient showed great potential [8]. Lastly, design of a Smart Droop Nose is being explored as a seamless high lift device for the leading edge of an A380. Knowing that laminarity shows great potential for step changes in drag reduction, aerodynamic efficiency is being researched and tested through the use of seamless outer wing surfaces [9]. All of the programs mentioned have strongly suggested and provided a means of pursuing this type of technology. The results show great promise to effectively design for the optimization of a variety of flight conditions. The design of this version of the VCCW was related to AFRL efforts to develop multidisciplinary air vehicle design methods that include air vehicle noise as a design parameter [3,10-12]. The VCCW discussed in this paper was designed and built to demonstrate a seamless skin VCCW concept on the benchtop and in the wind tunnel. The wing has not been optimized for flight weight or flight structural requirements, which allowed the use of nontraditional materials and fabrication methods. The VCCW was developed for wind tunnel testing at low speed and moderate Reynolds number similar to the environment of a small unmanned aerial system. A companion paper [13] contains additional details about the wing design beyond what is included in this paper. This paper describes wind tunnel testing of the VCCW in the AFRL Vertical Wind Tunnel (VWT). The objective of the VCCW wind tunnel testing was to first, demonstrate the ability to design, build, and vary the wing section camber using compliant mechanism technology and a single outer skin under aerodynamic load. Secondly, measure the change in aerodynamic lift of the wing as the camber is changed, and finally, understand the behavior of the wing and its unique flexible structure under aerodynamic load. Lessons learned from this effort will be fed back into the design process to improve future morphing wing designs at AFRL. A comparison of aerodynamic force predictions of the actual measured wing shape is compared to experimental measurements. A 3D panel code was used for the aerodynamic force predictions included in this document; the reader is referred to Miller et al. [14] for a coupled FEA/CFD analysis of the VCCW design.
II. Wing Design Design and analysis of the prototype VCCW followed a natural progression from conceptual design of the compliant mechanism to final design of the wing installation in the VWT. The entire process was guided by a limited number of basic requirements, which opened the design space to non-traditional materials and manufacturing processes. These non-traditional solutions enabled both rapid prototyping and the integration of lightweight compliant materials into the final wing design. The objective of the wing design was to achieve a 6% change in maximum camber with standard NACA four digit series profile targets. The NACA profile targets chosen were NACA 2410 thru NACA 8410. These section profile targets each have a 10% maximum thickness and a constant location of maximum camber (at 40% chord), but with a variation in maximum camber from 2 to 8%. The fundamental design requirement for the compliant mechanism was to start with an initial NACA 2410 profile and use a single actuation force to deform the profile to a NACA 8410. Wing deformation is then produced using a series of compliant mechanisms, which act independently to produce either uniform or non-uniform camber across the span. Note that the specification of NACA profiles is simply to provide well-known targets for the initial and final shapes (and the corresponding aerodynamics). Custom initial and final shapes, as well as intermediate shapes, would be accommodated by essentially the same design and analysis process. Nonlinear finite element analysis (FEA) models were developed and utilized for a host of trade studies to evaluate shape under aerodynamic load, materials, and structural strength. Results of these studies helped guide the design from the conceptual compliant mechanism to a scaled prototype utilized for bench testing and small-scale wind tunnel testing to the full scale VCCW installed and tested in the VWT. A one foot span section of the wing was built to validate the design, construction methods, and actuation control of the wing. Continuous profile shape change across the span and variable profile change along the span was demonstrated on the benchtop and in a small developmental wind tunnel. The final wing design resulted in a non-typical wing structure consisting of a rigid main spar, flexible ribs, composite skin, and embedded actuators. The wing profile shape is controlled by miniature electronic actuators. The actuators are distributed in the wing to allow variation of the wing maximum camber along the span (e.g. generating variation of section lift across the span). Actuators are controlled independently using a microcontroller that commands and achieves desired positioning using a position feedback signal. The VWT entry was planned to demonstrate a lift coefficient change with profile shape change at a low to moderate angle of attack range (up to α = 8°) with a freestream velocity of 25.7 m/s (50 knots). The prototype VCCW is a straight rectangular wing with nominal dimensions of 1.82 m (6 ft) span and 0.61 m (2 ft) chord. The 3 American Institute of Aeronautics and Astronautics
short span results in a low aspect ratio model. The span was chosen to enable wing installation in other wind tunnels, and it also resulted in an appropriate sized model for AFRL’s VWT. Table 1 lists the nominal design parameters and tunnel conditions.
III. Experimental Setup
Table 1. Design parameters and VWT flow conditions. Parameter Value c 2 ft (0.61 m) b 6 ft (1.82 m) S 12 ft2 (1.115 m2) AR 3 U∞ 50 knots (25.72 m/s) Re 1.05 x 106 q∞ 0.057 psi (393.53 Pa) α -4 to 8°
The prototype VCCW was tested in AFRL’s VWT at WrightPatterson AFB. The VWT has a 3.65m (12 ft) diameter open jet test section and offers continuous flow speeds up to approximately 46 m/s (150 ft/s). Several unique diagnostics methods were used to understand the performance of the flexible compliant mechanism wing under aerodynamic load. This section describes the experimental setup and methods used to test the wing in the AFRL VWT facility.
A. Test Fixture The AFRL VWT facility was chosen for testing the VCCW because of its large size, ability to operate at the desired flow conditions, and good optical access. The test section of the VWT is accessible to test personnel, with plenty of room to mount optical measurement equipment, light sources, and other equipment used in this test to perform surface oil flow visualization and DIC. Installation of the prototype VCCW in the VWT required a new mounting fixture because a model of this nature (size and flexible structure) had not previously been installed and tested in the VWT. In addition to needing a mounting fixture to hold the model in the flowstream, there was no aerodynamic force measurement system available to test an article of this size in the VWT. A test fixture with embedded load cells is under development, but analysis, calibration, and validation of the aerodynamic force measurement system is ongoing and not described in this paper. Instead, local section lift coefficient was calculated by surface static pressure measurements made near the wing midspan. A CAD rendering of the wing mounted in the test section of the VWT is shown in Figure 2. Note that the flow is vertically upward (positive Load Cell #2 x-direction). The 1.82 m (6 ft) span by 0.61 m (2 ft) chord wing is centrally mounted within the 3.65 m (12 ft) Load Cell #1 diameter test section. Custom, identical fixtures at each end are secured to the outside of the test a.) section walls and pipe supports extend from the fixtures inward to support the wing and permit adjustments to angle of attack. Attachment of the wing to each pipe support is through an adaptor plate connected to the main spar of the wing and a load cell sandwiched between the adaptor plate and a coupler at the end of the pipe. Note that all wing loads are transferred through two load cells (one load cell at each end of the wing) as illustrated in Flow Direction Figure 2a. The load cells are capable b.) of measuring all six axes (forces and moments in x, y, and z directions). Figure 2. a.) Wing test fixture design shown in VWT test section. B.) Again, the load cells measurements are Actual wing installed test fixture. not included here. 4 American Institute of Aeronautics and Astronautics
Finite element models of the fixturing and certain structural members of the wing were developed for analysis of deformations and stresses. The fixture design illustrated in Figure 2a is fabricated from structural steel in readily available sizes and the FEA results show that it meets the strength and stiffness requirements of the tunnel installation. Since measurements were being made on a wing with a variable section profile shape, local angle of attack changes depending on actuator position. Unless otherwise noted, the angle of attack used in this paper is defined as the angle of attack relative to the horizontal surface of the spar (αspar). The angle of the spar relative to the flowstream (αspar) was directly measured by an inclinometer with accuracy of +/- 0.05°. B. Pressure Measurements The wing was instrumented with pressure taps located on the top and bottom surface of the wing at midspan as well as two additional outboard locations on the upper surface only. The installation of pressure taps were arranged in a diagonal at 15 degrees to avoid tripping the boundary layer and effecting downstream readings. The arrangement of the pressure taps along the span was switched from diagonal to a staggered layout after 0.3 x/c due to internal structural interference. Thirty-two taps were used at midspan: 20 on the upper surface and 12 on the lower surface. The tubing used for the ports had an internal diameter of 0.04 in installed flush to the skin surface. A 20 inH20 pressure scanner (ESP-32HP) was used for surface static pressure measurements. Pressures were recorded at 50 samples/sec for 20 seconds.
Figure 3. Layout of the static pressure taps.
C. Surface Oil Flow Visualization Surface oil flow visualization was used to understand the flow over the upper surface of the wing. This method typically uses a mixture of oil and dye that covers the wetted surface of the article being studied and has been used for decades. Previous researchers have used oil flow visualization to study a variety of two and three dimensional flows, for example, airfoils [15], turbine blades, and pump blades [16]. Once the air flow is turned on, the oil and dye mixture moves across the surface in the direction of the shear stress. Over time a lasting pattern forms on the surface that is different than the original pattern giving a qualitative indication of the direction and magnitude of the surface shear stress. The resulting flow pattern gives no indication of temporal flow behavior. Zierke and Straka note that interpretation of flow visualization patterns is fairly subjective [16]. Complicating the interpretation of the flow pattern is the effect of strong pressure gradients on the oil flow. Selig and McGranahan [15] have shown the usefulness of surface oil flow visualization in identifying laminar boundary layer separation and turbulent reattachment in two-dimensional flows. In our experiment the flow patterns are complex, especially in transitional regions showing strong variation of the flow pattern along the span. Several of the surface oil flow images are presented to illustrate the variation of the flowfield along the VCCW span and the effect of the non-rigid morphing structure on the flow. In our setup, a thin layer of oil was rubbed onto the test article, and then a mixture of mineral oil and fluorescent dye (oil-glow 44) was sprayed onto the wing creating a thin uniform coating. The test article was illuminated with a black light which provided a high contrast image of the oil flow over the wing surface. Two high resolution Nikon digital SLR cameras mounted in the test section were used to capture images immediately after application of the oil and dye mixture, then at a periodic interval with flow turned on until a steady state oil pattern was reached. One camera captured the entire wing upper surface, and one captured half the span at a higher spatial 5 American Institute of Aeronautics and Astronautics
resolution. The combination of low flow speeds and vertical model orientation required a few trials to get the correct mixture and application of oil on the model.
D. 3D Digital Image Correlation (DIC) and Photography The profile shape change of the wing in still air was measured Lamps using 3D optical scanning methods that are detailed in our companion paper [13]. The shape change of the wing under aerodynamic load was also of interest. Two different methods were used to understand the wing shape and wing structural deformation under aerodynamic load. First, two high resolution PCO4000 cameras (4008 x 2672 pixels) were mounted to the test Cameras fixture just outside of the flow stream. The two cameras Flow Direction provided views of the forward wing tip section and aft wing tip section. Combining the two simultaneous camera images resulted in a composite image that Figure 4. 3D Digital Image Correlation Equipment in the VWT test showed the entire wing tip profile. section. The composite image of the wing tip profile provided a quick and valuable check of the wing profile shape as well as video showing section camber change under aerodynamic load. Unfortunately the view from the wing tip only shows the profile shape of the skin at the wing tip, which is outside of the outboard rib and unsupported. This makes the section profile captured by the wing tip cameras a poor representation of the actual profile shapes throughout the rest of the wing span, especially in the higher camber configurations. The second method utilized two 5MP (2448 x 2050 pixels) cameras mounted normal to the wing just outside of the flowstream. The two cameras provided views of the 6’ span and 2’ chord wing section, The cameras were positioned to maximize the avilaible view of the leading edge given the constrains of the wind tunnel. These camera images were used in conjunction with GOM - ARAMIS software to determine in-plane and out of plane full-field deformations. Images were acquired with no air flow and under aerodynamic load. The comparison of these images provides full-field deformations due to the aerodynamic load. As in many research efforts there are tools available to acquire valuable and necessary information but not always of the appropriate configuration. The application of the DIC was a new tool for the wind tunnel, meaning not all conditions were met for optimal performance. Specifically, the distance between cameras was limited to the existing mounting bar, where a 25° angle between cameras is optimal but only 22° could be obtained. This affects the out of plane displacement accuracy which in this case the optimal accuracy would be 20µm. Given the very large displacements of the wing section, even by reducing the accuracy to 40µm the ability to collect full-field measurements provides great benefit. In-plane suffers less from the camera angle issue so accuracy is
