Embedded Files
My undergraduate senior capstone project was a self arranged research project for Dr. James A. Liburdy, a Fluid Dynamics professor at the Oregon State University College of Mechanical, Industrial, and Manufacturing Engineering (MIME). One of his areas of research is Micro Air Vehicle (MAV) flight. MAV flight is characterized by slow speeds and small wing chord lengths (typically, although not exclusively, airspeeds under 40 mph, and chord lengths around 10-30cm). For the flight regime of MAVs, the highest-lift airfoil is usually very thin (small chamber) but prone to large over-wing vortices which dynamically move the center of lift to the rear of the wing as the vortex grows and travels downstream. This vortex has a low pressure center, and is partially responsible for the high lift generated. However, because of the dynamic movement of the center of lift, flight with such wings is often very unstable, and therefore unsuitable for any application requiring a stable platform such as cameras. Dr. Liburdy was interested in researching the ability to control this vortex to increase vehicle stability by means of a dynamically controlled leading edge flap independent from the angle of attack.
Along with my two team members David Haley and Kate Shreeve who agreed to join me on this project, I designed and built such a wing to be mounted to a sting balance (a rod that models can be mounted to to measure lift, drag, pitch, yaw, and roll moments for aerodynamic research) in the 4ft x 5ft OSU low speed wind tunnel. The following are a list of characteristics of the wing and control system:
Dimensions:
Chord Length: 20cm
Span: 77cm
Chamber: 7.7 mm
Flap length: 30% of chord
Mass: 640 g
5:1 Elliptical leading and trailing edge profiles.
Carbon fiber laminate and frame for lightweight and rigidity
The anticipated vortex shedding frequency was close to that of the wing's natural frequency, so care had to be taken to maximize spanwise stiffness. The weight of the wing was limited by the weight limits of the sting balance.
The flap was driven by RC servo's embedded in the wingtips and controlled using LabVIEW.
Through the LabVIEW interface, the user can control the flap frequency, waveform, and angular limits. The flap could also be fixed to a single user-defined angle.
Lift, drag, and flap angle data could be viewed in real time and acquired using LabVIEW.
Figure 1 shows the completed wing mounted to the sting balance in OSU's low speed wind tunnel. Flow splitter plates on either side ensure two-dimensional flow over the wing, allowing the wing to be used to validate simplified numerical models.
Fig 1. Wing and flow splitter plates mounted in the wind tunnel
Design of the wing was driven to make it both as light and stiff as possible to avoid natural frequencies in the range of vortex shedding frequencies, and to protect the sting balance sensors from being overloaded. A number of materials were considered, but carbon fiber proved to have the highest stiffness-to-weight ratio, and was therefore chosen. Additional modal analysis confirmed that it would provide adequate stiffness. The wing was built using a composite sandwich structure, with Divinycell HT110 foam core sandwiched between two .025" thick carbon fiber laminates. Figure 2 shows the flap construction. The outer surface of the elliptically profiled leading and trailing edges was made with 1/64th in thick plywood, due its greater flexibility over carbon fiber, allowing it to match the 5:1 required elliptical profile. A .060" diameter unidirectional carbon fiber rod was placed in the leading and trailing edges to function as both a hard edge to protect the plywood laminates from damage, as well as provide the tight curvature required at the leading and trailing edge. The wing is mounted to the sting balance by an aluminum insert, which is held in place between two carbon fiber square tubes inside the wing (see fig. 3). Steel pins lock the insert to the rods, anchoring it in place and precluding the risk of delamination under high loading.
Fig 2. Composite flap construction showing the carbon fiber square tube, foam core, plywood laminate, and carbon fiber rod in the leading edge. The plywood extends past the square tube to meet the fiberglass tube on the wing at a tangent at all flamp angles.
Fig 3. Composite wing construction showing the aluminum insert, foam core, plywood trailing edge, flap hinges, and the fiberglass tube used to provide an aerodynamic seal and to maintain continuous upper wing surface geometry.
The flap hinges were made from off-the-shelf RC aircraft hinges, glued into the foam core of the wing. The hinges could easily separate to allow the flap to be detached. To maintain an aerodynamic seal between the upper and lower surfaces of the wing, and to maintain a smooth profile over the upper surface of the wing, a fiberglass tube was placed between the flap and the wing, concentric with the flap hinge axis and tangent to the upper surface. As the flap moved down, the curved surface of the tube becomes visible, maintaining smooth and continuous geometry (once again to simplify comparisons to numerical models). Figure 4 shows the servo mounted in a recessed cavity in the bottom surface of the wing, , near the wing tips between the two carbon fiber square tubes. Flow visualization studies were expected to take place in the center of the wing, thus mounting the servos near the wing tips precluded interference with the 2D airflow in the center of the wing.
Fig 4. Flap and servo control linkage along with the bonded carbon fiber and plywood laminates and foam core that make up the structural components of the wing.
The user interface is shown below in Figures 5 & 6. It consists of a flap control interface which allows the user to specify an angular range for flap motion, the frequency of flapping, and the waveform (sinusoidal, sawtooth, triangle, etc). For static testing, the user inputs only a single flap angle. The data acquisition interface displays the wing's lift, drag, L/D ratio, roll and pitching moments, and flap angle live and allows the user to start and stop data acquisition. Figure 7 shows the LabVIEW block diagram.
Fig 5, 6, 7. LabVIEW user interface and block diagram for flap control and data acquisition, as well as live data monitoring.
Subsequent testing of the wing in the low speed wind tunnel was performed to validate the design. Figure 8 shows a negative picture of smoke lines in a 2D vertical plane at a moderate wing angle of attack and downward flap angle. As evidenced by the smooth smoke lines, the flow remains laminar in this configuration. Smoke lines were created by applying mineral oil to a tensioned steel wire using a cotton swab, and subsequently heating the wire by passing electrical current through it. The oil forms tiny droplets on the wire that burn up when heated, generating smoke that is swept away by the airstream.
Fig 8. Picture of smoke lines, enhanced to make the smoke more visible
Figure 9 shows how a moderate angle of attack with zero flap angle results in the generation of erratic vortices over the wing. These vortices are largely responsible for the unstable flight of MAVs. This picture also shows the heated wire burning the tiny droplets of mineral oil. Oscillating the flap at constant frequency (shown in fig. 10), the wing is capable of generating large distinct vortices in a controlled manner, much, in fact, in the way that hummingbirds do to maintain hover. As this was the primary goal of this project, it was cause for much celebration!
Fig 9. Erratic overwing vortices at 1.5 m/s, AoA = 14°
Fig 10. Controlled overwing vortex generation at 1.5 m/s, AoA = 14°
Following completion of this project, I performed additional work on my own to characterize the wing's static lift performance and make enhancements to the user interface (already shown above) and data acquisition capabilities of the wind tunnel facility. Some of this work is summarized below, but a full report is available here: Follow-up work on MAV wing. Figures 11-13 show lift as a function of both wing angle of attack, and flap angle. All plots were generated using MATLAB. Future research would perform characterization of dynamic performance.
Fig 11. Each curve of the lift performance at 10.0 m/s shows the angle of attack at which a stall occurs for each flap setting. This plot best depicts how a large flap angle will delay the onset of a stall (a sudden decrease in lift).
Fig 12. A surface plot of the lift performance at 10.0 m/s shows a level region at small flap angles and high angles of attack, indicating it might be possible to achieve relatively consistent lift in these configurations.
Fig 13. Required flap angle for maximum lift (lbs) at 10.0 m/s
Based on the plots in figures 11-13, the following was concluded: in order to maintain the maximum amount of lift, as the wing angle of attack is increased, so should the flap downward angle, by an amount greater than the increase in wing angle of attack (see fig. 14). the wing continued to be used to perform studies in micro air vehicle flight for several years at least, and has been a source of several other follow-up senior capstone projects. This project provided me with a lot of experience in the design of MAV wings, and the use of a wind tunnel to perform characterization of aerodynamic performance.
Fig 14. Maximum lift configurations for flap angle θ and angle of attack (AoA) α
Google Sites
Report abuse