The military have also identified the need for an automated refueling process9. Typically air-to-air refueling is performed manually; however, the docking process ...
Two-Dimensional Localized Flow Control Using Distributed, Biomimetic Feather Structures: A Comparative Study Christopher J. Blower*a and Adam M. Wickenheisera Department of Mechanical & Aerospace Engineering, The George Washington University, Washington, DC, 20052
a
ABSTRACT This paper presents the development of a bioinspired flight control system and a characterization of its performance when operating in turbulent and gusting airflow conditions. This design consists of a skeletal structure with a network of feather-like panels installed on the upper and lower surfaces, extending beyond the trailing edge. Each feather is able to deform into and out of the boundary layer, thus permitting local airflow manipulation. The gust load sensing is predominately performed near the leading edge of the airfoil, and the reaction forces are generated by the feathers located at the trailing edge. For this study, the focus presents a benchmark case of the NACA 4412 airfoil with the standard 20% trailing edge flap design operating in a gusting, turbulent airflow. COMSOL Multiphysics is used to model the flow field and the fluid-structure interactions using Direct Numerical Simulation. The dynamics of the gusting model are developed using MATLAB and LiveLink connected to COMSOL to enable unsteady, turbulent simulations to be performed. Discrete and continuous gusts are simulated at various airfoil angles of attack. Additionally, the airfoils’ aerodynamic performance is comparatively analyzed between time-varying and steady-state turbulence models. This paper discusses how these two-dimensional, time-varying turbulent and gusting airflow simulation results can be developed and integrated into a LQR closed-loop feed back flight control system. Keywords: morphing, hierarchical control, bio-inspired, gust alleviation
1. INTRODUCTION Since the first manned flight, onboard systems have extended aircrafts’ ability to operate in unfamiliar environments. The introduction of the Global Positioning System (GPS), stormscopes and autopilot functions have aided pilots with their work load and offered safety when operating in turbulent conditions. Each of these systems is interlinked with the aircraft’s main flight computer and allows the aircraft and the pilot to access the flight path and therefore take the necessary precautions for inbound turbulent conditions. Despite advances in sensing and automatic control, turbulent airflow has been identified as a dominant factor that requires considerable analysis during aircraft design1. Consequently, an improved gust alleviation system could offer increased stability and maneuverability, with the potential of additional structural weight savings. In military applications, the benefit of using Unmanned Aerial Vehicles (UAVs) has been demonstrated by a yearly growth in missions undertaken during the last decade. The benefits include the removal of pilots from war-zones, removal of onboard life support systems, and increased endurance and maneuverability capabilities. Although these advantages lead to the increased use of UAVs in military applications, turbulent weather conditions remain a significant hazard to these aircraft2–3. The cause of these incidents is a lack of sensory feedback to the pilots through the visual display screens available, where live video footage and flight data are presented. The limitations of the data displayed have lead to accidents during low-level flight where turbulence causes sudden changes in the aircrafts’ altitude and orientation2. UAVs have been used for reconnaissance missions due to their capability to operate over the target location for extended periods of time. To optimize the quantity and quality of the data collected during reconnaissance, low-level flight is required. However, natural and manmade structures on the Earth’s surface can induce turbulent characteristics in the airflow4. In turn, the aircraft operating in these regions experience sudden and severe gusts that can result in rapid Bioinspiration, Biomimetics, and Bioreplication, edited by Raúl J. Martín-Palma, Akhlesh Lakhtakia, Proc. of SPIE Vol. 7975, 79750L · © 2011 SPIE · CCC code: 0277-786X/11/$18 · doi: 10.1117/12.882310
Proc. of SPIE Vol. 7975 79750L-1
disturbances in orientation, position, and airspeed. The airflow interaction with ground-based obstructions varies due to numerous factors including size, shape and porosity5. Research performed by the Department of Land, Air, and Water Resources has analyzed the modes of turbulence generation through interaction with forested and jungle based regions. These studies have determined that the forests canopy is the main source of turbulent flow due to the uneven surface with which the air interacts. Additionally, the tree species also affects the profile of the turbulent flow due their drag characteristics and the resultant circulation flow that exists in the body of air in close proximity to the tree line6–8. To continue using UAVs within these locations, a gust alleviation system is required to maximize the aircraft’s stability while reducing the risk of crashing. To achieve this capability, the data collection systems onboard UAVs require an increase in local sensing resolution to reduce aircraft loss when experiencing highly irregular flows. The military have also identified the need for an automated refueling process9. Typically air-to-air refueling is performed manually; however, the docking process is considered hazardous due to the narrow margins of error available, restricted turbulence evasion capabilities, and a reduced spatial awareness2,10. By automating this process, an interaction between the two aircraft is required to allow both vehicles to share flight data and operate as a singular unit11. This approach would offer improved efficiency and decreased risk while in close formation flight and reduced turbulence evasion reaction times. Over the past two decades multiple air flow manipulation systems have been researched and developed to alleviate gust disturbances and reduce stall in turbulence. NASA has developed a LIDAR-based system designed to preemptively predict the inbound airflow characteristics by determining the velocity of dust particles. The system has been declared successful when operating in a minimum of 8km of clear visibility12. However, due to the size and weight of the LIDAR, the system is available only on commercial transports13 and consequently inapplicable to small-scale UAVs flying in close quarters. The alleviation of turbulence through the implementation of active flow control devices such as synthetic jet actuators (SJAs) has widened the flight envelope for many aircraft. The improvement in flight capabilities has been predominately achieved by delaying wing stall during high angle of attack maneuvers14–15. Additionally, several groups have researched the use of SJAs to maintain full flight control capabilities during operation in severe flight conditions through the generation of an oscillating flow that delays flow separation16–17. This will one day lead to aircraft that “can maneuver without mechanical control surfaces”14. However, research concerning SJAs focuses predominantly on the system’s ability to extend the flight capabilities and has yet to be applied directly for gust alleviation. The integration of vortex generators has become a common feature on multiple aircraft currently in operation. These devices manipulate the airflow over the wing’s surface to create micro-vortices that prevent boundary layer separation and delay wing stall18. This capability can be achieved while operating in turbulent bodies of air with high eddy dissipation19, consequently improving flight smoothness. In addition, a bio-inspired eddy flap system has been implemented on numerous aircraft. This system deflects flaps into the airflow passing over the upper wing surface to prevent the boundary layer separation point from travelling forward towards the leading edge20. This method is inspired and developed from the passive techniques used by birds to alleviate gusts. Bio-inspiration has also been a factor in morphing aircraft design and airflow separation delay devices. The concept of geometric morphing in-flight has been in development for many years and can be seen in military aircraft such as the Tornado & F-14 Tomcat. However, the advancements in morphing aircraft have developed dramatically in recent years due to progress in smart materials and distributed control systems. Several research groups have applied bio-inspiration to specialized mission objectives that aircraft today are yet to achieve. Groups at MIT & University of Toronto have successfully achieved controlled flight of ornithopters21–22. DeLaurier and Dietl have developed an ornithopter to hover similar to a humming bird23–24. Wickenheiser and Garcia have developed a bio-morphing aircraft with perching capabilities, thus allowing the aircraft to land in places once unobtainable by aircraft25. All these approaches have been achieved through studying bird flight; however, to date, a successful real-time gust and turbulence alleviation systems using control techniques and aeroelastic deformation used by birds have yet to be achieved. This paper presents the development and results of time dependent turbulent and gusting simulations for a bio-inspired, feather-like distributed gust alleviation system (GAS) originally discussed by Blower and Wickenheiser26. The integration of an electromechanical feather-based GAS across the wing surface has the opportunity to offer new flight capabilities including increased stability in turbulent flow regions and enhanced maneuverability. Airflow manipulation through wing morphing and feather deflections has been proven successful by the adaptive wing structure of birds. Their
Proc. of SPIE Vol. 7975 79750L-2
advanced fligght capabilitiess are visible thrrough decreaseed turning radiii and pitch andd roll rates that far exceed the abilities of today’s airrcraft. During flight, a bird has h the ability to t perform conntinuous and neearly instantanneous adjustments to its wing profile to allow flight along its inteended path durring a wide speectrum of weaather and envirronmental condditions27. This technicaal approach vaaries from prevvious GAS devvelopments ass bio-mimicry of nature’s feeedback loops has h been implementedd as a hierarchiccal control schheme controllinng the electrom mechanical feathhers. The proposedd GAS implem ments bio-mimeetic feathers ovver the entire wing w to perform m the necessaryy airflow deflecctions as seen in Figurre 1. The feath hers’ coverage of the wing alllows the replaccement of the standard flaps and ailerons, therefore t enabling the GAS to perfo orm the role of o both flight control and gust alleviationn. Blower andd Wickenheiseer26 have performed CFD simulation ns to compare the t flight charaacteristics of a singular trailinng flap, as typiically seen on modernday aircraft, to t their gust alleviation systeem design that implements the dual upper annd lower trailinng edge flaps shown s in Figure 2.
Fig gure 1: Wing witth feathers installled for the biom mimetic gust allevviation system
Figure 2:: NACA 4412 w. w standard singular trailing edge flap (above) andd the gust alleviaation systems duual trailing edge flap 2 configguration (below)26
t create and maintain m the airrfoil’s profile when w located inn their rest position, thereby allowing a The feathers are designed to the necessarry lift forces and a load suppport to be maaintained. Thhe GAS’s holllow skeletal wing w design has h been implementedd to enable each h feather to rottate into and out o of the wingg’s profile, perm mitting numeroous wing profiiles to be achievable thhrough featherr positioning. By enabling this capabilityy, the feathers can be rotatedd to allow parrticularly strong gusts to transpire th hrough the airrframe with litttle impedancee. Each featheer is integratedd into its own electromechanical control c modulle with the abbility to sensee turbulent loaadings and perrform actuatioon through thee use of piezoelectric transducers and a microcontrrollers. In adddition, the featther modules are a designed too have multifuunctional capabilities. Because the gust g alleviationn feathers coveer the entire wiing, the aileronns have been reemoved; conseequently, the momentss required to allow a the aircrraft to roll are achieved throough deflectionn of the trailinng-edge featheers. The developmentt of this dual integration i of flight control and gust allevviation system ms into the winng utilizes bioomimetic characteristiccs that maximizze the efficienccy of the system ms.
Proc. of SPIE Vol. 7975 79750L-3
This paper discusses the generation of the time-dependent turbulent airflow simulations with the integration of discrete and continuous gust models. The simulations performed enable the comparison of turbulent models to stationary simulations and experimental data appearing in the literature to ensure accuracy of the simulations is maintained through a range of flight conditions. Additionally, these time-varying simulations are used in the development of the closed-loop feedback flight control system. Subsequently, the stationary and time-varying simulation results are presented and the resultant coefficients of lift and drag are analyzed to ensure consistency through the development of the GAS model. 2. CFD MODELING This section describes the modeling and analysis process used to compare time-dependent gusting wind models to previously developed stationary solutions and historical, experimental data28. The development of a time-varying wind model is the initial step in the construction of the gust alleviation system’s closed-loop feedback controller. The wing model has been based on a NACA 4412, as there is a strong base of historical data available to confirm the results of the CFD simulations. The wing has been initially designed with a chord of 30 cm, from which a 20% flap length has been selected due to data being available for lift and drag coefficients (Cl and Cd, respectively). The gust alleviation system is implemented in the COMSOL Multiphysics environment. This software is chosen for its capabilities to solve computational fluid dynamics (CFD), fluid-structure interactions (FSI), structural design and mechanisms, piezoelectric transducers modeling, and optimization simultaneously. However, the focus of this paper is orientated around the CFD and FSI capabilities in addition to the software’s LiveLink function, where COMSOL can be interlinked with MATLAB to allow custom functions to be integrated into the simulation. This capability enables closed-loop control systems to be designed to the custom needs of this gust alleviation system. Our goal is to develop a closed-loop feedback system that measures the unsteady wing loadings being experienced during gusting to determine the required control effort to return the aircraft to its original flight path. This development process has been broken down into several stages that are discussed below. 2.1 Turbulent Stationary Modeling Initially, simulations are performed using stationary k-ε turbulence modeling under steady-state conditions in order to compare our results against historical and JavaFoil (a popular airfoil analysis package) data. The simulations are performed at various wind velocities, angles of attack, and trailing edge deflections. The results to date have indicated close correlation to both series of data with angles of attack between -5° and +10°. At angles of attack above and below these limits, discrepancies between the computational results and the historical data have been identified26. The turbulent models generated by COMSOL are presenting lift, drag, and pitching moment (Cl, Cd, and Cm) data that indicates that the boundary layer remains attached to the wing surface while at these angles of attack. From the historical data presented by Abbott and Doenhoff28, boundary layer separation and wing stalling should begin around an angle of attack of 10°. Consequently, investigations are under way to resolve the cause of the variation. The CFD simulationshave been set with a nominal speed of 12.17 m/s, giving a Reynolds number of 250000. Additionally, k-ε turbulence model parameters have been set to mimic the airflow characteristics within the wind tunnel used to generate the data in Ref. 28; these parameters are defined by the following equations:
k=
3 ( U I T )2 2
(1)
k3 2 LT
(2)
ε = Cμ
34
where U is the airspeed, IT is the turbulent intensity, which on a low turbulence wind tunnel can be assumed to be 0.004 (Ref. 29), LT is the turbulent length scale, and Cμ is the model constant for a flow through a pipe, given by Cµ = 0.09 (Ref. 30). The value for LT is determined from the following equation:
LT = 0.07L
(3)
where L is the width of the wind tunnel at the testing point; for this initial analysis the wind tunnel width has been defined as 1m. This dimension is selected to ensure the simulations generate comparable data without risk of influence of the tunnel walls on the flow over the airfoil. In addition, the walls of the defined control volume have been defined
Proc. of SPIE Vol. 7975 79750L-4
with slip characteristics, thereby minimizing boundary layer development that may disturb the airflow aroundthe wing profile. 2.2 Turbulent Time-Dependent Modeling After the stationary turbulence models have been developed, the simulations are extended to time-varying flows. Simulations are performed for a period of 10s with results given in time increments of 0.1s. The simulations are designed to initially start at a flight speed of 0 m/s and increaseto 12.17 m/s over a period of 0.5s, as shown in Figure 3. This allows the simulations to evolve into a fully developed flow over the remaining time. These simulations are used to confirm that the time-varying solutions matched the stationary flow results from previously performed simulations26.
Velocity [m/s]
20 15 10 5 0
0
1
2
3
4
5 Time [s]
6
7
8
9
10
Figure 3: Velocity vs. time plot displaying the CFD airflow acceleration from rest to a steady-state speed of 12.17 m/s
2.3 Gusting Wind Model Wind gusts are simulated using the Matlab LiveLink, which allows custom time-varying functions to be integrated into COMSOL simulations. Initially, a half sine wave is implemented to analyze the velocity field as a gust passes over the wing over the period of 30s. At this stage, gusts are implemented as changes in wind speed only; changes in wind incidence angle are beyond the scope of this work. The gust is initially designed to be discrete with a total gust speed of +6 m/s, consequently resulting in a maximum airflow of 18 m/s being experienced by the airfoil at t=5.5s, as seen in Figure 4. To develop a continuous gusting model, a Fourier series of three sine waves are combined to give a simplified continuous gust model based off of a sinusoidal Gaussian random process model31. For consistency, the simulations are all performed with a time step of 0.1s for 10s and with an initial acceleration from rest to cruise speed (12.17 m/s) over a period of 0.5s. Figure 5 demonstrates the change in wind speed over the 10-second period. A customized gust function is written in Matlab and implemented in COMSOL to create a time-dependent inlet velocity profile. This specific model is applied to enable direct comparison between the gusting and stationary models; the coefficients of lift and drag are compared to determine the quasi-steadiness of the flow dynamics. Simulations are performed with fixed trailing edges and varying angles of attack between ±5°.
Velocity [m/s]
20 15 10 5 0
0
1
2
3
4
5 Time [s]
6
7
8
Figure 4: Singular gusting wind model with varying inbound airflow velocity
Proc. of SPIE Vol. 7975 79750L-5
9
10
Velocity [m/s]
20 15 10 5 0
0
1
2
3
4
5 Time [s]
6
7
8
9
10
Figure 5: Periodic gusting wind model with varying inbound airflow velocity
3.4 Flap Deflections The flap deflection model is an open-loop system that enables a variation in the trailing edge angle during the timedependent turbulent simulations. This function is developed using the Matlab Live-Link to enable the trailing edge to deflect between the predetermined range of ±10˚ during real-time simulation. This capability enables comparative studies to be performed between the stationary model data and the time-dependent varying trailing edge deflection data at specified angles of deflection. Additionally a means of wing body rotation through angle of attack and pitching moment has been created using Matlab. This function enables the wing to deflect in response to changes in pressure and viscous stress distributions and the angle of rotation of the trailing edge. This function is integrated into the COMSOL time-dependent models to enable the simulations to output the typical changes in Cl, Cd and Cm that exist in the real world. Consequently, this produces the aerodynamic look-up tables for the closed-loop feedback system that enables the aircraft to return to the desired trim position after turbulence, or a discrete gust, has been sensed.
4. RESULTS A series of stationary turbulent models are created to determine aerodynamic characteristics of the aircraft wing over a wide range of flight conditions. During the simulations, the wing’s angle of attack, trailing edge deflection, and velocity are varied to create a field of data that can be used to compare COMSOL results to historical data. The experimental data, as seen in Figure 6, shows evidence of boundary layer separation at high angles of attack due to the decrease in Cl. The required angle of attack to cause boundary layer separation increases in accordance to the trailing edge deflection, which delays stalling with the increase downward deflection. The drag polar plot from the experimental data presents a turbulent plot trend with an increase in Cl for a small variation in Cd; this is visible for each of the simulations when the angle of attack is between -5˚ and +10˚. Additionally the graphs depicted below in Figures 7 and 8 demonstrate the variance of trailing edge deflection versus angle of attack and velocity versus angle of attack, respectively. Each simulation is performed with the trailing edge deflection being moved in one degree increments. The graphs demonstrate that the Cl for both the historical data and COMSOL follow similar trends throughout the angle of attack range simulated. The variation in Cl due to the change in trailing edge deflection in Figure 7 are confirmed by the flight characteristics discussed by Abbott and Doenhoff28, where the data demonstrate that a downward rotation of the flap results in an increase in the Cl and vice versa for an upward rotation. These COMSOL simulations display a linear variation in Cl with respect to the angle of flap deflection that follows the general trends of small angle deflection. The drag polar plot demonstrates trends that are comparable to those of the experimental data. The vertical shift between each drag polar shows the resultant change in Cl for each flap deflection angle. Each non-zero flap position demonstrates an increase in Cd due to the increase in wing profile that is exposed to the oncoming wind. However beyond the limits of -5˚ and +10˚, the COMSOL simulations predict less separation than data appearing in the literature, and no significant loss in Cl is incurred. The drag polar plot demonstrates a vertical displacement with respect to the change in flap displacement. Additionally, an increase in Cd can be identified with increased deflection angles, both positive and negative. This simulation data complies with that of the historical data; however, the shape of the polars diverges at high angles of attack. The data obtained from COMSOL portrays a drag polar similar to a low Reynolds number laminar flow; investigations are underway to determine the cause of the variation between the two data sets.
Proc. of SPIE Vol. 7975 79750L-6
2.5
2.5
2
2
1.5
1.5
Coefficient of Lift
Coefficient of Lift
The Cl vs. angle of attack plot for a fixed trailing edge and varying velocity, shown in Figure 8, demonstrates a consistent Cl profile throughout the velocity variations. However, a small horizontal displacement is evident with respect to the increase in Reynolds number that complies with the trends presented by Abbott and Doenhoff28. This displacement along the horizontal axis can be seen particularly when Cl=0. The increase in velocity, or Mach number, can be seen to result in the translation of the lift curve slope. Consequently, to maintain the same Cl characteristic, the angle of attack of the wing must be decreased; this trend complies with experimental data28. Similarly, the drag polar plots demonstrate an increase in Cd with increased angles of attack; this is due to the front facing cross-sectional that comes into contact with the inbound gust consequently resulting in an increase in induced drag. This effect is supported by the trends seen in Figure 6 and 7. The plots can be seen to shift in the horizontal direction with an increase in velocity; this is expected as the induced drag decreases with respect to velocity, whereas form drag increases at a slower rate. This is only possible when in the lower regions of the airfoils power curve and in these simulations the data demonstrates the airfoil is operating in this region.
1 0.5 0 -0.5 -1 -1.5 -10
1 0.5 0 -0.5 -1
-5
0 5 Angle of Attack
10
-1.5 0
15
0.05 0.1 0.15 Coefficient of Drag
0.2
2.5
2.5
2
2
1.5
1.5
Coefficient of Lift
Coefficient of Lift
Figure 6: NACA 4412 Cl vs. angle of attack (left) and drag polar plot (right) using experimental data from Ref. 28 with a constant velocity of 12.17 m/s – trailing-edge flap deflection angle + = 10˚, Δ = 5˚, □ = 0˚, ○ = –5˚, * = –10˚
1 0.5 0 -0.5 -1 -1.5 -10
1 0.5 0 -0.5 -1
-5
0 5 Angle of Attack
10
15
-1.5 0
0.05 0.1 0.15 Coefficient of Drag
0.2
Figure 7: NACA 4412 Cl vs. angle of attack (left) and drag polar plot (right) using COMSOL with a constant velocity of 12.17 m/s – trailing-edge flap deflection angle + = 10˚, Δ = 5˚, □ = 0˚, ○ = –5˚, * = –10˚
Proc. of SPIE Vol. 7975 79750L-7
2.5
2
2
1.5
1.5
Coefficient of Lift
Coefficient of Lift
2.5
1 0.5 0 -0.5 -1
1 0.5 0 -0.5 -1
-1.5 -10
-5
0 5 Angle of Attack
10
15
-1.5 0
0.05 0.1 0.15 Coefficient of Drag
0.2
Figure 8: NACA 4412 Cl vs. angle of attack (left) and drag polar plot (right) using COMSOL with a fixed trailing edge and a varying velocity – airflow velocity * = 9.34 m/s, ○ = 12.17 m/s, □ = 14.61 m/s, Δ = 17.04 m/s, + = 19.48 m/s
The gusting airflow model is developed using a Fourier series that allows varying gusts to be experienced in the simulations without sudden changes in the airflow velocity. The gusting model has been designed to create gradual variations in airflow velocity; this enables the changes in the lift and drag coefficients to be studied in a fully developed state during this stage of the research. However, the requirement of sudden changes in airflow velocity will be required in the future, as the GAS will be required to react to multiple inbound gusts that may be experienced simultaneously and from a range of incidence angles. The simulations are performed for 10 seconds, where the first 0.5s consists of the initial velocity increase from 0 to 12.17 m/s. Once the flow reaches the targeted cruise speed, the gusting wind model is initiated and produces several gusts within this time span, as seen in Figure 5. The forces experienced on the wing during each 0.1s time step are extracted from COMSOL pressure and viscous stress distributions and normalized to yield the Cl and Cd to be plotted with respect to time during the gusting simulations. Several simulations are performed at various angles of attack to allow comparisons to be performed under different operating conditions. The three simulations presented below are modeled with the airfoil at an angle of attack of +5˚, 0˚ and -5˚; see Figure 9 and 10 for Cl and Cd, respectively.
Coefficient of Lift
The values for Cl and Cd that are produced during the first simulation where the angle of attack is set to 0˚ comply with the values given by the stationary simulations. The increase in Cl is expected from the increased angle of attack of +5˚, and similarly the proportional decrease for the respective decrease in angle of attack is found. The analysis of the Cd complies with the data from the stationary simulations, indicating that the time scale of the transient aerodynamics is much shorter than the transit time of the free stream through the flow field. The variation in Cd is at a smaller order of magnitude than that of the Cl; however, this is still reasonable as the increase in induced drag is expected when the frontal area of the wing is increased as the airfoil pitches.
0.6 0.4 0.2 0 -0.2 0
1
2
3
4
5 Time [s]
6
7
8
9
Figure 9: Gusting wind model, lift force plot with a NACA 4412 airfoil at varying angles of attack – angle of attack Δ = +5° , □ = 0°, ○= -5°
Proc. of SPIE Vol. 7975 79750L-8
10
Coefficient of Drag
0.08 0.06 0.04 0.02 0 0
1
2
3
4
5 Time [s]
6
7
8
9
10
Figure 10: Gusting wind model, drag force plot with a NACA 4412 airfoil at varying angles of attack – angle of attack Δ = +5° , □ = 0°, ○= -5°
5. CONCLUSIONS AND FUTURE WORK The gust alleviation system model has been developed using a time-dependent k-ε turbulence model, where the airflow increases from 0 to 12.17 m/s and becomes a fully developed flow. Each of the time-varying simulations performed are shown to be closely correlated to the steady-state simulations developed previously. These simulations enable the progression of the turbulent air flow model to be extended to allow gusts to be incurred. The basic sinusoidal Gaussian random process model has been implemented to create multiple velocity magnitude gusts over a period of 10 seconds. The simulations performed include several gusts within this time frame of which the data at each time step closely follow the data gathered from the stationary and time-varying turbulence models. The trends seen during the gusting model provide a good basis to proceed with varying wind incidence models. A closed-loop feedback system is to be designed to assess the current perturbation of the aircraft wing’s position and orientation during flight in a turbulent airflow and to perform the necessary control surface adjustments to allow the wing to return to its original undisturbed position. The model is intended to measure the changes in pressure created by the inbound flow on the wing’s surface. The feedback loop will continually adjust the position of the trailing edge flap to counteract the changes in Cl, Cdand Cm when gusts of varying magnitude are experienced by the wing. The gusting model will be later developed to vary in angle of incidence once this step has been completed, thus allowing the simulations to represent airflow disturbances closer to those experienced in the real world. The development of the feedback loop, to date, has been constructed in several components. Initially two airfoil models have been developed in Matlab that have the capability to rotate with respect to angle of attack and trailing edge deflection. Once completed, the COMSOL LiveLink will enable the forces experienced on the wing to be exported to Matlab, where the necessary response will be calculated and then relayed back to COMSOL for implementation. The closed-loop feedback system will be developed for both the standard single trailing-edge wing model and the feathered wing. This system will be based on the control scheme of a Linear Quadratic Regulator (LQR). This will enable the control schemes of both the standard and biomimetic wing design to be compared and consequently enable the full extent of the benefits of the gust alleviations dual flap trailing edge design to be determined. This feedback loop will require additional constraints to be integrated into the wing model, as each flap will be designed to deflect individually or in clusters depending on the turbulent forces experienced by the wing. These constraints will prevent contact between neighboring feathers and will prevent transfer of forces between feathers that would cause inaccurate gust readings to be registered by the sensors.
REFERENCES [1] [2] [3] [4] [5]
Advisory Group for Aerospace Research and Development, “Gust load prediction and alleviation on a fighter aircraft,” AGARD, Report R-728, (1986). McCarley, J.S. and Wickens C.D., “Human factors concerns in UAV flight,” FAA, Report, (2004). Williams, W., and Harris, M., “The Challenges of Flight-Testing Unmanned Air Vehicles,” Proceedings: Systems Engineering, Test & Evaluation Conference, (2002). Cionco, R.M., Vaucher, V.T., D’Arcy, S., and Bustillos, M., “Near-building turbulent intensities, fluxes, and vortices,” US Army Research Laboratory, Report J6.4, (2004). Belcher, S.E., and Hunt, J.C.R., “Turbulent flow over hills and waves,” Annual Review of Fluid Mechanics, 30, pp. 507-538, (1998).
Proc. of SPIE Vol. 7975 79750L-9
[6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26] [27] [28] [29] [30] [31]
Shaw, R.H, and Schumann, U., “Large-eddy simulation of turbulent flow above and within a forest,” Department of Land, Air, and Water Resources, Report, (1992). Coutts, M.P. and Grace, J., “Wind And Trees”, Cambridge University Press, Cambridge, UK, (1995). Katul, G., Hsieh, C.I., Kuhn, G., and Ellsworth, D., “Turbulent eddy motion at the forest-atmosphere interface,” Journal of Geophysical Research, Vol. 102, pp. 409-413, (1997). Strategy Page, Aerial refueling for UAVs, [Online: 07/07/2009],[Cited: 07/06/2010], URL: www.strategypage.com. Sinsley, G.L., Long, L.N., Niessner, A.F. and Horn, J.F., “Intelligent Systems Software for Unmanned Air Vehicles,” AIAA, Report 2008-0871, (2008). How, J., King, E. and Kuwata, Y., “Flight demonstrations of cooperative control for UAV teams,” AIAA, Report 2004-6490, (2004). Soreide, D., Bogue, R.K., Ehernberger, L.J., and Bagley, H., “Coherent LIDAR turbulence measurement for gust load alleviation,” NASA, Report 104318, (1996). Soreide, D., Bogue, R.K., Ehernberger, L.J., Hannon, S.M., and Bowdle, D.A., “Airborne coherent LIDAR for advanced in-flight measurements (ACLAIM) flight testing of the LIDAR sensor,” NASA, Report 8.18, (2000). Barron Associates, Adaptive control for synthetic jet actuators, [Online: 2010], [Cited: 06/10/2010], URL: www.barron-associates.com/adaptive-control-of-synthetic-jet-actuators/. Duvigneau, R., and Visonneau, M., “Optimization of a Synthetic Jet Actuator for Aerodynamic Stall Control,” Computers & Fluids, Vol. 35, pp. 624, (2006). Breuker, R.D., and Abdalla, M., “Aeroelastic Control and Load Alleviation using Optimally Distributed Synthetic Jet Actuators,” Proceedings: Structures, Structural Dynamics, and Materials Conference, Report: 2007-2134, (2007). McGowan, AM.R.,Horta, L.G., Harrison, J.S., Raney, D.L., “Research Activities within NASA’s Morphing Program,” NASA, Report ADP010487, (2000). Tilmann, C.P., Langan, K.J., Betterton, J.G., and Wilson, M.J., “Characterization of pulsed vortex generator jets for active flow control,” Proceedings: RTO AVT, Report RTO MP-051, (2000). Cheung, P., Lam, C.C., and Chan, P.W., “Estimating turbulence intensity along flight paths in terrain-disrupted airflow using anemometer and wind profiler data,” Proceedings: Conference of Mountain Meteorology, Report P2.29, (2008). Patone, I.G, Muller, W., Bannasch, R., and Rechenberg, I.I., Bionik und Evolutionstechnik – Flexible flaps for separation control on a wing with low aspect ratio, [Online: 05/29/1997], [Cited: 12/23/2010], URL: www.bionik.tu-berlin.de/user/giani/vortrag/sld002.htm Cory, R.E., [Supermaneuverable Perching], MIT Doctoral Thesis, Cambridge, (2010). DeLaurier, J.D., “The Development and Testing of a Full-Scale Piloted Ornithopter,” Canadian Aeronautics and Space Journal, Vol. 45, pp.72, (1999). DeLaurier, J.D., “Ornithopter Research,” Proceedings: Bioflight Workshop, (2003). Dietl, J.M., Garcia, E., “Stability in Hovering Ornithopter Flight,” Proceedings: SPIE, Vol. 6930, (2008). Wickenheiser, A.M., and Garcia, E., “Optimization of Perching Maneuvers through Vehicle Morphing,” Journal of Guidance, Control and Dynamics, pp. 815, (2008). Blower, C.J., and Wickenheiser, A.M., “Biomimetic Feather Structures for Localized Flow Control and Gust Alleviation on Aircraft Wings,” Proceedings: International Conference on Adaptive Structures and Technologies, (2010). Moore, N., Birds, bats and insects hold secrets for aerospace engineers, [Online: 02/04/2008], [Cited: 10/03/2010], URL: ns.mich.edu/htdocs/releases/story.php?id=6312. Abbott, I.H., and Doenhoff, A.E.V., [Theory of Wing Sections], Dover Publications, New York, (1959). Nader, G., Santos, C.D., Jabardo, P.J.S., Cardoso, M., Taira, N.M., and Pereora, M.T., “Characterization of low turbulence wind tunnel,” Proceedings of World Congress: Metrology for a Sustainable Development, Report, (2006). COMSOL, [CFD Module User Guide], COMSOL, Burlington, (2010). Hoblit, F.M., [Gust Loads on Aircraft: Concepts and Applications], AIAA Education Series, Washington, (1988).
Proc. of SPIE Vol. 7975 79750L-10
