Search

Experimental measurements of the flow structure evolving in the separated flow over an NACA 0009 wing at 12° angle of attack were obtained with particle image velocimetry, surface pressures, and force transducer measurements of the lift coefficient and pitching moment coefficient. Phase-averaged two-dimensional velocity field measurements provide details of the separated shear layer evolution following a four-pulse burst sequence from a synthetic jet actuator. The flow field development is quite similar to the observations made by Brzozowski, et al. (2010), who used a pulsed-combustion actuator that is orders of magnitude stronger than the synthetic jet. Proper orthogonal decomposition of the PIV data sets showed that the combination of the time-varying coefficients modes 1 and 2 correlate with the negative of the lift coefficient response. The surface pressure signals were correlated with the roll up and convection of the large-scale vortex structure that follows the actuator burst input. A spatially localized region of high pressure occurs below and slightly behind a "kink" that forms in the shear layer. A localized region of high surface pressure that follows the kinked region correlates with the lift reversal that occurs within 2.0t^+ after the burst signal was triggered.

This work formulates a switched linear modeling procedure to understand and predict the unsteady aerodynamic forces arising from rapid pitching motion of a NACA 0012 airfoil at a Reynolds number of 50,000. The system identification procedure applies a generalized dynamic mode decomposition algorithm to time-resolved wind tunnel measurements of the lift and drag forces, as well as the pressure at six locations on the suction surface of the airfoil. Linear state space models are identified for 5-degree pitch-up and pitchdown maneuvers within an overall angle of attack range of 0◦–20◦. The identified models accurately capture the effects of flow separation and leading-edge vortex formation and convection. It is shown that switching between different linear models can give accurate prediction of the nonlinear behavior that is present in high-amplitude maneuvers. The models are accurate for a wide range of motions, including pitch-and-hold, sinusoidal, and pseudo-random pitching maneuvers. Providing the models access to a subset of the measured data channels can allow for improved estimates of the remaining states via the use of a Kalman filter, which could be of use for aerodynamic control applications.

Wind tunnel experiments were conducted on two symmetric airfoils with different thickness ratios as a test of the ability of Goman-Khrabrov-type models to predict the lift coefficient history during pitching maneuvers. The primary difference between the two airfoils was that the thin airfoil had no hysteresis in its lift curve during quasi-steady maneuvers, but static hysteresis was observed with the thick airfoil over a range of 16 < α < 22. Both airfoils exhibited dynamic hysteresis in the lift coefficient when the wing was pitching. The existence of static hysteresis had a strong effect on the lift response during periodic pitching, and must be accounted for in the model. A modified version of the Goman-Khrabrov model was introduced that captured the static hysteresis behavior and the large-scale features of the dynamic hysteresis for the thick airfoil. Large-scale variations in the lift trajectory at different pitching frequencies are reproduced by the modified model, even when the thick airfoil is in a deep stall condition. Some deviations between model and experiment are observed at the highest angles of attack, which are attributed to the dynamic stall vortex and trailing edge vortex shedding.

The lift coefficient schedule required for a flight vehicle to fly a neutral energy loop maneuver is computed using a parametric optimization procedure. The effects of gust period, phase between gust components, and the maximum allowable lift coefficient on the minimum gust amplitude required for neutral energy flight are examined. The question of whether the desired lift schedule is actually achievable is answered by considering the unsteady fluid dynamics. We show that the Froude number is the ratio of the fluid dynamic time scale to the vehicle dynamics time scale. The Froude number is relatively constant for flight vehicles ranging from small flying insects to large transport aircraft when they fly at minimum drag conditions. Limitations on achievable performance with active flow control are considered using simple models for separated flow and closed-loop control.

Airfoil aerodynamic loads are expected to have quasi-steady, linear dependence on the history of input disturbances, provided that small-amplitude bounds are observed. We explore this assertion for the problem of periodic sinusoidal streamwise gusts, by comparing experiments on nominally 2D airfoils in temporally sinusoidal modulation of freestream speed in a wind tunnel vs. sinusoidal displacement of the airfoil in constant freestream in a water tunnel. In the wind tunnel, there is a streamwise unsteady pressure gradient causing a buoyancy force, while in the water tunnel one must subtract the inertial load of the test article. Both experiments have an added-mass contribution to aerodynamic force. Within measurement resolution, lift and drag, fluctuating and mean, were in good agreement between the two facilities. For incidence angle below static stall, small-disturbance theory was found to be in good agreement with measured lift history, regardless of oscillation frequency. The circulatory component of fluctuating drag was found to be independent of oscillation frequency. For larger incidence angles, there is marked departure between the measured lift history and that predicted from Greenberg's formula. Flow visualization shows coupling between bluff-body shedding and motion-induced shedding, identifiable with lift cancellation or augmentation, depending on the reduced frequency. Isolating the buoyancy effect in the wind tunnel and dynamic tares in the water tunnel, and theoretical calculation of apparent-mass in both cases, we arrive at good agreement in measured circulatory contribution between the two experiments whether the flow is attached or separated substantiating the linear superposition of the various constituents to total lift and drag, and supporting the idea that aerodynamic gust response can legitimately be studied in a steady freestream by oscillating the test article.

Measurements of the time rate of energy exchange between a two-dimensional wing and the freestream flow were obtained during static and dynamic maneuvers. The change in the freestream speed in a wind tunnel is used as an indirect method for determining the time rate of energy exchange. A static calibration is first obtained by measuring the reduced freestream speed while changing the angle of attack of the wing from α = 0 o to 20 o . Two dynamic cases were examined, a vertically oscillating (plunging) maneuver with the wing at a fixed angle of attack, and a stationary wing with an oscillating freestream. The measured time rate of energy transfer and the period-averaged energy from flow to the wing is compared to resulting change in freestream speed. The static calibration of the change in flow speed to the rate of energy exchange is ineffective at predicting the change in energy during the dynamic cases. Although qualitative agreement between velocity time series and the time rate of energy change are observed.

Linear unsteady aerodynamic models are obtained for a NACA 0006 airfoil in a wind tunnel experiment at Reynolds number 65,000. A family of system identification maneuvers, based on a canonical step-up, hold maneuver, are specifically designed to extract relevant unsteady information from phase averaged wind tunnel measurements. Linear models based on these maneuvers are constructed using OKID/ERA for the case of pitching about α0 =0 ◦ and α0 =1 0 ◦ , as well as the case of pure plunge. It is shown that the identified models are more accurate than Theodorsen’s model for all maneuvers, especially those at larger base angle of attack.

Experimental measurements of the time varying lift, drag, and pitching moment acting on a nominally two-dimensional wing undergoing plunging motions in an oscillating freestream flow were obtained with a six-component force balance. Rates of energy transfer between the wing and the flow were computed at reduced frequencies in the range expected for micro aerial vehicle flight. Comparisons with classical theories for unsteady aerodynamics were made for the cases of pure plunging motion and pure oscillating freestream. The phase of the vertical plunging motion relative to the freestream flow was changed to investigate the effect on the rate of energy exchange between the flow and the wing. The maximum energy transfer to the wing occurred near a phase of 0 o , and the minimum occurred near a phase of 150 o . At a lower angle of attack of α = 5 o the energy transfer to the wing was sufficient to produce a net energy increase, but in the post-stall region, α = 15 o , the rate of energy increase was not sufficient to produce a net positive energy increase.

Numerical simulations of the natural and actuated unsteady flow over a three-dimensional low-aspect ratio wing are performed using Lattice Boltzmann method. The LBM simulations match the flow conditions and the detailed wing geometry from previous experiments, including the actuators that are installed internally along the leading edge of the wing. The present study focuses on the transient lift response to short-duration square-wave actuation, for the wing in a uniform flow at five different angles of attack. Overall, both mean and unsteady numerical results show good agreement with the experimental data, in particular at the post-stall angle of attack 19 � , where the maximum lift enhancement occurs. At that angle of attack, the effects of the actuation strength and duration are investigated. In general, the lift response to a single pulse increases with increasing actuator mass-flow rate and pulse duration.

Vortical structures in the separated flow and near-wake regions behind a threedimensional wing in an oscillating freestream and a pitching two-dimensional flat plate are documented using particle image velocimetry. The phase-averaged vorticity field is used to compute the circulation around the wing. Three different angles of attack were studied corresponding to fully attached, partially attached and fully separated flows. The effect of the freestream oscillation frequency on the circulation around the semi-circular wing is compared to measurements of the fluctuating lift. In a separate experiment the flow field on the suction surface of the two-dimensional flat plate was documented during a pitch-up, hold, then pitch-down maneuver in a steady flow. The spanwise variation in the flow structure over the wing was documented.

The unsteady lift of a low Reynolds number wing in an oscillating freestream is documented in terms of its amplitude and phase. The phase variation of the lift relative to the freestream velocity shows a larger phase difference than predicted by classical unsteady flow theory. A constant time delay between the lift and the actuator was observed to be τ^+ = t_(delay)U/c = 5.3 when normalized by the freestream speed and chord. Feed forward control of pulsed-jet actuators is used to modulate the lift coefficient of the wing, in an attempt to suppress the lift oscillations. Suppression of the fluctuating lift at the fundamental frequency was partially successful, but additional "noise" was added to harmonics of the lift signal by the controller.

Active flow control is used to modify the lift, drag and pitching moments on a semicircular wing with aspect ratio, AR = 2, and chord Reynolds number is 68,000. The wing is mounted on a pitch/plunge sting mechanism that responds to the instantaneous loads and moments acting on the wing. The leading edge of the airfoil contains 16 spatially localized actuators that can be independently controlled. Smoke wire visualization, surface pressure and six-component force balance measurements are used to characterize the effects of openloop forcing. The lift coefficients on the steady wing are enhanced with the actuation, similar to the effect of dynamic stall vortex lift enhancement that occurs during a pitch up maneuver. Surface pressure measurements are being used to construct a flow model for use in feedback control. Progress toward the goal of designing a feedback controller to stabilize the flight of the model in an oscillatory freestream is discussed.

The goal of this work is to develop low order dynamical systems models for the unsteady lift and drag forces on small wings in various modes of flight, and to better understand the physical characteristics of unsteady laminar separation. Velocity field and body force data for a flat plate at static angle of attack and in sinusoidal pitch and plunge maneuvers are generated by 2D direct numerical simulations using an immersed boundary method at Re = 100. The lift of a sinusoidally plunging plate is found to deviate from the quasi-steady approximation at a reduced frequency of k = 0.5 over a range of Strouhal numbers. Lagrangian coherent structures illustrate formation and convection of a leading-edge vortex in sinusoidal pitch and plunge. A phenomenological ODE model with three states is shown to reproduce the lift on a flat plate at a static angle of attack above the stall angle. DNS for a 3D pitch-up maneuver of a rectangular plate at Re = 300 shows the effect of aspect ratio on vortical wake structure and lift. Wind tunnel experiments of a wing in single pitch-up and sinusoidal pitch maneuvers are compared with a dynamic model incorporating time delays and relaxation times to produce hysteresis.

Experiments in a supersonic wind tunnel studied the behavior of acoustic tones introduced by open-loop forcing, and the ability of a closed-loop controller to suppress acoustic tones. The performance of pulsed-blowing and zero-net-mass actuators was documented. By independent changes in the wind tunnel static pressure and actuator supply pressure, we showed that the amplitude of disturbances input from the actuator is proportional to the pressure difference across the actuator, and scaled with the freestream dynamic pressure. The cavity used in this experiment at M = 1.86 responded linearly to the disturbance input from the actuators. Acoustic tone amplitudes in the cavity were linearly proportional to actuator input, and nonlinear interactions between different acoustic modes (such as the formation of combination modes) were not observed. Given that the system responded linearly to open-loop forcing and the amplitude of disturbances from the actuator exceeds background disturbance levels, then closed-loop control of the Rossiter tones should be feasible using standard linear control design tools. A simple gain-phase adjustable feedback controller is used to demonstrate linear closed-loop control.

Open-loop active flow control is used to modify the spanwise distribution of circulation around an NACA 0012 and flat plate wing. The leading edge on both airfoils and tip regions of the NACA airfoil contain spatially localized actuators that can be independently controlled in terms of amplitude and frequency, allowing the spanwise distribution of circulation to be modified. Different orientations of the pulsed-blowing actuators were used to provide upstream, downstream, in-line with the flow, and outward span components of actuation. The actuation effectiveness was documented using force balance measurements of the lift and drag, smoke-wire visualization, surface pressure measurements and PIV velocity field measurements. Actuation with an upstream component is shown to be far more effective in reducing the separated region than actuation in the streamwise direction. Initial measurements of the change in circulation on the suction surface of the airfoil indicate that spatially localized forcing produces global changes over the wing, primarily associated with the reduction in size of the separated flow region.

Open-loop active flow control is used to induce swirl in the flow behind a row of inlet guide vanes known as Smart Vanes. Smart IGV’s are aligned with the flow, and employ the Coanda effect to produce flow vectoring. Experiments at Illinois Institute of Technology and full-scale proof-of-concept experiments in the U.S. Air Force Research Laboratory’s Compressor Research Facility at Wright Patterson AFB demonstrated 14 degrees of vectoring could be achieved at 85% design speed.

Active flow control of cavity tones has been successful in controlling Rossiter tones in both subsonic and supersonic flows. While closed-loop control promises the lowest actuator power requirements and the ability to adapt to changing flow conditions, so far it has only been demonstrated on subsonic cavities. Suppression of Rossiter tones in supersonic cavities has been demonstrated using open-loop techniques. Progress and issues related to closedloop control and controlling supersonic cavities are reviewed.

Feedback control is used to suppress oscillations in the subsonic flow past a rectangular cavity. A heuristic feedback law is introduced into 2D direct numerical simulations, using zero-net-mass forcing at the leading edge of the cavity and a pressure sensor at the wall, and the oscillations are reduced by 13dB. Reduced-order models are obtained from fullorder direct numerical simulations, and used to design observers to reconstruct the full flow information from a single pressure measurement at the wall, and these observers are shown to be more eective than Linear Stochastic Estimation (LSE), which is commonly used for this purpose. Initial results of a new experiment are also given, and a model for designing zero-net-mass actuators with a desired bandwidth is presented, along with experimental data which supports the model predictions.

This paper presents a series of experiments aimed at model-based feedback control of oscillations in the compressible flow past a rectangular cavity. Zeronet-mass air injection is used with feedback from pressure transducers located inside the cavity to suppress the oscillations, and measurements of the shear layer above the cavity are made for both the natural and suppressed state. A system identification experiment is performed, and used to determine a linear model that describes the oscillations, and this model is used to derive feedback laws for acoustic suppression. The model accurately predicts the behavior of dierent feedback controllers, including frequencies where suppression occurs, as well as new resonant frequencies that may be introduced by the controller.

Experiments using active control to reduce oscillations in the flow past a rectangular cavity have uncovered surprising phenomena: in the controlled system, often new frequencies of oscillation appear, and often the main frequency of oscillation is split into two sideband frequencies. The goal of this paper is to explain these effects using physics-based models, and to use these ideas to guide control design. We present a linear model for the cavity flow, based on the physical mechanisms of the familiar Rossiter model. Experimental data indicates that under many operating conditions, the oscillations are not self-sustained, but in fact are caused by amplification of external disturbances. We present some experimental results demonstrating the peak-splitting phenomena mentioned above, use the physics-based model to study the phenomena, and discuss fundamental performance limitations which limit the achievable performance of any control scheme.

An experimental investigation aimed at controlling multiple acoustic modes in a cavity has been conducted. Closed-loop control is used to suppress and/or enhance individual resonant modes in the cavity at Mach numbers ranging from 0.2 to 0.55. The energy distribution among the resonant acoustic modes is shown to depend on cavity depth. A connection between the various Rossiter modes was established using narrow bandpass filters to control specific resonant modes. When the second Rossiter mode is suppressed, then the first and third modes are enhanced and vice versa. Furthermore, suppressing the second Rossiter mode caused the fourth mode to be suppressed also. Enhanced energy in a mode indicates that it exists for a greater percentage of the time of the data record. Multiple acoustic modes could be suppressed simultaneously with the appropriate configuration of the feedback control system. Introduction Acoustic resonances in aircraft weapons bay cavities can reach sound pressure levels of 170 dB at Mach numbers as low as 0.5. Such large pressure fluctuations result in early structural failures within the aircraft and stores, create high drag on the aircraft, and generate unpredictable loads on the stores during separation. Systematic investigations of the nature of the acoustic resonances within aircraft cavities have been conducted since the 1950’s’Z2. Nomenclature C, = Pressure coefficient (p-p,)/9 D = Cavity depth f = Frequency L = Cavity length M = Freestream Mach number PI = Wind tunnel static pressure 9 = Dynamic pressure % pU2 St = Strouhal number, fL/U U = Freestream speed W = Cavity width In practical applications the high sound pressure levels are commonly reduced using passive devices such as serrated spoilers at the upstream edge of the cavity and ramps at the downstream wall. However, the performance of passive sound suppression devices are known to be sensitive to Mach numbe? limiting the range of their effectiveness, and enhancing certain acoustic tones at off-design conditions. Ideally the cavity control system should adapt to changing flight conditions to maintain a minimum sound pressure level over the entire flight envelope. The potential for active control of weapons bay cavities was recognized by the U.S. Air Force several years ago and led to the formation of the Active Robust ConTrol of Internal Cavities (ARCTIC) group. This consortium of Department of Defense laboratories, industries, and universities was formed in 1996 as the ’ Professor, Associate Fellow AIAA T Currently, Assistant Professor, Mechanical Engineering Department, Santa Clara University, Santa Clara CA. Copyright