3rd AIAA Flow Control Conference, Jun. 5–8, 2006, San Francisco,CA
Flow Control Using Plasma Actuators and Linear/Annular Plasma Synthetic Jet Actuators Arvind Santhanakrishnan∗ and Jamey D. Jacob† Dept. of Mechanical Engineering, University of Kentucky, Lexington, KY 40506, U.S.A
and Yildirim B. Suzen‡ Mechanical Engineering & Applied Mechanics, North Dakota State University, Fargo, ND 58105, U.S.A This paper investigates the use of dielectric barrier discharge plasma actuators in low Re flow control applications. Three different actuator geometries have been tested: a conventional design using two rectangular strip electrodes (the linear actuator) that produces a nearly two-dimensional horizontal wall jet upon actuation, and two new designs that render the plasma induced flow in the form of a vertical jet that can be either three-dimensional using an annular electrode array actuator construction - the plasma synthetic jet actuator, PSJA or nearly two dimensional using a modified linear actuator construction consisting of two exposed electrodes and one embedded electrode, the L-PSJA. The modification in actuator design can be used to broaden its applicability and enhance the flow control effects. 2-D PIV measurements are used to characterize the operation of these actuators in quiescent flow, a flat plate boundary layer, and flow over a circular cylinder. In quiescent flow, these actuators add momentum to the residual fluid with significant velocity fluctuations. The interaction of the plasma induced flow with a mean flow is shown to vary with the actuator geometry. The PSJA and L-PSJA geometries enhance the penetration of the plasma induced jet as compared to the linear actuator. The actuators act as an active boundary layer trip, the effectiveness of which is seen to decrease with increasing freestream velocity. While the PSJA affects the global flowfield, the L-PSJA and linear actuator affect primarily the near wall region. The linear actuator is observed to be a better configuration for flow control on a circular cylinder as opposed to the L-PSJA.
I.
Introduction
he subject of flow control is currently of major importance in fluid dynamics research, and several T techniques to alter and manipulate a flowfield have been explored. Some examples of passive flow control methods (where an input of external energy is not required) that have been used in existing literature 1
are: boundary layer trips,2 roughness elements,3 ejector nozzles4 and surface perturbations.5, 6 Active flow control, which is specifically of interest to this paper, requires an input of external energy and has the advantage of being controllable, viz., it can be turned on or off as required. Some examples of active flow control methods include acoustic excitation,7 continuous or pulsed suction and blowing8–10 and surface motion.11–13 Detailed reviews of active and passive flow methods can be found elsewhere.14–16 Plasma actuators, also known as dielectric barrier discharge actuators (or OAUGDPTM , one atmosphere uniform glow discharge plasma17 ) typically refer to an asymmetric arrangement of two electrodes (one exposed to the atmosphere and the other embedded under the aerodynamic surface) separated by dielectric material (see Fig. 1). Under input of a high voltage, high frequency AC, a region of dielectric barrier discharge plasma is created in the interfacial air gap. This plasma region drives the residual fluid in the form ∗ Ph.D.
Candidate; Student Member AIAA;
[email protected]. Professor; Senior Member AIAA;
[email protected]. ‡ Assistant Professor; Senior Member AIAA;
[email protected]. c 2006 by the authors. Published by the American Institute of Aeronautics and Astronautics, Inc. with Copyright permission. † Associate
of a horizontal, zero-net mass flux (ZNMF) wall jet. Plasma actuators can be readily employed as active flow control devices, and have been shown to control boundary layer separation through their addition of near-wall flow momentum.18, 19 Likewise, synthetic jets are also ZNMF jets that are formed from the working fluid of the system in which they are applied, and have been shown to be useful active flow control devices.20–23 They are commonly produced by using an oscillating diaphragm mounted in a cavity that is embedded flush on the aerodynamic surface of interest. The diaphragm is driven in such a way that fluid is alternatively ejected from or sucked into the cavity in a periodic manner. The jet is created by the advection and interaction of discrete vortical structures. A comprehensive review on synthetic jets can be found elsewhere.24 It should be noted that the above two active flow control devices operate on fundamentally different mechanisms. There are possibilities to enhance the flow control obtained by plasma actuators by mimicking the behavior of synthetic jets, and thereby expand their applicability. The present work aims to investigate a novel concept that combines the features of both plasma actuators and synthetic jets, in order to construct a potentially more efficient flow control device. A.
Previous Work
A brief background on plasma actuators with some recent developments in the state-of-the-art will be presented in this section. The articles by Rivir et al.,25 Suchomel et al.,26 and Corke and Post27 present detailed reviews of the developments in plasma flow control, and can be referred to for further information. Malik et al.28 first used DC corona discharge (or “ion wind”) to manipulate flat plate boundary layers. They reported drag reduction in the order of 20% for freestream velocities up to 30 m/s and 15 kV applied voltage. They concluded that drag reduction was largely voltage-dependent, and that it would be necessary to use multiple discharges to obtain higher efficiencies. El-Khabiry and Colver29 numerically investigated corona discharge induced drag reduction in flat plates at low Re. They observed that drag reduction diminished with increasing freestream velocity and decreasing potential difference between the wire electrodes. Roth et al.30 used glow discharge surface plasma to control boundary layer flow on flat plates. They observed that asymmetric spanwise electrode strip configurations produced thrust, while symmetric spanwise electrode strip configurations increased drag. They attributed these effects to be the result of a combination of mass transport and vortical structures induced by electrohydrodynamic body forces. Some applications of plasma actuators that have been investigated include: lift enhancement on airfoils,31, 32 as virtual flaps and slats for airfoil flow control,33 low pressure turbine blade separation control,18, 34 phase synchronization of the KVS from multiple cylinders,35 circular cylinder wake vortex control,36 and to control the coherent structures in a planar, weakly compressible, free shear layer.37 A study that is of interest to this paper is the application of plasma actuators for landing gear noise reduction by Thomas et al.38 They mounted multiple actuator elements along the circumference of both upstream and downstream parts of a generic landing gear strut with cylindrical cross-section. Reduction of both drag and vortex shedding was obtained with steady actuation, and drag reduction was found to be be inversely related to Re. Alternate strategies for noise control using base blowing (to control vortex shedding) and vectoring the upstream wake were discussed. While most of the above applications are restricted to incompressible flows, some high speed applications such as in axisymmetric jet forcing using arc plasma based configurations have been reported.39 Some studies have shown that regions of low intensity plasma can have large impacts on the surrounding flow field.19 Enloe et al.40, 41 presented numerous findings on the behavior of plasma actuators. Based on large scale integral measurements of thrust output, voltage and plasma emission measurements, and simulations, the authors make several interesting conclusions, including that the power input, P , to the plasma is nonlinear with the voltage drop, ∆V , across the dielectric, and that both the maximum induced velocity and thrust are proportional to input power. Their final conclusion is that the plasma induces an electrostatic body force on the surrounding fluid that is proportional to the net charge density and the strength of the electric field. They observed that the direction of the plasma-induced flow can be tailored by the electrode arrangement, which is important to the current study. Several factors including input power, input voltage, input frequency, electrode geometry, actuator orientation, dielectric material, dielectric thickness, freestream Re, pressure gradient, plasma chemistry and humidity of air have been shown to affect the behavior of a plasma actuator. Baughn et al.42 measured velocity profiles at several locations upstream and downstream of a plasma actuator, and calculated the body force using a control volume momentum balance approach. They observed that the force was not affected in the presence of an imposed crossflow (for freestream velocities in the order few meters per second), and 2 of 30 American Institute of Aeronautics and Astronautics Paper 2006-3033
the extent of the force was limited to well within the boundary layer. Also, the body force production decreased with input AC frequency (between 5 kHz and 20 kHz) for nearly constant input power. Porter et al.43 measured the body force produced on steady operation of a plasma actuator, and examined the effects of individually varying input AC voltage and frequency. They conclude that the time-averaged body force is: (a) linearly proportional to input AC frequency between 5 kHz and 20 kHz (for constant input voltage), and (b) non-linear with the input voltage (for constant frequency). Further, they observed that the actuator “pushes” (with higher magnitude) and “pulls” (with lower magnitude) the fluid in opposing directions during each cycle. Balcer et al.44 used plasma actuators on a flat surface with adverse pressure gradient resembling the suction surface of a Pack-B turbine blade, and the exposed electrode was oriented sixty degrees to freestream direction. While an increase in near-wall velocities was observed on actuation, the skewed electrode orientation did not introduce any substantial longitudinal vorticity in the boundary layer. Roth and Dai45 present detailed investigations on the effects of changing dielectric material, electrode geometry, input frequency and input voltage on the flow velocity induced by a plasma actuator. It was observed that the choice of dielectric material affects: (a) plasma volume, (b) distribution of electric field lines (dictated by the dielectric constant value), and (c) dielectric heating power loss (which in turn, is proportional to input AC frequency and electrode area). In general, a material combining all these properties was recommended: higher dielectric constant value (e.g. alumina), higher dielectric strength (e.g. Kapton) and lower heating loss factor (e.g. quartz and Teflon). Compared to quartz, Teflon was found to generate a higher value of induced flow velocity for lesser power input and was used as the dielectric material for their remaining parametric studies. The width of the embedded electrode was found to have negligible effect on both the maximum induced velocity and input power per unit length of the actuator. The separation distance between the edge of the exposed and embedded electrodes (horizontal distance between rightmost edge of upper electrode and leftmost edge of lower electrode) in Fig. 1 was adjusted to have variable gap, zero gap, and overlap between electrodes, with all other parameters being the same. Interestingly, this variation had a significant effect on the induced velocity, and a 1-2 mm gap width was found to be optimum for their actuator. The maximum induced velocity was found to increase with input voltage (at constant frequency) and input frequency (at constant voltage) to a certain level, after which it reached a constant value and later started to drop with any increase. VanDyken et al.46 observed that a thicker dielectric would be able to handle higher input voltage, and an optimum input AC frequency (to obtain maximum body force) existed for a specific dielectric. Enloe et al.47 speculated that the lifetime of the plasma was related to the presence of negative oxygen ions in the air. Anderson and Roy48 showed that the skin friction coefficient produced by a plasma actuator mounted on a flat plate varied with the relative humidity of air. While a vast majority of PFC literature is primarily experimental in nature, several methodologies for numerical modeling of plasma actuators have been investigated. Most of the current models typically use an electrohydrodynamic (EHD) body force term to simulate the effect of PFC actuator on the external flow. Shyy et al.49 considered a linear approximation for the electric field distribution within a plasma region consisting of a triangular area located above the embedded electrode. The time-averaged electrostatic body force acting on the fluid was computed and then added to the 2-D steady, incompressible, Navier-Stokes equations to solve for the entire flowfield. A similar approach of specifying the charge distribution and spatial distribution of electric field has been used recently to investigate effects of both steady and unsteady plasma actuation on a 3-D flow.50 Suzen et al.51, 52 computed the body force as the product of net charge density and electric field, and this was introduced as a source term in the Navier-Stokes equations. The electric field was obtained from Maxwell’s equations as the gradient of scalar electric potential, and another equation for charge density was solved. Hall et al.53 adopted a potential flow approach to model the behavior of a plasma actuator, where a doublet element was used to mimic the actuator-induced velocity field. Orlov and Corke54, 55 used the body force expression developed by Enloe et al.,41 and a lumped electric circuit model was used to compute the electric potential and volume of plasma. In contrast to the simplified treatment used in the above models, some researchers have used a more elaborate particle physics and plasma chemistry based approach. Font et al.56–58 used a particle-in-cell direct-simulation-Monte-Carlo (PIC-DSMC) method to model the interaction between various ionic species, and numerically solve for electric potential. The model developed by Roy et al.59 solves for drift-diffusion form of equations governing dynamics of ions, electrons and fluid. Shang60 solved the time-dependent Maxwell’s equations along with particle concentration equations to compute the electromagnetic field distribution of a dielectric barrier discharge. Besides the beneficial attributes of ease of construction, high operational bandwidth, large control authority33 and excellent dynamic response,38 the actuator geometry can in itself be varied to create complex
3 of 30 American Institute of Aeronautics and Astronautics Paper 2006-3033
flow structures.61 This has important implications on the use of plasma actuators in flow control.51, 62 B.
Background
Linear actuators, as shown in Fig. 1, while are useful devices, are limited by the extent of the wall jet in affecting the flowfield. Figure 2(a) shows the flowfield induced by a linear actuator in quiescent flow. The width (or in this case, height) of the jet is confined to less than 1 cm away from the wall. This is seen more clearly in the velocity profiles of the horizontal velocity u (Fig. 2(b)) plotted for 8 stations, including the actuator interface near the region of densest plasma. It is observed that at the interface, the velocity is still relatively small, and the magnitude at this point is largely due to the vertical downward component. As we proceed to the right of the actuator, the velocity increases as we travel downstream until we reach a maximum of around 100 cm/s located 3 cm away from the interface. Essentially, these actuators, whether single or multiple phased arrays, affect primarily the boundary layer flow on a scale of 0(1 cm) as shown by the results above. To create more impact for flow control, the penetration of the jet would have to be increased, especially in higher speed situations. Further details on this actuator performance can be found elsewhere.61, 62 To enhance the impact of the actuator on the flowfield, a novel plasma actuator design consisting of an annular electrode array, the plasma synthetic jet actuator (PSJA, see Fig. 3(a)) was investigated. This particular configuration creates a plasma ring upon actuation (Fig. 3(b)), and the plasma-induced flow is in the form of a vertical “synthetic” jet. A direct advantage of the PSJA over conventional synthetic jets is that the actuator configuration can be easily reversed to act as suction devices (see Fig. 4 for the power actuator schematic for both blowing and suction configurations). Figure 5 shows some sample characteristics for 10 Hz pulsed operation of the PSJA in quiescent flow, where the diameters of exposed electrode (do ) and embedded electrode (di ) were 25.4 mm and 12.7 mm (with gap w= 0) respectively, and the results were obtained from ensemble-averaged PIV measurements made phase-locked to actuator pulsing frequency fp . The basic flowfield obtained at jet evolution time t=24 ms on pulsed operation is shown in Fig. 5(a), and a starting vortex ring is clearly seen near the actuator base. The vertical extent of the jet is at least 4.5 cm (limited by the size of the observation window), which is higher than the jet width obtained from the linear actuator. The maximum axial jet velocity at 10 Hz pulsing of this PSJA geometry is about 100 cm/s (Fig. 5(b)), and it is interesting to note that the axial velocity is not equal to zero even at the streamwise (coordinate normal to the actuator surface) extent of the FOV due to the additional interactive entrainment between the advecting vortex rings created in the flowfield. Thus, it can be seen that it is possible to increase both the peak velocity and penetration depth of the plasma-induced ZNMF jet by adopting the PSJA design, and thereby expand the applicability of plasma actuators in general. A detailed experimental investigation of the PSJA in quiescent flow can be found in the article by Santhanakrishnan and Jacob.63 It must be noted that the PSJA, akin to a circular synthetic jet, would introduce three-dimensional structures in the flow. In order to make this actuator easier to analyze in two-dimensional flow situations (such as laminar flow over a flat plate), the linear actuator can be modified to have two exposed electrodes and one embedded electrode (see Fig. 6). This design induces a vertical jet in the same way as the PSJA described above, but resembles a synthetic jet emanating from a rectangular orifice, which can be considered to be nominally two-dimensional.64 This particular design, the linear plasma synthetic jet actuator (hereon referred to as L-PSJA), is in effect the limiting case of an elliptical actuator when the major axis goes to infinity. To match the dimensions of the PSJA with the L-PSJA, the embedded electrode width must be equal to di , while the exposed electrode width must be equal to (do -di )/2 (without including any gap between the electrodes). This paper aims to provide a comparative study of the characteristics of the conventional linear plasma actuator, the PSJA and the L-PSJA in several flow situations, namely: quiescent flow, actuator mounted on a flat plate parallel to incoming crossflow (to remove effects of both pressure gradient and curvature), and separation control of the flow over a circular cylinder (to include effects of both pressure gradient and curvature).
4 of 30 American Institute of Aeronautics and Astronautics Paper 2006-3033
II. A.
Experimental Arrangements
Actuator construction and Power Supply
The plasma actuator electrodes tested consisted of conductive copper tape material, shaped in the form of annular arrays for the PSJA and as rectangular strips for both the L-PSJA and linear actuators. A 0.025 inch thick alumina ceramic slab was used as the dielectric for the quiescent flow and flat plate boundary layer experiments, while Kapton was used as the dielectric material for cylinder flow control measurements due to its flexibility. The diameters of exposed (do ) and embedded (di ) electrodes of the PSJA tested herein were 25.4 mm (1 inch) and 12.7 mm (0.5 inch) respectively, with either no air gap (w=0) or a 1-2 mm overlap. The dimensions of the L-PSJA used in both quiescent flow and flat plate experiments were chosen to match the PSJA, and the width of the exposed electrodes and embedded electrode of the L-PSJA were 6.35 mm (0.25 inch) and 12.7 mm (0.5 inch) respectively, with no air gap (w=0) and a spanwise length of around 90 mm. For quiescent flow measurements on the linear actuator, the width of the exposed and embedded electrodes were about 6 mm, with spanwise lengths of around 90 mm. The actuator arrangement (and coordinate system used for analysis) for quiescent flow and flat plate boundary layer modification experiments is shown in Fig. 7. For the cylinder flow measurements, the exposed and embeddded electrodes for both L-PSJA and linear actuator were 5 mm in width (spanwise length of 510 mm), as the dimensions of the cylinder used was not large enough to accommodate the L-PSJA dimensions that matched the PSJA. The results for quiescent flow and flat plate boundary layer experiments were obtained for a nominal 5 kV amplitude, 2.8 kHz frequency square wave AC input, operated with a 50% duty cycle (unless mentioned otherwise). A nominal 2 kV amplitude, 1.1 kHz frequency square wave AC input operated with a 50% duty cycle was used for the cylinder flow experiments. A function generator is used to provide a sinusoidal input to a Kepco BPM-01 100 V power supply. The output drives an non-inductively matched step-up transformer, whose output is sent to the electrodes. Two step-up transformers were used during the course of the experiments conducted herein, one capable of 6 kV output with 1-250 V RMS input at 1-7 kHz (Industrial Test Equipment Company, referred to as T-1 hereon) and the other capable of 5.5 kV output with 40 V RMS input at 6 kHz (Corona Magnetics, Inc., referred to as T-2 hereon). For measurements made with T-2, the exposed and embedded electrodes were driven with a 180◦ phase difference, while the embedded electrode was grounded whenever T-1 was used. The input power was monitored using a Tektronix TDS channel oscilloscope, while a high voltage probe and current monitor were used for voltage and current measurements, respectively, to determine input power. The PSJA and L-PSJA results presented herein are limited exclusively to the blowing configuration. B.
Particle Image Velocimetry
Measurements were made using both phase-locked and time-averaged PIV. The laser sheet for the PIV measurements was generated from a 50 mJ double-pulsed Nd:YAG laser with a maximum repetition rate of 15 Hz, and pulse separation was varied from 50-300 µs based on the freestream velocity. Wherever phaselocked PIV was conducted, the image pair realizations were acquired at the same rate as the actuator pulsing frequency, and measurements were made along several times of the actuator duty cycle, as shown in Fig. 8. A 10 bit Kodak Megaplus CCD camera with a 1008x1018 pixel array was used for capturing images. Uniform seeding was accomplished using 1 micron oil droplets inserted in the test chamber. For each PIV run, 122 images were recorded for processing resulting in a minimum of 61 vector and vorticity fields from which to generate mean flow field and statistics. The Wall Adaptive Lagrangian Parcel Tracking (WaLPT) DPIV algorithm was employed for processing the raw data. A detailed description of the algorithm and its formulation can be found elsewhere.65 In the WaLPT algorithm, seeding is tracked as fluid parcel markers and tracks both their translations and deformations. During this tracking, fluid particles registered by individual CCD pixels are advected with individually estimated velocities and total accelerations. The velocity field needed to initialize the WaLPT process is obtained from a standard DPIV algorithm which uses multiple passes, integer window shifting, and adjustable windows. Both the WaLPT and DPIV algorithms employ a rigorous peak-detection scheme to determine velocity vectors and use the local velocity gradient tensor to identify spurious velocity vectors. The WaLPT algorithm works well in the flow field of a vortex that is characterized by high deformation rates where traditional DPIV algorithms are plagued by biasing and limited dynamic range. No smoothing algorithms or other post-processing techniques are employed on the data. Vorticity, being a component of
5 of 30 American Institute of Aeronautics and Astronautics Paper 2006-3033
the velocity gradient tensor, is calculated spectrally at each grid point as an intrinsic part of the WaLPT algorithm. The raw images were processed as image pairs in 32x32 interrogation areas to give 61 tensors containing the flow information, i.e the velocities and velocity gradients. The sampling rate of the present PIV system was limited to a nominal value of 15 Hz. Post processing was performed in MATLAB.
III. A.
Results and Discussion
Actuator in Quiescent Flow
The quiescent flow measurements for the linear actuator and L-PSJA will be examined in detail in this section. Detailed experimental results of the PSJA in quiescent flow can be found elsewhere.63 The results were obtained from time-averaged PIV measurements on these actuators, and are specific to alumina ceramic being used as dielectric material. To provide a quiescent environment, the actuator inserts were placed over a layer of non-conductive material, such as acrylic or plexiglass, and covered with a transparent enclosure. A comparison of some of the quiescent flow characteristics of these actuators in steady operation will be presented in this section. Figure 9 shows vorticity contours overlaid with streamlines obtained from time averaged PIV measurements for steady operation of the linear actuator and L-PSJA in quiescent flow. The upstream influence of the linear actuator is moderate, and the flow at x=1.2 cm (location of the actuator interface) resembles a potential sink. The thickness or width of the wall jet induced downstream of the plasma is about 0.4 cm. In contrast, the flowfield induced by the L-PSJA resembles a steady jet (Fig. 9(b)), and the jet height is 3.5 cm (the streamwise extent of the entire jet is not shown, and is limited by the FOV in this case). The flowfield of the L-PSJA resembles that of the PSJA in steady operation. An examination of the jet characteristics reveals that the peak value of the mean axial (or vertical) jet velocity denoted by V obtained from the L-PSJA is about 45 cm/s (Fig. 10(b)), which is close to the equivalent PSJA maximum jet velocity value of 62 cm/s. The jet appears to reach self-similarity at distances as close as 0.5di downstream, as the profiles of mean axial velocity V (normalized with local maximum axial velocity Vmax ) collapse reasonable well for distances 0.5di and beyond (Fig. 10(a)). The entrainment of residual fluid by the plasma is therefore limited to streamwise distances less than 0.5di . The peak value of local maximum mean axial velocity (Vmax ) occurs close to the actuator surface at 0.5di , which is the same location for the PSJA in steady operation. However, the peak value of the L-PSJA linearly decreases with downstream distance, and streamwise jet spreading is still seen at the end of FOV (Vmax =15 cm/s at y=3di ). This is in contrast to the PSJA, where the local maximum mean axial velocity reaches a nearly constant value after 0.5di .63 The mean and RMS horizontal velocity profiles for both the linear actuator and L-PSJA are shown in Figs. 11-12, where the RMS profiles have been non-dimensionalized with the maximum mean horizontal velocity. For the linear actuator (Fig. 11(a)), the maximum ZNMF wall jet velocity is around 40 cm/s, which is less than the value of 90 cm/s obtained using a Kapton dielectric (see Fig. 2(b)), and occurs at y=0.1 cm from the surface. The total absolute value of the maximum mean horizontal velocity for the L-PSJA is around 40 cm/s (Fig. 11(b), around 25 cm/s at x=0 and 15 cm/s at x=-2∆), and the mean horizontal velocity is distributed in an almost symmetric manner on either sides of the jet centerline. It is interesting to note that starting from x=-2∆ (upstream), the horizontal velocity changes direction as we move downstream, i.e. the profiles oscillate about the x-axis (Fig. 11(b)). This is because of the entrainment of residual fluid by the plasma from either side of the jet centerline, and the width of this entrainment region is around 0.4 cm, which is almost the same as the width of wall jet produced by the linear actuator (Fig. 11(a)). For both the linear actuator and L-PSJA, large fluctuations are present in the jet. The maximum value of the horizontal RMS velocity for linear actuator (Fig. 12(a)), reaches 100% of maximum mean horizontal wall jet velocity (U ) value , while the L-PSJA maximum RMS velocity is around 85% (Fig. 12(b)). The y-axis locations of u0max follows the respective locations for the mean horizontal velocity. The maximum RMS velocity value for the linear actuator occurs downstream of the actuator interface, while it occurs at the jet centerline for the L-PSJA. The non-negligible turbulent fluctuations produced by the actuators will be examined in detail in the following section. B.
Flat Plate Boundary Layer Modification
To remove the effects of curvature and pressure gradient, a flat plate at zero inclination to the freestream was used to study the PSJA and L-PSJA characteristics. The flat plate was 12 inches in length, and the
6 of 30 American Institute of Aeronautics and Astronautics Paper 2006-3033
Table 1. Flat plate experiments.
Case A B C
U∞ [cm/s] 85 125 175
ReL 9 · 103 13 · 103 18 · 103
actuator was mounted inside a recess 0.025 inch deep cut at its center (Fig. 7). A 1.27 mm (0.05 inch) diameter wire was used to trip the incoming boundary layer, and was located at the leading edge of the flat plate. A low-turbulence open-circuit blow-down wind tunnel was used for PIV measurements. This is driven by a 7.5 hp motor and has a 0.2 m x 0.4 m test section with a length of 1.5 m. The results presented herein are specific to actuators using the alumina ceramic dielectric. Based on the quiescent flow measurements performed earlier on the PSJA,63 the actuator was operated at a pulsing frequency fp =10 Hz in addition to a baseline steady operation case (where the maximum PSJA jet velocity was 62 cm/s). The measurements were conducted for three different cross-stream velocities: 0.85 m/s (case A), 1.25 m/s (case B) and 1.75 m/s (case C), as denoted in Table 1. These values were chosen to investigate the interaction of the jet with crossflows that were lower, nearly same and higher in strength as the peak jet velocity at 10 Hz pulsing. The actuator center was located at a distance of 14.8 cm from the leading edge of the flat plate, the test Reynolds numbers based on this length ReL were approximately 9 · 103 , 13 · 103 and 18 · 103 . The basic flowfield will be examined first, followed by a discussion of derived flow characteristics. Figure 13 shows vorticity contours overlaid with streamlines from time-averaged PIV measurements for steady operation of the PSJA in crossflow. It can be seen that the behavior of the plasma synthetic (ZNMF) jet in a crossflow is similar to a conventional (non-ZNMF) jet. The suction of fluid by the plasma region just upstream of the actuator center is not predominant as it is opposed by the incoming flow. A clear separation streamline can be seen to start from near the actuator center for all the three velocities, and it divides the flowfield into a global region with the mean flow, and an inner region where the effects of the plasma-synthetic jet are observed. The amount of penetration, viz., the cross-stream distance to which the inner region prevails is affected by the incoming flow. On comparing Figs. 13(a)-(c), it can be seen that the penetration of the jet decreases with increasing crossflow velocity, as expected. While a clear two-region flow is observed for case A (Fig. 13(b)) and case B (Fig. 13(b)), the inner region is almost non-existent for case C ((Fig. 13(c))). The starting vortex ring has advected out of the FOV for steady operation of the actuator for all the three cases. Figures 14-16 show vorticity contours overlaid with streamlines obtained from phase-locked PIV measurements for 10 Hz pulsed operation of the PSJA in crossflow, and several times of evolution of the jet are shown. For case A (Fig. 14), the starting vortex is clearly seen within the FOV for all times except t=83 ms. It should be noted that the last two times (57 ms and 83 ms) examine the portion of the actuation cycle when the plasma is turned off. The separation streamline originates at 14.6 cm from the leading edge of the flat plate for times from 18 ms to 24 ms, after which the location shifts downstream to 14.7 cm for times from 36 ms to 57 ms. The penetration of the jet is around 0.5 cm for t=18 ms, then increases and stays constant at 0.6 cm up to t=36 ms, after which it peaks to around 0.8 cm for t=48 ms toward the end of the actuation cycle. At t=57 ms, where the plasma is turned off, the penetration of the jet reduces to 0.5 cm, and the flow returns to baseline freestream flow at t=83 ms. The starting vortex advects with the flow 18 ms to 36 ms, after which it loses structure due to dissipation in the global flow region. A new starting vortex incarnation starts to form near the actuator base at 24 ms, and the resulting interactive entrainment with the earlier vortex can be attributed to the increase in penetration depth from 24 ms to 57 ms. It is interesting to note that at t=18 ms, the left hand side of the starting vortex ring (that is nearest facing to the incoming flow) is almost completely weakened by the freestream, while the right hand side of the ring causes the fluid rotation in the inner fluid region. In this regard, a visual comparison of the inner region flow for times 18 ms and 48 ms shows that while the inner region streamlines are parallel to the freestream for 18 ms (due to negative vorticity added by the right hand side of the ring), the inner region streamlines for 48 ms are inclined to the freestream. This suggests that the strength of the right hand side of the vortex ring decreases
7 of 30 American Institute of Aeronautics and Astronautics Paper 2006-3033
with increasing time. Furthermore, the area of the inner region (with separated flow) increases with time from 18 ms to 48 ms. Since the side of the vortex ring with rotational sense opposing the freestream causes the fluid to separate, the strength of the left hand side vortex would have to increase with increasing time. This asymmetry in the vortex ring strength is believed to be important to effecting jet penetration with the mean flow. For case B (Fig. 15), the starting vortex ring is seen at all times when the actuator is on (18 ms to 48 ms). At t=18 ms, the jet penetration is around 0.4 cm, after which it increases to a maximum value of around 0.5 cm at t=24 ms, and then decreases and stays nearly constant at 0.4 cm up to t=48 ms. The separation streamline at 1 m/s starts just downstream of the actuator center location (x=15 cm), compared to x=14.8 cm for case A. In general, the reverse flow region is lesser in extent for the 1 m/s case. For case C (Fig. 16), the jet penetration is around 0.2 cm, and it stays constant for all times when the actuator is on. Also, the separation streamline has moved downstream to 15 cm from the 1 m/s location, as expected. The reverse flow region is non-existent at this particular velocity. The effectiveness of the actuator in affecting a crossflow is limited by the value of the penetration depth of the jet, and it was found to be less than 1 cm for all the cases tested herein. It has been observed that varying the input power might affect this value to some extent, and will be investigated in future experiments. Figure 17 shows vorticity contours overlaid with streamlines obtained from time-averaged PIV measurements for steady operation of L-PSJA in crossflow, and the three different freestream cases are compared. A stationary cross-stream vortex centered at around x=15.7 cm is seen at all three velocities. The penetration of the plasma synthetic jet is effected through this standing vortex, and it is about 0.3 cm for case A, and about 0.2 cm for cases B and C. Note that any such cross-stream vortical structures were absent in the case of steady operation of the PSJA (Fig. 13). The suction effect of the upstream facing plasma region in the L-PSJA is seen to start at around x=14.5 cm for all three velocities. The center of the vortex is located at the outer edge of exposed electrode (edge farthest to the freestream and away from the plasma region), and it is created as a result of near-wall fluid entrainment by the plasma region. Figures 18-20 show vorticity contours overlaid with streamlines obtained from phase-locked PIV measurements for 10 Hz pulsed operation of the L-PSJA, and three times of jet evolution corresponding approximately to starting, middle and “plasma-off” phases of the actuation cycle (respectively, in the order of increasing time) are shown. For case A, at t=18 ms (Fig. 18(a)), fluid is sucked in upstream of the actuator center at x=14.6 cm, and a cross-stream vortex centered at x=15.7 cm is seen. With increase in time to 36 ms, the size of the vortex is increased, indicating additional entrainment by the plasma. The location of the vortex remains nearly stationary, however. At t=57 ms (portion of cycle where the plasma is turned off, Fig. 18(c)), the size of the vortex increases to effect a jet penetration of around y=0.8 cm, and the vortex location has shifted downstream and upward. This is higher than the equivalent case of the PSJA (Fig. 14(e)), suggesting the mechanism of the L-PSJA in affaecting a flowfield, on account of its two-dimensionality - is not similar to the PSJA. For L-PSJA, case B (Fig. 19), the sequence of events are the same as in case A, albeit with a reduction in the cross-stream vortex size (and therefore, reduced jet penetration), as expected. It is interesting to note that for all velocities, the L-PSJA induced cross-stream vortex is “trapped” for both jet evolution times (18 ms and 36 ms, when the plasma is on) observed herein. A similar near-wall “trapped” vortex ring was observed in the case of pulsed operation of the PSJA.63 This was a secondary vortical structure, however, as opposed to the only primary vortex seen in the L-PSJA. For L-PSJA, case C (Fig. 20), the effect of the plasma is diminished, but the standing vortex is still seen during both times when the plasma is on. In general, at the end of the actuation cycle, the standing vortex is released into the freestream, and advects downstream of the actuator and in the positive cross-stream direction. A similar cross-stream vortex structure has been observed on the pulsed operation of a linear actuator.61 It is clear that the near-wall fluid entrainment by the plasma is singularly responsible for both creating and sustaining the standing vortex. In the case of a linear actuator, there is a single plasma region spanwise that creates this downstream facing cross-stream vortex. The L-PSJA produces two such spanwise plasma regions on actuation (see Fig. 6). Based on symmetry arguments, it was expected that the plasma region of the L-PSJA that is located closest to the incoming flow would have to create another cross-stream vortex, opposite in sign to the observed “trapped” vortex. Such a structure was not observed in the experiments, however, and it appears that the incoming flow completely eradicates any such upstream facing rotation. Hence, it is conjectured that the effect of the upstream facing plasma region would be in adding a finite circulation to the vortex. A comparison of the circulation strengths of the cross-stream vortex created by the linear and L-PSJA actuators would have to be examined to see if the upstream facing plasma region has any effect on
8 of 30 American Institute of Aeronautics and Astronautics Paper 2006-3033
the flowfield at all. (Similar behavior has been studied for linear plasma actuators.66 ) Figures 22-23 show the velocity profiles of mean and RMS components of streamwise velocity respectively, and the individual profile curves are separated by a distance of ∆=3.3 mm. Both the PSJA and L-PSJA characteristics in steady operation are presented in each of these figures. The baseline flow characteristics is shown in Figure 21 for comparison. The PSJA profiles will be examined first, and then compared to those of the L-PSJA. For PSJA case A, the flow separates at the actuator center due to the jet interaction, and starts to recover at 2∆ downstream (Fig. 22(a)). The influence of the jet on the crossflow for case A extends to a distance of y=0.5 cm, and the increase in mean velocity is very near the wall (at y=0.1 cm for x=4∆ downstream, Fig. 22(a)). With increase in freestream velocity to case B (PSJA), the flow separates immediately downstream of the actuator center, and the reverse flow region is smaller compared to case A (Fig. 22(c)). The flow does not recover to the baseline profile, however, even at 4∆ downstream. The effect of the jet is weakened for case C of the PSJA (Fig. 22(e)), and there is no reverse flow. It is interesting to note that for all these cases, the profiles with steady actuation are shifted upward in the cross-stream direction, at distances downstream of the actuator center indicating the jet penetration. However, the peak in the velocity stays at nearly constant cross-stream location even as we move downstream, in contrast to what is observed in the linear actuator.62 Also, at distances upstream of the actuator center, the profiles are only slightly deviated, which is also observed in the case of the linear actuator. 62 For L-PSJA case A (Fig. 22(b)), the flow separates at distance 2∆ downstream of the actuator center (compared to r=0 for PSJA, Fig. 22(a)), indicating that although the PSJA and L-PSJA are designed along similar lines, their interaction with the flowfield is different. The peak in velocity prior to separation occurs at 1∆ downstream of the actuator center, and the jet cross-stream height at that point is about 0.1 cm. However, the overall peak velocity occurs post-separation at 3∆ downstream, and the jet height has increased to about 0.4 cm, which is almost the same as the equivalent case of PSJA (Fig. 22(a)). It is interesting to note that for case A, the peak in velocity reaches the freestream value for L-PSJA (Fig. 22(b)), as compared to about 70% of freestream value for PSJA (Fig. 22(a)). For L-PSJA case B (Fig. 22(d)), the peak velocity value is lowered to 90% of the freestream value at a distance of y=0.4 cm and x=3∆ downstream due to the increased freestream, still higher than the equivalent case of the PSJA (Fig. 22(c)). Note that the profiles for L-PSJA do not recover to the baseline profile (Figs. 21(a) and (c)). The peak velocity for case C of L-PSJA (Fig. 22(f)) is around 80% of freestream, and there is a reverse flow region at 3∆ downstream due to the presence of the standing cross-stream vortex, compared to the fully attached equivalent case of the PSJA (Fig. 22(e)). The profiles recover to the case C baseline profile (Fig. 21(e)) at 4∆ downstream. The separation location remains constant (for all velocities) at 3∆ downstream, which is also the streamwise location of the L-PSJA peak velocity, where there is a nearly stationary cross-stream vortex (Fig. 17). At the lowest freestream velocity for PSJA, the value of fluctuating velocity is increased within the viscous layer (Fig. 23(a)) to about 50% of its maximum value, after which it decays to the baseline case at distances y=0.7 cm and upward. For the PSJA, the peak fluctuating value tends to be nearly at the same crossstream distance at different downstream distances. The increase in RMS velocity from the baseline profile, though finite, is lowered with increase in freestream velocity (Figs. 23(c) & (e)), as expected. The increase in RMS velocity of the L-PSJA reaches 100% of its maximum value at all velocities (Figs. 23(b), (d) and (f)) at the 3∆ downstream location of the standing vortex. At 4∆ downstream, the L-PSJA RMS velocity approaches the baseline profile for all three velocities (Figs. 21(b), (d) and (f)). In general, the growth of the L-PSJA RMS velocity seems to proceed outward in the flow (viz., in the positive cross-stream and streamwise directions), which is the direction of motion of the cross-stream vortex in phase-locked observations discussed earlier (Figs. 18-20). The cross-stream location of the maximum RMS velocity of the L-PSJA (at x=3∆ downstream in Figs. 21(b), (d) and (f)) is lowered with increase in freestream velocity, indicating reduced jet penetration. For both actuators, the peak in the fluctuating velocities occur at the same cross-stream distances as the peak in their respective mean velocities. Figures 24-25 show the displacement (δ ∗ ) and momentum (θ) thicknesses calculated using their standard integral definitions.67 For every case considered, the values of δ ∗ and θ have been normalized to the initial values determined from the first profile station. For steady operation of the PSJA in case A, the displacement thickness (Fig. 24(a)) for all three velocities increases from the start of the plasma region (around x=-1∆, due to the adverse pressure gradient introduced by the plasma synthetic jet) and peaks at the center of the actuator. The displacement thickness values decrease after the center and reach a constant level. The increase in displacement thickness is about 40% (case A), 30% (case B) and 20% (case C) higher than their respective initial values. The PSJA adds additional energy to the flowfield, and hence the momentum
9 of 30 American Institute of Aeronautics and Astronautics Paper 2006-3033
thickness values are lowered in the vicinity of the plasma for all three speeds. At the center of the actuator, an impressive 40% reduction in momentum thickness is observed for case A (Fig. 24(a)), while it is lowered to about 10% for case B. The actuator seems to have negligible effect on the momentum thickness for case C. Downstream of the actuator center, the momentum thickness values increase, similar to what is observed in a linear actuator.62 For cases B and C, the momentum thickness returns to its initial value at downstream station x=3∆. For case A, a 10% reduction in momentum thickness is still seen at the farthest downstream station. For steady operation of the L-PSJA (Fig. 24(b)) at all velocities, the presence of the “trapped” crossstream vortex results in an increase of both displacement and momentum thicknesses from downstream of the start of the plasma region (at x=-1∆). The suction of the plasma is seen for cases A and B upstream of x=1∆, resulting in a reduction of momentum thickness values decrease up to this distance. The displacement thickness for cases A and C, however, do not change from the initial value up to a distance of x=-1∆. Note that θ is reduced by 30% for case A, compared to about 10% for case B, while case C is not affected by the actuator prior to x=-1∆. For all three velocities, both δ ∗ and θ do not return to their respective intial values. Pulsing the PSJA at 10 Hz (Fig. 25(a)) results in about 60% momentum thickness reduction for case A, 20% higher than steady operation of the PSJA for the same freestream velocity (Fig. 24(a)). For 10 Hz pulsed operation of the L-PSJA (Fig. 25(b)), both displacement and momentum thickness values are lowered, although the trend in which they vary is exact similar to steady operation of the same actuator (Fig. 24(b)). The action of the PSJA is thus similar to an active boundary layer trip, and when compared to a linear actuator,62 the PSJA produces a larger reduction of the momentum thickness, albeit with a larger increase in displacement thickness. The L-PSJA also acts as an active boundary layer trip, but the presence of the “trapped” vortex results in it affecting longer distances downstream. While the momentum addition of the PSJA is directed toward the global flowfield, the momentum addition of the L-PSJA is directed nearer the wall. The effectiveness of this tripping action for both PSJA and L-PSJA is seen to decrease with increase in Re. The crossflow interaction of the L-PSJA is observed to be different from the PSJA, and is actually in close resemblance to that of a linear actuator.While the PSJA, on account of its three-dimensionality, would have to create a vortex ring on actuation, the L-PSJA would have to create two-dimensional “vortex tubes” (see Fig. 6). In the case of L-PSJA, the vortex with rotational sense opposing upstream is weakened completely by the flow. However, the strength of the other vortex (facing downstream) is increased by the crossflow, which is similar to the standing vortex observed on a linear actuator.61 The circulation strength of the upstream facing vortex (that is weakened by the flow) will have to be added to the downstream facing standing vortex to keep the rate of change of circulation constant, in accordance with Kelvin’s theorem. This is in direct contrast to the mechanism of the PSJA, where the side of the vortex ring facing upstream is strengthened with time as discussed earlier (observations on Fig. 14). Thus, the effect of the PSJA would be perceived more on the global freestream, as opposed to the near-wall effects of both the L-PSJA and linear actuator. C.
Flow Over a Circular Cylinder
To study the applicability of the plasma and plasma synthetic jet actuators for separation control, the laminar flow over a circular cylinder was chosen as a test problem. The effects of both pressure gradient and curvature are included in this case, and it is well known from theory that laminar separation occurs at around 82◦ to the freestream for diameter based Reynolds numbers less than 30 · 103 .67 The diameter (D) and length (L) of the cylinder used for experiments was 38.1 mm (1.5 inches) 203 mm (8 inches) respectively, and it was transparent for clean optical access. Two different freestream velocties were investigated, and the diameter based Reynolds numbers (ReD ) were 5·103 (U∞ =2 m/s) and 10·103 (U∞ =4 m/s). The experiments were conducted in the low-turbulence open-circuit blow-down wind tunnel described in the previous section. The results presented hereon were obtained from time averaged PIV measurements, and four windows of observation were used to obtain the flowfield information from the center of the cylinder x=0 to x=11D downstream in the wake. The linear actuator and L-PSJA configurations were tested, with Kapton used as the dielectric material. As the flow was treated to be nominally two-dimensional, the PSJA configuration was not tested. The cylinder was covered with a layer of Kapton, on which the actuators were constructed. Two actuators were mounted downstream of the laminar separation point, at 90◦ and 270◦ to the freestream for symmetric forcing of the cylinder wake. A schematic of the cylinder, actuator arrangement and power supply is shown in Fig. 26. The transformer T2 was used to power the actuators for this series of experiments, 10 of 30 American Institute of Aeronautics and Astronautics Paper 2006-3033
and the input voltage supplied to the exposed and embedded electrodes of the actuators were driven with 180◦ phase difference. In addition to steady actuator operation, pulsing the actuator at non-dimensional pulsing frequency f + equal to Strouhal number St (non-dimensional wake shedding frequency) value of 0.2 was considered, where f + is given by f + = fp D/U∞
(1)
where fp is the pulsing frequency of the actuator. Figures 27-28 show vorticity contours overlaid with streamlines from time averaged PIV measurements for both the linear actuator and L-PSJA. In each of these figures, the baseline flow with actuator off is shown for comparison. For ReD =5 · 103 , the reverse flow region in the wake extends to around 2D downstream (Fig. 27(a)), and a set of counter-rotating vortices are seen. The reverse flow region is reduced to about 1D downstream on steady operation of the linear actuator at this ReD (Fig. 27(b)). The wall jet induced by the plasma region of the two linear actuators force the separation streamlines to move closer to the surface, resulting in tightening of the wake. Note that the counter-rotating wake vortices actually reverse their sense of rotation on actuation, compared to the baseline case. The reverse flow region is reduced on steady operation of the L-PSJA (Fig. 28(b)), but does not have as much impact as the linear actuator. The wake width is not changed on steady operation of L-PSJA at ReD =5 · 103 , and the wake vortices do not change their rotational sense. At this Re, pulsing at f + =St does not seem to promote any major improvement from steady actuation in the flow for both the actuators (Figs. 27(c) and 28(c)), other than changing the rotational sense of the wake vortices in case of the linear actuator. Though not shown, the reverse flow region for the baseline case at ReD =10 · 103 extends to about 10 cm downstream. Steady operation of the linear actuator at this Re results in marginal reduction of the reverse flow region and changing the rotation sense of the wake vortices, and pulsing at f + =0.2 does not produce any major changes (compared to steady operation) in the flowfield. At this higher speed, both steady and pulsed operation of the L-PSJA reduces the reverse flow region, if only slightly. Figures 29 and 30 show the cross-stream variation of velocity deficits in the cylinder wake for ReD 5 · 103 and 10 · 103 respectively, and the baseline defect is compared to that of steady (f + =0) and pulsed (f + =0.2) actuator operation. For each ReD , the deficits of the linear actuator and L-PSJA are compared for two streamwise stations at 1 diameter and 10 diameters downstream. For ReD of 5 · 103 , operation of the linear actuator (in both steady and pulsed modes) actually tightens the near wake, as seen from the narrow velocity defects (reduced cross-stream width) in Fig. 29(a), although the actual defect value is itself increased slightly from baseline value. Operation of the L-PSJA (both steady and pulsed modes) does not alter the velocity defect from the baseline curve (Fig. 29(b)). At downstream distance of x=10D, the pulsed operation of the linear actuator reduces the velocity defect marginally from the baseline case (Fig. 29(c)), while steady operation of the L-PSJA reduces the velocity defect from the baseline case (Fig. 29(d)). For ReD of 10 · 103 , the linear actuator still promotes tightening of the near wake at x=1D (Fig. 30(a)), while the L-PSJA does not produce any major changes in the near wake (Fig. 30(b)). At this Re, both steady and pulsed operation of the linear actuator reduce the deficit at 10 diameters downstream (Fig. 30(c)), with pulsed operation producing the lowest velocity deficit. The effectiveness of the L-PSJA at ReD =10 · 103 is lowered so that the actuator operation does not produce any significant changes in the velocity deficits at x=10D downstream (Fig. 30(d)). In general, the effectiveness of both these actuators for separation control decrease with increase in Re. Overall, the linear actuator seems to be more promising in cylinder flow separation control than the L-PSJA. The difference in the observed flow is because of the fundamental difference in the induced flowfield of the linear actuator and L-PSJA. While a linear actuator in this arrangement actually accelerates the flow by adding momentum in the downstream direction in the near wall region, the L-PSJA adds momentum by penetrating the mean flow.
IV.
Summary
This paper presented a comparative study of three different dielectric barrier discharge plasma actuator geometries that tailor the flow in the form of horizontal wall jets or vertical synthetic jets. Each actuator was observed to affect the flowfield in a different manner. The annular plasma synthetic jet actuator (PSJA) was observed to interact with a crossflow in a similar manner as a conventional (non-ZNMF) jet. The interaction of the linear plasma synthetic jet actuator (L-PSJA) with a crossflow resulted in the formation of a cross11 of 30 American Institute of Aeronautics and Astronautics Paper 2006-3033
stream vortex downstream of the plasma region, similar to the linear actuator. The PSJA was observed to affect the mean flowfield, as opposed to the L-PSJA and linear actuators that affect the near-wall flow. The effectiveness of the plasma synthetic jet in penetrating the mean flow was found to decrease with increase in Re. In the case of flow over a circular cylinder, the linear actuator geometry was more effective in reducing the wake deficits than the L-PSJA.
Acknowledgments This work was supported in part by a grant from Kentucky NASA EPSCoR under the direction of Drs. Richard and Karen Hackney.
References 1 Gad-El-Hak,
M., “Flow Control: The Future,” Journal of Aircraft, Vol. 38, 2001, pp. 402–418. C. A., Selig, M. S., and Broeren, A. P., “Boundary layer trips on airfoils at low Reynolds numbers,” AIAA Paper 1997–0511, 35th AIAA Aerospace Sciences Meeting, Reno, NV, Jan. 1997. 3 Kerho, M. F. and Bragg, M. B., “Airfoil Boundary Layer Development and Transition with Large Leading-Edge Roughness,” AIAA Journal, Vol. 35, No. 1, 1997, pp. 75–84. 4 McQuilling, M., Hollon, B., and Jacob, J. D., “Active Separation Flow Control in a Low Pressure Turbine Blade Cascade Model,” AIAA Paper 2003–0615, 41st Aerospace Sciences Meeting and Exhibit, Reno, NV, Jan. 2003. 5 Miklosovic, D., Murray, M., Howle, L., and Fish, F., “Leading-Edge tubercules delay stall on humpback whale (Megaptera novaeangliae) flippers,” Phys. Fluids, Vol. 16, No. 5, May 2004, pp. 39–42. 6 Santhanakrishnan, A. and Jacob, J. D., “Effect of Regular Surface Perturbations on Flow Over an Airfoil,” AIAA Paper 2005–5145, 35th AIAA Fluid Dynamics Conference and Exhibit, Toronto, Ontario, June 2005. 7 Zaman, K. B. M. Q., Bar-Sever, A., and Mangalam, S. M., “Effect of Acoustic Excitation on the Flow Over a Low Re Airfoil,” J. Fluid Mech., Vol. 182, 1987, pp. 127–148. 8 Seifert, A., Bachar, T., Koss, D., Shepshelovich, M., and Wygnanski, I., “Oscillatory Blowing: A Tool to Delay BoundaryLayer Separation,” AIAA Journal, Vol. 34, 1993, pp. 2052–2060. 9 Seifert, A., Darabi, A., and Wygnanski, I., “Delay of Airfoil Stall by Periodic Excitation,” Journal of Aircraft, Vol. 33, 1996, pp. 691–698. 10 Amitay, M., Smith, D. R., Kibens, V., Parekh, D. E., and Glezer, A., “Aerodynamic Flow Control Over an Unconventional Airfoil Using Synthetic Jet Actuators,” AIAA Journal, Vol. 39, No. 3, 2001, pp. 361–370. 11 Modi, V. J., “On the Moving Surface Bounday-Layer Control,” AIAA Paper 2000–2238, Fluids 2000 Conference and Exhibit, Denver, CO, June 2000. 12 Lee, T. and Gerontakos, P., “Investigation of Flow Over an Oscillating Airfoil,” J. Fluid Mech., Vol. 512, 2004, pp. 313– 341. 13 Pern, N. J., Jacob, J. D., and LeBeau, R. P., “Characterization of Zero Mass Flux Flow Control for Separation Control of an Adaptive Airfoil,” AIAA Paper 2006–3032, to be presented at the 3rd AIAA Flow Control Conference, San Francisco, CA, June 2006. 14 Gad-El-Hak, M., “Control of Low-Speed Airfoil Aerodynamics,” AIAA Journal, Vol. 28, 1990, pp. 1537–1552. 15 Gad-El-Hak, M., Flow Control: Passive, Active, and Reactive Flow Management, Cambridge University Press, New York, 2000. 16 Stanewsky, E., “Adaptive Wing and Flow Control Technology,” Progress in Aerospace Sciences, Vol. 37, 2001, pp. 583– 667. 17 Roth, J. R., Sherman, D. M., and Wilkinson, S. P., “Boundary Layer Flow Control With A One Atmosphere Uniform Glow Discharge Plasma,” AIAA Paper 98–0328, 36th AIAA Aerospace Sciences Meeting and Exhibit, Reno, NV, Jan. 1998. 18 List, J., Byerley, A., McLaughlin, T., and Dyken, R. V., “Using a Plasma Actuator to Control Laminar Separation on a Linear Cascade Turbine Blade,” AIAA Paper 2003–1026, 41st AIAA Aerospace Sciences Meeting and Exhibit, Reno, NV, Jan. 2003. 19 Jacob, J. D., Rivir, R., Carter, C., and Estevadeordal, J., “Boundary Layer Flow Control Using AC Discharge Plasma Actuators,” AIAA Paper 2004–2128, 2nd AIAA Flow Control Conference, Portland, OR, June 2004. 20 Amitay, M., Smith, B. L., and Glezer, A., “Aerodynamic Flow Control Using Synthetic Jet Technology,” AIAA Paper 98–0208, 36th AIAA Aerospace Sciences Meeting and Exhibit, Reno, NV, Jan. 1998. 21 Rediniotis, O. . K., Ko, J., X.Yue, and Kurdilla, A., “Synthetic Jets, their Reduced Order Modeling and Applications to Flow Control,” AIAA Paper 99–1000, 37th AIAA Aerospace Sciences Meeting and Exhibit, Reno, NV, Jan. 1999. 22 Mittal, R. and Rampunggoon, P., “On the Virtual Aeroshaping Effect of Synthetic Jets (Brief Communication),” Phys. Fluids, Vol. 14, No. 4, 2002, pp. 1533–1536. 23 Chiekh, M. B., Bera, J. C., and Sunyach, M., “Synthetic Jet Control for Flows in a Diffuser: Vectoring, Spreading and Mixing Enhancement,” Journal of Turbulence, Vol. 32, 2003, pp. 1–12. 24 Glezer, A. and Amitay, M., “Synthetic Jets,” Annual Rev. of Fluid Mech., Vol. 34, 2002, pp. 503–529. 25 Rivir, R., White, A., Carter, C., Ganguly, B., Forelines, A., and Crafton, J., “Turbine Flow Control, Plasma Flows,” AIAA Paper 2003–6055, 41st AIAA Aerospace Sciences Meeting and Exhibit, Reno, NV, Jan. 2003. 2 Lyon,
12 of 30 American Institute of Aeronautics and Astronautics Paper 2006-3033
26 Suchomel, C., Wie, D. V., and Risha, D., “Perspectives on Cataloguing Plasma Technologies Applied to Aeronautical Sciences,” AIAA Paper 2003–3852, 34th AIAA Plasmadynamics and Lasers Conference, Orlando, FL, June 2003. 27 Corke, T. C. and Post, M. L., “Overview of Plasma Flow Control: Concepts, Optimization, and Applications,” AIAA Paper 2005–563, 43rd AIAA Aerospace Sciences Meeting and Exhibit, Reno, NV, Jan. 2005. 28 Malik, M., Weinstein, L., and Hussani, M., “Ion Wind Drag Reduction,” AIAA Paper 83–0231, 21st AIAA Aerospace Sciences Meeting and Exhibit, Reno, NV, Jan. 1983. 29 El-Khabiry, S. and Colver, G., “Drag Reduction by DC Corona Discharge Along an Electrically Conductive Flate Plate for Small Reynolds Number Flow,” Phys. Fluids, Vol. 9, No. 3, May 1997, pp. 587–599. 30 Roth, J. R., Sherman, D. M., and Wilkinson, S. P., “Electrodynamic Flow Control with a Glow-Discharge Surface Plasma,” AIAA Journal, Vol. 38, No. 7, 2000, pp. 1166–1172. 31 Corke, T. C., Jumper, E. J., Post, M., Orlov, D., and McLaughlin, T. E., “Application of Weakly-Ionized Plasmas as Wing Flow-Control Devices,” AIAA Paper 2002–0350, 40th AIAA Aerospace Sciences Meeting and Exhibit, Reno, NV, Jan. 2002. 32 Corke, T. C., Mertz, B., and Patel, M. P., “Plasma Flow Control Optimized Airfoil,” AIAA Paper 2006–1208, 44th AIAA Aerospace Sciences Meeting and Exhibit, Reno, NV, Jan. 2006. 33 Corke, T. C., He, C., and Patel, M. P., “Plasma Flaps and Slats: An Application of Weakly Ionized Plasma Actuators,” AIAA Paper 2004–2127, 2nd AIAA Flow Control Conference, Portland, OR, June 2004. 34 Hultgren, L. and Ashpis, D., “Demonstration of Separation Delay with Glow-Discharge Plasma Actuators,” AIAA Paper 2003–1025, 41st AIAA Aerospace Sciences Meeting and Exhibit, Reno, NV, Jan. 2003. 35 Asghar, A. and Jumper, E. J., “Phase Syncrhonization of Vortex Shedding from Multiple Cylinders Using Plasma Actuators,” AIAA Paper 2003–1028, 41st AIAA Aerospace Sciences Meeting and Exhibit, Reno, NV, Jan. 2003. 36 McLaughlin, T. E., Munska, M. D., Vaeth, J. P., Dauwalter, T. E., Goode, J. R., and Siegal, S. G., “Plasma-Based Actuators for Cylinder Wake Vortex Control,” AIAA Paper 2004–2129, 2nd AIAA Flow Control Conference, Portland, OR, June 2004. 37 Siegenthaler, J., Jumper, E. J., and Asghar, A., “A Preliminary Study in Regularizing the Coherent Structures in a Planar, Weakly Compressible, Free Shear Layer,” AIAA Paper 2003–0680, 41st AIAA Aerospace Sciences Meeting and Exhibit, Reno, NV, Jan. 2003. 38 Thomas, F. O., Kozlov, A., and Corke, T. C., “Plasma Actuators for Landing Gear Noise Reduction,” AIAA Paper 2005–3010, 11th AIAA/CEAS Aeroacoustics Conference, Monterey, CA, May 2005. 39 Samimy, M., Adamovich, I., Kim, J., Webb, B., Keshav, S., and Utkin, Y., “Active Control of High Speed Jets Using Localized Arc Filament Plasma Actuators,” AIAA Paper 2004–2130, 2nd AIAA Flow Control Conference, Portland, OR, June 2004. 40 Enloe, C. L., McLaughlin, T. E., VanDyken, R. D., Kachner, K. D., Jumper, E. J., and Corke, T. C., “Mechanisms and Responses of a Single Dielectric Barrier Discharge Plasma Actuator: Plasma Morphology,” AIAA Journal, Vol. 42, No. 3, 2004, pp. 589–594. 41 Enloe, C. L., McLaughlin, T. E., VanDyken, R. D., Kachner, K. D., Jumper, E. J., Corke, T. C., Post, M., and Haddad, O., “Mechanisms and Responses of a Single Dielectric Barrier Discharge Plasma Actuator: Geometric Effects,” AIAA Journal, Vol. 42, No. 3, 2004, pp. 595–604. 42 Baughn, J. W., Porter, C. O., Peterson, B. L., McLaughlin, T. E., Enloe, C. L., Font, G. I., and Baird, C., “Momentum Transfer for an Aerodynamic Plasma Actuator with an Imposed Boundary Layer,” AIAA Paper 2006–168, 44th AIAA Aerospace Sciences Meeting and Exhibit, Reno, NV, Jan. 2006. 43 Porter, C. O., Baughn, J. W., McLaughlin, T. E., Enloe, C. L., and Font, G. I., “Temporal Force Measurements on an Aerodynamic Plasma Actuator,” AIAA Paper 2006–104, 44th AIAA Aerospace Sciences Meeting and Exhibit, Reno, NV, Jan. 2006. 44 Balcer, B. E., Franke, M. E., and Rivir, R. B., “Effects of Plasma Induced Velocity on Boundary Layer Flow,” AIAA Paper 2006–875, 44th AIAA Aerospace Sciences Meeting and Exhibit, Reno, NV, Jan. 2006. 45 Roth, J. R. and Dai, X., “Optimization of the Aerodynamic Plasma Actuator as an Electrohydrodynamic (EHD) Electrical Device,” AIAA Paper 2006–1203, 44th AIAA Aerospace Sciences Meeting and Exhibit, Reno, NV, Jan. 2006. 46 VanDyken, R. D., McLaughlin, T. E., and Enloe, C. L., “Parametric Investigations of a Single Dielectric Barrier Plasma Actuator,” AIAA Paper 2004–846, 42nd AIAA Aerospace Sciences Meeting and Exhibit, Reno, NV, Jan. 2004. 47 Enloe, C. L., McLaughlin, T. E., VanDyken, R. D., and Fischer, J. C., “Plasma Structure in the Aerodynamic Plasma Actuator,” AIAA Paper 2004–844, 42nd AIAA Aerospace Sciences Meeting and Exhibit, Reno, NV, Jan. 2004. 48 Anderson, R. and Roy, S., “Preliminary Experiments of Barrier Discharge Plasma Actuators using Dry and Humid Air,” AIAA Paper 2006–369, 44th AIAA Aerospace Sciences Meeting and Exhibit, Reno, NV, Jan. 2006. 49 Shyy, W., Jayaraman, B., and Andersson, A., “Modeling of Glow Discharge-Induced Fluid Dynamics,” J. App. Phys., Vol. 92, No. 11, Dec. 2002, pp. 6434–6443. 50 Visbal, M. and Gaitonde, D. V., “Control of Vortical Flows Using Simulated Plasma Actuators,” AIAA Paper 2006–505, 44th AIAA Aerospace Sciences Meeting and Exhibit, Reno, NV, Jan. 2006. 51 Suzen, Y. B., Huang, P. G., Jacob, J. D., and Ashpis, D. E., “Numerical Simulation of Plasma Based Flow Control Applications,” AIAA Paper 2005–4633, 35th AIAA Fluid Dynamics Conference and Exhibit, Toronto, Ontario, June 2005. 52 Suzen, Y. B. and Huang, P. G., “Simulations of Flow Separation Control Using Plasma Actuators,” AIAA Paper 2006– 877, 44th AIAA Aerospace Sciences Meeting and Exhibit, Reno, NV, Jan. 2006. 53 Hall, K. D., Jumper, E. J., Corke, T. C., and McLaughlin, T. E., “Potential Flow Model of a Plasma Actuator as a Lift Enhancement Device,” AIAA Paper 2005–783, 43rd AIAA Aerospace Sciences Meeting and Exhibit, Reno, NV, Jan. 2005. 54 Orlov, D. M. and Corke, T. C., “Numerical Simulation of Aerodynamic Plasma Actuator Effects,” AIAA Paper 2005– 1083, 43rd AIAA Aerospace Sciences Meeting and Exhibit, Reno, NV, Jan. 2005.
13 of 30 American Institute of Aeronautics and Astronautics Paper 2006-3033
55 Orlov, D. M., Corke, T. C., and Patel, M. P., “Electric Circuit Model for Aerodynamic Plasma Actuator,” AIAA Paper 2006–1206, 44th AIAA Aerospace Sciences Meeting and Exhibit, Reno, NV, Jan. 2006. 56 Font, G. I., “Boundary Layer Control with Atmospheric Plasma Discharges,” AIAA Paper 2004–3574, 40th AIAA/ASME/SAE/ASEE Joint Propulsion Conference and Exhibit, Fort Lauderdale, FL, July 2004. 57 Font, G. I. and Morgan, W. L., “Plasma Discharges in Atmospheric Pressure Oxygen for Boundary Layer Separation Control,” AIAA Paper 2005–4632, 35th Fluid Dynamics Conference and Exhibit, Toronto, Ontario, June 2005. 58 Font, G. I., Jung, S., Enloe, C. L., McLaughlin, T. E., Morgan, W. L., and Baughn, J. W., “Simulation of the Effects of Force and Heat Produced by a Plasma Actuator on Neutral Flow Evolution,” AIAA Paper 2006–167, 44th AIAA Aerospace Sciences Meeting and Exhibit, Reno, NV, Jan. 2006. 59 Roy, S., Kumar, H., Gaitonde, D. V., and Visbal, M., “Effective Discharge Dynamics for Plasma Actuators,” AIAA Paper 2006–374, 44th AIAA Aerospace Sciences Meeting and Exhibit, Reno, NV, Jan. 2006. 60 Shang, J. S., “Electromagnetic Field of Dielectric Barrier Discharge,” AIAA Paper 2005–5182, 36th AIAA Plasmadynamics and Lasers Conference, Toronto, Ontario, June 2005. 61 Ramakumar, K. and Jacob, J. D., “Flow Control and Lift Enhancement Using Plasma Actuators,” AIAA Paper 2005– 4635, 35th AIAA Fluid Dynamics Conference and Exhibit, Toronto, Ontario, June 2005. 62 Jacob, J. D., Ramakumar, K., Anthony, R., and Rivir, R. B., “Control of Laminar and Turbulent Shear Flows Using Plasma Actuators,” TSFP 4-225, 4th International Symposium on Turbulence and Shear Flow Phenomena, Williamsburg, VA, June 2005. 63 Santhanakrishnan, A. and Jacob, J. D., “On Plasma Synthetic Jet Actuators,” AIAA Paper 2006–0317, 44th AIAA Aerospace Sciences Meeting and Exhibit, Reno, NV, Jan. 2006. 64 Smith, B. L. and Glezer, A., “The Formation and Evolution of Synthetic Jets,” Phys. Fluids, Vol. 10, No. 9, 1998, pp. 2281–2297. 65 Sholl, M. and Savas, O., “A Fast Lagrangian PIV Method for Study of General High-Gradient Flows,” AIAA Paper 1997–0493, 35th AIAA Aerospace Sciences Meeting, Reno, NV, Jan. 1997. 66 Ramakumar, K., Active Flow Control of Low Pressure Turbine Blade Separation Using Plasma Actuators, Master’s thesis, University of Kentucky, Lexington, Kentucky, May 2006. 67 Schlichting, H. and Gersten, K., Boundary Layer Theory, Springer-Verlag, New York, 8th ed., 2000.
Paper v. 1.8, May 26, 2006.
Figure 1. Cross section of the linear actuator flush mounted on an arbitrary surface, the direction of the wall jet induced by the plasma is shown.
14 of 30 American Institute of Aeronautics and Astronautics Paper 2006-3033
4 4
x=2 cm upstream x=1 cm upstream x=0 (at interface) x=0.5 cm downstream x=1 cm downstream x=1.5 cm downstream x=2 cm downstream x=3 cm downstream
3.5 3
3
y [cm]
y [cm]
2.5
2
2 1.5 1
1
0.5 1
2
3
0
4
x [cm]
(a)
0
20
40 60 Velocity [cm/s]
80
100
(b)
Figure 2. Linear actuator in quiescent flow: (a) streamlines of plasma-induced flow; actuator interface is located at the 2 cm tick, (b) velocity profiles.
(a)
(b)
Figure 3. (a) Schematic of PSJA: top view (left), and cross section (right), (b) Plasma ring created on actuation.
Figure 4. PSJA in blowing (left) and suction (right) configurations.
15 of 30 American Institute of Aeronautics and Astronautics Paper 2006-3033
(a)
(b)
Figure 5. PSJA in quiescent flow, pulsed at fp =10 Hz: (a) phase-locked vorticity contours overlaid with streamlines at jet evolution time t=24 ms, (b) Streamwise variation of local maximum mean axial velocity, three different jet evolution times are compared.
Figure 6. PSJA (left) and L-PSJA (right) in blowing configurations, the direction of the induced jet is shown in each case.
Figure 7. Flat plate and actuator arrangement.
16 of 30 American Institute of Aeronautics and Astronautics Paper 2006-3033
Figure 8. Phase-locked PIV: actuator duty cycle is shown by the solid square wave, and the dashed square wave indicates image capture with time delay t.
(a) Linear actuator.
(b) L-PSJA.
Figure 9. Averaged PIV results showing vorticity contours overlaid with streamlines for steady operation of actuator in quiescent flow.
(a) Cross-stream distribution of mean axial velocity.
(b) Streamwise distribution of local maximum mean axial velocity.
Figure 10. Jet characteristics for steady operation of L-PSJA in quiescent flow.
17 of 30 American Institute of Aeronautics and Astronautics Paper 2006-3033
(a) Linear actuator.
(b) L-PSJA.
Figure 11. Mean velocity profiles for steady operation of actuator in quiescent flow (∆=3.3 mm, x=0 corresponds to location of actuator interface in the linear actuator, and center of the actuator in the L-PSJA).
(a) Linear actuator.
(b) L-PSJA.
Figure 12. RMS velocity profiles for steady operation of actuator in quiescent flow (∆=3.3 mm, x=0 corresponds to location of actuator interface in the linear actuator, and center of the actuator in the L-PSJA).
(a) Case A.
(b) Case B.
(c) Case C.
Figure 13. Averaged PIV results showing vorticity contours overlaid with streamlines for steady operation of PSJA in crossflow. Flow direction is from left to right.
18 of 30 American Institute of Aeronautics and Astronautics Paper 2006-3033
(a) t=18 ms.
(b) t=24 ms.
(c) t=36 ms.
(d) t=48 ms.
(e) t=57 ms.
(f) t=83 ms.
Figure 14. Averaged PIV results showing vorticity contours overlaid with streamlines for PSJA in crossflow case A, with measurements phase-locked to actuator pulsing frequency fp =10 Hz. Flow direction is from left to right.
(a) t=18 ms.
(b) t=24 ms.
(c) t=36 ms.
(d) t=48 ms.
(e) t=57 ms.
(f) t=83 ms.
Figure 15. Averaged PIV results showing vorticity contours overlaid with streamlines for PSJA in crossflow case B, with measurements phase-locked to actuator pulsing frequency fp =10 Hz. Flow direction is from left to right.
19 of 30 American Institute of Aeronautics and Astronautics Paper 2006-3033
(a) t=18 ms.
(b) t=24 ms.
(c) t=36 ms.
(d) t=48 ms.
(e) t=57 ms.
(f) t=83 ms.
Figure 16. Averaged PIV results showing vorticity contours overlaid with streamlines for PSJA in crossflow case C, with measurements phase-locked to actuator pulsing frequency fp =10 Hz. Flow direction is from left to right.
20 of 30 American Institute of Aeronautics and Astronautics Paper 2006-3033
(a) Case A.
(b) Case B.
(c) Case C.
Figure 17. Averaged PIV results showing vorticity contours overlaid with streamlines for steady operation of L-PSJA in crossflow. Flow direction is from left to right.
(a) t=18 ms.
(b) t=36 ms.
(c) t=57 ms.
Figure 18. Averaged PIV results showing vorticity contours overlaid with streamlines for L-PSJA in crossflow case A, with measurements phase-locked to actuator pulsing frequency fp =10 Hz.
(a) t=18 ms.
(b) t=36 ms.
(c) t=57 ms.
Figure 19. Averaged PIV results showing vorticity contours overlaid with streamlines for L-PSJA in crossflow case B, with measurements phase-locked to actuator pulsing frequency fp =10 Hz.
(a) t=18 ms.
(b) t=36 ms.
(c) t=57 ms.
Figure 20. Averaged PIV results showing vorticity contours overlaid with streamlines for L-PSJA in crossflow case C, with measurements phase-locked to actuator pulsing frequency fp =10 Hz.
21 of 30 American Institute of Aeronautics and Astronautics Paper 2006-3033
(a) Case A.
(b) Case A.
(c) Case B.
(d) Case B.
(e) Case C.
(f) Case C.
Figure 21. Baseline flow (actuator off ), mean and RMS velocity profiles at different streamwise locations (∆=3.3 mm, r=0 corresponds to actuator center location x=14.8 cm).
22 of 30 American Institute of Aeronautics and Astronautics Paper 2006-3033
(a) Case A, PSJA.
(b) Case A, L-PSJA.
(c) Case B, PSJA.
(d) Case B, L-PSJA.
(e) Case C, PSJA.
(f) Case C, L-PSJA.
Figure 22. Mean velocity profiles for steady operation of actuator at different streamwise locations (∆=3.3 mm; r=0 (for PSJA) and x=0 (for L-PSJA) correspond to actuator center location x=14.8 cm).
23 of 30 American Institute of Aeronautics and Astronautics Paper 2006-3033
(a) Case A, PSJA.
(b) Case A, L-PSJA.
(c) Case B, PSJA.
(d) Case B, L-PSJA.
(e) Case C, PSJA.
(f) Case C, L-PSJA.
Figure 23. RMS velocity profiles for steady operation of actuator at different streamwise locations (∆=3.3 mm; r=0 (for PSJA) and x=0 (for L-PSJA) correspond to actuator center location x=14.8 cm).
24 of 30 American Institute of Aeronautics and Astronautics Paper 2006-3033
(a) PSJA.
(b) L-PSJA.
Figure 24. Displacement, δ ∗ (solid black line) and momentum, θ (dashed red line) thicknesses for cases A, B, and C; actuator in steady operation; ∆=3.3 mm, ∆=0 corresponds to actuator center location at x=14.8 cm.
(a) PSJA.
(b) L-PSJA.
Figure 25. Displacement, δ ∗ (solid black line) and momentum, θ (dashed red line) thicknesses for cases A, B, and C; actuator pulsed at 10 Hz; ∆=3.3 mm, ∆=0 corresponds to actuator center location at x=14.8 cm.
25 of 30 American Institute of Aeronautics and Astronautics Paper 2006-3033
Figure 26. Schematic of two L-PSJAs mounted on the circular cylinder at 90◦ and 270◦ to the incoming freestream (the arrangment is the same for two linear actuators, with the absence of exposed electrodes EE2); thicker red lines indicate connections from the exposed electrode, thinner black lines indicate connections from the embedded electrode.
26 of 30 American Institute of Aeronautics and Astronautics Paper 2006-3033
(a) Baseline, actuator off.
(b) Steady operation.
(c) 10 Hz pulsing (f + =0.2). Figure 27. Averaged PIV results showing vorticity contours overlaid with streamlines for linear actuator mounted on circular cylinder for ReD =5 · 103 .
27 of 30 American Institute of Aeronautics and Astronautics Paper 2006-3033
(a) Baseline, actuator off.
(b) Steady operation.
(c) 10 Hz pulsing (f + =0.2). Figure 28. Averaged PIV results showing vorticity contours overlaid with streamlines for L-PSJA mounted on circular cylinder for ReD =5 · 103 .
28 of 30 American Institute of Aeronautics and Astronautics Paper 2006-3033
(a) Linear actuator, x=1D downstream.
(b) L-PSJA, x=1D downstream.
(c) Linear actuator, x=10D downstream.
(d) L-PSJA, x=10D downstream.
Figure 29. Cross-stream variation of velocity deficits for ReD =5 · 103 .
29 of 30 American Institute of Aeronautics and Astronautics Paper 2006-3033
(a) Linear actuator, x=1D downstream.
(b) L-PSJA, x=1D downstream.
(c) Linear actuator, x=10D downstream.
(d) L-PSJA, x=10D downstream.
Figure 30. Cross-stream variation of velocity deficits for ReD =10 · 103 .
30 of 30 American Institute of Aeronautics and Astronautics Paper 2006-3033
