3rd AIAA Flow Control Conference AIAA 2006-3023

5 - 8 June 2006, San Francisco, California

Reattachment of a Separated Boundary Layer on a Flat

Plate in a Highly Adverse Pressure Gradient Using a Plasma
Actuator

Isaac G. Boxx* and Richard B. Rivir†

Air Force Research Laboratory, Wright Patterson AFB, OH, 45433, USA

Jeffery M. Newcamp ‡
United States Air Force, Robins AFB, GA, 31098, USA

and

Nathan M. Woods§
Wright State University, Dayton , Ohio 45354, USA

An experimental study was performed to examine the phase-dependent response

characteristics of a dielectric barrier discharge plasma flow control actuator. The actuator
was investigated on a fully separated flat plate boundary layer with an adverse freestream
pressure gradient distribution. The actuator was positioned downstream of the boundary
layer separation. Phase-locked particle image velocimetry (PIV) was used to obtain two-
dimensional velocity field measurements at thirty equally spaced phase-angles along the
sinusoidal voltage input to the actuator. A subtle phase-dependence was observed in the
response of the streamwise (U-velocity) component of boundary layer velocity field. No
similar phase-dependence was observed in the vertical direction (V-component of velocity).

Nomenclature
ρ = Freestream air density (kg/m3)
c = Blade Chord Length (m)
Cp = (Pt – Ps)/(1/2 ρu2)
Ps = Local static pressure (N/m2)
Pt = Total pressure (N/m2)
Rec = ρUinfc/μ
Tu = u ′ 2 + v′ 2 U 2 +V 2
U = Local Streamwise Velocity (m/s)
Uinf = Freestream Velocity at Infinity (m/s)
V = Local freestream vertical velocity (m/s)
Vinf = Freestream vertical velocity at infinity (m/s)

I. Introduction

T HE increasing demands for greater performance and efficiency in low-pressure turbine blades has lead to higher
and higher airfoil loading. A limiting parameter for blade loading is the increased level of boundary layer

*
NRC Research Associate, AFRL/PRTT, 1950 Fifth St, Wright-Patterson AFB, OH 45423, AIAA Member.
†
AFRL Fellow, Propulsion Directorate, 1950 Fifth St, Wright-Patterson AFB, OH 45423, AIAA Fellow.
‡
USAF, Robins AFB, GA 31098, AIAA Member
§
Graduate Research Assistant, Department of Mechanical Engineering, AIAA Member.

1
American Institute of Aeronautics and Astronautics

This material is declared a work of the U.S. Government and is not subject to copyright protection in the United States.
separation at low Reynolds numbers. In an effort to maintain high blade loading, forestall flow-separation or
reattach already separated flows over the LPT at low-Reynolds numbers, various passive and active flow-control
mechanisms have been investigated.
Plasma excitation of the wall region offers a method of manipulating the near-wall velocity profile and to induce
boundary layer reattachment. The effects of electrostatic fields on fluid flows have been demonstrated over many
years. Velkolf (1962) investigated electric field effects on heat transfer and pressure distributions for stagnation and
flat plate boundary layer flows. Roth et al. (1998) found that a dielectric barrier discharge (DBD) with an
asymmetric electrode configuration mounted on a flat plate could reduce drag and also the overall boundary layer
thickness by inducing local acceleration of fluid near the wall. Since then, a number of researchers have
characterized the effect of these actuators on flow over aerodynamic surfaces. Corke et al. (1997) studied flush
mounted and subsurface plasma actuators as a means to introduce controlled disturbances into flow over
axisymmetric bodies in supersonic flows of Mach numbers 3.5 and 6. Post et al. (2003) studied plasma actuators as
a means of controlling flow separation over a NACA 663 – 018 airfoil over Reynolds numbers ranging from 77,000
to 460,000. They demonstrated an 8 degree increase in maximum angle of attack, accompanied by a full pressure
recovery after stall. Post et al. (2004) also studied the effect of plasma actuators on an oscillating NACA-0015
airfoil and were able to achieve a higher cycle-integrated lift.
Several investigations of the application of the DBD actuators to separated low-pressure turbine (LPT) cascade
flows have been reported. Hultgren et al., (2003) studied an array of asymmetric electrode DBD plasma actuators
mounted on a flat plat in a simulated pressure field of the suction side of a Pak-B LPT blade at Reynolds numbers
ranging from 50,000 to 300,000. They concluded their phased array DBD plasma actuator was an effective device
for separation control on the LPT blade. List et al. (2003) studied an asymmetric electrode DBD plasma actuator on
a linear cascade of Langston turbine blades and found the actuators could reduce profile loss by 14% at low
Reynolds numbers (Re = 30,000). Huang et al. (2003) studied plasma actuators positioned at various chord
locations on the surface of a Pak-B profile in a linear turbine cascade over Reynolds numbers ranging from
Rec = 10,000 to 100,000. They found the boundary layer flow reattachment point induced by the plasma actuator
was highly sensitive to freestream turbulence and Reynolds number.
Although the studies above have produced useful insight into the effect of plasma actuators on a variety of
boundary layer flows, they have relied almost entirely upon time-averaged measurements. While such measurements
are useful in studying global effects and trends they also inevitably obscure the periodic nature of the physical
mechanism through which these actuators force the near-wall boundary layer flow. Time-averaged measurements
are particularly poor in determining what (if any) phase of the actuator cycle is dominant in affecting change in local
and global flowfield characteristics. For example, it has been noted by previous researchers (Rivir et al., 2004,
Enloe et al., 2004, Orlov et al., 2005) that clusters of very short duration (≈ 10s of ns) current spikes form during
the ignition phase of the DBD actuator cycle. Peak current during these events can be several orders of magnitude
higher than the peak-to-peak current variation elsewhere in the cycle. Figure 1 shows a characteristic voltage and
current trace measured in this study. A better understanding of what effect these local current spikes have on the
forcing effectiveness of the actuator may be useful in future efforts to optimize these devices for specific flow-
control applications.
The objective of the current study was to determine
what if any phase-relation exists between the sinusoidal
input signal of a DBD actuator and the response of the
flow over the actuator. This was accomplished through
the acquisition of a statistically meaningful set of
phase-averaged velocity field measurements over a
complete cycle of a DBD actuator in a well-
characterized flow-field. The flow-field was a fully
separated flat-plate boundary layer in a freestream
pressure distribution designed to simulate that of a
generic low-pressure turbine blade. The primary
diagnostic used in this study was high spatial-
resolution, phase-locked PIV. Two-dimensional
velocity field measurements were acquired at each of
thirty points along a complete cycle of the sinusoidal Figure 1 - Voltage and Current over DBD cycle
input wave of a DBD plasma actuator.

2
American Institute of Aeronautics and Astronautics
 II. Experimental Apparatus
This experiment used the same facility and experimental apparatus described in Boxx et al. (2006). The
experimental apparatus is described in greater detail in that paper.

A. Plasma Actuator

The DBD flow-control actuator used in this

study is shown in Figure 1. It consisted of two
copper electrodes separated by a 1.56 mm thick
layer of fiberglass laminate. The upper and lower
electrodes are set in an asymmetric configuration
with the upper electrode on the upstream side and
the lower one on the downstream. Plasma
generation occurs in the region where the electrodes
run parallel to one another. This interface line is
taken as the origin of the x-y coordinate system in
this study.
The electrodes were driven with 3 kHz sine-
wave A.C. signal. The signal was produced with an
adjustable A.C. power supply (Compact Power -
Titan Series™) connected to the primary coils of a Figure 2 - Schematic diagram of the dielectric barrier
pair of high-voltage step-up transformers discharge flo w-control actuator configuration
(manufactured by Industrial Test Equipment Co.)
which raised its voltage to 5-6 kV (peak
positive/negative).
The voltage potential across the transformers was measured using a pair of 1000× voltage-attenuation probes
(Tektronics Model P6015A). Current in the system was measured with a current coil (Pearson Model 4100). In
order to correct for induction and capacitance inherent in the power-conditioning system, a baseline case (using a
series of low-impedance pure resistors in place of the actuator) was also measured using the same instruments.

B. Low-Speed Wind Tunnel

The experiment was conducted in the

low-speed wind-tunnel section of the
Turbine Aero Thermal Basic Research
Facility at the Air Force Research
Laboratory. The facility has a rectangular
test-section measuring 0.38 × 0.25 m.
Temperature of the flow through the tunnel
was regulated and set to 26.6 deg C. (80 F)
using a water-cooled chiller unit. Flow
through the facility was seeded with
propylene-glycol/water droplets (nominally
4 μm diameter) from a theatrical fog
generator (Rosco Model 4500).
The actuator described above was
mounted on a flat plate located in the test-
Figure 3 - Lo w speed wind tunnel and PIV setup
section of the tunnel. A recess milled into
the upper surface of the plate ensured the
actuator remained flush with the surface. The active (i.e. the plasma-generating) region of the actuator spanned only
the center 95 mm of the plate in order to avoid wall-effects from the sides of the tunnel. A contoured block of
polystyrene foam was mounted on the upper surface of the test section. The contour was designed to generate a
pressure distribution approximating that of suction-side of a generic aft-loaded, low-pressure turbine blade. In order
to keep flow attached over the contour a vacuum was applied to a 0.25” long slot which spans the width of the block

3
American Institute of Aeronautics and Astronautics
downstream of the throat of the contoured section. The vacuum was generated using a highly-throttled, 2HP
commercial shop-vac system.

C. PIV System
The PIV system used a dual-head, frequency-doubled, flashlamp-pumped Nd:YAG laser (New Wave Solo-120)
and an adjustable light-arm to deliver sheet-illumination to the test-section. Light scattered from propylene-glycol
seeding droplets was imaged with a high-framerate, 1024 × 1024 pixel resolution CMOS camera (Photron-APX).
The camera was triggered in a two-frame burst mode at a 10 Hz repetition rate to produce a frame-straddling PIV
system. The camera was equipped with a 200 mm lens (Nikon, AF-Nikkor) operating at f/11. The field of view of
the PIV system was 17.9 × 17.9 mm. Each pixel corresponds to 17.45 μm in physical space for approximately 1-to-1
imaging. The camera and lasers were synchronized using two pulse/delay-generator timing boxes (Stanford
Research Systems DG-535 and Quantum Composers 9300 Series respectively). The images were processed with a
commercially available adaptive window offset cross-correlation algorithm (Dantec FlowManager). The final
window size was 32 × 32 pixels with 50% overlap for a final spatial resolution of ≈ 0.56 mm and vector placement
every ≈ 0.28 mm. The data was then post-processed using
in-house codes.

D. Boundary Conditions
The driver signal for the DBD actuator in this study
was set to 3 kHz and had a peak-to-peak potential
difference of 12 kV. The local freestream velocity above
the actuator location was 1.5 m/s and Tu = 4.6%
( Tu = u '2 + v'2 / U 2 + V 2 , where U and V are the local
freestream velocity components in the x and y directions).
The static-pressure distribution along the test-section
wall was measured at fourteen points using a pressure
transducer (GE - Druck LPM 9481, 0.2” H2O full-scale
range). These pressures were used to compute the
Cp distribution shown in the plot presented in Figure 4. Cp
was defined as C = (P − P ) (1 ρU 2 ) , where U is the Figure 4 – C distribution measured along test
p t s 2 p
freestream velocity upstream of the actuator. As noted section.
above, suction was applied to the contoured upper surface
of the wind-tunnel test-section in order to prevent flow-
separation there. The plot shown in Figure 4 also shows the
Cp distribution for the case where no suction was applied
to the upper surface. It can be seen from the similarity of
the two profiles that although the applied suction resulted
in an attached flow over the upper surface, it did not result
in a substantial alteration of the freestream pressure
characteristics. Comparing our measured Cp distribution
to that of a generic low-pressure turbine airfoil, we
determined that the contoured test-section produced and
effective chord length of 0.35m, resulting in a simulated
chord-Reynolds number of 23,500 for this study.
In order to better characterize the nature of the flow
Figure 5 – Streamlines of flow through the test-section
through the test-section, we also performed a series of
in the vicinity of the actuator. The dashed lines show
wide-field PIV measurements using the system described
the approximate location laser-sheet cutoff, where
above. The streamlines derived from these measurements particle-dropout was seen to result in a region of lower
are shown in Figure 5. Overlaid on these streamlines are SNR.
the relative positions of the contoured upper section of the St li St li
wind-tunnel (including the slot where the vacuum was applied) and the outer bounds of the laser sheet illumination.
We note that near the edges of the laser sheet, the signal-to-noise ratio of the measurements dropped off quickly,
resulting in the misshapen streamlines shown there in the image.
It is clear from these streamlines that the suction applied to the contoured upper section of the tunnel induced a
small but noticeable (≈ 0.2m/s) velocity in the vertical direction. Although this added vertical velocity was

4
American Institute of Aeronautics and Astronautics
undesirable it was necessary in order to maintain stability and uniformity in the test-section. The vertical velocity
induced by the vacuum provided an additional challenge to flow-reattachment and thus allowing a more rigorous
test of the capabilities of the flow-control actuator.

III. Results and Discussion

(a) (b)

(c) (d)
Figure 6 - Representative velocity fields measured in the boundary layer. (a) and (b) show the U- and
V-velocity components measured in the initially separated boundary layer respectively.
(c) and (d) show this same for the actuated boundary layer. Units are in m/s.

A. Velocity Field Measurements

Figure 6 shows the measured velocity fields in the streamwise and normal directions for the initially separated
boundary layer and a representative sample of the velocity fields for the actuated boundary layer. Before discussing
the phase-dependent response of the boundary layer velocity field, it is instructive to examine the general response
and behavior of the boundary layer when acted upon by the DBD actuator.
Figure 6a and 6b show the U- and V-components of velocity for the case of the un-actuated boundary layer. In
this figure, the axes are referenced such that the origin corresponds to the center of the two electrodes on the
actuator. Both the vectors and contours of U- velocity show that the boundary layer is fully separated in the un-
actuated case. Although the separation begins upstream of the PIV window, it is straightforward to estimate its
location to be ≈ 12 - 15mm upstream of the origin by tracing the lowest U-velocity contour to its intersection with
the streamwise axis. The contours of V-velocity in Figure 6b illustrate the previously mentioned vacuum-induced
V-velocity. Although previous studies have indicated that placing the actuator immediately upstream of the
separation is most effective for preventing flow-separation over an airfoil we find that placing it in the stagnant near-
wall fluid of an already separated boundary layer gives the clearest contrast in the PIV measurements and thus
provides further insight into the nature of the actuator / boundary-layer interaction.

5
American Institute of Aeronautics and Astronautics
 Comparing the measured velocity fields shown in Figure 6a and 6b with those in 6c and 6d the effect of the
actuator is clear. The U-velocity contours show the actuator affects a re-attachment of the boundary layer
downstream of its location. Consistent with previous research, we see a region of high (relative) velocity fluid
jetting along the wall downstream of the actuator. Above this region exists an area of lower-velocity but still
attached boundary layer flow. Upstream of the actuator however there is a standing separation bubble. Previous
research in the present study (Boxx et al., 2006) has shown that this separation bubble decreases both in strength and
spatial magnitude with increasing voltage and/or power input to the actuator. In order to prevent unnecessary wear
on the dielectric barrier material we decided against increasing the power input to the levels required to eliminate
this separation bubble altogether.
The measured V-velocity field shown in Figure 6d indicates the extent to which the actuator influences the
velocity normal to the aerodynamic surface. Figure 6d shows that the actuator counteracts the initial vertical
vacuum-induced vertical velocity and induces a net downward velocity throughout most of the field of view. This
downward velocity peaks immediately downstream the origin, i.e. near the interface of the actuator electrodes. This
corresponds to the location of plasma generation by the actuator and the most intense concentration of E-field lines
over the actuator. A secondary peak in the V-velocity field occurs approximately 4 mm upstream of the origin.
Comparison of Figures 6c and 6d shows that this secondary peak in downward pointing (negative) V-velocity is in
the region of the aforementioned upstream separation bubble. Figure 6d also shows an increased velocity in the
vertical direction associated with the jetting region downstream of the actuator. It is not clear if this is the result of
charged particles following the field lines of the actuator or if it’s a thermal effect induced by heating caused by the
plasma. The spatial extent of the actuators influence upon the vertical velocity field is significantly larger than that
of its effect on the streamwise velocity.

B. Phase-localized measurement of U-velocity in the upper boundary-layer region

(a) (c)

(b) (d)
Figure 7 – Measured heights of U-velocity contours vs. the phase of the input signal to the actuator with
increasing downstream distance. (a) Representative voltage and current trace measured to
actuator (b) 90% peak U-velocity (c) 80% peak U-velocity (d) 70% peak U-velocity.
Legend – Black = 0mm, red = 2.7mm, blue = 5.4mm, magenta = 8.3mm, green = 9.9mm

6
American Institute of Aeronautics and Astronautics
 Previous research has focused upon the effect of these actuators upon the CP distribution and near-wall velocity
field modification (Boxx et al., 2006). The PIV data for each phase angle was threshholded to determine the 70, 80
and 90% freestream velocity contours. The height of each of these contours was then extracted from the data at
several downstream locations and plotted against their corresponding phase-angles. These plots are shown above, in
Figure 7. Comparing the height of the velocity contours to the corresponding phase angles of the input signal to the
actuator reveals a subtle correlation. There is a considerable change in the U-component of velocity associated with
upwards rising zero-cross of the actuator input signal. There is also a (much smaller) change in U-velocity
associated with the downward falling zero-cross of the input voltage signal to the actuator. These phase-localized
changes in the height of the U-velocity contours are most evident for the case of the 90% contour, shown in
Figure 7b but can be seen in all three contours plotted in Figure 7. The reason for the observed changes in U-
velocity component with phase-location becomes clear when comparing the data to measured current traces rather
than the input voltage signal.
Figure 7a. shows a representative voltage and current trace for the actuator at the conditions examined in this
study. Consistent with previous researcher observations, we note clusters of large (relative) magnitude spikes in the
measured current signal associated with the upward rising and downward falling zero-cross of the input voltage
signal. As noted in a previous section, researchers have linked these current spikes to the ignition and quenching of
the plasma across the DBD actuator. As can be seen in Figure 7a, the current associated with these clustered spikes
and be orders of magnitude greater than the current flowing during other phases of the cycle. The data shows that
this increased current density changes the actuator effectiveness during the portion of the cycle where they occur.
Figure 7 shows that the increased current density associated with the ignition and quenching events has a subtle,
though observable influence on the upper reaches of the boundary layer and its coupling with the freestream flow.
The net effect of this phase-localized change in the U-velocity profile is negative. Rather than drawing high-
velocity freestream fluid closer to the surface during the events, these current spikes have the opposite effect. The
boundary-layer thickness increases during these events, suggesting that that these spikes inhibit the effective
operation of the actuator. Increased thermal dissipation caused by increased current density may be the cause of the
observed increase in boundary layer thickness.
It should be noted at this point that a limitation of
the experimental apparatus may have biased the
magnitude of the observed variation resulting in a lower
magnitude change than would have otherwise been
measured. This limitation is a result of the 110μs time-
separation of the laser pulses in the PIV setup needed to
resolve the relatively low-velocity flow. This
corresponds to approximately one third the time-scale of
the 3-kHz voltage driver signal to the actuator. As can
be seen in Figure 8, the effect of taking a running
average of a sine wave over one third of a cycle has the
effect of reducing the magnitude of a signal by
smoothing or blurring the data. As Figure 7 shows the
flow responds to the forcing of the actuator at the
frequency of the input signal there should be a
smoothing effect associated with the 1/3rd cycle delta-t Figure 8 – Effect of 1/3rd cycle running average taken over
of the PIV measurement. Combining this with the over the period of a sine wave.
localized clusters of current spikes the time-delay of the
PIV system may result in a smaller magnitude
difference in the measured velocity response.

C. Phase-localized measurement of V-velocity field

Figure 9 shows the line-contours of V-velocity derived from representative points in the PIV measurements. The
points used correspond to that shown in Figures 6c and 6d. It is clear from this figure that (unlike for the U-velocity
field) the contours vary in two dimensions. The histograms of the V-component of velocity were compared over the
entire field of view of the PIV system. This allowed for the slight two-dimensional variations of given V-velocity
contours while still capturing the characteristics of the flow-field.

7
American Institute of Aeronautics and Astronautics
 Figure 10 shows the histograms for the V-
component of velocity for the initial, separated
boundary layer as well as for several sequential phase-
locations along the actuator input cycle. Figure 10
illustrates the differences in the V-velocity fields
before and after the application of the actuator forcing.
The initial histogram for that case is dual-peaked, with
one peak at zero (corresponding to the stagnant fluid in
the separation region) and a larger peak near 0.2 m/s.
The histogram shows a uniformly positive V-velocity
field for the unforced case. The histograms for the
actuated case in Figure 10 are dual peaked. There is
one peak at 0 m/s and a second peak between 0.04 -
0.08 m/s. Figure 9 shows the positive peak
corresponds to fluid in the upper boundary layer and
freestream flow. Although there is significant vertical
velocity in the upward direction associated with the
jetting downstream of the actuator the spatial extent
over which this occurs in the PIV window is small and
does not appear to have a major influence on the Figure 9 – Contours of V-velocity in the actuated
histogram profile. The measurement was limited by boundary layer.
the width of the PIV window and the V-component
velocity associated with the jetting region will
continue to grow with increasing distance
downstream.
The histogram profile for the actuated boundary
layer case indicates that V-velocity imparted by the
actuator acts over a wide area of the PIV window.
This can be seen by comparing it to the histogram
profile of the un-forced case. The profile of the
separated boundary layer cuts off sharply at zero
velocity, the histogram profile for the actuated case
drops off much more slowly.
Figure 11 shows the histograms of V-velocity
for each of the thirty phase-angles in this study.
The four plots that are shown each correspond to a
quarter cycle of the input signal. In all cases the
histogram has a dual-peak profile with a slowly
decaying tail on the negative side. The magnitude Figure 10 – Histogram of V-velocity in the boundary layer. Solid black
of the histogram peaks while not identical do not line corresponds to the histogram of the initial, separated
vary significantly with the phase of the cycle. The boundary layer. The remaining curves represent samples
rate of decay of the tails of the histogram show of the actuated boundary layer V-velocity field at sequential
small variations. phase angles.

CONCLUSIONS
Detailed experimental measurements of the global- and phase-dependent response of a separated boundary layer
to the forcing of a dielectric-barrier discharge plasma flow control actuator were presented. The data showed a
subtle phase-dependence in the U-component of velocity in the upper boundary layer to forcing by the actuator.
This phase-dependence is linked to the clusters of current-spikes which occur during distinct portions of the a.c.
discharge cycle and has the effect of increasing the boundary layer thickness. No similar phase-dependence was
observed in the measured data for the V-component of velocity.

8
American Institute of Aeronautics and Astronautics
 (a) (b)

(c) (d)
Figure 11 – Histograms of V-velocity measured in PIV window for each quarter of the
actuator input cycle. a, b, c and d., correspond to the first, second, third and forth
quarter of the voltage / current cycle shown in Figure 7a.

ACKNOWLEDGMENTS
This work was performed under sponsorship from the Air Force Office of Scientific Research. The views and
conclusions contained herein are those of the authors and should not be interpreted as necessarily representing the
official policies or endorsements, expressed or implied, of the Air Force Office of Scientific Research or the U.S.
Government. The authors would like thank Campbell Carter for providing hardware and Lt. Balcer for his
contributions to the design and construction of the experimental apparatus.

References
Corke, T.C., Cavalieri, D.A. “Controlled experiments on instabilities and transition to turbulence in supersonic
boundary layers”. AIAA 97-1817. 1997.

Enloe, C.L., McLaughlin, T.E., VanDyken, R.D., Kachner, D., Jumper, E.J., Corke, T.C. “Mechanisms and
Responses of a Single Dielectric Barrier Plasma Actuator: Plasma Morphology”, AIAA Journal, Vol. 43(3). 2004.

Huang, J., Corke, T.C., Thomas, F.O. “Plasma actuators for separation control of low pressure turbine blades”,
AIAA 2003-1024.

Hultgren, L.S., Ashpis, D.E. “Demonstration of separation delay with glow-discharge plasma actuators”. 41st
Aerospace Sciences Meeting and Exhibit, Reno NV. AIAA-2003-1025. 2003.

9
American Institute of Aeronautics and Astronautics
List, J., Byerley, A.R., McLaughlin, T.E., van Dyken, R.D. “Using a plasma actuator to control laminar separation
on a linear cascade turbine blade”. 41 Aerospace Sciences Meeting and Exhibit. Reno, NV. 2003.

McLaughlin, T.E., Munska, M.D., Vaeth, J.P., Dauwalter, T.E., Good, J.R., Siegel, S.G., “Plasma-based actuators
for cylinder wake vortex control”. AIAA 2004-2129. 2004.

Post, M. L., Corke, T.C. “Separation control on high angle of attack airfoil using plasma actuators”. 41st Aerospace
Sciences Meeting and Exhibit, Reno NV. AIAA 2003-1024. 2003.

Post, M.L., Corke, T.C. “Separation control using plasma actuators – Stationary and oscillating airfoils”. AIAA
2004-0841. 2004.

Roth, J.R., Sherman, D.M, Wilkinson, S.P. “Boundary layer flow control with a one atmosphere uniform flow
discharge surface plasma” . AIAA 98-0328. 1998.

Roth, J.R., Madhan, R.C.M, Yadav, M., Rahel, J., Wilkinson, S.P. “Flow field measurements of paraelectric,
peristaltic, and combined plasma actuators based on the one atmosphere uniform glow discharge plasma
(OAUGDPTM)” . 2004-845. 42nd AIAA Aerospace Sciences Meeting and Exhibit, Reno NV. 2004

Velkoff, H. R., “Investigation of the effects of electrostatic fields on heat transfer and boundary lasers”, Ph.D.
Dissertation. Ohio State University, 1962.

10
American Institute of Aeronautics and Astronautics