AIAA JOURNAL

Vol. 56, No. 1, January 2018

Control of Dynamic Stall over a NACA 0015

Airfoil Using Plasma Actuators

Achal Singhal,∗ David Castañeda,∗ Nathan Webb,† and Mo Samimy‡

The Ohio State University, Columbus, Ohio 43235
DOI: 10.2514/1.J056071
Dynamic stall is observed in numerous applications, including sharply maneuvering fixed-wing aircraft,
biomimetics, wind turbines, and most notably, rotorcraft. The associated unsteady loading can lead to
aerodynamic flutter and mechanical failure in the system. The present work explores the ability of nanosecond
pulse-driven dielectric barrier discharge plasma actuators to control dynamic stall over a NACA 0015 airfoil.
Downloaded by KUNGLIGA TEKNISKA HOGSKOLEN KTH on December 31, 2018 | http://arc.aiaa.org | DOI: 10.2514/1.J056071

The Reynolds number, reduced frequency, and excitation Strouhal number were varied over large ranges:
Re
167;000–500;000, k
0.025–0.075, and Ste
0–10, respectively. Surface pressure measurements were
taken for each combination of Reynolds number, reduced frequency, and excitation Strouhal number. Phase-
locked particle image velocimetry measurements were acquired for select cases. It was observed that the trends of
effect of Ste were similar for all combinations of Reynolds number and reduced frequency, and three major
conclusions were drawn. First, it was observed that low Strouhal number excitation (Ste < 0.5) results in
oscillatory aerodynamic loading in the stalled stage of dynamic stall. This oscillatory behavior was gradually
reduced as Ste increased and was not observed beyond Ste > 2. Second, all excitation resulted in earlier flow
reattachment. Last, it was shown that excitation, especially at high Ste , resulted in reduced aerodynamic
hysteresis and dynamic stall vortex strength. The decrease in the strength of the dynamic stall vortex is achieved by
the formation of large-scale structures induced by the excitation that bleed the leading-edge vorticity before the
ejection of the dynamic stall vortex. At sufficiently high excitation Strouhal numbers (Ste ≈ 10), the dynamic stall
vortex was completely suppressed.

Nomenclature λ = nondimensional swirling strength,H λc∕u∞

CL = lift coefficient, −∫ 1−1 Cp sin
θ ds Ξ = damping coefficient, −
1∕πΔα2  CM dα
CM = moment coefficient, −∫ 1−1 Cp ×

1∕4 sin θ − x sin θ − ϕ = phase, deg
y cos θ ds
Cp = pressure coefficient,
p − p∞ ∕q∞ I. Introduction
c, L = airfoil chord, 203, mm
f
k
=
=
frequency, Hz
reduced frequency, πfc∕u∞ T HE primary focus of this work is exploring flow physics and
control of dynamic stall using nanosecond pulse-driven
dielectric barrier discharge (NS-DBD) plasma actuators. Dynamic
p = static pressure, Pa
po = stagnation pressure, Pa stall occurs in a variety of applications, including wind turbines and
p∞ = freestream static pressure, Pa sharply maneuvering fixed-wing aircraft, and occurs in rotorcraft
q = dynamic pressure, Pa, 1.05 ×
po − p∞ , where 1.05 is primarily due to the change in rotor angle of attack α required to
the tunnel calibration constant maintain lift symmetry [1,2]. Dynamic stall occurs when the rate of
Re = Reynolds number, u∞ c∕ν change of α is fast enough to maintain attached flow past the static
Ste = excitation Strouhal number, fc∕u∞ stall angle of attack. As the airfoil motion slows and begins to pitch
U = velocity vector,
u; v, m/s down, a vortex forms at the leading edge, convected over the airfoil,
U = nondimensional velocity vector, U∕u∞ and shed. This dynamic stall vortex momentarily increases the lift
u∞ = freestream velocity, m/s production. Once it has convected past the airfoil and shed, the flow is
x = streamwise position, mm fully stalled. This process results in large oscillatory loads that can
y = vertical position, mm significantly reduce the life of the blades [3]. This problem has
α = angle of attack, deg attracted much attention in the rotorcraft community because it is
β = moment stall magnitude desirable to eliminate the oscillatory loads.
Δα = motion amplitude, deg Previous research has focused on the use of flow control devices to
θ = surface normal angle mitigate the dynamic stall vortex. These devices consist of both
λ = swirling strength, Hz passive (such as geometric modifications to the airfoil) and active
(such as momentum injection) techniques [4–8]. For practical
application, it is pertinent to ensure such a device can be incorporated
Presented as Paper 2017-1687 at the 55th AIAA Aerospace Sciences
into the blade of a rotorcraft. Adding geometric modifications, such
Meeting, Grapevine, TX, 9–13 January 2017; received 14 February 2017; as vortex generators, would result in heavier rotor blades and may not
revision received 7 July 2017; accepted for publication 16 August 2017; easily be incorporated into an existing rotorcraft fleet. Furthermore,
published online 19 September 2017. Copyright © 2017 by the American geometric modifications tend to work well at design conditions, but
Institute of Aeronautics and Astronautics, Inc. All rights reserved. All requests become ineffective or detrimental at off-design conditions. Active
for copying and permission to reprint should be submitted to CCC at www. techniques aim to provide efficacy over a broad operating range.
copyright.com; employ the ISSN 0001-1452 (print) or 1533-385X (online) to Active techniques typically involve momentum injection [4].
initiate your request. See also AIAA Rights and Permissions www.aiaa.org/ Although potentially effective, it is not a desirable solution due to the
randp. increased mechanical complexity, weight, and cost. Recently,
*Graduate Student, Department of Mechanical and Aerospace Engineering.
Student Member AIAA. dielectric barrier discharge (DBD) plasma actuators have been
†
Research Engineer, Aerospace Research Center. Member AIAA. explored as a potential alternative to provide momentum injection.
‡
John B. Nordholt Professor of Mechanical and Aerospace Engineering, These devices consist of two copper electrodes separated by a
Director of Aerospace Research Center; [email protected]. Fellow AIAA dielectric barrier and use the electrohydrodynamic effect to generate a
(Corresponding Author). secondary flow near the surface to energize the boundary layer,
78
 SINGHAL ET AL. 79

thereby maintaining minimally complex construction. These devices is defined to be positive when it results in the pitch-down motion of
are generally referred to as AC-DBD actuators, because they operate the nose, as defined in Fig. 1.
using an alternating current. The airfoil oscillation mechanism was designed to maximize
The primary disadvantage of momentum injection techniques is optical access and modeled after that of Greenblatt and Wygnanski [16].
that, at high speeds, more momentum is required to maintain efficacy. It consists of two acrylic disks secured in aluminum rings. The ends of
Meeting this increasing demand for high Reynolds number the airfoil are mounted to these disks. Thin ball bearings are used to
application flows is often impractical. The focus of this paper will be mount the airfoil support rings to the tunnel sidewalls. The aluminum
on the use of NS-DBD plasma actuators. These actuators are rings are driven by a servo via timing belts, as shown in Fig. 2.
constructed identically to AC-DBD plasma actuators. However, The motion of the airfoil a
t in the remainder of the paper is
instead of an AC signal, high-voltage, short-duration (approximately described by three parameters: the mean angle αm , amplitude Δα, and
100 nanoseconds) DC pulses are used to create localized rapid reduced frequency (k  πfc∕u∞ , where f is the physical oscilla-
heating, which results in the formation of a compression wave at the tion frequency of the airfoil). The angle of attack is defined as
DC pulse frequency [9–13]. If introduced in the proper location a
t  am − Δa ⋅ cos
2ku∞ t∕L, where t is the time in seconds.
and at the proper frequency, the localized heating can be used to During experimentation, αm  10 deg and Δα  10 deg.
excite natural flow instabilities, which then grow and roll up into A single actuator, powered by a custom in-house manufactured
large flow structures. This is the flow control mechanism for the pulse generator, was placed on the airfoil to explore the efficacy
Downloaded by KUNGLIGA TEKNISKA HOGSKOLEN KTH on December 31, 2018 | http://arc.aiaa.org | DOI: 10.2514/1.J056071

NS-DBD plasma actuators, and because it is based on excitation of of excitation. The actuator (shown in Fig. 3) is constructed of two
instabilities, relatively low power input is required and the 0.09-mm-thick copper tape electrodes; the exposed high-voltage
actuation remains effective at high speeds. These actuators excite (HV) electrode is 6.35 mm wide and the covered ground electrode is
the flow instabilities in a manner similar to another class of plasma 12.70 mm wide. The dielectric layer is composed of three layers of
actuators called localized arc filament plasma actuators, which Kapton tape, each 0.09 mm thick with a dielectric strength of 10 kV.
have been used for excitation of instabilities in several flows with The total thickness of the entire actuator is 0.45 mm. The actuator was
instabilities [14,15]. placed on the suction side of the airfoil with the electrode junction
NS-DBD plasma actuators have been shown to reattach flow in at x∕c  0.01 (see Fig. 1) and covers the entire span of the airfoil.
stalled airfoils for a variety of airfoil shapes and flow velocities in a
As previously mentioned, the actuators are driven by a nanosecond
static configuration [9–12]. The static stall control work includes
pulse to produce perturbations. This placement was motivated by the
flow control of an airfoil in fully reversed flow [11]. This work aims to
location of the receptivity region of the shear layer instabilities, which
extend the existing body of work on NS-DBD flow control to
is the location of maximum sensitivity to perturbations.
unsteady flows.
Static pressure measurements on the airfoil surface were acquired
using three Scanivalve digital pressure sensor arrays (DSA-3217).
II. Experimental Arrangement A total of 35 taps are located on the surface of the airfoil. The pressure
Experiments were performed in the recirculating wind tunnel coefficient was phase averaged over 4 sets of 16,384 samples
located at the Gas Dynamics and Turbulence Laboratory, within acquired at 400 Hz near the airfoil centerline. The sectional lift and
the Aerospace Research Center at The Ohio State University [9–12]. moment coefficients CL and CM were calculated from these data and
The tunnel has an optically clear acrylic test section measuring used to determine the effect of excitation on the dynamic stall
61 × 61 cm in cross section and 122 cm in length. A NACA 0015 process. Figures containing pressure data, or results calculated from
airfoil with a 203 mm chord was used in the experimentation. pressure data, include error bars, unless their presence would make
Although not a typical cross section for rotorcraft blades, it has been the figure difficult to read. These bars are measures of statistical
well characterized in static flow, and for this first effort it was deemed uncertainty and represent the 95% confidence interval.
appropriate. The experimental arrangement used in this work is Particle image velocimetry (PIV) was the primary diagnostic
briefly recalled here. More details are found in Singhal [13]. technique. Two-component planar PIV data were acquired for both
Two coordinate systems are used throughout this paper. The first the flow surrounding the suction side of the airfoil and behind the
coordinate system has its origin at the aerodynamic leading edge, as trailing edge. Olive oil seed particles were injected upstream of the
shown in Fig. 1. The axis of the first system lies along the chord line test section. An approximately 2-mm-thick laser sheet illuminated
and the coordinate is normalized by the chord length, denoted x∕c. the measurement plane at 60.4% of the airfoil span. Two com-
A positive coordinate indicates the aerodynamic suction side and a mercially available LaVision cameras (Imager Pros) and software
negative coordinate indicates the aerodynamic pressure side. This (DaVis) were used to acquire 5 sets of 100 images pairs for each
system is in the airfoil reference frame and is used to specify the excited case. For the baseline cases, a single set of 500 images was
location of onboard instrumentation, actuators, and the flow acquired. Acquisition was locked to the airfoil motion to provide
separation line. The second coordinate system is a two-dimensional phase-averaged flowfield information. Swirling strength was used as
grid aligned with the test section. These coordinates are also the vortex identification technique, as detailed by Adrian et al. [17],
normalized by the chord, denoted x∕L and y∕L (L, instead of c, is and was nondimensionalized using the freestream velocity and the
used to distinguish between the two normalized x coordinates). This chord length. In the presentation of the PIV results, the airfoil is
system is in the test section reference frame and is used to define the plotted in black along with an arc. The arc is a manifestation of an
location and direction of velocity data. The origin of this coordinate optical obtrusion and is not indicative of any flow features.
system is at the point of rotation, or x∕c  0.25. A moment M PIV acquisition was phase locked to the motion of the airfoil
and is described by the phase ϕ. It is easier to consider the cameras
to be acquiring snapshots of the flowfield at the same rate as
the oscillation frequency of the airfoil. Then, the equation of the
airfoil motion (relative to the cameras) becomes a
t  am −
Δα ⋅ cos
2ku∞ t∕L  ϕ. Ideally, phase and angle of attack are
injective; however, due to small errors in the motion, this may not be
the case. As such, the phase will always be provided in discussions
involving PIV results. For convenience, the angle of attack will also
be provided.
It is also important to note that, for a given excitation Strouhal
number, any given excitation pulse will always occur at the same
phase for any cycle of the airfoil motion. As such, in the excited PIV
Fig. 1 Schematic of experimental coordinate systems showing the results, PIVacquisition is phase locked to both the airfoil motion and
origins. the actuation.
 80 SINGHAL ET AL.
Downloaded by KUNGLIGA TEKNISKA HOGSKOLEN KTH on December 31, 2018 | http://arc.aiaa.org | DOI: 10.2514/1.J056071

Fig. 2 Photographs of the oscillating mechanism.

Fig. 3 NS-DBD plasma actuator schematic. Vertical dimension exaggerated for clarity.

III. Error Analysis

A detailed analysis of the error is provided by Singhal [13].
For brevity, the main conclusions of the error analysis in terms of
the motion profile, pressure measurements, and PIV are provided in
this section. In particular, the error will be described by 95%
confidence intervals (abbreviated as “ci” or “CI” in the figures) for
Re  300;000 and k  0.05.
For a majority of the data presented herein, the desired motion is
described by k  0.05, am  10 deg, and Δα  10 deg. The actual
motion is described by k  0.05  0.00, αm  10.20  0.05 deg,
and Δα  10.03  0.02 deg. As such, the cycle-to-cycle repeat-
ability was approximately 0.07 deg.
Pressure measurements used pressure taps connected to a pressure
transducer via tubing. Empirical correlations were used to account for
the gain and phase delay introduced by the tubing length. The effect
of these empirical correlations, as well as the repeatability of the
motion profile, can be characterized by the confidence interval,
which is shown in Fig. 4, for each excitation Strouhal number. In this Fig. 4 Normalized confidence interval for Re
300;000 and k
0.05.
figure, the mean confidence interval is normalized by the baseline
(Ste  0) lift peak or the magnitude of the moment stall. As indicated
in the figure, the confidence intervals are nearly three orders of
magnitude lower than the lift peak or the magnitude of the moment
stall, which indicate high repeatability.
When analyzing the phase-locked PIV data, it is important to
understand the uncertainty of the data. Shown in Fig. 5 is the root
mean square of a particular phase where the flow is completely
separated. In this figure, it is observed that the root mean square is
on the order of the freestream velocity in the separation region.
Three distinct processes could cause this: velocity fluctuations
in the separation region, repeatability error in the airfoil motion,
and error inherent in the PIV process. The latter two can be
characterized by the 95% confidence interval of the data, which is
shown in Fig. 6. The confidence interval is an order of magnitude
smaller than the root mean square (Fig. 5), indicating that the
majority of the fluctuations are indeed flow related and not an Fig. 5 Root mean square of the streamwise velocity at Re
300;000
artifact of the experimental setup. and k
0.05.
 SINGHAL ET AL. 81

PIV results were acquired at one combination of Reynolds number

and reduced frequency. This case was Re  300;000 and k  0.05.
PIV data were obtained for only two excitation Strouhal numbers
cases (Ste  0.3 and 9.9), in addition to the baseline case. These
conditions were chosen as representative cases after an analysis of the
pressure data. Phase-locked (to the airfoil motion) PIV data were
collected at 17 phases for the baseline case and 9 phases for the
excited cases.

A. Baseline Results
Figure 8 shows the baseline phase-averaged lift and moment
coefficients. In this figure, the five stages of dynamic stall, as outlined
by Corke and Thomas [4], are numbered. The process of dynamic
stall begins as the airfoil starts pitching up. In this stage, the flow is
Fig. 6 Confidence interval of the streamwise velocity at Re
300;000 attached and the lift increases steadily with α. When α exceeds the
and k
0.05. static stall α, the lift continues to increase; this is the second stage of
Downloaded by KUNGLIGA TEKNISKA HOGSKOLEN KTH on December 31, 2018 | http://arc.aiaa.org | DOI: 10.2514/1.J056071

dynamic stall. In literature, the delayed separation is attributed to the

airfoil motion and/or the formation of a closed separation bubble near
the leading edge of the airfoil [4].
IV. Results and Discussion In the second stage, the vorticity at the leading edge accumulates
and the dynamic stall vortex begins to form. This accumulation of
Before examining the effect of excitation, the baseline flow results
vorticity is better visualized by the PIV results, shown in Fig. 9. In this
were documented. Although the actuators are thin (0.0022c), the
figure, the phase of motion ϕ and angle of attack are shown in the
physical presence of the actuator near the leading edge does impact bottom left of each image.
the flow. Shown in Fig. 7 are lift and moment curves for two different As α increases, the airfoil undergoes the third stage of dynamic
cases at Re  167;000. The red curves indicate that the actuator was stall [4]: the ejection of this vorticity in the form of the dynamic stall
mounted on the airfoil (but not active). The blue curves are the cases vortex. The ejection of this vortex is attributed to vortex-induced
without an actuator present. In this figure, the motion is described by separation [18]. As observed by Mulleners and Raffel [18], the
αm  10 deg, Δα  10 deg, and k  0.050. As demonstrated by dynamic stall vortex formation is accompanied by counter-rotating
the plots, the lift peak and moment stall due to the dynamic stall vortex vortices on the airfoil’s surface. These vortices travel upstream,
are slightly delayed. Furthermore, the moment stall is weakened by the pushed by the dynamic stall vortex, resulting in the detachment of the
presence of the actuator. Thus, excited data will be compared with data dynamic stall vortex. This phenomenon is termed vortex-induced
acquired with the actuator mounted to the airfoil but turned off separation [19]. Vorticity data clearly indicate the presence of
(referred to as baseline or Ste  0) throughout this paper. counter-rotating vorticity near the surface (Fig. 10). Swirling strength
Flows were investigated for various Reynolds numbers (Re  results, shown in Fig. 11, also show the two sets of vortices. Figure 11
167;000; 300;000; 500,000) and reduced frequencies (k  0.025; also shows the dynamic stall vortex convecting over the airfoil at
0.05; 0.075). Twenty excitation Strouhal numbers were tested for ϕ  73.5 deg (α  19.6 deg). However, the vortex does not appear
each combination of Reynolds number and reduced frequency. The to be well defined, due to cycle-to-cycle variation in the phase of the
only motion profile tested was a sinusoidal waveform with α m  vortex.
10 deg and Δα  10 deg. These parameters place the motion The shape of the dynamic stall vortex in the PIV results can be
within the regime of deep stall. The results presented herein are at explained by the results in Fig. 12. The figure illustrates the
Re  300;000 and k  0.05. The trends for each combination of suction side pressure coefficient as it varies chordwise (horizontal
Reynolds number and reduced frequency are similar with respect to axis) and in time (vertical axis). Vortices produce pressure spikes
the excitation Strouhal number, and no significant information is lost that move along the chord and can be identified [20] by red streaks
by the omission of the other combinations of Reynolds number and in the surface pressure. The dynamic stall vortex is shown by the
reduced frequency. green dashed line on the left plot. On the right plot, the phase

Fig. 7 Lift and moment curves at Re
167;000 and k
0.05 without an actuator (blue) and with a plasma actuator mounted on the airfoil but not
turned on (red). The darker and lighter colors represent pitch-up and pitch-down motion, respectively.
 82 SINGHAL ET AL.
Downloaded by KUNGLIGA TEKNISKA HOGSKOLEN KTH on December 31, 2018 | http://arc.aiaa.org | DOI: 10.2514/1.J056071

Fig. 8 Baseline phase-averaged lift and moment coefficients at Re
300;000 and k
0.050.

Fig. 9 Close up of baseline phase-averaged vorticity maps at Re
300;000 and k
0.050.

Fig. 10 Baseline phase-averaged normalized vorticity at Re
300;000 and k
0.050.

ϕ  73.5 deg (α  19.6 deg) is shown by the blue dashed line, simultaneously. After the dynamic stall vortex has convected, the
which does not exactly follow the red streak. This supports the airfoil is fully stalled. This marks the fourth stage of the dynamic stall
conclusion that the dynamic stall vortex is convecting over the process [4] and corresponds to the sharp drop in the lift observed in
airfoil as it is pitching. However, the width of the red streak, not Fig. 8. The final stage of dynamic stall is flow reattachment. The flow
being narrow and sharp, indicates that the phase at which the reattachment is a stochastic process, in which random features of the
dynamic stall vortex convects is not consistent from cycle-to- shear layer result in large cycle-to-cycle variation in aerodynamic
cycle. This is attributed to the stochastic nature of dynamic stall loads during recovery [4,22,23]. The time of reattachment can be
process. determined from the pressure data.
The convection of the dynamic stall vortex results in an additional
increase in lift, as shown by the lift peak near α  19 deg in Fig. 8.
The vortex shedding also results in a sharp decrease in pitching B. Excited Results
moment, because the vortex moves the airfoil center of pressure After characterizing the baseline flow, a sweep of excitation
downstream as it convects [21]. It is known that moment stall Strouhal numbers was used to perform an assessment of the NS-DBD
precedes lift stall [4]; however, due to the limited temporal and spatial actuator’s potential as a dynamic stall flow control device. Figure 13
resolution of the pressure taps, they appear to occur nearly displays the phase-averaged lift coefficient for the baseline and
 SINGHAL ET AL. 83
Downloaded by KUNGLIGA TEKNISKA HOGSKOLEN KTH on December 31, 2018 | http://arc.aiaa.org | DOI: 10.2514/1.J056071

Fig. 11 Baseline phase-averaged normalized swirling strength at Re
300;000 and k
0.050.

Fig. 12 Baseline suction side pressure coefficient at Re
300;000 and k
0.050, showing the (left) dynamic stall vortex convection time and (right) PIV
phase acquisition time (right).

excitation results for Re  300;000 and k  0.05. Although a total 1. Effects of Excitation in the Stalled Regime
of 20 different excitation Strouhal numbers were tested only four The first trend, that excitation introduces oscillatory behavior into
(Ste  0; 0.3; 0.96, and 9.9), which are representative, will be shown the stalled regime of dynamic stall, is indicative of the formation of
to keep the figures uncluttered. large coherent flow structures in the shear layer over the separated
The trend of the effects of excitation on the overall progression of region. As shown by previous work with the actuators [9,11,12], low
dynamic stall appears to be similar for different combinations of Strouhal number excitation leads to the formation of coherent
Reynolds number and reduced frequencies. Each combination structures at poststall α. If the excitation pulse timing is plotted with
follows three primary trends: the moment and lift curves for the low excitation Strouhal number
1) Low excitation Strouhal numbers (Ste < 0.5) result in case, it is easy to observe that every pulse results in the formation of a
oscillatory lift and drag following the dynamic stall vortex shedding. structure (Fig. 14). The static stall research previously conducted in
This behavior smooths out as the excitation Strouhal number our laboratory has demonstrated that the flow is receptive to
increases. excitation over a range of Strouhal numbers [13]. This is also
2) All excited cases result in earlier flow reattachment than the observed in the unsteady flow results, particularly at low excitation
baseline. Strouhal numbers, where the large structures were also detected using
3) Excited cases (in general) result in reduced lift and moment PIV. The swirling strength for the stalled phases at a low (left) and
hysteresis and decreased dynamic stall vortex strength. high (right) Strouhal number excitation is shown in Fig. 15. At high
These trends will be further discussed in the next three sections. excitation Strouhal numbers, this oscillatory behavior is not
 84 SINGHAL ET AL.
Downloaded by KUNGLIGA TEKNISKA HOGSKOLEN KTH on December 31, 2018 | http://arc.aiaa.org | DOI: 10.2514/1.J056071

Fig. 13 Phase-averaged lift and moment coefficients for various excitation Strouhal numbers at Re
300;000 and k
0.050.

Fig. 14 Phase-averaged lift and moment coefficients for Ste
0.3 at Re
300;000 and k
0.050.

Fig. 15 Phase-averaged normalized swirling strength at Re
300;000 and k
0.050 for (left) Ste
0.35 and (right) Ste
9.9.

observed. This is expected given that higher excitation Strouhal number case, one can observe a large structure near 40% chord over
numbers result in the formation of smaller structures. These the airfoil. In the high excitation Strouhal number case, there are
structures also form further upstream on the airfoil and quickly several smaller structures convecting over the airfoil. It also bears
develop, disintegrate, and dissipate [9,11,12]. This too is well mentioning that, due to the limited spatial and temporal resolution of
captured by PIV data, shown in Fig. 15. In the low excitation Strouhal pressure measurements, high-frequency events would be filtered out.
 SINGHAL ET AL. 85

Thus, although similar signature oscillations, but with much lower formation of coherent structures entrains high-momentum flow from
amplitude, may exist for the high Strouhal number cases, they would the freestream and results in momentary reattachment [9,11,12]. The
not manifest in these data. formation of small structures (such as those produced by high
excitation Strouhal numbers) only results in partial reattachment
2. Effect of Excitation on Reattachment [9,11,12]. Earlier flow reattachment reduces the aerodynamic
The second trend, that excitation results in earlier reattachment, is hysteresis.
easily observed in Fig. 13. It can be seen by the increased lift on the In Fig. 16, the reattachment α (which is defined as α such that the
pitch-down motion of the airfoil, the lift on the pitch-down motion for coefficient of lift is equal on the down- and upstroke) is plotted as a
excited cases exceeds the lift on the pitch-up motion sooner than the function of the excitation Strouhal number. At low excitation
baseline case. This phenomenon is related to the first trend. The Strouhal numbers (Ste < 2), there is a stronger dependence between
the excitation Strouhal number and the reattachment α. This is
attributed to the timing of the actuator pulse relative to the motion of
the airfoil. If vortex-induced flow attachment does not occur before
the airfoil α is less than the critical α, then reattachment does not occur
until the next pulse. Thus, timing is especially critical at low
excitation Strouhal numbers, because the period between pulses is
Downloaded by KUNGLIGA TEKNISKA HOGSKOLEN KTH on December 31, 2018 | http://arc.aiaa.org | DOI: 10.2514/1.J056071

relatively large. However, at high excitation Strouhal numbers

(Ste > 6), the flow remains attached, and thus there is less
dependence between the reattachment α and the Strouhal numbers.
Flow attachment is best observed by the suction side pressure
coefficient. Shown in Fig. 17 is the suction side pressure coefficient
for the baseline (left) and Ste  9.9 (right). In the baseline case, from
20 deg ↑ to 15 deg ↓ (the stalled stage of dynamic stall), the pressure
coefficient is relatively constant with respect to the chordwise
position. When excitation is applied (right figure), we can observe the
emergence of a pressure peak (indicated by the dark colors near
x∕c  0) in the stalled stage of dynamic stall, indicating that the flow
is attached.

3. Effect of Excitation on Aerodynamic Hysteresis

The final primary trend is that excitation results in decreased lift
and moment hysteresis and decreased dynamic stall vortex strength.
The first part of this trend is related to the generation of coherent
structures in the stalled stage of dynamic stall and earlier flow
reattachment: The excitation perturbations grow and roll up into
large-scale structures, which entrain high-momentum freestream
fluid into the separated region, energizing and reattaching the flow.
Fig. 16 Reattachment α versus the excitation Strouhal number at Re
This results in increased lift production and thus reduces the lift
300;000 and k
0.050. hysteresis for all excitation Strouhal numbers. At high excitation

Fig. 17 Suction side pressure distribution at Re
300;000 and k
0.050 for (left) Ste
0 baseline and (right) Ste
9.9.
 86 SINGHAL ET AL.

coefficient is reduced, particularly at much higher excitation

Strouhal numbers. This is attributed to the constant partial reattach-
ment seen in high Strouhal number excitation, as well as the
decreased strength of the dynamic stall vortex. As the strength of
the dynamic stall vortex decreases, the magnitude of the moment
peak also decreases [6], further reducing the magnitude of the
negative damping coefficient.
Phase-locked PIV measurements were used to understand why the
strength of the dynamic stall vortex was decreased by excitation.
Figure 20a shows the normalized swirling strength for various phases
(at angle of attack near the peak lift) at Re  300;000 and k  0.05
for the pitch-up motion. The images are ordered by the excitation
Strouhal number columnwise, that is, the left, middle, and right
columns contain the data for Ste  0, Ste  0.35, and Ste  9.9,
respectively. Baseline lift curves shown in Fig. 20b provide a
reference for when the PIV data were acquired with respect to the lift
Downloaded by KUNGLIGA TEKNISKA HOGSKOLEN KTH on December 31, 2018 | http://arc.aiaa.org | DOI: 10.2514/1.J056071

Fig. 18 Negative damping (red) and positive damping (green) shown on and moment coefficients.
moment coefficient curve for Re
300;000, k
0.05, and Ste
0. As discussed previously, the ejection of the dynamic stall vortex
is preceded by the accumulation of vorticity at the leading edge. This
is shown by the high swirling strength in Fig. 20 at Ste  0 and
ϕ  44.5; 53.0; 64.2 deg (α  17–19 deg). At ϕ  73.5 deg
(α  19.6 deg), it is seen that the dynamic stall vortex has been
ejected and is convecting downstream. Because of the unsteady
nature of dynamic stall, the vortex is not well defined. However, when
coupled with pressure results, it is apparent that the vortex is
convecting over the airfoil at this angle of attack. Compared with the
baseline case, the accumulation of vorticity is much smaller in the
excited cases, particularly for the high excitation Strouhal number.
At Ste  9.9, this accumulation of vorticity is replaced by a stream of
small, coherent vortices, which appear to remove the accumulating
vorticity near the leading edge. Thus, it follows that, as the number of
vortices increases (i.e., as Ste increases), more vorticity is removed
from the leading edge and the dynamic stall vortex should be weaker
(as observed in the results).
To an extent, this is also observed in Fig. 20 at Ste  0.35 and
ϕ  73.5 deg (α  19.6 deg), where the swirling strength of the
dynamic stall vortex is considerably weaker than its baseline
counterpart. However, at the highest excitation Strouhal number
(Ste  9.9), the formation and ejection of the dynamic stall vortex
is no longer observed. Although one might think that the dynamic
Fig. 19 Negative damping coefficient Ξ− versus the excitation Strouhal stall vortex could have shed between the captured phases, suction
number Ste for Re
300;000 and k
0.05. side pressure coefficient data (Fig. 21) suggests that this is not the
case. In Fig. 21, vortex shedding is marked in the figure by green
lines for the dynamic stall vortex and purple lines for the vortices
Strouhal numbers, when the vortices are small and the induced resulting from excitation. At the high excitation Strouhal number
unsteady loading due to the convection and shedding of vortical case, the pressure spikes associated with vortex convection are not
structures is small, the moment hysteresis is significantly reduced, as observed. Thus, the high Strouhal number excitation seems to have
shown in Fig. 13. Note that, at low excitation Strouhal numbers, the depleted the leading-edge vorticity, completely suppressing the
moment hysteresis is larger than at high excitation Strouhal numbers, dynamic stall vortex.
due to the formation of large structures. Closer examination of pressure results (Fig. 13) indicates that there
An important measure of the aerodynamic hysteresis is the is a fairly monotonic relationship between the excitation Strouhal
damping coefficient. In particular, regions of negative damping (in number and the lift peak, at least in the limited cases shown in the
the stalled stage of dynamic stall) can lead to aerodynamic flutter [4]. figure. This is consistent with the prior discussion of the formation of
Negative damping is associated with clockwise loops in the moment vortices, namely, that the formation of structures due to excitation
coefficient curve [4], as shown in Fig. 18. The damping coefficient is removes accumulated vorticity from the leading edge, thereby
defined as the closed-loop integral of the moment coefficient curve. weakening the dynamic stall vortex and reducing the magnitude of
However, to understand the effect of the actuation on the negative the lift and moment peaks.
damping, the integration is only performed over the closed loop of the Because of the poor temporal resolution in the phase-locked PIV
clockwise trajectory of the moment coefficient. This quantity is data, it is difficult to directly determine the strength of the dynamic
referred to as the negative damping coefficient Ξ− . stall vortex. As a simple alternative, the reduction in the magnitude of
Shown in Fig. 19 is the negative damping coefficient versus the moment stall due to excitation is shown in Fig. 22. The reduction in
excitation Strouhal number. It is observed that low excitation the moment coefficient is normalized by the baseline value. The
Strouhal numbers result in a worse negative damping coefficient than figure indicates that, in general, as the excitation Strouhal number
the baseline. As indicated by Fig. 13, low Strouhal number excitation increases, the magnitude of the peak moment coefficient decreases.
results in the formation of large structures that cause large-amplitude This continues until the formation of the dynamic stall vortex is
unsteadiness during the pitch down of the airfoil motion. This suppressed, at which point the moment coefficient value seems to
unsteadiness increases the magnitude of the negative damping plateau with respect to the excitation Strouhal number, as the limited
coefficient. However, as the excitation Strouhal number increases points show in the figure. However, a more direct metric should be
beyond around Strouhal number of 0.3 (the natural shedding obtained with time-resolved PIV, in which the dynamic stall vortex
Strouhal number), the magnitude of the negative damping and its boundaries could be better defined.
 SINGHAL ET AL. 87
Downloaded by KUNGLIGA TEKNISKA HOGSKOLEN KTH on December 31, 2018 | http://arc.aiaa.org | DOI: 10.2514/1.J056071

Fig. 20 Phase-averaged normalized swirling strength at Re
300;000 and k
0.050.

Fig. 21 Suction side pressure coefficient results for Re
300;000, k
0.05, and various excitation Strouhal numbers.
 88 SINGHAL ET AL.

attributed to the decreased vorticity accumulation at the leading edge

during the pitch-up motion. Excitation results in the formation of
vortices before the ejection of the dynamic stall vortex. These vortices
remove some of the accumulated vorticity. At high excitation
Strouhal numbers, this results in the suppression of the dynamic stall
vortex.
Although the experimental results clearly demonstrate the efficacy
of the actuators, reducing the sensor noise and uncertainty in the
repeatability of the airfoil motion requires that a new mechanism
should be designed and built for improving the results. Higher phase
resolution in the phase-locked PIV will provide higher data fidelity
and may reveal a well-defined relationship between the excitation
Strouhal number and the dynamic stall vortex strength. Phase-locked
PIV data collected at other combinations of Reynolds numbers and
reduced frequencies will better support the generalization of trends
presented in this paper.
Downloaded by KUNGLIGA TEKNISKA HOGSKOLEN KTH on December 31, 2018 | http://arc.aiaa.org | DOI: 10.2514/1.J056071

Fig. 22 Reduction in moment coefficient peak versus excitation

Strouhal number at Re
167;000 and k
0.05. Acknowledgments
The support of this research by the U.S. Air Force Research
V. Conclusions Laboratory Collaborative Center for Aerospace Sciences at the Ohio
State Aerospace Research Center with Miguel Visbal is greatly
Dynamic stall is present in many applications, including rotorcraft,
appreciated. The fellowship supports for Achal Singhal from the
sharply maneuvering fixed-wing aircraft, and wind turbines.
NASA/Ohio Space Grant Consortium and for David Castañeda from
Associated unsteadiness and the potential for aerodynamic flutter has the Fulbright-Colciencias Convocatoria from Colombia are also
made it a problem of interest to the aerodynamics and flow control greatly appreciated.
communities. Nanosecond pulse-driven dielectric barrier discharge
(NS-DBD) plasma actuators have previously demonstrated the
ability to reattach separated flow in a variety of static stall conditions. References
This work, which is a first step, has sought to explore the use of these [1] Leishman, J. G., Principles of Helicopter Aerodynamics, 2nd ed.,
actuators to control dynamic stall in an airfoil and to provide a better Cambridge Univ. Press, Cambridge, England, U.K., 2006, pp. 423–464.
understanding of dynamic stall flow physics. [2] McCroskey, W. J., “The Phenomenon of Dynamic Stall,” NASA TM
A single NS-DBD plasma actuator was mounted to the airfoil 81264, Moffett Field, CA, 1981.
near the leading edge, and baseline pressure for αm  10 deg and [3] Ham, N. D., “Aerodynamic Loading on a Two-Dimensional Airfoil During
Δα  10 deg was taken for each combination of three Reynolds Dynamic Stall,” AIAA Journal, Vol. 6, No. 10, 1968, pp. 1927–1934.
numbers and three reduced frequencies. Pressure data were acquired doi:10.2514/3.4902
[4] Corke, T. C., and Thomas, F. O., “Dynamic Stall in Pitching Airfoils:
for all nine combinations of Reynolds number and reduced frequency. Aerodynamic Damping and Compressibility Effects,” Annual Review of
Phase-locked particle image velocimetry (PIV) measurements were Fluid Mechanics, Vol. 47, No. 1, 2015, pp. 479–505.
acquired at Re  300;000 and k  0.05. The baseline pressure and doi:10.1146/annurev-fluid-010814-013632
PIV results compare well with the stages of dynamic stall in the [5] Joo, W., Lee, B., Yee, K., and Lee, D., “Combining Passive Control
literature: attached flow, the accumulation of vorticity at the leading Method for Dynamic Stall Control,” Journal of Aircraft, Vol. 43, No. 4,
edge, the ejection of the accumulated vorticity in the form of the 2006, pp. 1120–1128.
dynamic stall vortex, stalled flow, and reattachment. PIV results doi:10.2514/1.17957
[6] Heine, B., Mulleners, K., Jourbert, G., and Raffel, M., “Dynamic Stall
indicated that the ejection of the dynamic stall vortex was due to
Control by Passive Disturbance Generators,” AIAA Journal, Vol. 51,
vortex-induced separation, a counterclockwise rotating stream of No. 9, 2013, pp. 2086–2097.
vorticity generated over the airfoil, moving upstream during the pitch- doi:10.2514/1.J051525
up motion and resulting in the ejection of the dynamic stall vortex. [7] Weaver, D., McAlister, K. W., and Tso, J., “Suppression of Dynamic
Over all the combinations of Reynolds numbers and reduced Stall by Steady and Pulsed Upper-Surface Blowing,” NASA TP 3600,
Strouhal numbers, the observed trends were remarkably similar. Moffett Field, CA, 1996.
From these results, three major trends were observed. The first was [8] Post, M. L., and Corke, T. C., “Separation Control Using Plasma
that excitation at low Strouhal number results in the formation of Actuators: Dynamic Stall Vortex Control on Oscillating Airfoil,” AIAA
Journal, Vol. 44, No. 12, 2006, pp. 3125–3135.
large-scale flow structures in the stalled regime, which result in doi:10.2514/1.22716
unsteadiness in the lift and moment curves. As the excitation Strouhal [9] Little, J., Takashima, K., Nishihara, M., Adamovich, I., and Samimy,
number increases, this unsteadiness decreases and eventually M., “Separation Control with Nanosecond-Pulse-Driven Dielectric
disappears. As observed in static stall flow control experiments with Barrier Discharge Plasma Actuators,” AIAA Journal, Vol. 50, No. 2,
NS-DBD plasma actuators, low Strouhal number excitation results in 2012, pp. 350–365.
the formation of large coherent structures, whereas high Strouhal doi:10.2514/1.J051114
number excitation results in the formation of small structures that [10] Rethmel, C., Little, J., Takashima, K., Sinha, A., Adamovich, I., and
quickly develop, break down, and result in only partial reattachment Samimy, M., “Flow Separation Control Using Nanosecond Pulse
Driven DBD Plasma Actuators,” International Journal of Flow Control,
of the flow. Thus, high Strouhal number excitation results in Vol. 3, No. 4, 2011, pp. 213–232.
smoother lift and moment coefficient curves. doi:10.1260/1756-8250.3.4.213
In a second trend, excitation also results in earlier reattachment due [11] Clifford, C., Singhal, A., and Samimy, M., “Flow Control over an Airfoil
to the formation of coherent structures that entrain the high- in Fully Reversed Condition Using Plasma Actuators,” AIAA Journal,
momentum freestream flow and reattach the flow. Once α of the Vol. 54, No. 1, 2016, pp. 141–149.
airfoil decreases past the static stall angle, the flow remains attached. doi:10.2514/1.J054157
Last, excitation results in the reduction of the lift and moment [12] Esfahani, A. G., Singhal, A., Clifford, C., and Samimy, M., “Flow
Separation Control over a Boeing Vertol VR-7 Using NS-DBD Plasma
hysteresis. This is reflected in the increased damping coefficient. This
Actuators,” 54th AIAA Aerospace Sciences Meeting, AIAA Paper
is partially due to the reattachment of the flow in the stalled regime. 2016-0843, 2016.
The other major factor is the decreased dynamic stall vortex strength, [13] Singhal, A., “Unsteady Flow Separation Control over a NACA 0015
which subsequently results in a decreased peak lift and moment Using NS-DBD Plasma Actuators,” M.S. Thesis, Dept. of Mechanical
coefficients. The decreased strength of the dynamic stall vortex is and Aerospace Engineering, The Ohio State Univ., Columbus, OH, 2016.
 SINGHAL ET AL. 89

[14] Samimy, M., Kim, J. H., Kastner, J., Adamovich, I., and Utkin, Y., [20] Frankhouser, M. W., “Nanosecond Dielectric Barrier Discharge Plasma
“Active Control of High-Speed and High-Reynolds-Number Jets Using Actuator Flow Control of Compressible Dynamic Stall,” M.S. Thesis,
Plasma Actuators,” Journal of Fluid Mechanics, Vol. 578, May 2007, Dept. of Mechanical and Aerospace Engineering, The Ohio State Univ.,
pp. 305–330. Columbus, OH, 2015.
doi:10.1017/S0022112007004867 [21] Harris, F. D. J., Tarzanin, F. J. J., and Fisher, R. K., “Rotor High Speed
[15] Samimy, M., Webb, N., and Crawley, M., “Excitation of Free Shear- Performance, Theory Versus Test,” Journal of the American Helicopter
Layer Instabilities in High-Speed Flows,” AIAA Journal (in review). Society, Vol. 15, No. 3, 1970, pp. 35–44.
[16] Greenblatt, D., and Wygnanski, I., “Dynamic Stall Control by Periodic doi:10.4050/JAHS.15.35
Excitation: Part 1: NACA 0015 Parametric Study,” Journal of Aircraft, [22] Shih, C., Lourenco, L. M., van Dommelen, L. L., and Korthapalli, A.,
Vol. 38, No. 3, 2001, pp. 430–438. “Unsteady Flow past an Airfoil Pitching at a Constant Rate,” AIAA
doi:10.2514/2.2810 Journal, Vol. 30, No. 5, 1992, pp. 1153–1161.
[17] Adrian, R. J., Christensen, K. T., and Liu, Z. C., “Analysis and doi:10.2514/3.11045
Interpretation of Instantaneous Turbulent Velocity Fields,” Experiments [23] Wernert, P., Geissler, W., Raffel, M., and Kompenhans, J.,
in Fluids, Vol. 29, No. 3, 2000, pp. 275–290. “Experimental and Numerical Investigations of Dynamic Stall on a
doi:10.1007/s003489900087 Pitching Airfoil,” AIAA Journal, Vol. 34, No. 5, 1996, pp. 982–989.
[18] Mulleners, K., and Raffel, M., “The Onset of Dynamic Stall Revisited,” doi:10.2514/3.13177
Experiments in Fluids, Vol. 52, No. 3, 2012, pp. 779–793.
doi:10.1007/s00348-011-1118-y A. Naguib
Downloaded by KUNGLIGA TEKNISKA HOGSKOLEN KTH on December 31, 2018 | http://arc.aiaa.org | DOI: 10.2514/1.J056071

[19] Obabko, A., and Cassel, K., “Detachment of the Dynamic Stall Vortex Associate Editor
Above a Moving Surface,” AIAA Journal, Vol. 40, No. 9, 2002,
pp. 1811–1822.
doi:10.2514/2.1858
 This article has been cited by:

1. Stuart I. Benton, Miguel R. Visbal. Evaluation of Thermoacoustic-based Forcing for Control of Dynamic Stall . [Citation]
[PDF] [PDF Plus]
2. Mo Samimy, Nathan Webb, Michael Crawley. 2018. Excitation of Free Shear-Layer Instabilities for High-Speed Flow
Control. AIAA Journal 56:5, 1770-1791. [Abstract] [Full Text] [PDF] [PDF Plus]
3. Hiroaki Fukumoto, Hikaru Aono, Taku Nonomura, Akira Oyama, Kozo Fujii. Large-Eddy Simulations of Flow Control
Effects of a DBD Plasma Actuator at Various Burst Frequencies on a Dynamic Flowfield around a Pitching NACA0012
Airfoil at Reynolds Number of 256,000 . [Citation] [PDF] [PDF Plus]
4. Claudia Durasiewicz, Ashish Singh, Jesse C. Little. A Comparative Flow Physics Study of Ns-DBD vs Ac-DBD Plasma
Actuators for Transient Separation Control on a NACA 0012 Airfoil . [Citation] [PDF] [PDF Plus]
Downloaded by KUNGLIGA TEKNISKA HOGSKOLEN KTH on December 31, 2018 | http://arc.aiaa.org | DOI: 10.2514/1.J056071