momentum transfer and flow induction in a dielectric barrier discharge plasma actuator

During active momentum transfer in the acceleration region, the periodic inflation of the wall jet thickness was synchronized with the decreasing phase of the applied high-voltage input,

T Abe, M Takagaki, and K Yamada

Citation: AIP Advances 2, 042150 (2012); doi: 10.1063/1.4768802

View online: http://dx.doi.org/10.1063/1.4768802

View Table of Contents: http://aip.scitation.org/toc/adv/2/4

Published by the American Institute of Physics

AIP ADVANCES 2, 042150 (2012)

Momentum transfer and flow induction in a dielectric

barrier discharge plasma actuator

T Abe,1M Takagaki,2and K Yamada1

1Institute of Space and Astronautical Science/JAXA, 3-1-1, Yoshinodai, Sagamihara,

Kanagawa 229, Japan

2Tohoku University, Miyagi 980-8579, Japan

(Received 9 March 2012; accepted 9 November 2012; published online 21 November 2012)

The flow induction process resulting from a dielectric barrier discharge plasma ac-tuator was experimentally investigated by a flow visualization method This method enables a detailed examination of the flow induction process The overall behavior

as well as the detailed temporal and spatial behavior of the flow induction process was clarified The induced flow originated from the vicinity of an exposed electrode edge, and similar to a wall jet, it smoothly spread over the surface until a vortex was generated at the tip of the wall jet Even after the vortex disappeared, the wall jet continued to extend This wall jet was classified into two distinct regions The first region contained the wall jet extending from the electrode edge to the location where the vortex was generated In this region, active momentum coupling occurred; therefore, the flow was actively induced During active momentum transfer in the acceleration region, the periodic inflation of the wall jet thickness was synchronized with the decreasing phase of the applied high-voltage input, which suggested differ-ent mechanisms of momdiffer-entum transfer during the increasing and decreasing phases

of the high-voltage input The second region was an extension of the first one and was

formed through the inertia of the flow originating from the first region Copyright

2012 Author(s) This article is distributed under a Creative Commons Attribution 3.0 Unported License [http://dx.doi.org/10.1063/1.4768802]

I INTRODUCTION

In recent years, surface dielectric barrier discharge (DBD) plasma actuators have attracted much attention, because they can be applied to flow controllers without moving parts Furthermore, consid-erable experimental and theoretical efforts have been made to understand the working mechanism of these actuators.1 8In a surface DBD plasma actuator, a surface electric discharge is generated under

an atmospheric condition, and this leads to momentum transfer to the ambient gas Increasing the effectiveness of momentum transfer is essential for enhancing the performance of plasma actuators Therefore, we must understand the mechanism of momentum transfer Many studies have already attempted this However, the mechanism of momentum transfer needs to be investigated further Recent studies related to DBD plasma actuators, including those on their working mechanism and applications, were comprehensively reviewed by Moreau.9

Theoretically, it is believed that the collision of ions in the discharge area with neutral molecules

in the ambient gas is the main process of momentum transfer to neutral molecules and flow induction

in the ambient gas Among many theoretical models for the DBD plasma actuator, a fluid model proposed by Boeuf1is attractive because of its simplicity, and in fact Nishida and Abe have showed that the three fluid model (the electron, a positive ion and a negative ion) could simulate the characteristics of the discharge plasma structure and the thrust force (the averaged body force field) observed in previous experimental studies.18

Experimentally, the momentum transfer process has been investigated using several phenomena related to momentum transfer The behavior of electric discharge in a DBD plasma actuator was investigated, and its temporal behavior was clarified.4 , 8 These studies are important because it is

believed that momentum transfer is caused by electric discharge Furthermore, the acoustic wave emission associated with a DBD plasma actuator was investigated.6 , 10 The intense acoustic noise emission in a DBD plasma actuator can be caused by momentum transfer to the ambient gas; therefore, information on momentum transfer can be obtained by analyzing such emissions In the investigation, the relationship between the acoustic pressure wave and the high-voltage input responsible for the electric discharge was clarified In particular, it was found that an intense acoustic wave is not generated uniformly during the entire phase of the alternating high voltage, but it is generated specifically in the decreasing phase of the high-voltage input.10

Recently, Enloe et al measured the temporal behavior of the total force produced by a plasma

actuator as a result of the momentum transfer to the ambient gas.11Their result clearly showed that the total force produced is not uniformly generated during the increasing phase of the alternating high voltage but depends on the phase of the alternating high-voltage input.12 This dependence

is qualitatively in agreement with the measurement result of acoustic waves that also exhibit a dependence on the phase of the high-voltage input.10

Regarding the flow induction caused by momentum coupling, Balcon et al clarified the overall

development of the induced flow using the particle image velocimetry (PIV).13According to their result, a wall jet originating from an exposed electrode edge develops along the surface on which the exposed electrode is mounted, and a vortex is ejected before a steady-state wall jet is established

A similar behavior can also be observed even in a DBD plasma actuator driven by repetitive high-voltage nanosecond pulses with a DC or low-frequency sinusoidal bias.14In addition to the overall development, the non-stationary local flow velocity in the wall jet was measured, and it was suggested that momentum coupling depends on the phase of the alternating high voltage; that is, the local flow speed along the surface varies synchronizing with the alternating high voltage.15 – 17

In the present study, we attempt to clarify the momentum transfer process from a viewpoint

of the flow induction To this end, we study the flow induction process from an initial to a nearly steady state using a high-speed schlieren imaging technique In particular, we focus on the temporal behavior of the flow induction (its variation with the phase of the alternating high voltage input) because this may provide valuable information about the relationship between momentum coupling and the manner of electric discharge

II EXPERIMENTAL SETUP

The entire experimental setup is shown in Fig.1 To investigate the influence of the ambient gas pressure on a DBD plasma actuator, the actuator is placed in an evacuation chamber with a diameter and width of 1.01 m and 1.15 m, respectively In the present experiment, we use air as the ambient gas The ambient gas pressure for the DBD plasma actuator can be varied from 1000 hPa

to 400 hPa by evacuating the chamber The present plasma actuator consists of a pair of electrodes staggered in parallel with a dielectric plate between them, as shown in Fig.2 The dielectric plate

is composed of glass epoxy (glass-fiber-reinforced epoxy) and has a chord, span, and thickness of

72, 200, and 1.8 mm, respectively Identical pieces of copper foil with a chord, span, and thickness

of 15 mm, 100 mm, and 40μm, respectively, are typically used as electrodes The rear electrode is

covered with a multi-layered Kapton tape to prevent electric discharge, whereas the front electrode

is exposed to the ambient gas There is no gap between the staggered electrodes An alternating high voltage that is generated by amplifying a low-voltage signal using a high-voltage amplifier (10/40A; Trek Inc.) is applied across the electrodes, whereas the covered electrode is electrically grounded

A ballast register is not used because the discharge in the present study is stable without one The low-voltage signal is generated using a low-voltage general-purpose digital signal generator that enables generating an alternating voltage with an arbitrary waveform The voltage applied to the electrodes and the electric current passing through the electrodes are measured using a high-voltage probe (P6015A9; Techtronix) and an electric current probe (701933; Yokogawa Electric Corp.), respectively All the data are recorded on a PC and are analyzed after the measurements

A high voltage having the three different types of triangular waveforms shown in Fig 3 is

applied to the DBD plasma actuator Here T pis a time interaval for one period The first waveform

is modified so that the decreasing part is steeper than the increasing part The second waveform is

042150-3 Abe, Takagaki, and Yamada AIP Advances 2, 042150 (2012)

FIG 1 Schematic of the experimental setup.

Insulato

Insulator

H.V.Electrod H.V.Electrode GND Electrode(Back Side GND Electrode(Back Side) Dielectric Plat

Dielectric Plate

FIG 2 Schematic of the DBD plasma actuator.

oppositely modified so that the increasing part is steeper than the decreasing part The last waveform

is a regular triangular waveform with equally steep increasing and decreasing parts The high voltage has a period and amplitude of 2 kHz and 10 kV, respectively

In the present experiment, we investigate the temporal development of the induced flow at an early phase of DBD plasma actuation in still air In other words, we observe its temporal development during the first few tens of AC cycles of the high-voltage input To this end, we measure the density variation associated with the flow induction using the high-speed schlieren imaging method The schlieren imaging system is shown Fig.1 A laser beam generator (diode-pumped solid-state green laser, Omicron Laserage Laserprodukte GmbH, model: FKLA800 green) acts as a light source The light from the generator is expanded by a beam expander and focused on the pinhole plate, and the light from the pinhole is reflected on a concave mirror to generate a parallel light beam The DBD plasma actuator is placed in the parallel light beam such that the span direction of the DBD plasma actuator is parallel to the light beam Therefore, the schlieren image produced by the present setup reflects the density variation integrated along the span-wise direction of the plasma actuator After

FIG 3 Waveforms of the high voltage.

passing through the DBD plasma actuator, the parallel light beam is refocused by another concave mirror At the focusing point, a knife edge is placed to hide a portion of the light and to generate a schlieren image The knife edge in the schlieren imaging system is set parallel to the DBD plasma actuator plate The light passing through the knife edge is introduced into an image-converter high-speed camera (Hyper Vision HPV-1, SHIMAZU) The size of the image captured using the camera

is 312× 260 pixels The frame rate can be set as 32, 64, 125, or 250 μs/frame; the exposure time for

all frame rates is set at 32μs, except for the frame rate of 32 μs for which the exposure time is set

at 16μs A total of 102 frames are captured for each record The camera recording is triggered by

the high-voltage signal applied on the DBD plasma actuator Therefore, sequential images from the very beginning of the DBD plasma actuator operation can be recorded A band pass filter is placed

in front of the camera to block all light except the laser beam

To visualize the flow induced by the DBD plasma actuator, the DBD plasma actuator is not treated in any special manner, but, slightly warmed air naturally generated by the electric discharge

042150-5 Abe, Takagaki, and Yamada AIP Advances 2, 042150 (2012)

FIG 4 Sequential photographs of schlieren images over the rst 44 cycles of the alternating high-voltage input.

is used.14 Such air can be visualized by the schlieren system because it has lower density and refractivity

III RESULTS

A Overall behavior and the acceleration region

The overall temporal behavior of the flow induced by the DBD plasma actuator is shown in Fig.4where sequential photographs of the schlieren images are shown Hereafter, it is assumed that the alternating high voltage with waveform 1 is applied and the ambient gas pressure is 1000 hPa, unless otherwise stated In the figures, a pair of the horizontal indicators are drawn to represent the vertical scale and the upper one of them specifies the location of the exposed electrode edge adjacent

to the buried electrode The elapsed time since the application of the high voltage is indicated under each photograph

In the sequence of the schlieren images shown in Fig.4, the bright region representing lighter air originates near the electrode edge and spreads over the surface underneath which another electrode

is buried It is expected that the lighter air generated near the electrode edge by the electric discharge moves along with the ambient gas Considering this streak behavior, we can conclude that the bright area observed in the schlieren images is the wall jet induced by the DBD plasma actuator and this wall jet includes a component of induction toward the surface Starting from the electrode edge, the wall jet continues to extend along the surface, while its vertical expansion remains rather limited; the thickness near the edge is around 0.5 mm and gradually increases along the cord-wise direction This behavior continues until around 2 ms At around 2 ms, the tip of the wall jet extends up to 2.5 mm away from the electrode edge and a vortex generation appears, as can be clearly observed at the tip of the wall jet shown in Fig.4

The vortex generation in the course of the wall jet development may be attributed to the halt of the flow acceleration appearing until then and the deceleration of wall jet at the front tip is caused

by the ambient gas accumulated in front of it Then, the vortex gradually grows, inducing a vortical flow around it, and moves away while weakening Even after the vortex disappears, the wall jet

FIG 5 Temporal development of the wall jet length.

continues to extend along the surface In the last photograph in the sequence, the wall jet length is approximately 10 mm, whereas the wall jet height gradually increases from approximately 0.4 mm

at the electrode edge to approximately 1 mm at the tip of the wall jet The temporal development

of the wall jet length that is determined from the schlieren images is shown in Fig.5 The wall jet length increases rapidly until around 2 m-sec but increases moderately afterwards The irregularity

in the wall jet length is observed around 2m-sec, which is caused by the emergence of the vortex The overall flow behavior described above is consistent with the previously reported results.13 , 14 Summarizing all the above-mentioned observations, we hypothesize that the physical region of the wall jet is classified into two distinct regions In the first region, the wall jet extends from the electrode edge to the location at which vortex generation starts This wall jet region forms until around 2 ms and extends to a distance of 2.5 mm from the electrode edge In the second region, the wall jet extends beyond the first region In this wall jet region, the flow speed may decrease, and consequently, the wall jet thickness gradually increases In the first region, it can be assumed that the flow is actively accelerated by the momentum transfer which is believed to occur in the electric discharge plasma In fact, as shown in Fig.6, the area of the first region almost coincides with the electric discharge area In the figure, the region of the wall jet observed before the vortex generation

is compared with the electric discharge area in various ambient gas pressure environment In the figure, still images of the discharge region are captured from a side view to make them comparable

to the schlieren images; these images are captured using a CCD camera with an exposure time of the order of a few seconds Note that both the discharge area and the first region of the wall jet increase as the pressure of the ambient gas decreases Hereafter we call the first region of the wall jet as an acceleration region

B Temporal characteristics of the wall jet

To critically observe the temporal behavior of the wall jet evolving during an early phase up to 6.4 ms, we investigate the streak camera images for the slit region perpendicular to the surface at several locations along the surface This early phase covers vortex emergence The streak camera images are constructed using the sequences of the schlieren images; that is, the shlieren image through a slit region set vertically to the surface at a certain location is stacked vertically in time line Figure7shows these images at several locations ranging from near the electrode edge to beyond the vortex generation location The location of the surface is indicated by the right vertical indicator of

042150-7 Abe, Takagaki, and Yamada AIP Advances 2, 042150 (2012)

FIG 6 Comparison of the wall jet with the discharge area at various ambient pressures.

the spatial scale under the streak camera image Under each figure, the corresponding coordinate x

is indicated In the panel for x= 0.2 mm, the waveform of the high-voltage input is indicated for reference The high voltage starts at zero and then periodically increases and decreases As shown by the images in Fig.7, the appearance of the wall jet is delayed for locations away from the electrode edge, because the flow begins at the electrode edge and extends along the surface In these images,

we can observe the temporal variation of the wall jet thickness In particular, in the image near the electrode edge (e.g., x= 0.2 mm), we can observe a clear periodic inflation of the wall jet thickness This inflation is synchronized with the alternating high-voltage input during the decreasing phase This periodic inflation of the wall jet thickness can be clearly observed at locations up to 2.2 mm that coincides with the vortex generation location It should be noted that this region of the wall jet coincides with the acceleration region On the contrary, at locations beyond the vortex generation location, in other wards, in the second region of the wall jet, no obvious oscillations can be observed;

At 3.2 and 4.2 mm, the wall jet is separated from the surface and then it reappears near the surface

As shown in Fig.4, this behavior indicates that after the vortex disappears, the wall jet appears as

an extension of the wall jet generated at the electrode edge

The present periodic behavior of the wall jet is expected to be affected by the waveform of the applied high-voltage input To confirm this, we applied three types of high voltages, as shown

in Fig.3 The effect of the waveform is shown in Fig.9, in which the streak image of the flow at

x= 0.2 mm is shown For waveforms 2 and 3, we observe a similar periodic inflation of the wall jet, although the degree of periodic inflation is slightly altered and is lower than that for waveform

FIG 7 Streak images for schlieren images at several locations along the surface over the first 12 cycles of the alternating high-voltage input.

FIG 8 Schematics for the acceleration region in the wall jet.

1 Nonetheless, the inflation of the wall jet during the decreasing phase is a common feature of wall jets induced in DBD plasma actuators

The present observation mentioned above can be summarized schematically in Fig.8 The figure suggests the acceleration region in the wall jet, of which height periodically oscillates synchronizing with the AC high voltage The increase of the height of the acceleration region occurs in the decreasing phase of the AC high voltage

042150-9 Abe, Takagaki, and Yamada AIP Advances 2, 042150 (2012)

FIG 9 Streak images for schlieren images near the electrode edge (x = 0.2 mm) for various waveforms.

The periodic inflation of the wall jet thickness is not an isolated evidence that the flow induction

is synchronized with the alternating high-voltage input or the electric discharge temporal behavior

In fact, a previous experiment by Forte et al showed a periodic variation of the flow velocity in the

wall jet: the parallel component of the flow velocity increased during the decreasing phase of the alternating high voltage.16 The temporal behavior of the total force produced by the DBD plasma actuator, which is caused by the flow induction, provides additional evidence: the measured total force varies periodically and is synchronized with the alternating high-voltage input.11To interpret such temporal behavior, Font12 proposed a theoretical model for the momentum exchange region appearing in the electric discharge area In the theory, the momentum exchange region varies with the phase of the alternating high-voltage input, and Font concluded that the region must be thicker during the decreasing phase compared with that during the increasing phase This theoretical model may explain the temporal behavior of the wall jet observed in the present experiment In addition, a recent theoretical investigation by Nishida and Abe18 clarified the periodic variation of the acceleration region resulting from the electric discharge, and they showed that the thickness (away from the surface) of the acceleration region varies depending on the phase of the alternative high voltage and is larger in the decreasing phase than in the increasing phase This result also agrees with the observations in this study

IV CONCLUSIONS

The flow induction process in a DBD plasma actuator was experimentally investigated by the flow visualization method First, the overall behavior of the wall jet development was confirmed

to agree with that previously reported.13,14 That is, the induced flow originated near the exposed electrode edge, and similar to a wall jet, it smoothly spread over the surface away from the electrode edge During the extension of the wall jet, a vortex formed at its tip After vortex shedding, the wall jet continued to further extend In response to this wall jet formation, the ambient gas near the electrode edge was sucked toward this edge and into the wall jet

In addition to the overall behavior of the induced flow, further investigation of the induced flow behavior observed in this study suggested that the wall jet thus formed could be classified into two regions: (1) the wall jet expanding from the location of the electrode edge to that of vortex generation, and (2) the part of the wall jet that extends beyond the first region In the first region, active momentum coupling occurs; therefore, the flow is actively induced However, the second region is formed by the inertia of the flow originating from the first region In the first region, the temporal behavior of its area was found to periodically vary, synchronizing with the alternating