Figure
1.
Schematic of the actuator used for ionic wind velocity and thrust measurements.
| Citation: |
Jingwen FAN, Huijie YAN, Ting LI, Yurong MAO, Jiaqi LI, Jian SONG. Surface charge characteristics in a three-electrode surface dielectric barrier discharge[J]. Plasma Science and Technology, 2024, 26(11): 115403. DOI: 10.1088/2058-6272/ad7821
|
The surface charge characteristics in a three-electrode surface dielectric barrier discharge (SDBD) are experimentally investigated based on the Pockels effect of an electro-optical crystal. The actuator is based on the most commonly used SDBD structure for airflow control, with an exposed electrode supplied with sinusoidal AC high voltage, a grounded encapsulated electrode and an additional exposed electrode downstream supplied with DC voltage. The ionic wind velocity and thrust can be significantly improved by increasing DC voltage although the plasma discharge characteristics are virtually unaffected. It is found that the negative charges generated by the discharge of the three-electrode structure accumulate on the dielectric surface significantly further downstream in an AC period compared to the actuator with a two-electrode structure. The negative charges in the downstream region increase as the DC voltage increases. In addition, the DC voltage affects the time required for the positive charge filaments to decay. The positive DC voltage expands the ionic acceleration zone downstream to produce a greater EHD force. The amplitude of the DC voltage affects the electric field on the dielectric surface and is therefore a key factor in the formation of the EHD force. Further research on the surface charge characteristics of a three-electrode structure has been conducted using a pulse power to drive the discharge, and the same conclusions are drawn. This work demonstrates a link between surface charge characteristics and EHD performance of a three-electrode SDBD actuator.
Surface dielectric barrier discharge plasma actuator is a flow control device using atmospheric pressure discharge. Due to the simple and lightweight structure, fast time response, low energy consumption, and not changing the original aerodynamic shape of the aircraft, it has been extensively researched in the aviation field of suppressing boundary layer separation, aerodynamic drag reduction, and delaying laminar turbulence transition [1–3]. Two asymmetric electrodes are separated by an insulated dielectric, with one electrode encapsulated by insulated materials and the other exposed to air. When a sufficient AC voltage is applied between the electrodes, a layer of plasma is generated. The charged particles in the plasma collide with neutral particles under the traction of an electric field to transfer momentum, forming the electrohydrodynamic (EHD) body force that acts on quiescent air and induces ionic wind for flow control [4].
The relatively low ionic wind velocity will diminish the flow control effect in practical applications. Many studies have been conducted to improve the EHD performance. The thrust could be increased by nearly 50% by replacing the conventional aluminum foil electrode with a mesh electrode [5]. Increasing the width of the encapsulated electrode expands the electric field to improve the ionic wind velocity [6]. Using filamentary exposed electrode has also been proven to be an effective way that can significantly improve the EHD effect [7]. It has been shown that the best EHD performance is achieved when a 10 μm filament is used as the exposed electrode [8]. Benard and Moreau [9] investigated the effect of the sine, square, positive ramp, and negative ramp waves on the actuator performance and found that the square wave produced the greatest thrust. The results of Yan et al [10] showed that when a pulse signal is superimposed on the peak and valley of the AC voltage, the pulse-induced breakdown occurs, with an increase in thrust of 100%–300%. In 2016, McGowan et al [11] proposed a DC-biased pulse voltage that achieved several times greater thrust than AC voltage, and the actuator has well applied in compressor stall control. In addition, the proposal of a three-electrode SDBD structure is also an effective method without increasing power consumption. In 2008, Moreau et al [12] added an exposed electrode in the downstream region of the AC discharge and applied a DC voltage to the third electrode, finding a significant increase in ionic wind velocity when the DC voltage was positive. They explained this phenomenon by the fact that the positive polarity of the third electrode accelerated the negative charges produced during the negative half period of the AC discharge. In 2013, the third electrode was used to regulate the surface charge by Cristofolini et al [13], which also improved the actuator performance.
The current common explanation for the formation of the EHD force and the ionic wind is the momentum transfer caused by the collision between charged particles and neutral particles. In a typical AC-SDBD, the plasma exhibits different discharge characteristics in the positive and negative half periods. The EHD force is mainly generated in the AC negative half period, and the main components inducing airflow acceleration are oxygen negative ions [14, 15]. Previous studies have shown that the charge accumulation is an important factor affecting discharge characteristics and actuator performance [16, 17]. The charge accumulation can significantly change the charge distribution and the dynamic behavior of charged particles. Moreover, the accumulated charges have a memory effect on the continuous development of the DBD [18]. The better performance can be achieved by effectively regulating the surface charge.
In recent years, the visualization techniques for studying surface charge using the electro-optical crystal have been greatly developed. In this work, a measurement system based on the optical polarization method is adopted, which was initially used to study the charge generated by volume dielectric barrier discharge [19–22]. The measurement principle of this technique was reviewed by Zhang et al [23]. In 2011, Takeuchi et al [24] used the system for the first time in SDBD to measure the surface charge. In 2021, the evolution of the surface charge in AC-SDBD was acquired by Mitsuhashi et al [25] using the optical polarization method. Sato et al [26] compared the surface charge distribution of different SDBD structures with DC-biased repetitive pulses. Our group investigated the distribution and evolution properties of the surface charge in SDBD with the conventional two-electrode structure using the optical polarization method [27–29]. Although some studies have been conducted on the link between surface charge properties and EHD performance, there is a lack of research on the mechanism by which the three-electrode structure effectively improves the EHD performance. Therefore, this work aims to investigate the dynamic behavior of the surface charge in a three-electrode plasma actuator by the optical polarization method.
In this work, we investigate the accumulation and decay properties of the surface charge using AC and pulse power to drive the SDBD and adding DC voltage to the third electrode. It is worth mentioning that in this configuration, the positive DC voltage significantly enhances the ionic wind, while the negative DC voltage is detrimental to the enhancement of the ionic wind and even results in a sliding discharge. Therefore, the DC voltages we apply in this work are positive.
The actuators used in this work are based on the most commonly used SDBD structure for airflow control, with a pair of electrodes asymmetrically arranged on both sides of the dielectric plate supplied with AC high voltage and grounded separately and an additional exposed electrode downstream supplied with DC voltage. Two different structures of the SDBD actuators are used in this work for the purposes of aerodynamic characteristics and surface charge study respectively. All experiments were carried out in the open air. The experiment was completed during a few weeks period of time, during which the temperature and humidity remained at around 25 °C and 40%, respectively by air conditionings.
The structure of the actuator used for ionic wind velocity and thrust measurements is shown in figure 1. The dielectric plate made of quartz glass is of the size of 80 × 80 × 1 mm3. The electrodes are made of aluminum foils with the thickness of 0.06 mm. The exposed electrodes 1 and 3 are of the same area of 50 × 7 mm2. The electrode 2 with an area of 50 × 14 mm2 is encapsulated with a glass plate of the same size of the dielectric plate to avoid unwanted discharge on the lower surface. The upper boundary position of the electrode 1 is defined as y = 0 mm.
The structure of the actuator used for surface charge measurement is shown in figure 2. The dielectric plate made of quartz glass is of the dimension of 80 × 80 × 2 mm3. The dielectric has a 20 × 20 × 1 mm3 area hollowed out in the center and embedded with a Bi12SiO20 (BSO) crystal inside which is used to measure the surface charge accumulation through Pockels effect. The two exposed electrodes are made of aluminum foils with the thickness of 0.06 mm and the area of 50 × 7 mm2. The encapsulated electrode is a piece of glass plate covered with a conductive indium tin oxide (ITO) film on top. The transparent ITO film allows light to pass through the dielectric to study the surface charge. The upper edge of the electrode 1 is aligned with the bottom of the ITO film.
The electrical circuit is shown in figure 3. The electrode 1 is connected to the AC high voltage generated by the plasma generator (CORONA-Lab CTP-2000K). The electrode 2 is grounded. The electrode 3 is connected in series with a 1 MΩ resistor and then connected to a DC power (DW-N303-1ACH2) and a 10 nF capacitor connected in parallel. The resistor is used to protect the DC power from abnormal discharge and the capacitor is used to reduce the disturb of AC voltage on the DC power. In section 3.4, a pulse voltage is applied to the electrode 1. The pulse power was described in detail in reference [27]. The signals measured by the voltage probe (Tektronix P6015A) and the current probe (Pearson 2877) are recorded with a digital oscilloscope (Tektronix DPO 4104).
The ionic wind induced by the actuator produces a change in air pressure in the discharge area. The wind velocity measurement device consists of an air pressure probe and a pressure sensor (SETRA model 268, 0–50 Pa, 4–20 mA output and 0.2 Pa accuracy). During the experiment, the low-pressure inlet end of the sensor is open to the atmosphere, and the high-pressure inlet end is connected to the pressure probe shown in figure 4 via a pitot tube. The pressure probe consists of two pieces of quartz glass (0.2 mm thickness) and does not interfere with the discharge. The pressure probe can move in different directions driven by a stepper motor and transmit the air pressure at different spatial locations to the pressure sensor to realize the accurate measurement of the ionic wind velocity. The time-averaged ionic wind velocity is derived from Bernoulli equation [30]:
| p=p0+12ρv2, | (1) |
where p is the total pressure, p0 is the static pressure, ρ = 1.29 kg/m3 is the air density at standard atmospheric pressure, and v is the airflow velocity.
The ionic wind generates a thrust in the opposite direction by recoil, which is measured using an electronic balance (GvG JJ224BC, 220 g range, 0.1 mg accuracy). The balance is placed under a metal cover to prevent electromagnetic waves generated by the discharge from interfering with the measurement. The actuator is mounted vertically on the balance through an insulated bracket, with the exposed electrode below and the encapsulated electrode above, as shown in figure 5, so that the ionic wind blows vertically upwards (away from the balance). When the SDBD actuator is set up on the balance, two pieces of thin copper wires with a diameter of about 0.1 mm and a length of about 20 cm are used to connect the fixed voltage terminals to the actuator electrodes slackly from the side (vertical to the ionic wind direction). During the force measurement, the displacement of the balance tray is almost imperceptible (less than 0.01 mm). So the soft thin wires have almost no additional deformation stress on the balance to disturb the body force measurement. The recoil force is applied to act on the balance, and the thrust can be read from the balance. In addition, the balance is connected to a computer so that the force data can be recorded continuously through a computer program at a rate of about 5 Hz. For each measurement, the thrust force data was recorded only when the balance reported a stable reading. The time-averaged thrust and error bars are obtained from the mean value and statistical standard deviation separately from about 100 data samples in a time duration of about 20 s.
The accumulation and decay characteristics of the surface charge are investigated through the Pockels effect of the BSO crystal. The difference in the refractive index of the electro-optic crystal between the slow and fast axis results in a phase difference. The 634 nm monochromatic light emitted from the led passes through the Kohler system to form a 25 mm diameter parallel beam, then passes through a polarizer and a 1/4 wave plate to form an elliptic polarized beam. The light intensity changes when the beam passes through the BSO crystal during the discharge because of the phase difference resulted by the charge accumulation on the crystal. The light finally passes through another polarizer (used as analyzer) and imaged by a high-speed video camera (Phantom VEO 410L). The brightness distribution of the image reflects the information of surface charge distribution. The surface charge measurement system is shown in figure 6. More detailed principles of the optical polarization can be found in reference [23]. The details of the experimental setup can be found in our previous work [27–29].
The phase difference Δφ can be described as:
| Δφ=2πλn30γ41UBSO, | (2) |
where λ is the wavelength of the LED light, n0 is the normal refractive index, and γ41 is the Pockels coefficient. The voltage drops UBSO across the BSO crystal consists of two components: the applied voltage Uapp and the voltage Uσ caused by the charge accumulation. The following relationship is satisfied by the phase difference Δφ, the incident light intensity I0 and transmitted light intensity I:
| II0=sin2(Δφ+π22)=12(1+sinΔφ). | (3) |
The spatiotemporal surface charge density σ is calculated from the formula:
| σ=ε0εBSOdBSO[IUBSO−IU=0kI0−Uapp], | (4) |
where ε0 is the vacuum permittivity, εBSO is the relative permittivity of the BSO crystal, dBSO is the thickness of the crystal. IU=0 is the intensity captured at Uapp = 0.
To keep the image in sync with the discharge, the high-speed camera is triggered at the zero point of rising edge of the sinusoidal AC waveform using another oscilloscope (RIGOL DS1000Z). In section 3.4, the camera is triggered at the falling edge of the pulse voltage. Under existing equipment conditions, we need to balance the temporal and spatial resolution. Finally, the sample rate and exposure time of the camera are set to 40000 fps and 24.507 μs, respectively. The image resolution is 256 × 360 which covers an area of 9.2 × 12.9 mm2.
The discharge images are processed to remove the defects of the crystal itself to present clear charge distribution. Figure 7 shows the image processing. The original discharge image and the background image without discharge are shown in figures 7(a) and (b), respectively. According to formula (4), there is a linear relationship between charge density and light intensity. Dividing the gray value of each position in figure 7(a) by that at the same position in figure 7(b) can obtain the surface charge image in figure 7(c). A more intuitive charge image is obtained by further coloring processing, as shown in figure 7(d). The colored scale defines the polarity and density of the charge, and the charge density is given in unit of nC/cm2. All the discharge images in this paper were processed in the same way.
In this study, the peak-to-peak value and frequency of the AC voltage are 12 kV and 4 kHz, respectively. The DC voltages are set to 0, 2, 4, and 6 kV, respectively.
The voltage and current waveforms are shown in figure 8. For the convenience of comparison, the current curves at different DC voltages are offset to 0, 20, 40, and 60 mA, respectively. In a typical AC-SDBD, the positive half cycle exhibits dispersed and strong current pulses, while the negative half cycle exhibits dense and weak current pulses. However, the pulse number and amplitude of the discharge current do not vary significantly with the DC voltage. Moreover, as shown in the discharge image captured with a Digital Single Lens Reflex camera (Nikon D7000) in figure 9, there is no difference in the discharge morphology and plasma extension length with different DC voltages. Therefore, the change in discharge characteristics of the SDBD is almost negligible with this three-electrode structure and driving method [12].
In addition, we calculate the power consumption of the actuator with different DC voltages. It should be noted that the power consumption for AC and DC dual high-voltage drive consists of both AC power input and DC power input. The DC voltage produces only a very small DC current which is hard to be measured because of the limited current resolution of 0.01 mA in our experiment. The results in reference [12] also show that the power consumption caused by a DC bias of 20 kV with a current resolution of 0.05 mA is almost zero. Therefore, the DC power consumption in the experiment is almost negligible. For AC voltage, the power consumption can be calculated using the voltage-current integration:
| P=1T∫T0V(t)I(t)dt, | (5) |
where V(t) and I(t) are the changes of the voltage and current with time respectively, and T is the AC voltage period. When the DC voltages are 0 kV, 2 kV, 4 kV, and 6 kV, the AC power consumptions of the actuator are 14.37 W/m, 14.15 W/m, 14.00 W/m, and 14.20 W/m, respectively. Due to the lack of significant differences in discharge current and voltage with different DC voltages, the power consumption is basically the same. This indicates that the DC voltage in the three-electrode structure does not significantly increase the power consumption of the actuator.
In order to investigate the EHD performance of the three-electrode SDBD structure, the time-averaged ionic wind velocity and thrust at different DC voltages are measured in the experiment. The probe is set to move from y = 0 to y = 30 mm for each measurement of the induced ionic wind distribution. The lower end of the probe is placed at z = 0 mm (upper surface of the dielectric).
The time-averaged ionic wind velocity along the y direction at different DC voltages is displayed in figure 10. The wind velocity increases with the discharge extension. After reaching the maximum value, the wind velocity starts to decrease. The influence of the DC voltage on the ionic wind velocity in different regions is different. In the AC discharge region near the electrode 1, the DC voltage has little effect on the ionic wind. In the downstream region, the ionic wind velocity at different DC voltages increases with the DC voltage. Since the plasma extension length and the discharge current are almost unchanged, this implies that the region where the DC voltage is acting is not the AC discharge region, but the downstream region away from the plasma. The downstream region away from the plasma has a weak electrical conductivity, and the DC voltage significantly modulates the surface charge and enhances the electric field for ionic acceleration, which results in a significant change in the airflow velocity in this region. In other words, the change in airflow velocity with the DC voltage can be attributed solely to the change in surface charge distribution. The higher the DC voltage, the stronger the acceleration effect on the negative ions.
Figure 11 shows the increase in thrust with increasing DC voltage. This phenomenon is also easily explained. In fact, the EHD force is a spatial volumetric force [31]. The thrust is the result from the integration of the airflow over the entire space. The ionic wind velocity in the y direction shows a tendency to increase with the DC voltage. The tendency of the thrust to increase with the DC voltage is more pronounced because the thrust is the result of the integration of the entire spatial airflow field. The results of the thrust further indicate that the variation in surface charge has an important effect on the EHD performance.
To understand the influence of the surface charge on the EHD performance of the three-electrode SDBD actuator, the surface charge characteristics at different DC voltages are investigated. The actuator structure with a BSO crystal used in this section is shown in figure 2. The electrical circuit is described in section 2.2. An AC voltage is applied to the electrode 1 to generate discharge and a DC voltage is applied to the electrode 3 to modulate the surface charge.
The discharge morphology at BSO dielectric is a little different from that at quartz plate. The dielectric constant of BSO is larger than that of normal quartz glass, resulting that the discharge filament on a BSO surface is a little shorter than that on a quartz glass plate. But the discharge mechanism and the charge transport mechanism are the same. It is feasible to use the results from a BSO structure to qualitatively analyze the surface charge characteristics of a conventional SDBD.
Ten surface charge images can be acquired in an AC cycle (250 μs). Figure 12 shows the shooting times corresponding to ten discharge images in an AC voltage cycle. The charge images during t3 (50–75 μs) and t8 (175–200 μs) at different DC voltages are displayed in figures 13(a)–(d). There are obvious differences in the surface charge distribution between the two half cycles of AC discharge. The positive half cycle accumulates mainly positive charges, which are filamentary and sparsely distributed, while the negative half cycle accumulates mainly negative charges, which are uniformly distributed. In the negative half cycle, the positive charge filaments generated in the previous positive half cycle do not disappear. The negative ions from the discharge move downstream and deposit on the dielectric surface away from the discharge for higher DC voltages as shown in figures 13(c) and (d).
Since the positive charge filaments in the filamentary discharge are randomly generated, we statistically average the charge density values at the same position in the charge images for 30 consecutive AC cycles at different DC voltages and obtain the surface charge images for an AC cycle, as shown in figure 14. For an AC discharge cycle, the extension of positive charge filaments is significantly inhibited at higher DC voltages. In addition, negative charges increase significantly and are deposited mainly in the downstream region. This phenomenon is caused by the positive DC voltage at the electrode 3. The positive DC voltage exerts electrostatic repulsive forces on the filamentary discharge, suppressing the formation and propagation of positive charge filaments. Whereas, the positive DC voltage has an electrostatic attraction on the negative ions emitted by discharge, causing the negative charges to migrate downstream and deposit on the dielectric surface away from the discharge. In addition, the positive DC voltage promotes the generation of negative charges by glow-like discharge.
To clarify the distribution regularity of the surface charge, the variation of the sum of all charge densities on the x-axis with respect to the y-distance is obtained by summing up all density data in the x direction for the same y position, as shown in figure 15. Figure 15 is divided into four regions according to the variation of the surface charge density, region I for 0 mm < y < 1.3 mm, region II for 1.3 mm < y < 4 mm, region III for 4 mm < y < 12 mm, and region IV for 12 mm < y < 14 mm. In region I, the edge of the electrode 1 accumulates mainly positive charges, which decrease slightly and then gradually increase along the plasma extension direction, with the maximum value occurring at about y = 1.3 mm. In region II, the positive charge density decreases until it becomes closed to 0 nC/cm2. In region III, the charge density is almost 0 nC/cm2 for VDC = 0 kV, and the negative charge accumulation becomes more obvious for the higher DC voltage. In region IV, positive surface charges gradually accumulate. In fact, the positive charges at the edge of the electrode 3 in the surface charge image are not caused by the discharge, but by the applied DC voltage.
It is necessary to mention why a positive charge peak exists at the edge of the electrode 1, and decreases slightly and then increases with the plasma extension. The charge at the edge of the electrode 1 is mainly influenced by the space charges in the sheath-like structure. The sheath-like structure formed in the glow-like discharge is known as the cathode layer, in which positive ions move towards the electrode while electrons migrate downstream [14, 31–33]. Since the mobility of electrons is much greater than that of ions, there are additional positive spatial charges in the cathode layer and therefore a positive charge at the edge of the electrode 1. The migration of the negative charges emitted by the discharge causes a slight decrease in the positive charge density. The plasma end will continuously gain positive charge accumulation without being immediately neutralized in the subsequent negative half cycle due to the difference in discharge pattern between positive and negative half cycles, thus forming positive charge accumulation in regions I and II. The charge accumulation can significantly affect the electric field in the discharge plasma region from figure 15.
We believe that the difference in charge density distribution at different DC voltages may be the main reason for the variation in ionic wind velocity and thrust. The electric field required to accelerate negative ions downstream is determined by both the electric field parallel to the surface of the BSO crystal caused by the DC voltage and the surface charge caused by the charge accumulation. The positive charge in region I can accelerate negative ions directionally, which is also the ionic acceleration region of the conventional two-electrode structure. Negative ions moving downstream react with the positive charges on the dielectric surface in a neutralization reaction, creating a charge density distribution in region II. The positive charge in region IV extends the ionic acceleration region to the electrode 3, attracting negative charges to migrate downstream and be deposited downstream. The higher the DC voltage, the more negative charges are deposited downstream.
The negative charge accumulation formed by the glow-like discharge creates a shielding electric field on the dielectric surface, which does not favor the downstream movement of the negative charges to generate the EHD force. The positive charge at the electrode 3 attracts negative charges from the electrode 1 to migrate downstream, suppressing the shielding electric field by reducing the negative charge accumulation in the plasma region, thereby effectively enhancing the EHD effect. The DC voltage provides a stronger electric field for ionic acceleration, resulting in a stronger EHD force. In section 3.2, the ionic wind velocity increases significantly with the increase of DC voltage in the downstream region away from the discharge. This feature indirectly explains the fact that the EHD force extends downstream.
The influence of the surface charge distribution on the actuator performance in the three-electrode SDBD structure can be summarized as follows: on one hand, the positive voltage at the electrode 3 attracts the negative charges to migrate downstream and suppresses the shielding effect caused by the negative charge accumulation in the discharge plasma region; on the other hand, the positive voltage at the electrode 3 expands the ionic acceleration region to the third electrode, and the higher DC voltage provides a stronger electric field for the ionic acceleration.
The research above indicates that the positive DC voltage significantly alters the distribution of the surface charge, which leads to differences in the propagation of the positive streamer at different DC voltages. Perhaps the decay rate of the positive charges in the streamer channel is also related to the DC voltage. Therefore, we further investigate the decay properties of the surface charge.
Previous studies have shown that the decay rate of the surface charge is affected by the location of the charge [27]. With the location further away from the exposed electrode, the surface charge decay rate becomes slower. Positive surface charges closer to the exposed electrode are easily neutralized by negative charges in the subsequent negative discharge, making it difficult to study the decay at the tail of the positive charge filaments. We therefore concentrate on studying the decay at the head of the positive charge filaments, which is further away from the electrode 1 and can exist for a longer time. Four positive charge filaments with similar length and head density are selected at different DC voltages. The selected head region is 0.24 × 0.24 mm2, circled with black frames, as shown in figure 16.
Taking the moment when the shooting ends in figure 16 as 0 μs, the evolution of the total charge density at the head of the four filaments for 20 consecutive AC cycles is given in figure 17. The charge density at the head shows an exponential decay pattern [26]. The decay rate is faster for higher DC voltages. After four AC discharge cycles (1000 μs), the filament head has decayed most of the charges for VDC = 6 kV, followed by 4 kV, and about half of the charges for VDC = 2 kV. The charge decay is the slowest for VDC = 0 kV, with the filament head still maintaining a high charge density.
Figure 18 shows the images of the charge filaments evolving with time at different DC voltages. The tail of the charge filaments decays faster, while the middle and head decay slower. The negative charges must be neutralized by the positive surface charges at the tail before reaching the middle and head. Consequently, the decay at the middle and head of the charge filaments is slower. The decay at the head of the charge filaments at different DC voltages is displayed in table 1. The higher the DC voltage, the faster the decay at the head of the charge filaments. In the case of 3125 μs, for example, most of the charges at the head have decayed for VDC = 6 kV and VDC = 4 kV, with only 10.2% and 16.5% remaining, respectively. The surface charges have decayed by about 70.9% for VDC = 2 kV and by about 47.0% for VDC = 0 kV.
The neutralization of positive and negative charges dominates the decay process of the positive surface charges. Negative charges include electrons and negative ions generated by discharge and drifting from space above the dielectric. The positive DC voltage attracts the negative charges to migrate downstream and neutralize the positive charges on the dielectric surface. In addition, the positive DC voltage also promotes glow-like discharges to generate more negative charges. Therefore, the higher the DC voltage, the faster the positive charges decay.
| DC (kV) | 125 μs | 625 μs | 3125 μs |
| 0 | 20.7% | 24.0% | 47.0% |
| 2 | 29.3% | 45.8% | 70.9% |
| 4 | 38.9% | 64.0% | 83.5% |
| 6 | 41.0% | 68.8% | 89.8% |
DownLoad:
CSV
Compared to the milliampere order of discharge current generated by AC-SDBD, nanosecond pulsed SDBD (NP-SDBD) can generate the ampere order of discharge current, which can inject more power into the actuator. NP-SDBD can generate mechanical, thermal, acoustic and other physical disturbances, making it more promising for flow control. In this section, the EHD effect and surface charge characteristics of the NP-SDBD with a three-electrode structure are investigated using a nanosecond pulse power.
The electrical circuit is shown in figure 3. The electrode 2 is grounded and the electrode 1 is connected to the pulse power supply to generate discharge. The electrode 3 is applied with DC voltage. Due to the important role of the negative ions in the momentum transfer to generate the EHD force, a negative polarity pulse is adopted in this experiment. The pulse voltage amplitude Vp is −6 kV. The output pulse width and pulse repetition frequency (PRF) are set to 200 ns and 4 kHz, respectively. The discharge voltage and current are collected when the DC voltage VDC = 0, 2, 4, and 6 kV. For ease of comparison, the current curves at different DC voltages are offset to −4, −2, 0, and 2 A, respectively, as shown in figure 19. Similar to the results in section 3.1, the pulse number and amplitude of the discharge current do not vary significantly with the DC voltage. The macroscopic discharge images for pulse driving are shown in figure 20. There is no significant change in the morphology of the plasma with different DC voltages, and the extension length is slightly smaller than that driven by AC voltage. The effect of adding the third electrode on the discharge characteristics is negligible in this driving mode.
Figures 21 and 22 show that the ionic wind velocity and thrust increase as the DC voltage increases. The improvement in the ionic wind velocity remains in the downstream region. However, the time-averaged ionic wind velocity and thrust cannot be generated by NP-SDBD for VDC = 0 kV. The ionic wind velocity and thrust for pulse drive are much smaller than that for AC drive. In reality, the topology of the airflow field induced by pulse discharge is much more complex. However, the wind velocity measurement we use is only the time-averaged distribution along a certain direction and has some limitations.
The high-speed camera is triggered on the falling edge of the pulse voltage and takes images from the first pulse discharge. Figure 23 shows the surface charge images of the NP-SDBD at different DC voltages. The negative pulse discharge leads to the negative charge accumulation at the edge of the electrode 1, and the phenomenon is more pronounced at higher DC voltages. As the pulse number increases, the negative charges increase and tend to move downstream. Figure 24 shows the evolution of the total negative charge density at different DC voltages. The distribution of the negative surface charge reaches a steady state after about 1000 pulses (250 ms). This is consistent with the result that the charge saturation time is hundreds of milliseconds for PRF = 5 kHz in reference [26].
The same treatment was done as before by summing up all the surface charge densities in the x direction for the same y-distance. The variation of the sum of all charge densities on the x-axis with respect to the y-distance for the 5000th pulse at different DC voltages is shown in figure 25. Based on the variation of the charge density, figure 25 can be divided into three regions, region I for 0 mm < y < 0.8 mm, region II for 0.8 mm < y < 12 mm, and region III for 12 mm < y < 14 mm.
The negative pulse discharge causes negative charges to accumulate in region I, and the charge density reaches a maximum as the discharge plasma extends. In region II, as it moves away from the discharge region, the charge density becomes almost 0 nC/cm2 for VDC = 0 kV. More negative charges are deposited downstream for higher DC voltages. The positive charge in region III provides a directional electric field for the downstream movement of the negative charges and reduces the negative charge accumulation in the plasma region generated by the negative pulse discharge. Therefore, effective ionic wind and thrust are achieved by using the three-electrode NP-SDBD structure. Although the means we used are limited, it is significant to study the association between the EHD performance and surface charge characteristics in a three-electrode SDBD structure.
The BSO crystal is used to measure the surface charge in the three-electrode SDBD structure in this study. The propagation of the streamer discharge on the surface of BSO crystal will not be as far as that on the surface of quartz glass due to the large difference in dielectric constant between BSO crystal and quartz glass, but the internal mechanism is the same. This study reveals the charge distribution characteristics of the three-electrode SDBD actuator by the optical polarization method. The relationship between the EHD performance and the surface charge distribution is researched by varying the DC voltage at the third electrode for AC-SDBD and NP-SDBD.
The third electrode effectively modulates the surface charge distribution and is crucial in forming the ionic wind and EHD force, whether using AC or pulse power supply. However, the surface charge distribution characteristics of the two discharge patterns are significantly different due to the difference between AC and pulse signals. The commonality is that the negative charges can migrate downstream and deposit mainly in the downstream region under the attraction of the positive DC voltage, which also proves that the increase in ionic wind velocity is mainly in the downstream region.
The conventional two-electrode SDBD actuator can be divided into a strong ionization zone (the cathode layer) and an ionic acceleration zone (the discharge extension zone) on the dielectric surface, but the acceleration zone is small. The ionic acceleration zone of the three-electrode SDBD actuator is extended to the downstream exposed electrode by applying positive DC voltage to the third electrode. The positive DC voltage suppresses the shielding effect caused by the negative charge accumulation and significantly enhances the EHD effect. The EHD performance of the SDBD actuator can be well improved by modulating the transport process of charged particles by appropriately changing the charge distribution.
This work was supported by National Natural Science Foundation of China (Nos. 51777026 and 11705075).
| [1] |
Schatzman D M and Thomas F O 2010 AIAA J. 48 1620 doi: 10.2514/1.J050009
|
| [2] |
Roth J R, Sherman D M and Wilkinson S P 2000 AIAA J. 38 1166 doi: 10.2514/2.1110
|
| [3] |
Grundmann S and Tropea C 2007 Exp. Fluids 42 653 doi: 10.1007/s00348-007-0256-8
|
| [4] |
Mahdavi H and Sohbatzadeh F 2019 Phys. Scr. 94 085204 doi: 10.1088/1402-4896/ab0e17
|
| [5] |
Abe T et al 2008 AIAA J. 46 2248 doi: 10.2514/1.30985
|
| [6] |
Qi X H et al 2016 Plasma Sci. Technol. 18 1005 doi: 10.1088/1009-0630/18/10/07
|
| [7] |
Hoskinson A R, Hershkowitz N and Ashpis D E 2008 J. Phys. D: Appl. Phys. 41 245209 doi: 10.1088/0022-3727/41/24/245209
|
| [8] |
Debien A, Benard N and Moreau E 2012 J. Phys. D: Appl. Phys. 45 215201 doi: 10.1088/0022-3727/45/21/215201
|
| [9] |
Benard N and Moreau E 2012 Appl. Phys. Lett. 100 193503 doi: 10.1063/1.4712125
|
| [10] |
Yan H J et al 2016 J. Phys. D: Appl. Phys. 49 295203 doi: 10.1088/0022-3727/49/29/295203
|
| [11] |
McGowan R et al 2016 Pulsed-DC plasma actuator characteristics and application in compressor stall control In: Proceedings of 2016 54th AIAA Aerospace Sciences Meeting San Diego: California, USA doi: 10.2514/6.2016-0394
|
| [12] |
Moreau E, Sosa R and Artana G 2008 J. Phys. D: Appl. Phys. 41 115204 doi: 10.1088/0022-3727/41/11/115204
|
| [13] |
Cristofolini A, Neretti G and Borghi C A 2013 J. Appl. Phys. 114 073303 doi: 10.1063/1.4817378
|
| [14] |
Soloviev V R 2012 J. Phys. D: Appl. Phys. 45 025205 doi: 10.1088/0022-3727/45/2/025205
|
| [15] |
Kim W et al 2007 Appl. Phys. Lett. 91 181501 doi: 10.1063/1.2803755
|
| [16] |
Deng J B et al 2010 J. Phys. D: Appl. Phys. 43 495203 doi: 10.1088/0022-3727/43/49/495203
|
| [17] |
Ren C H et al 2023 Plasma Sources Sci. Technol. 32 025004 doi: 10.1088/1361-6595/acb4b9
|
| [18] |
Pan C et al 2020 IEEE Trans. Dielectr. Electr. Insul. 27 1951 doi: 10.1109/TDEI.2020.008960
|
| [19] |
Stollenwerk L, Laven J G and Purwins H G 2007 Phys. Rev. Lett. 98 255001 doi: 10.1103/PhysRevLett.98.255001
|
| [20] |
Wild R, Benduhn J and Stollenwerk L 2014 J. Phys. D: Appl. Phys. 47 435204 doi: 10.1088/0022-3727/47/43/435204
|
| [21] |
Dosoudilová L et al 2015 J. Phys. D: Appl. Phys. 48 355204 doi: 10.1088/0022-3727/48/35/355204
|
| [22] |
Tschiersch R, Bogaczyk M and Wagner H E 2014 J. Phys. D: Appl. Phys. 47 365204 doi: 10.1088/0022-3727/47/36/365204
|
| [23] |
Zhang B et al 2021 High Volt. 6 608 doi: 10.1049/hve2.12073
|
| [24] |
Takeuchi N et al 2011 J. Electrost. 69 87 doi: 10.1016/j.elstat.2011.01.001
|
| [25] |
Mitsuhashi K et al 2021 Plasma Sources Sci. Technol. 30 04LT02 doi: 10.1088/1361-6595/abefa7
|
| [26] |
Sato S et al 2021 J. Phys. D: Appl. Phys. 54 455203 doi: 10.1088/1361-6463/ac1b5d
|
| [27] |
Yu S Q et al 2022 J. Phys. D: Appl. Phys. 55 125201 doi: 10.1088/1361-6463/ac4184
|
| [28] |
Li T et al 2022 Plasma Sources Sci. Technol. 31 055016 doi: 10.1088/1361-6595/ac676e
|
| [29] |
Li J Q et al 2023 J. Appl. Phys. 133 063301 doi: 10.1063/5.0134362
|
| [30] |
Yan H J et al 2015 J. Appl. Phys. 117 063302 doi: 10.1063/1.4907992
|
| [31] |
Boeuf J P et al 2007 J. Phys. D: Appl. Phys. 40 652 doi: 10.1088/0022-3727/40/3/S03
|
| [32] |
Soloviev V R and Krivtsov V M 2009 J. Phys. D: Appl. Phys. 42 125208 doi: 10.1088/0022-3727/42/12/125208
|
| [33] |
Gibalov V I and Pietsch G J 2012 Plasma Sources Sci. Technol. 21 024010 doi: 10.1088/0963-0252/21/2/024010
|
| [1] | Simin ZHOU (周思敏), Xiutao HUANG (黄修涛), Minghai LIU (刘明海). Electrical model and experimental analysis of a double spiral structure surface dielectric barrier discharge[J]. Plasma Science and Technology, 2019, 21(6): 65401-065401. DOI: 10.1088/2058-6272/ab0814 |
| [2] | Songru XIE (谢松汝), Yong HE (何勇), Dingkun YUAN (袁定琨), Zhihua WANG (王智化), Sunel KUMAR, Yanqun ZHU (朱燕群), Kefa CEN (岑可法). The effects of gas flow pattern on the generation of ozone in surface dielectric barrier discharge[J]. Plasma Science and Technology, 2019, 21(5): 55505-055505. DOI: 10.1088/2058-6272/aafc50 |
| [3] | Yanqin LIU (刘彦琴), Guangning WU (吴广宁), Guoqiang GAO (高国强), Jianyi XUE (薛建议), Yongqiang KANG (康永强), Chaoqun SHI (石超群). Surface charge accumulation behavior and its influence on surface flashover performance of Al2O3-filled epoxy resin insulators under DC voltages[J]. Plasma Science and Technology, 2019, 21(5): 55501-055501. DOI: 10.1088/2058-6272/aafdf7 |
| [4] | Zilu ZHAO (赵紫璐), Dezheng YANG (杨德正), Wenchun WANG (王文春), Hao YUAN (袁皓), Li ZHANG (张丽), Sen WANG (王森). Volume added surface barrier discharge plasma excited by bipolar nanosecond pulse power in atmospheric air: optical emission spectra influenced by gap distance[J]. Plasma Science and Technology, 2018, 20(11): 115403. DOI: 10.1088/2058-6272/aac881 |
| [5] | Donglai WANG (王东来), Tiebing LU (卢铁兵), Yuan WANG (王源), Bo CHEN (陈博), Xuebao LI (李学宝). Measurement of surface charges on the dielectric film based on field mills under the HVDC corona wire[J]. Plasma Science and Technology, 2018, 20(5): 54008-054008. DOI: 10.1088/2058-6272/aaac26 |
| [6] | Cheng PAN (潘成), Ju TANG (唐炬), Dibo WANG (王邸博), Yi LUO (罗毅), Ran ZHUO (卓然), Mingli FU (傅明利). Decay characters of charges on an insulator surface after different types of discharge[J]. Plasma Science and Technology, 2017, 19(7): 75503-075503. DOI: 10.1088/2058-6272/aa6436 |
| [7] | Jianyu FENG (冯建宇), Lifang DONG (董丽芳), Caixia LI (李彩霞), Ying LIU (刘莹), Tian DU (杜天), Fang HAO (郝芳). Hollow hexagonal pattern with surface discharges in a dielectric barrier discharge[J]. Plasma Science and Technology, 2017, 19(5): 55401-055401. DOI: 10.1088/2058-6272/aa594a |
| [8] | QI Xiaohua (齐晓华), YANG Liang (杨亮), YAN Huijie (闫慧杰), JIN Ying (金英), HUA Yue (滑跃), REN Chunsheng (任春生). Experimental Study on Surface Dielectric Barrier Discharge Plasma Actuator with Different Encapsulated Electrode Widths for Airflow Control at Atmospheric Pressure[J]. Plasma Science and Technology, 2016, 18(10): 1005-1011. DOI: 10.1088/1009-0630/18/10/07 |
| [9] | LIU Wenzheng(刘文正), LI Chuanhui(李传辉). Study on the Generation Characteristics of Dielectric Barrier Discharge Plasmas on Water Surface[J]. Plasma Science and Technology, 2014, 16(1): 26-31. DOI: 10.1088/1009-0630/16/1/06 |
| [10] | HONG Yi (洪义), LU Na (鲁娜), PAN Jing (潘静), LI Jie (李杰), WU Yan (吴彦). Discharge Characteristics of an Atmospheric Pressure Argon Plasma Jet Generated with Screw Ring-Ring Electrodes in Surface Dielectric Barrier Discharge[J]. Plasma Science and Technology, 2013, 15(8): 780-786. DOI: 10.1088/1009-0630/15/8/12 |
Figures(25) / Tables(1)
Supported by: Beijing Renhe Information Technology Co., Ltd. 
| DC (kV) | 125 μs | 625 μs | 3125 μs |
| 0 | 20.7% | 24.0% | 47.0% |
| 2 | 29.3% | 45.8% | 70.9% |
| 4 | 38.9% | 64.0% | 83.5% |
| 6 | 41.0% | 68.8% | 89.8% |
DownLoad:
CSV
