Non-starting phase flow field dynamic regulation and control device and method for continuous sliding dielectric barrier discharge

Abstract

The invention discloses a non-starting phase flow field dynamic regulation and control device and method for continuous sliding dielectric barrier discharge, and belongs to the field of flow control. The regulation and control device comprises a power supply module, an electrode assembly, a dielectric substrate assembly, a flow field sensor and a dynamic control module, the electrode assembly is composed of three asymmetrically arranged electrodes, and the three electrodes are arranged on the dielectric substrate assembly in an inverted delta shape; the flow field sensor module monitors rotation rate, strain rate and speed fluctuation characteristics of a non-starting phase flow field in real time and application model surrounding flow field parameters and performance indexes, and the dynamic control module regulates and controls the power supply module to output a composite excitation waveform through a voltage-frequency-performance parameter matching algorithm by using a deep learning technology according to sensor data. The regulation and control purpose is realized. Quantitative flow control of the flow field is realized through the specific mathematical relationship between the voltage and the velocity field and by utilizing the velocity characteristic of the sliding dielectric barrier discharge induced flow in the continuous mode.

Description

Technical Field

[0001] The present invention relates to the field of flow control technology, and specifically to a device and method for dynamically controlling the flow field in a non-starting phase of a continuous sliding dielectric barrier discharge, which is suitable for active flow control in scenarios such as low / high-speed aircraft and engine air inlets, and is especially suitable for precise control of flow field stability and direction in a non-starting phase (dynamic equilibrium stage). Background Art

[0002] Plasma flow control technology achieves active intervention in flow characteristics through the coupling of electric field and flow field. Although traditional surface arc discharge exciters can change the flow field through thermal or electromagnetic effects, their application has the following limitations: Directional fixedness: The convection direction generated by traditional exciters is perpendicular to the electrode surface and cannot be dynamically adjusted with flow conditions, making it difficult to cope with complex flow field environments. Transient control limitations: Existing technologies focus more on the suppression of flow separation in the startup phase (initial transient stage), but lack effective means for continuous regulation of the non-startup phase (dynamic equilibrium stage). Material and energy consumption limitations: Traditional devices mostly use photosensitive resins or ceramic materials, which have insufficient high temperature resistance, and the intermittent working mode of the pulse power supply leads to high energy consumption.

[0003] As a new type of discharge technology, sliding dielectric barrier discharge adjusts the electric field distribution through the dynamic sliding of the dielectric layer, and has the advantages of continuous operation and fast response speed. However, the existing technology has not yet combined sliding dielectric barrier discharge with the dynamic regulation of the non-starting phase flow field, resulting in insufficient flow control efficiency and adaptability. Summary of the invention

[0004] Purpose of the invention: The purpose of the present invention is to provide a device and method for dynamically controlling the non-starting phase flow field of a continuous sliding dielectric barrier discharge, which at least partially solves the problems of continuous stability control of the non-starting phase flow field, dynamic adaptive adjustment of the flow field direction and intensity, high energy consumption and insufficient high temperature resistance of materials.

[0005] Technical solution: In order to achieve the above-mentioned invention object, the present invention adopts the following technical solution:

[0006] In the first aspect, a non-starting phase flow field dynamic control device of a continuous sliding dielectric barrier discharge comprises: a power module, an electrode assembly, a dielectric substrate assembly, a flow field sensor and a dynamic control module, wherein the power module comprises an AC power supply and a DC power supply, and outputs a composite excitation waveform; the dielectric substrate assembly comprises a first material substrate and a second material dielectric layer with a gradient thickness arranged on the first material substrate; the electrode assembly consists of three asymmetrically arranged electrodes, wherein the first electrode and the second electrode are separately arranged on the upper surface of the second material dielectric layer, and the third electrode is arranged on the lower surface of the second material dielectric layer, and the three electrodes are arranged in an inverted "pin" shape and are bonded to the surface of the second material dielectric layer; the flow field sensor module comprises a first sensor module for real-time monitoring of the rotation rate, strain rate and velocity fluctuation characteristics of the non-starting phase flow field, and a second electrode for real-time monitoring of the rotation rate, strain rate and velocity fluctuation characteristics of the non-starting phase flow field, and a third electrode for real-time monitoring of the rotation rate, strain rate and velocity fluctuation characteristics of the non-starting phase flow field, and a flow field sensor module ... and a second sensor module for real-time monitoring of the flow field parameters and performance indicators around the application model; the dynamic control module, based on the sensor data received from the flow field sensor module, utilizes the dynamic balance equation based on the coupling relationship between the electric field and the induced flow velocity and the deep learning prediction unit based on the recursive neural network, and controls the composite excitation waveform output by the power supply module through the voltage-frequency-performance parameter matching algorithm, and adjusts the alternating period of the induced flow rotation rate and the strain rate to achieve the control purpose, wherein the input of the voltage-frequency-performance parameter matching algorithm includes the fluctuation frequency ratio of the lateral velocity component and the longitudinal velocity component of the induced flow, the phase difference between the velocity field rotation rate and the velocity field strain rate, and the relationship between the velocity field fluctuation frequency ratio and the flow field velocity, and the output is the mathematical relationship of the optimal voltage V-flow field velocity u-application model performance parameter P combination parameter.

[0007] Furthermore, the electrode assembly is composed of three electrodes made of thin copper foil material, wherein the first electrode and the second electrode arranged on the upper surface of the second material dielectric layer have a first width, and the third electrode arranged on the lower surface of the second material dielectric layer has a second width, and the first width is different from the second width.

[0008] Furthermore, the thickness distribution of the second material dielectric layer with a thickness gradient satisfies the following functional relationship:

[0009]

[0010] Where L0 is the thickness of a single dielectric layer, L is the length of the electrode coverage area, x is the coordinate along the airflow direction, 0≤x≤L, [] Floor is the floor function.

[0011] Furthermore, the first material substrate is a polyetheretherketone dielectric substrate, and the second material dielectric layer is a polyimide film dielectric layer.

[0012] Furthermore, the lower surface of the second material dielectric layer is provided with micro heat dissipation channels, and the channel size and number satisfy the following relationship:

[0013]

[0014] where k fluid,heat is a coefficient related to fluid properties and heat dissipation requirements calibrated by experiments, W total is the total width of the first material substrate, d is the channel depth, and n is the number of channels.

[0015] Furthermore, the dynamic equilibrium equation based on the coupling relationship between the electric field and the induced flow velocity specifically includes:

[0016] According to the experiment, the expression between the electric field intensity and the induced flow velocity is:

[0017]

[0018] Where E is the electric field strength vector, u d is the induced flow velocity field vector, k is the coupling coefficient, C k is a constant; is an inverted triangular operator, representing the gradient; according to electrodynamics, the relationship between the variable output voltage value and the induced flow velocity is obtained as follows:

[0019]

[0020] Wherein, Δ represents the Laplace operator, and V is the output voltage value of the AC power controlled by the dynamic control module;

[0021] The relationship between the induced flow velocity, the performance parameters of the application model, and the mainstream velocity around the application model is:

[0022] u d =f(u m ,P)

[0023] Among them, u m is the mainstream velocity around the application model, P is the performance parameter of the application model, and the specific form of expression f depends on the form of the application model.

[0024] Furthermore, the control logic of the dynamic control module is as follows:

[0025] Receive data from the flow field sensor module, according to Processing, calculation of coupling coefficient k and constant C k ;

[0026] Fit the relationship between voltage and flow rate V = αu x +βu y +γ, according to the coupling coefficient k and the constant C k Calculate the coefficients α, β, and γ;

[0027] The deep learning prediction unit based on recursive neural network is called to solve the relationship between the induced flow velocity fluctuation frequency ratio f1 and the flow field velocity u through spectrum analysis and deep learning:

[0028]

[0029] u=g(f1,Ω,S)

[0030] Where g is the relationship learned by the recurrent neural network, u x,frequency 、u y,frequency are the induced flow transverse and longitudinal velocity components u x and u y The frequency of fluctuations;

[0031] And the velocity field rotation rate and strain rate phase difference Δφ:

[0032] Δφ=arg(Ω)-arg(S)

[0033] Where Ω is the velocity field rotation rate, S is the velocity field strain rate;

[0034] Finally determine the performance parameter P:

[0035] P=f(u,u x ,u y ,Ω,S,Δφ).

[0036] In a second aspect, a method for dynamically controlling a non-starting phase flow field of a continuous sliding dielectric barrier discharge comprises the following steps:

[0037] S1. A composite excitation waveform is applied by an AC power supply and a DC power supply to generate a sliding dielectric barrier discharge plasma excitation between electrodes, wherein the electrodes include three asymmetrically arranged electrodes, and the three electrodes are arranged in an inverted "pin" shape on a dielectric substrate assembly, wherein the dielectric substrate assembly includes a first material substrate and a second material dielectric layer with a gradient thickness arranged on the first material substrate, wherein the first electrode and the second electrode are separately arranged on the upper surface of the second material dielectric layer, and the third electrode is arranged on the lower surface of the second material dielectric layer, and the three electrodes are all in contact with the surface of the second material dielectric layer;

[0038] S2, the first sensor module collects the rotation rate, strain rate and velocity fluctuation characteristics of the non-starting phase flow field in real time, and the second sensor module measures the flow field parameters and performance indicators around the application model in real time;

[0039] S3. The dynamic control module, based on the sensor data received from the first and second sensor modules, uses the dynamic equilibrium equation based on the coupling relationship between the electric field and the induced flow velocity and the deep learning prediction unit based on the recurrent neural network, and regulates the composite excitation waveform output by the power supply module through the voltage-frequency-performance parameter matching algorithm, adjusts the alternating period of the induced flow rotation rate and the strain rate, and achieves the purpose of dynamic flow field regulation. The input of the voltage-frequency-performance parameter matching algorithm includes the fluctuation frequency ratio of the transverse velocity component and the longitudinal velocity component of the induced flow, the phase difference between the velocity field rotation rate and the velocity field strain rate, and the relationship between the velocity field fluctuation frequency ratio and the flow field velocity. The output is the mathematical relationship of the combined parameters of the optimal voltage V - flow field velocity u - application model performance parameter P.

[0040] In a third aspect, an electronic device includes: a processor, a memory, and a computer program, where the computer program is stored in the memory and is configured to be executed by the processor. When the computer program is executed by the processor, the following steps are implemented:

[0041] Based on the sensor data received from the flow field sensor module, use the dynamic equilibrium equation based on the coupling relationship between the electric field and the induced flow velocity and the deep learning prediction unit based on the recurrent neural network, and regulate the composite excitation waveform output by the power supply module through the voltage-frequency-performance parameter matching algorithm, adjust the alternating period of the induced flow rotation rate and the strain rate, and achieve the purpose of dynamic flow field regulation. The input of the voltage-frequency-performance parameter matching algorithm includes the fluctuation frequency ratio of the transverse velocity component and the longitudinal velocity component of the induced flow, the phase difference between the velocity field rotation rate and the velocity field strain rate, and the relationship between the velocity field fluctuation frequency ratio and the flow field velocity. The output is the mathematical relationship of the combined parameters of the optimal voltage V - flow field velocity u - application model performance parameter P;

[0042] The flow field sensor module includes a first sensor module that real-time monitors the rotation rate, strain rate, and velocity fluctuation characteristics of the non-starting phase flow field, and a second sensor module that real-time monitors the flow field parameters and performance indicators around the application model;

[0043] The induced flow is generated by the electrode assembly driven by the power supply module. The power supply module includes an AC power supply and a DC power supply and outputs a composite excitation waveform. The electrode assembly consists of three electrodes with an asymmetric layout. The three electrodes are arranged in an inverted "pin" shape on the dielectric substrate assembly. The dielectric substrate assembly includes a first material substrate and a second material dielectric layer with a thickness gradient arranged on the first material substrate. The first electrode and the second electrode are separately arranged on the upper surface of the second material dielectric layer, and the third electrode is arranged on the lower surface of the second material dielectric layer. All three electrodes are attached to the surface of the second material dielectric layer.

[0044] Fourth aspect, a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the following steps are implemented:

[0045] According to the sensor data received from the flow field sensor module, using the dynamic balance equation based on the coupling relationship between the electric field and the induced flow velocity and the deep learning prediction unit based on the recurrent neural network, through the voltage-frequency-performance parameter matching algorithm, the composite excitation waveform output by the power supply module is regulated, the alternating period of the induced flow rotation rate and the strain rate is adjusted, and the purpose of dynamic regulation of the flow field is achieved. The input of the voltage-frequency-performance parameter matching algorithm includes the fluctuation frequency ratio of the lateral velocity component and the longitudinal velocity component of the induced flow, the phase difference between the velocity field rotation rate and the velocity field strain rate, and the relationship between the velocity field fluctuation frequency ratio and the flow field velocity. The output is a mathematical relationship of the combined parameters of the optimal voltage V - the flow field velocity u - the application model performance parameter P;

[0046] The flow field sensor module includes a first sensor module for real-time monitoring of the rotation rate, strain rate, and velocity fluctuation characteristics of the non-starting phase flow field, and a second sensor module for real-time monitoring of the flow field parameters and performance indicators around the application model;

[0047] The induced flow is generated by the electrode assembly driven by the power supply module. The power supply module includes an AC power supply and a DC power supply and outputs a composite excitation waveform; the electrode assembly is composed of three electrodes with an asymmetric layout, and the three electrodes are arranged in an inverted "pin" shape on the dielectric substrate assembly. The dielectric substrate assembly includes a first material substrate and a second material dielectric layer with a thickness gradient arranged on the first material substrate. The first electrode and the second electrode are separately arranged on the upper surface of the second material dielectric layer, and the third electrode is arranged on the lower surface of the second material dielectric layer. All three electrodes are attached to the surface of the second material dielectric layer.

[0048] Beneficial effects:

[0049] (1) In the present invention, a composite excitation waveform is applied through the AC power supply and the DC power supply to generate a sliding dielectric barrier discharge plasma excitation between the electrodes. The rotation rate, strain rate, and velocity fluctuation characteristics of the non-starting phase flow field, as well as the flow field and performance parameters around the application model, are measured in real time by the sensor module. The dynamic control module adjusts the AC power supply frequency and the DC power supply voltage according to the dynamic balance equation and the deep learning algorithm, and adjusts the alternating period of the induced flow rotation rate and the strain rate to achieve the purpose of dynamically regulating the flow field. Continuous discharge is achieved through the composite waveform power supply, avoiding the energy intermittency problem of the traditional pulse mode; through the continuous mode and real-time feedback control, high-precision dynamic adjustment of the non-starting phase flow field is achieved, the response time is shortened to 5 ms, and the dynamic regulation ability is strong.

[0050] (2) The three electrodes of the present invention are arranged in an inverted "pin" shape on a dielectric substrate assembly, and a non-uniform electric field distribution is induced by a second material dielectric layer with a gradient thickness, which directionally enhances the high-frequency fluctuation component of the induced flow velocity and improves the flow separation suppression efficiency. The composite excitation waveform and gradient dielectric layer design reduce energy consumption by 30% and improve control efficiency.

[0051] (3) The present invention provides a micro heat dissipation channel on the lower surface of the second material dielectric layer to improve the heat resistance of the device and is suitable for extreme working conditions.

[0052] (4) The present invention can be extended to scenarios such as hypersonic aircraft wing surfaces and engine air inlets, and has broad application prospects. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] Figure 1 It is a schematic structural diagram of a non-starting phase flow field dynamic control device for continuous sliding dielectric barrier discharge disclosed in an embodiment of the present invention;

[0054] Figure 2 is a schematic diagram of the arrangement of the dielectric substrate assembly and the electrode assembly disclosed in this embodiment;

[0055] Figure 3 is an algorithm flow chart of a dynamic control module disclosed in an embodiment of the present invention;

[0056] Figure 4 is a schematic diagram of a device disclosed in an embodiment of the present invention installed on the leading edge of an airfoil;

[0057] Figure 5 is a schematic diagram of velocity distribution of a regulated flow field disclosed in an embodiment of the present invention;

[0058] Figure 6 It is a comparison diagram of the effects before and after the non-startup phase flow field regulation disclosed in an embodiment of the present invention. DETAILED DESCRIPTION

[0059] The technical solution of the present invention is further described below in conjunction with the accompanying drawings.

[0060] Embodiment 1

[0061] Reference Figure 1 and Figure 2 This embodiment provides a non-starting phase flow field dynamic control device for continuous sliding dielectric barrier discharge, including a power module 1, an electrode assembly 2, a dielectric substrate assembly 3, a flow field sensor 4 and a dynamic control module 5. The implementation of each part is described in detail below.

[0062] (I) Power module

[0063] The power module 1 includes two parts: an AC power supply and a DC bias power supply. It outputs a composite excitation waveform. Both the peak-to-peak voltage and the AC frequency can be parameterized. In this embodiment, the AC frequency range is 1-10kHz, the DC bias voltage range is (±) 1-20kV, and continuous mode discharge is supported; the peak-to-peak voltage is 10-18kV, and the flow field requirements are matched by dynamically adjusting the frequency and voltage.

[0064] The AC power module outputs a superimposed waveform of a high-frequency component and a low-frequency component, and the DC bias power module outputs a negative DC voltage, and the peak value of the output AC voltage and the peak value of the output negative DC voltage satisfy the following relationship:

[0065]

[0066] Among them, V ACPeak is the peak value of AC voltage, V DCPeak is the peak value of DC voltage, and a is a non-zero real number. V ACPeakIn is the peak value of the adjusted AC voltage.

[0067] (II) Electrode assembly and dielectric substrate assembly

[0068] The dielectric substrate assembly 3 comprises a dielectric substrate 3-2 of a first material and a dielectric layer 3-1 of a second material with a gradient thickness arranged on the dielectric substrate 3-2 of the first material. In the embodiment of the present invention, the dielectric substrate 3-2 of the first material is a polyetheretherketone dielectric substrate 3-2, and the dielectric layer 3-1 of the second material is a polyimide film dielectric layer 3-1; the electrode assembly 2 is composed of a thin copper foil, which is arranged on the dielectric substrate and is electrically connected to the power module 1. Specifically, refer to Figure 2 The electrode assembly 2 is composed of three thin copper foil electrodes with an asymmetrical layout. The first and second electrodes 2-1 and 2-2 with adjustable width are arranged on the upper surface of the polyimide film dielectric layer 3-1 with a thickness gradient, and are consistent with the gradient of the polyimide film dielectric layer 3-1 and closely fit. The thickness of the polyimide film dielectric layer with a thickness gradient changes nonlinearly from the edge of any electrode on the upper surface to the edge of another electrode on the upper surface. The third electrode 2-3 is arranged on the lower surface of the polyimide film dielectric layer 3-1 and embedded in the polyetheretherketone dielectric substrate 3-2. The first and second electrodes 2-1 and 2-2 are also called upper surface electrodes, and the third electrode 2-3 is also called lower surface electrode. The first and second electrodes 2-1 and 2-2 are generally arranged at the edge of the upper surface of the polyimide film dielectric layer 3-1, also called edge electrodes; correspondingly, the third electrode 2-3 is called the center electrode. According to the design requirements, the spacing between the three electrodes can also be changed according to the requirements, but the three electrodes are all in contact with the surface of the polyimide film dielectric layer 3-1.

[0069] According to an embodiment of the present invention, in the inverted "pin" - shaped layout of the electrodes made of three thin copper - foil materials, the two electrodes on the upper surface are arranged on the surface of a model (such as an aircraft), directly contacting the gas, and playing the role of directly breaking down the gas. However, the electrode on the lower surface is arranged inside the model or in a position where it does not directly contact the gas. Even if it is arranged in a position where it contacts the external gas, it does not play the role of directly breaking down the gas. The width of the lower - surface electrode is different from the widths of the two upper - surface electrodes.

[0070] The widths of the two electrodes on the upper surface are related to the discharge intensity (i.e., the severity of gas breakdown). However, if the width is too small, the accumulated heat locally in the actuator will ablate the electrode. Therefore, there is also a discriminant formula for the width of the upper - surface electrode: E WUp,min ≥E WUp,erosion where E WUp,min is the minimum value of the width of the upper - surface electrode, and E WUp,erosion is the minimum width of the upper - surface electrode at which the accumulated heat ablates the electrode and causes damage.

[0071] The role of the lower - surface electrode is to communicate the electric field between the two upper - surface electrodes. If the lower - surface electrode 2 - 3 is too short, the upper - surface electrodes 2 - 1 and 2 - 2 cannot break down the gas. Therefore, there is also a minimum width value for the lower - surface electrode. There is also a discriminant formula here: E WUnder,min ≥E WUp,breakdown where E WUnder,min is the minimum value of the width of the lower - surface electrode, and E WUnder,breakdown is the minimum width of the lower - surface electrode at which the accumulated heat ablates the electrode and causes damage.

[0072] The thickness distribution of the gradient polyimide - film dielectric layer satisfies the following functional relationship:

[0073]

[0074] where L0 is the thickness of a single - layer dielectric layer, L is the length of the electrode - covered area, x is the coordinate along the air - flow direction, Floor is the floor - function. In this embodiment, the thickness increases from the central electrode (0.1 mm) to the edge electrode (0.5 mm), optimizing the electric - field distribution through the dielectric - layer gradient.

[0075] In order to improve the heat - dissipation performance, micro - heat - dissipation channels are provided on the lower surface of the polyether - ether - ketone dielectric substrate 3 - 2, reducing the damage to the application model (such as an aircraft) caused by the accumulated heat of the actuator discharge. Based on the laminar - flow assumption and ignoring the end - effect, the relationship satisfied by the channel width, depth, and number is obtained:

[0076]

[0077] where k fluid,heatis a coefficient related to fluid properties and heat dissipation requirements calibrated by experiments, W total is the total width of the substrate, d is the depth of the heat dissipation channel, and n is the number of channels. In the embodiment, the heat dissipation channel is filled with aluminum nitride ceramic, which has a temperature resistance of 300° C. The dimensions of the heat dissipation channel are 0.2 mm in width and 0.5 mm in depth.

[0078] (III) Flow field sensor

[0079] The flow field sensor 4 includes two micro sensors distributed on the surface of the dielectric substrate component, the flow field non-contact measurement sensor 4-1 and the flow field measurement sensor 4-2. The flow field non-contact measurement sensor 4-1 and the flow field measurement sensor 4-2 are both integrated into the application model (such as an aircraft). The flow field non-contact measurement sensor 4-1 monitors the rotation rate, strain rate, and high-frequency fluctuation frequency of the lateral velocity and longitudinal velocity of the non-starting phase flow field in real time, and the flow field measurement sensor 4-2 monitors the flow field parameters and performance indicators around the application model in real time.

[0080] In this embodiment, the flow field non-contact measurement sensor 4-1 adopts a multi-modal optical fiber sensor array, integrates a Bragg grating (as a strain rate measurement unit) and a Doppler interferometer (as a velocity fluctuation detection unit), and contains a data fusion processing unit to collect the rotation rate, strain rate and lateral (u x )、Vertical(u y ) High-frequency fluctuation data of speed, with a sampling frequency of up to 1kHz.

[0081] (V) Dynamic Control Module

[0082] The dynamic control module 5 receives the data from the two sensors, and generates a control signal through a voltage-frequency matching algorithm, dynamically adjusts the AC power frequency and the DC bias voltage, and drives the electrode assembly 2 to induce the periodic alternation of the flow rotation rate and the strain rate, thereby realizing real-time flow control of the application model.

[0083] The dynamic control module 5 has a built-in dynamic balance equation algorithm. According to the experiment, the expression between the electric field intensity and the induced flow velocity is obtained as follows:

[0084]

[0085] Where E is the electric field strength vector, u d is the induced flow velocity field vector, k is the coupling coefficient, C k is a constant. is an inverted triangular operator, called nabla, the Hamiltonian operator, representing the gradient; is the divergence, which characterizes the active nature of the field.

[0086] According to electrodynamics, the relationship between electric potential and electric field strength is:

[0087]

[0088] get:

[0089]

[0090] Therefore, the relationship between the variable output voltage value and the induced flow velocity is:

[0091]

[0092] Among them, Δ represents the Laplace operator, and V is the output voltage value of the AC power controlled by the dynamic control module.

[0093] The relationship between the induced flow velocity, the performance parameters of the application model, and the mainstream velocity around the application model is:

[0094] u d =f(u m ,P)

[0095] Among them, u m is the mainstream speed around the application model, which is a static parameter. P is the performance parameter of the application model, which is the parameter to be paid attention to. The specific form of expression f depends on the form of the application model.

[0096] The core of the dynamic control module 5 is to control the electric field intensity distribution by controlling the electric potential, that is, the voltage, so as to affect the plasma clusters generated by the breakdown gas of the actuator, and use the plasma clusters to change the local flow of the model. Finally, the performance parameter P of interest is controlled. The dynamic control unit 5 uses a recursive neural network to learn the relationship between the performance parameter P and other parameters.

[0097] The dynamic control module 5 includes a Transformer-based recursive neural network (RNN) deep learning prediction unit, whose input includes the induced flow velocity component u x and u y The relationship between the fluctuation frequency ratio, the velocity field rotation rate-velocity field strain rate phase difference, the velocity field fluctuation frequency and the flow field velocity u is output as a mathematical relationship of the optimal voltage V-flow field velocity u-performance parameter P combination parameters.

[0098] The relationship between flow field velocity and voltage:

[0099] V=αu x +βu y +γ

[0100] Among them, α, β, and γ are constants obtained through experiments or data fitting.

[0101] The relationship between the induced flow velocity fluctuation frequency ratio and the flow field velocity u is solved through spectrum analysis and deep learning:

[0102]

[0103] u=g(f1,Ω,S)

[0104] Where g is the relationship learned by the neural network. Ω is the velocity field rotation rate, and S is the velocity field strain rate. u x,frequency 、u y,frequency Refers to the velocity component u x and u y The frequency of fluctuation.

[0105] Velocity field rotation rate and strain rate phase difference:

[0106] Δφ=arg(Ω)-arg(S)

[0107] Performance parameters P:

[0108] P=f(u,u x ,u y ,Ω,S,Δφ,V)

[0109] The expression of the loss function that measures the degree of agreement (error) between the prediction results of the network model and the real data is:

[0110]

[0111] Among them, V pred and P pred is the predicted voltage and performance parameter value, V true and P true is the corresponding true observation value, λ is the weighting coefficient of the performance parameter, and N refers to the number of samples.

[0112] When the loss function meets the specified accuracy (for example, the relationship Loss≤10 -6 ), the training ends, that is, the parameters reach the optimal value, and the parameters are output and set.

[0113] The parameters that the entire dynamic control module obtains through fitting and training and needs to output directly are voltage V, rotation rate Ω, and shear rate S, and the rest of the parameters are affected by them. After training and learning, the deep learning prediction unit finally determines the coefficients in these formulas and the calculation relationship, and makes predictions based on the trained network model and outputs the prediction results.

[0114] According to an embodiment of the present invention, the working method of the control device is as follows: first, discharge initiation and flow field induction: a composite waveform is applied by an AC power supply and a DC bias power supply to generate a continuous sliding dielectric barrier discharge (SL-DBD) between electrodes, ionizing the gas to generate a charged particle flow. Then, the non-starting phase dynamic regulation is entered: the sensor monitors the flow field parameters in real time, and the dynamic control module adjusts the power output according to the feedback.

[0115] Reference Figure 3 , the specific working process of the dynamic control module 5 is as follows:

[0116] (1) receiving the flow velocity, basic flow and other parameter data from two sensor modules 4-1 and 4-2 (i.e., flow field flow velocity measurement sensors);

[0117] (2) Process the data according to the simplified equation Processing is performed to calculate the coefficient k and constant C in the three-dimensional state k , which provides a basis for fitting the coefficients of V and velocity component equations in the next step;

[0118] (3) Fitting the relationship between voltage and flow rate V = αu x +βu y +γ, calculate the coefficients and input V into the recurrent neural network as a constant to store it for the subsequent calculation of the Loss function;

[0119] The coefficients α, β, and γ are calculated using the following relationship:

[0120] k=w T Φ(α,β)

[0121] where w T is the weight vector. The expression of Φ(α,β) is as follows:

[0122] Φ(α,β)=[1,α,β,α 2 ,αβ,β 2 ,α 3 ,α 2 β,…] T

[0123] C k and γ satisfy the following relationship:

[0124] γ=a0+a1C k +a2C k 2 +a3C k 3 +a4C k 4 +a5C k 5 +…a i Ck i

[0125] where a i are the fitting coefficients.

[0126] (4) Call the recursive neural network for training and give the relationship between the induced flow velocity fluctuation frequency ratio f1, flow field velocity u, velocity field rotation rate Ω, and velocity field strain rate S. The meaning of each velocity is as follows: u m u is the mainstream velocity around the application model; d is the induced flow velocity; u is the velocity of the mainstream after the excitation induced flow is coupled with the plasma excitation generated in the atmosphere after the exciter breaks down the gas.

[0127] (5) Process the parameters of fitting and training to obtain the relationship between performance parameters and parameters P = f(u,u x ,u y ,Ω,S,Δφ).

[0128] (6) According to the loss function The residual is used to judge the rationality of the parameters V and performance parameters P obtained through training. When the parameters are small enough, that is, Loss ≤ 10 -6 , at this time the training process ends, the trained parameters are output, and the parameters are set.

[0129] Setting parameters is to regulate the power supply: switch high-frequency or low-frequency mode according to the prediction results. When the strain rate is detected to be dominant, switch to the first frequency (low frequency) mode to stabilize the flow field; when the rotation rate is detected to be dominant, switch to the second frequency (high frequency) mode to enhance the vortex structure induction. The gradient dielectric layer distributes the non-uniform electric field to enhance the lateral velocity u x The high-frequency components of the flow are suppressed to suppress flow separation.

[0130] The dominant position of strain rate and rotation rate is different, and the fluctuation frequency of the flow field is different. When the rotation rate is dominant, the fluctuation frequency is high, and vice versa. This ultimately leads to different flow states of induced excitation, and different effects of excitation and regulation of the flow field.

[0131] The discriminant of dominance is as follows:

[0132]

[0133] Where Q is the discriminant symbol. When Q is greater than 0, the rotation rate dominates; otherwise, the strain rate dominates.

[0134] The calculation of rotation rate and strain rate is based on the flow velocity u components in the ox axis and oy axis directions.

[0135] In order to verify the control performance of the control device and the control algorithm, a non-startup phase flow field control effect comparison experiment was carried out. Comparison scenario: Particle image velocimetry (PIV) results comparison between the uncontrolled state (turbulent flow field) and the controlled state (stable flow field).

[0136] The experiment used NACA0015 airfoil flow separation suppression as the application scenario. The experiment was conducted in the Key Laboratory of Unsteady Aerodynamics and Flow Control of the Ministry of Industry and Information Technology of Nanjing University of Aeronautics and Astronautics. The device was installed on the leading edge of the airfoil, such as Figure 4 The electrode assembly is opposite to the incoming flow direction. The incoming flow velocity range is 15-25 m / s, corresponding to the Reynolds number range of 2×10 5 ~3×10 5 .

[0137] Start the continuous mode, the power module outputs a composite waveform (AC 5kHz + DC bias 10kV), triggering the sliding dielectric barrier discharge (SL-DBD).

[0138] Non-starting phase regulation: sensor 4-2 measures and processes the velocity signal of the incoming flow, and sensor 4-1 measures and processes the velocity of the SL-DBD excitation-induced flow, including the excitation rotation rate and shear rate;

[0139] The processed results of sensor 4-1 and sensor 4-2 are transmitted to the dynamic control module according to Further process the data to determine coefficients and constants;

[0140] Will be based on The result of the processed data is transmitted to the dynamic control module V = αu x +βu y +γ part, then fit it and transfer the fitting result to the recurrent neural network part;

[0141] The recursive neural network performs data training and prediction, and outputs the prediction results obtained based on the training to the performance parameter P prediction part. In this case, the performance parameter P is the lift-to-drag ratio K (the ratio of the lift coefficient to the drag coefficient);

[0142] The performance parameter P prediction part starts to measure the flow field velocity u and performance parameter K data under different input parameters (parameters input by the recursive neural network) in real time. At the same time, according to the formula Distinguish high-frequency / low-frequency patterns. Finally, give the fitting relationship and input the result into the loss function residual discriminant;

[0143] The loss function discriminant is used for judgment. If the 10e-6 condition is met, the parameters are output and set. Otherwise, the training is returned to continue.

[0144] The dynamic control module ends and starts to set parameters and perform airfoil flow separation control.

[0145] The experimental results show that the dynamic control module switches to the high-frequency mode (150Hz) to enhance the vortex structure induction ability; the gradient dielectric layer directionally enhances the lateral velocity u x component (50Hz), inhibiting the development of separation vortex and stabilizing the flow field, such as Figure 5 Velocity distribution of the flow field. When t=0ms, the excitation is not turned on, that is, there is no regulation. The rest are the dynamic effects of the excitation regulation flow field after the excitation is turned on. The results show that SL-DBD has the ability to suppress airfoil flow separation, promote flow reattachment, and enhance laminarity in the continuous mode.

[0146] Figure 6 is the quantitative control result after turning on the excitation. Where K is the lift-to-drag ratio, Re is the Reynolds number, U ∞ is the velocity of the incoming flow. δK is defined as:

[0147]

[0148] The wind tunnel test results show that the lift-to-drag ratio K of the airfoil is increased by 2.7 to 3.2 times after the excitation control is applied. The device achieves the expected control effect, promotes the reattachment of the flow, suppresses the flow separation, and verifies the control effectiveness and engineering practicability.

[0149] Embodiment 2

[0150] This embodiment provides a method for dynamically controlling a non-starting phase flow field of a continuous sliding dielectric barrier discharge, comprising the following steps:

[0151] S1. A composite excitation waveform is applied by an AC power supply and a DC power supply to generate a sliding dielectric barrier discharge plasma excitation between electrodes, wherein the electrodes include three asymmetrically arranged electrodes, and the three electrodes are arranged in an inverted "pin" shape on a dielectric substrate assembly, wherein the dielectric substrate assembly includes a first material substrate and a second material dielectric layer with a gradient thickness arranged on the first material substrate, wherein the first electrode and the second electrode are separately arranged on the upper surface of the second material dielectric layer, and the third electrode is arranged on the lower surface of the second material dielectric layer, and the three electrodes are all in contact with the surface of the second material dielectric layer;

[0152] S2, the first sensor module collects the rotation rate, strain rate and velocity fluctuation characteristics of the non-starting phase flow field in real time, and the second sensor module measures the flow field parameters and performance indicators around the application model in real time;

[0153] S3. The dynamic control module controls the composite excitation waveform output by the power module based on the sensor data received from the first and second sensor modules, using the dynamic balance equation based on the coupling relationship between the electric field and the induced flow velocity and the deep learning prediction unit based on the recursive neural network, through the voltage-frequency-performance parameter matching algorithm, and adjusts the alternating period of the induced flow rotation rate and the strain rate to achieve the purpose of dynamic flow field control. The input of the voltage-frequency-performance parameter matching algorithm includes the fluctuation frequency ratio of the lateral velocity component and the longitudinal velocity component of the induced flow, the phase difference between the velocity field rotation rate and the velocity field strain rate, and the relationship between the velocity field fluctuation frequency ratio and the flow field velocity. The output is the mathematical relationship of the optimal voltage V-flow field velocity u-application model performance parameter P combination parameter.

[0154] Furthermore, when the dynamic control module detects that the strain rate is dominant according to the prediction results, it switches to the first frequency (low frequency, such as 20 Hz) mode to stabilize the flow field; when the rotation rate is detected to be dominant, it switches to the second frequency (high frequency, such as 50 Hz) mode to enhance the vortex structure induction.

[0155] Embodiment 3

[0156] This embodiment provides an electronic device, including: a processor, a memory, and a computer program, wherein the computer program is stored in the memory and configured to be executed by the processor, and when the computer program is executed by the processor, the following steps are implemented:

[0157] According to the sensor data received from the flow field sensor module, the dynamic balance equation based on the coupling relationship between the electric field and the induced flow velocity and the deep learning prediction unit based on the recursive neural network are used to control the composite excitation waveform output by the power supply module through the voltage-frequency-performance parameter matching algorithm, and adjust the alternating period of the induced flow rotation rate and the strain rate to achieve the purpose of dynamic flow field control. The input of the voltage-frequency-performance parameter matching algorithm includes the fluctuation frequency ratio of the induced flow transverse velocity component and the longitudinal velocity component, the phase difference between the velocity field rotation rate and the velocity field strain rate, and the relationship between the velocity field fluctuation frequency ratio and the flow field velocity. The output is the mathematical relationship of the optimal voltage V-flow field velocity u-application model performance parameter P combination parameter;

[0158] The flow field sensor module includes a first sensor module for real-time monitoring of the rotation rate, strain rate and velocity fluctuation characteristics of the non-starting phase flow field, and a second sensor module for real-time monitoring of the flow field parameters and performance indicators around the application model;

[0159] The induced flow is generated by the electrode assembly driven by the power supply module. The power supply module includes an AC power supply and a DC power supply, and outputs a composite excitation waveform. The electrode assembly consists of three electrodes with an asymmetric layout. The three electrodes are arranged in an inverted "pin" shape on the dielectric substrate assembly. The dielectric substrate assembly includes a first material substrate and a second material dielectric layer with a thickness gradient arranged on the first material substrate. The first electrode and the second electrode are separately arranged on the upper surface of the second material dielectric layer, and the third electrode is arranged on the lower surface of the second material dielectric layer. All three electrodes are in contact with the surface of the second material dielectric layer.

[0160] Embodiment 4

[0161] This embodiment provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the following steps are implemented:

[0162] According to the sensor data received from the flow field sensor module, using the dynamic balance equation based on the coupling relationship between the electric field and the induced flow velocity and the deep learning prediction unit based on the recurrent neural network, through the voltage-frequency-performance parameter matching algorithm, the composite excitation waveform output by the power supply module is regulated, the rotation rate of the induced flow and the alternating period of the strain rate are adjusted, and the purpose of dynamic regulation of the flow field is achieved. The input of the voltage-frequency-performance parameter matching algorithm includes the fluctuation frequency ratio of the transverse velocity component and the longitudinal velocity component of the induced flow, the phase difference between the rotation rate of the velocity field and the strain rate of the velocity field, and the relationship between the fluctuation frequency ratio of the velocity field and the flow field velocity. The output is a mathematical relationship of the combined parameters of the optimal voltage V - flow field velocity u - application model performance parameter P.

[0163] The flow field sensor module includes a first sensor module that monitors the rotation rate, strain rate, and velocity fluctuation characteristics of the non-starting phase flow field in real time, and a second sensor module that monitors the flow field parameters and performance indicators around the application model in real time.

[0164] The induced flow is generated by the electrode assembly driven by the power supply module. The power supply module includes an AC power supply and a DC power supply, and outputs a composite excitation waveform. The electrode assembly consists of three electrodes with an asymmetric layout. The three electrodes are arranged in an inverted "pin" shape on the dielectric substrate assembly. The dielectric substrate assembly includes a first material substrate and a second material dielectric layer with a thickness gradient arranged on the first material substrate. The first electrode and the second electrode are separately arranged on the upper surface of the second material dielectric layer, and the third electrode is arranged on the lower surface of the second material dielectric layer. All three electrodes are in contact with the surface of the second material dielectric layer.

[0165] For the specific layout and calculation process implementation of the power supply module, electrode assembly, dielectric substrate assembly, flow field sensor, and dynamic control module, refer to the description in Embodiment 1, and will not be elaborated here.

Claims

1. A non-starting phase flow field dynamic control device for continuous sliding dielectric barrier discharge, characterized in that: include: A power module, an electrode assembly, a dielectric substrate assembly, a flow field sensor and a dynamic control module, wherein the power module includes an AC power supply and a DC power supply, and outputs a composite excitation waveform; the dielectric substrate assembly includes a first material substrate and a second material dielectric layer with a gradient thickness arranged on the first material substrate; the electrode assembly is composed of three asymmetrically arranged electrodes, wherein the first electrode and the second electrode are separately arranged on the upper surface of the second material dielectric layer, and the third electrode is arranged on the lower surface of the second material dielectric layer, and the three electrodes are arranged in an inverted "pin" shape and are bonded to the surface of the second material dielectric layer; the flow field sensor module includes a first sensor module for real-time monitoring of the rotation rate, strain rate and velocity fluctuation characteristics of the non-starting phase flow field, and a second sensor module for real-time monitoring of the flow field parameters and performance indicators around the application model; The dynamic control module, based on the sensor data received from the flow field sensor module, uses the dynamic balance equation based on the coupling relationship between the electric field and the induced flow velocity and the deep learning prediction unit based on the recursive neural network to control the composite excitation waveform output by the power module through the voltage-frequency-performance parameter matching algorithm, adjust the alternating period of the induced flow rotation rate and the strain rate to achieve the control purpose. The input of the voltage-frequency-performance parameter matching algorithm includes the fluctuation frequency ratio of the lateral velocity component and the longitudinal velocity component of the induced flow, the phase difference between the velocity field rotation rate and the velocity field strain rate, and the relationship between the velocity field fluctuation frequency ratio and the flow field velocity. The output is the mathematical relationship of the optimal voltage V-flow field velocity u-application model performance parameter P combination parameter.

2. The device according to claim 1, characterized in that The electrode assembly is composed of three thin copper foil electrodes, wherein the first electrode and the second electrode arranged on the upper surface of the second material dielectric layer have a first width, and the third electrode arranged on the lower surface of the second material dielectric layer has a second width, and the first width is different from the second width.

3. The device according to claim 1, characterized in that The thickness distribution of the second material dielectric layer with a thickness gradient satisfies the following functional relationship: Where L0 is the thickness of a single dielectric layer, L is the length of the electrode coverage area, x is the coordinate along the airflow direction, 0≤x≤L, [] Floor is the floor function.

4. The device according to claim 1, characterized in that The first material substrate is a polyetheretherketone dielectric substrate, and the second material dielectric layer is a polyimide film dielectric layer.

5. The device according to claim 1, characterized in that The lower surface of the second material dielectric layer is provided with micro heat dissipation channels, and the channel size and number satisfy the following relationship: where k fluid,heat is a coefficient related to fluid properties and heat dissipation requirements calibrated by experiments, W total is the total width of the first material substrate, d is the channel depth, and n is the number of channels.

6. The device according to claim 1, characterized in that The dynamic equilibrium equation based on the coupling relationship between the electric field and the induced flow velocity includes: According to the experiment, the expression between the electric field intensity and the induced flow velocity is: ▽·E=k▽·u d +C k Where E is the electric field strength vector, u d is the induced flow velocity field vector, k is the coupling coefficient, C k is a constant; ▽ is an inverted triangular operator, indicating the gradient; according to electrodynamics, the relationship between the variable output voltage value and the induced flow velocity is obtained as follows: -ΔV=k▽·u d +C k Wherein, Δ represents the Laplace operator, and V is the output voltage value of the AC power controlled by the dynamic control module; The relationship between the induced flow velocity, the performance parameters of the application model, and the mainstream velocity around the application model is: in d =f(u m ,P) Among them, u m is the mainstream velocity around the application model, P is the performance parameter of the application model, and the specific form of expression f depends on the form of the application model.

7. The device according to claim 6, characterized in that The control logic of the dynamic control module is as follows: Receive the data of the flow field sensor module, according to -ΔV=k▽·u d +C k Processing, calculation of coupling coefficient k and constant C k ; Fit the relationship between voltage and flow rate V = αu x +βu y +γ, according to the coupling coefficient k and the constant C k Calculate the coefficients α, β, and γ; The deep learning prediction unit based on recursive neural network is called to solve the relationship between the induced flow velocity fluctuation frequency ratio f1 and the flow field velocity u through spectrum analysis and deep learning: u=g(f1,Ω,S) Where g is the relationship learned by the recurrent neural network, u x,frequency 、u y,frequency are the induced flow transverse and longitudinal velocity components u x and u y The frequency of fluctuations; And the velocity field rotation rate and strain rate phase difference Δφ: Δφ=arg(Ω)-arg(S) Where Ω is the velocity field rotation rate, S is the velocity field strain rate; Finally determine the performance parameter P: P=f(u,u x ,u y ,Ω,S,Δφ)。 8. A method for dynamically controlling the non-starting phase flow field of a continuous sliding dielectric barrier discharge, characterized in that: The following steps are involved: S1. A composite excitation waveform is applied by an AC power supply and a DC power supply to generate a sliding dielectric barrier discharge plasma excitation between electrodes, wherein the electrodes include three asymmetrically arranged electrodes, and the three electrodes are arranged in an inverted "pin" shape on a dielectric substrate assembly, wherein the dielectric substrate assembly includes a first material substrate and a second material dielectric layer with a gradient thickness arranged on the first material substrate, wherein the first electrode and the second electrode are separately arranged on the upper surface of the second material dielectric layer, and the third electrode is arranged on the lower surface of the second material dielectric layer, and the three electrodes are all in contact with the surface of the second material dielectric layer; S2, the first sensor module collects the rotation rate, strain rate and velocity fluctuation characteristics of the non-starting phase flow field in real time, and the second sensor module measures the flow field parameters and performance indicators around the application model in real time; S3. The dynamic control module controls the composite excitation waveform output by the power module based on the sensor data received from the first and second sensor modules, using the dynamic balance equation based on the coupling relationship between the electric field and the induced flow velocity and the deep learning prediction unit based on the recursive neural network, through the voltage-frequency-performance parameter matching algorithm, and adjusts the alternating period of the induced flow rotation rate and the strain rate to achieve the purpose of dynamic flow field control. The input of the voltage-frequency-performance parameter matching algorithm includes the fluctuation frequency ratio of the lateral velocity component and the longitudinal velocity component of the induced flow, the phase difference between the velocity field rotation rate and the velocity field strain rate, and the relationship between the velocity field fluctuation frequency ratio and the flow field velocity. The output is the mathematical relationship of the optimal voltage V-flow field velocity u-application model performance parameter P combination parameter.

9. An electronic device, characterized in that: include: A processor, a memory, and a computer program, wherein the computer program is stored in the memory and configured to be executed by the processor, and when the computer program is executed by the processor, the following steps are implemented: According to the sensor data received from the flow field sensor module, the dynamic balance equation based on the coupling relationship between the electric field and the induced flow velocity and the deep learning prediction unit based on the recursive neural network are used to control the composite excitation waveform output by the power supply module through the voltage-frequency-performance parameter matching algorithm, and adjust the alternating period of the induced flow rotation rate and the strain rate to achieve the purpose of dynamic flow field control. The input of the voltage-frequency-performance parameter matching algorithm includes the fluctuation frequency ratio of the induced flow transverse velocity component and the longitudinal velocity component, the phase difference between the velocity field rotation rate and the velocity field strain rate, and the relationship between the velocity field fluctuation frequency ratio and the flow field velocity. The output is the mathematical relationship of the optimal voltage V-flow field velocity u-application model performance parameter P combination parameter; The flow field sensor module includes a first sensor module for real-time monitoring of the rotation rate, strain rate and velocity fluctuation characteristics of the non-starting phase flow field, and a second sensor module for real-time monitoring of the flow field parameters and performance indicators around the application model; The induced flow is generated by the electrode assembly driven by the power supply module. The power supply module includes an AC power supply and a DC power supply, and outputs a composite excitation waveform. The electrode assembly consists of three electrodes with an asymmetric layout. The three electrodes are arranged in an inverted "pin" shape on the dielectric substrate assembly. The dielectric substrate assembly includes a first material substrate and a second material dielectric layer with a thickness gradient arranged on the first material substrate. The first electrode and the second electrode are separately arranged on the upper surface of the second material dielectric layer, and the third electrode is arranged on the lower surface of the second material dielectric layer. All three electrodes are in contact with the surface of the second material dielectric layer.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by the processor, the following steps are implemented: According to the sensor data received from the flow field sensor module, using the dynamic balance equation based on the coupling relationship between the electric field and the induced flow velocity and the deep learning prediction unit based on the recurrent neural network, through the voltage-frequency-performance parameter matching algorithm, the composite excitation waveform output by the power supply module is regulated, the rotation rate of the induced flow and the alternating period of the strain rate are adjusted, and the purpose of dynamic regulation of the flow field is achieved. The input of the voltage-frequency-performance parameter matching algorithm includes the fluctuation frequency ratio of the transverse velocity component and the longitudinal velocity component of the induced flow, the phase difference between the rotation rate of the velocity field and the strain rate of the velocity field, and the relationship between the fluctuation frequency ratio of the velocity field and the flow field velocity. The output is a mathematical relationship of the combined parameters of the optimal voltage V - flow field velocity u - application model performance parameter P. The flow field sensor module includes a first sensor module that real-time monitors the rotation rate, strain rate, and velocity fluctuation characteristics of the non-starting phase flow field, and a second sensor module that real-time monitors the flow field parameters and performance indicators around the application model. The induced flow is generated by the electrode assembly driven by the power supply module. The power supply module includes an AC power supply and a DC power supply, and outputs a composite excitation waveform. The electrode assembly consists of three electrodes with an asymmetric layout. The three electrodes are arranged in an inverted "pin" shape on the dielectric substrate assembly. The dielectric substrate assembly includes a first material substrate and a second material dielectric layer with a thickness gradient arranged on the first material substrate. The first electrode and the second electrode are separately arranged on the upper surface of the second material dielectric layer, and the third electrode is arranged on the lower surface of the second material dielectric layer. All three electrodes are in contact with the surface of the second material dielectric layer.

Images