Plasma parameter probe-microwave collaborative diagnosis system

Abstract

The invention discloses a plasma parameter probe-microwave collaborative diagnosis system. The plasma parameter probe-microwave collaborative diagnosis system comprises a diagnosis probe, a reflecting antenna, a vector network analyzer and a computer terminal, according to the method, global and local information of plasma parameters is provided through the diagnosis probe and the reflecting antenna, inversion is performed on S11 data and plasma electron density of a monitoring point where the diagnosis probe is located through the trained neural network model, and plasma density distribution and plasma collision frequency are obtained. Cooperative diagnosis of the plasma is completed, and the density distribution of the plasma and the collision frequency of the plasma can be rapidly and accurately obtained.

Description

Technical Field

[0001] The invention relates to the technical field of plasma diagnosis, and in particular to a plasma parameter probe-microwave collaborative diagnosis system. Background Art

[0002] When an aircraft flies at a supersonic speed between 20 and 100 kilometers above the ground, its head surface will rub against the air violently to form a plasma sheath, which will cause attenuation and phase shift of electromagnetic waves, seriously affecting the communication quality and may even cause radio interruption, i.e. the "black barrier" phenomenon. In order to solve the "black barrier" phenomenon, it is necessary to diagnose the plasma sheath parameters, which generally include electron density and collision frequency.

[0003] Microwave diagnosis is one of the main methods of plasma diagnosis, which includes two types: invasive and non-invasive. Invasive microwave diagnosis mainly uses probes, including microwave resonance probes, plasma impedance probes, plasma absorption probes, etc. Non-invasive microwave diagnosis can be divided into microwave reflection method and microwave transmission method. Microwave reflection method uses the amplitude and phase obtained by the reflection of electromagnetic waves on the plasma interface to diagnose plasma parameters, and microwave transmission method uses the attenuation of electromagnetic waves through the plasma to infer the electron density and collision frequency.

[0004] Existing plasma diagnostic technology has the following disadvantages:

[0005] 1. The current plasma diagnostic methods are mainly divided into two categories: invasive diagnostic methods and non-invasive diagnostic methods. The invasive diagnostic methods mainly include electrostatic probes and microwave probes. The invasive diagnostic methods will cause certain contamination to the plasma sheath to be tested.

[0006] 2. When the probe is contaminated by impurities, the volt-ampere characteristics will be affected, and the diagnostic accuracy will be greatly affected by the measurement environment; and the probe will destroy the aerodynamic shape of the aircraft and affect the normal flight of the aircraft.

[0007] 3. The traditional microwave reflection method is developed from radar technology. It uses the cutoff phenomenon of electromagnetic waves below the characteristic frequency of plasma during propagation, and calculates the electron density at the point where total reflection occurs through the incident wave frequency. The traditional microwave reflectometer can diagnose the electron density distribution, but it requires that there is an obvious reflection interface in the electron density distribution, and it cannot diagnose the collision frequency at the same time.

[0008] 4. The diagnostic equipment is relatively complex because it is necessary to measure the arrival time of the reflected signal. The collision frequency of the plasma sheath on the surface of a hypersonic vehicle is generally high, and there is no obvious reflection cross section. The use of traditional microwave reflection methods will result in large diagnostic errors. Summary of the invention

[0009] In view of the above-mentioned deficiencies in the prior art, the present invention provides a plasma parameter probe-microwave collaborative diagnosis system which solves the problem that the existing plasma diagnosis technology is difficult to accurately diagnose the plasma density distribution and plasma collision frequency at the same time in actual scenarios.

[0010] In order to achieve the above-mentioned object of the invention, the technical solution adopted by the present invention is:

[0011] A plasma parameter probe-microwave collaborative diagnosis system is provided, which comprises a diagnosis probe, a reflection antenna, a vector network analyzer and a computer terminal; the diagnosis probe and the reflection antenna are arranged on the surface of an aircraft; the reflection antenna is connected to the signal input end of the vector network analyzer; the output ends of the vector network analyzer and the diagnosis probe are respectively connected to the computer terminal;

[0012] Reflection antenna and vector network analyzer to obtain S11 data caused by plasma;

[0013] a diagnostic probe for acquiring a probe current induced in the plasma environment;

[0014] The computer terminal is used to calculate the plasma electron temperature according to the probe current; calculate the plasma electron density at the monitoring point where the diagnostic probe is located according to the plasma electron temperature; load the trained neural network model, invert the S11 data and the plasma electron density at the monitoring point where the diagnostic probe is located through the trained neural network model, obtain the plasma density distribution and plasma collision frequency, and complete the collaborative diagnosis of the plasma.

[0015] Furthermore, the diagnostic probe and the reflective antenna are arranged coplanarly.

[0016] Furthermore, the expression for calculating the plasma electron temperature based on the probe current is:

[0017]

[0018] Where T e is the plasma electron temperature; e is the electron charge; V p is the shell point; I is the probe current; I i0 is the saturated ion current; U is the voltage; ln(.) represents the logarithmic function; d(.) represents the differential.

[0019] Furthermore, the expression for calculating the plasma electron density at the monitoring point where the diagnostic probe is located according to the plasma electron temperature is:

[0020]

[0021] Where n iis the plasma electron density at the monitoring point where the diagnostic probe is located; S is the collection area of ​​the diagnostic probe; k is the Boltzmann constant; m i It is the plasma quality at the monitoring point where the diagnostic probe is located.

[0022] Furthermore, the diagnostic probe comprises a metal rod and a ceramic tube sleeved on the outer surface of the metal rod, and the upper end of the metal rod passes through the ceramic tube; the material of the metal rod is tungsten, molybdenum or tantalum.

[0023] Furthermore, the neural network model adopts a fully connected neural network model.

[0024] Furthermore, the training process of the neural network model is:

[0025] A simulation model is constructed using simulation software. Different plasma parameters are set during the simulation process. The S11 data caused by the plasma is obtained through a reflection antenna and a vector network analyzer. The probe current caused in the plasma environment is obtained through a diagnostic probe. The plasma electron density corresponding to the data obtained by the diagnostic probe is calculated. The plasma collision frequency is obtained by querying the NRL manual. The S11 data and plasma electron density measured during the simulation process are used as inputs to the neural network model, and the density distribution parameters of the actual plasma and the plasma collision frequency during the simulation are used as labels to train the neural network model until the neural network model converges to obtain a trained neural network model.

[0026] Furthermore, the specific method of inverting the S11 data and the plasma electron density at the monitoring point where the diagnostic probe is located by the trained neural network model to obtain the plasma density distribution and plasma collision frequency is as follows:

[0027] The S11 data and the plasma electron density at the monitoring point where the diagnostic probe is located are used as the input of the trained neural network model, and the plasma collision frequency and electron density distribution parameters are obtained by inverting the trained neural network model; the electron density distribution parameters are substituted into the double Gaussian distribution model to obtain the density distribution of the plasma; the electron density distribution parameters include peak position, decrease coefficient, increase coefficient and electron density peak value.

[0028] The beneficial effects of the present invention are:

[0029] 1. The present invention provides global and local information on plasma parameters through diagnostic probes and reflective antennas, and uses a neural network model to complete in-situ diagnosis of plasma profiles, which can quickly and accurately obtain plasma density distribution and plasma collision frequency.

[0030] 2. The radiation aperture of the reflective antenna of the present invention is within 100 mm, the working bandwidth can cover the 4-26.5 GHz frequency band, it has high gain within the frequency band, and can withstand high temperatures. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] Figure 1 This is the structural diagram of the system;

[0032] Figure 2 Schematic diagram of the potential distribution around the electrostatic diagnostic probe and reference electrode;

[0033] Figure 3 Provide a schematic diagram for explaining the characteristic curve of the electrostatic diagnostic probe;

[0034] Figure 4 Design a flow chart for a neural network model;

[0035] Figure 5 It is the software interface of plasma parameter probe-microwave collaborative diagnosis system;

[0036] Figure 6 is a schematic diagram of the cross section of the plasma sheath;

[0037] Figure 7 Schematic diagram of the structure of the reflector antenna. DETAILED DESCRIPTION

[0038] The specific implementation modes of the present invention are described below so that those skilled in the art can understand the present invention. However, it should be clear that the present invention is not limited to the scope of the specific implementation modes. For those of ordinary skill in the art, as long as various changes are within the spirit and scope of the present invention as defined and determined by the attached claims, these changes are obvious, and all inventions and creations utilizing the concept of the present invention are protected.

[0039] like Figure 1 As shown, the plasma parameter probe-microwave collaborative diagnosis system includes a diagnosis probe, a reflection antenna, a vector network analyzer and a computer terminal; the diagnosis probe and the reflection antenna are arranged on the surface of the aircraft; the reflection antenna is connected to the signal input end of the vector network analyzer; the output ends of the vector network analyzer and the diagnosis probe are respectively connected to the computer terminal; the diagnosis probe and the reflection antenna are arranged coplanar;

[0040] Reflection antenna and vector network analyzer to obtain S11 data caused by plasma;

[0041] a diagnostic probe for acquiring a probe current induced in the plasma environment;

[0042] The computer terminal is used to calculate the plasma electron temperature according to the probe current; calculate the plasma electron density at the monitoring point where the diagnostic probe is located according to the plasma electron temperature; load the trained neural network model, invert the S11 data and the plasma electron density at the monitoring point where the diagnostic probe is located through the trained neural network model, obtain the plasma density distribution and plasma collision frequency, and complete the collaborative diagnosis of the plasma.

[0043] In the specific implementation process, the vector network analyzer can be connected through a microwave cable, and the collected data can be transmitted to the computer terminal in real time for processing, including setting, reading, and displaying the S11 data of the vector network analyzer. The specific process is as follows:

[0044] 1. Set the start frequency, end frequency and acquisition interval frequency:

[0045] Use LabVIEW's Front Panel to create Numeric Controls named "StartFrequency", "Stop Frequency", and "Frequency Interval". These controls will be used to input the start frequency, stop frequency, and sampling interval frequency.

[0046] 2. Collect S11 data:

[0047] Use VISA or GPIB function libraries in LabVIEW to configure and communicate with the vector network analyzer. Send commands to the vector network analyzer to set the start frequency, end frequency, and acquisition interval frequency. Use the VISA Read function to read the S11 data returned from the vector network analyzer.

[0048] 3. Display amplitude and phase:

[0049] Use LabVIEW chart controls, such as Waveform Chart or Graph, to display the read S11 data on amplitude and phase charts.

[0050] When the computer terminal is performing inversion, the System Exec VI of LabVIEW can be used to execute the executable file (EXE) of the external neural network model. The collected S11 data and the plasma electron density at the monitoring point where the diagnostic probe is located are passed as input parameters to the executable file of the neural network model. The inversion results of the plasma electron density distribution parameters are read from the executable file of the neural network model. The inversion results of the plasma electron density distribution parameters are brought into the double Gaussian distribution model to obtain the plasma electron density distribution results. The obtained plasma electron density distribution results are displayed on the chart using the chart control of LabVIEW, and the horizontal axis is the profile distance.

[0051] In the specific implementation process, the diagnostic probe is a refractory metal (tungsten, molybdenum, tantalum) wire or rod, which is generally covered with a ceramic tube, leaving only a small section exposed to the plasma. Sometimes a metal (stainless steel) tube is added to the outside of the ceramic tube for electrostatic shielding. When in use, a DC voltage is applied to the probe core, and the other end is connected to a reference electrode in the plasma, or to the conductive vacuum chamber wall and grounded. If the vacuum chamber is an insulator, a reference electrode with a certain area must be used to keep the potential on it stable. The applied voltage and the current are measured, and the IV characteristic curve can be obtained by scanning the applied voltage.

[0052] The voltage on the general characteristic curve is the voltage U applied in the experiment, not the potential V of the probe relative to the plasma. p U is V p and the voltage U of the plasma relative to the reference electrode to which the probe circuit ground is connected R If the reference electrode is the wall of the container, it is generally negatively charged to the plasma, U R is the height of the plasma sheath. p When plotting the variable, the curve should be horizontally shifted. On this basis, the probe current I of the diagnostic probe is:

[0053]

[0054] Among them I e0 is the saturation electron current; T e is the plasma electron temperature; e is the electron charge; V p is the shell point; I i0 is the saturated ion current; exp(.) represents the exponential function.

[0055] Then the expression for calculating the plasma electron temperature based on the probe current is:

[0056]

[0057] Where U is voltage; ln(.) represents logarithmic function; d(.) represents differential.

[0058] That is, the electron temperature is calculated from the slope of the middle part of the semi-logarithmic coordinate characteristic curve. This calculation does not require the zero point of the horizontal axis, so the voltage applied in the experiment is used to obtain the value of Figure 2 The characteristic curve data shown is sufficient.

[0059] The intersection of the characteristic curve and the horizontal axis, that is, the voltage when the circuit is open, is the floating potential, denoted as V FIt is different from the plasma sheath between the plasma potential. Due to the size and shape, this sheath is not equal to the sheath on the reference electrode, so the floating potential of the open-circuit probe is generally not equal to the reference electrode potential. The left side of the probe characteristic curve corresponds to the very negative voltage, which is a pure ion saturation flow. The ion flow flows to the probe surface at the ion sound speed. The corresponding ion current is as follows Figure 3 On this basis, the expression for calculating the plasma electron density at the monitoring point where the diagnostic probe is located according to the plasma electron temperature is:

[0060]

[0061] Where n i is the plasma electron density at the monitoring point where the diagnostic probe is located; S is the collection area of ​​the diagnostic probe; k is the Boltzmann constant; m i It is the plasma quality at the monitoring point where the diagnostic probe is located.

[0062] In this embodiment, a complex nonlinear fitting function is learned based on the S11 data obtained in advance and the plasma electron density at the monitoring point where the diagnostic probe is located, so that the plasma parameter information can be predicted. This problem is a typical regression problem, and the fully connected neural network is very effective in classification and regression problems, and is also the most basic deep learning network. Therefore, in this embodiment, the neural network model adopts a fully connected neural network model. The training process of the neural network model is:

[0063] The simulation model is constructed by simulation software. Different plasma parameters (including plasma density distribution parameters) are set during the simulation process. The S11 data caused by the plasma is obtained through the reflection antenna and vector network analyzer. The probe current caused in the plasma environment is obtained through the diagnostic probe. The plasma electron density corresponding to the data obtained by the diagnostic probe is calculated. The plasma collision frequency is obtained by querying the NRL manual. The S11 data and plasma electron density measured during the simulation process are used as inputs to the neural network model, and the actual plasma density distribution parameters and plasma collision frequency during the simulation process are used as labels to train the neural network model until the neural network model converges to obtain the trained neural network model. The overall design flow chart for parameter inversion using the neural network model is as follows: Figure 4 shown.

[0064] The S11 data and the plasma electron density at the monitoring point where the diagnostic probe is located are inverted by the trained neural network model to obtain the plasma density distribution and the plasma collision frequency. The specific method is as follows: the S11 data and the plasma electron density at the monitoring point where the diagnostic probe is located are used as the input of the trained neural network model, and the plasma collision frequency and the electron density distribution parameters are obtained by inverting the trained neural network model; the electron density distribution parameters are brought into the double Gaussian distribution model to obtain the plasma density distribution; the electron density distribution parameters include the peak position, the decrease coefficient, the increase coefficient and the electron density peak value.

[0065] like Figure 7 As shown, the reflective antenna adopts a four-ridged horn. By adopting a conical cavity structure to improve the matching of the waveguide section, the gap of the four-ridged structure is optimized to produce a capacitive effect, further improving the performance of the reflective antenna; outside the transition section, the horn shell adopts a larger angle, and inside the transition section, the metal ridge adopts a linear change in the transition section except for the exponential section. Since the antenna gain and beam width are related parameters, the design achieves a compromise between gain and beam width by controlling the length of the antenna aperture and the horn section. The gain of this reflective antenna is greater than 10dBi in the full frequency band, and the half-power beam width is less than 30 degrees in the frequency band below 5GHz, and the half-power beam width is less than 15 degrees in the frequency band above 5GHz.

[0066] In one embodiment of the present invention, the data of the diagnostic probe is collected by an acquisition card, and the probe data collected by the acquisition card and the antenna data collected by the vector network analyzer are transmitted to a computer terminal, and processed by the principle prototype software developed under the LabVIEW environment to obtain the inversion result. The software interface of the plasma parameter probe-microwave collaborative diagnosis system is shown in FIG. Figure 5 The left side is the setting and operation area, where you can set frequency parameters, control the vector network analyzer acquisition process, and perform inversion operations; the right side is the display area, where you can display the acquired S11 amplitude, phase, and inversion results.

[0067] In one embodiment of the present invention, the diagnostic results of the system for plasmas in different states are tested and recorded. The specific testing steps are as follows:

[0068] 1. Build a plasma parameter probe-microwave collaborative diagnosis system and connect the reflection antenna with the vector network analyzer;

[0069] 2. Turn on, preheat and evacuate all instruments and equipment in the system;

[0070] 3. Turn on the discharge plasma system to form a plasma sheath between the generated plasma beam and the antenna protection cover;

[0071] 4. Real-time online measurement of plasma density through a plasma probe array system;

[0072] 5. Set the starting frequency, cutoff frequency and sampling frequency in the inversion software interface, and collect the reflection coefficient S11 and transmission coefficient S21 through the vector network analyzer; record the interval time of each sampling as the time resolution of the diagnostic system;

[0073] 6. Collect data through computer terminals, invert plasma electron density distribution parameters and plasma collision frequency, and generate plasma electron density distribution.

[0074] Obtain various indicator parameters of the diagnostic system based on the collected test data.

[0075] Spatial resolution generally refers to the size or dimension of the smallest unit that can be distinguished in detail on a remote sensing image. In this embodiment, spatial resolution is used to describe the smallest interval unit of the plasma electron density profile.

[0076] For the plasma sheath on the surface of the aircraft, assume that its thickness is z max , with the aircraft surface as the 0 point, establish the coordinate axis perpendicular to the aircraft surface and outward, such as Figure 6 shown.

[0077] The diagnostic result obtained by the experiment is a set of data y = [ne 1 ,ne 2 ,......,ne j ,......,ne m ], corresponding to a set of points on the coordinate axis [z 1 , z 2 ,......,z j ,......,z m ], where z m =z max ,ne j That is, the z axis j The plasma electron density corresponding to the point. When obtaining the diagnosis results, the interval between the points on the z-axis is fixed, and the interval value is the spatial resolution of the diagnosis system.

[0078] Each experiment can obtain the diagnostic results of the diagnostic probe and the diagnostic results of the collaborative diagnostic system respectively. These two sets of results are used for numerical simulation. According to the diagnostic density distribution, a plasma model is established in the simulation software to simulate the transmission coefficient, and the transmission coefficient S21 obtained by simulation is obtained. The simulation results are compared with the actual measured S21 to compare the accuracy of the diagnostic results of the two diagnostic methods. In this process, the point-by-point statistical root mean square error is used as the accuracy judgment standard, and the calculation formula is:

[0079]

[0080] Where X sim,i is the simulation data; X exp,i is the measured data; n is the total number of data.

[0081] Perform multiple diagnosis on the plasma in the same state. Assume that the diagnosis is repeated n times. Each time the collaborative diagnosis system obtains a set of diagnosis results y i =[ne i1 ,ne i2 ,......,ne ij ,......,ne im ], where 1≤i≤n; 1≤j≤m, average the n groups of results to get y p =[ne p1 ,ne p2 ,......,ne pj ,......,ne pm ], for each diagnosis result y i ,definition The plasma electron density is compared with the mean ratio jitter, which verifies whether the ER of each diagnosis is lower than 20%. If it is lower than 20%, the surface diagnosis is accurate.

[0082] In summary, the present invention provides global and local information on plasma parameters through diagnostic probes and reflective antennas, and uses a neural network model to complete in-situ diagnosis of plasma profiles, which can quickly and accurately obtain plasma density distribution and plasma collision frequency.

Claims

1. A plasma parameter probe-microwave collaborative diagnosis system, characterized in that: It includes a diagnostic probe, a reflective antenna, a vector network analyzer and a computer terminal; the diagnostic probe and the reflective antenna are arranged on the surface of the aircraft; the reflective antenna is connected to the signal input end of the vector network analyzer; the output ends of the vector network analyzer and the diagnostic probe are respectively connected to the computer terminal; Reflection antenna and vector network analyzer to obtain S11 data caused by plasma; a diagnostic probe for acquiring a probe current induced in the plasma environment; The computer terminal is used to calculate the plasma electron temperature according to the probe current; calculate the plasma electron density at the monitoring point where the diagnostic probe is located according to the plasma electron temperature; load the trained neural network model, invert the S11 data and the plasma electron density at the monitoring point where the diagnostic probe is located through the trained neural network model, obtain the plasma electron density distribution and plasma collision frequency, and complete the collaborative diagnosis of the plasma.

2. The plasma parameter probe-microwave collaborative diagnosis system according to claim 1, characterized in that: The diagnostic probe is arranged coplanar with the reflective antenna.

3. The plasma parameter probe-microwave collaborative diagnosis system according to claim 1, characterized in that: The expression for calculating the plasma electron temperature based on the probe current is: Where T e is the plasma electron temperature; e is the electron charge; V p is the shell point; I is the probe current; I i0 is the saturated ion current; U is the voltage; ln(.) represents the logarithmic function; d(.) represents the differential.

4. The plasma parameter probe-microwave collaborative diagnosis system according to claim 2, characterized in that: The expression for calculating the plasma electron density at the monitoring point where the diagnostic probe is located based on the plasma electron temperature is: Where n i is the plasma electron density at the monitoring point where the diagnostic probe is located; S is the collection area of ​​the diagnostic probe; k is the Boltzmann constant; m i It is the plasma quality at the monitoring point where the diagnostic probe is located.

5. The plasma parameter probe-microwave collaborative diagnosis system according to claim 1, characterized in that: The diagnostic probe comprises a metal rod and a ceramic tube sleeved on the outer surface of the metal rod, and the upper end of the metal rod passes through the ceramic tube; the material of the metal rod is tungsten, molybdenum or tantalum.

6. The plasma parameter probe-microwave collaborative diagnosis system according to claim 1, characterized in that: The neural network model adopts a fully connected neural network model.

7. The plasma parameter probe-microwave collaborative diagnosis system according to claim 1, characterized in that: The training process of the neural network model is: A simulation model is constructed using simulation software. Different plasma parameters are set during the simulation process. The S11 data caused by the plasma is obtained through a reflection antenna and a vector network analyzer. The probe current caused in the plasma environment is obtained through a diagnostic probe. The plasma electron density corresponding to the data obtained by the diagnostic probe is calculated. The plasma collision frequency is obtained by querying the NRL manual. The S11 data and plasma electron density measured during the simulation process are used as inputs to the neural network model, and the density distribution parameters of the actual plasma and the plasma collision frequency during the simulation are used as labels to train the neural network model until the neural network model converges to obtain a trained neural network model.

8. The plasma parameter probe-microwave collaborative diagnosis system according to claim 1, characterized in that: The specific method of inverting the S11 data and the plasma electron density at the monitoring point where the diagnostic probe is located by using the trained neural network model to obtain the plasma density distribution and plasma collision frequency is as follows: The S11 data and the plasma electron density at the monitoring point where the diagnostic probe is located are used as the input of the trained neural network model, and the plasma collision frequency and electron density distribution parameters are obtained by inverting the trained neural network model; the electron density distribution parameters are substituted into the double Gaussian distribution model to obtain the density distribution of the plasma; the electron density distribution parameters include peak position, decrease coefficient, increase coefficient and electron density peak value.

Images