Electromagnetic parameter partition near-field scanning testing device suitable for non-flat composite material

Abstract

The invention relates to the field of electromagnetic parameter near-field scanning, in particular to an electromagnetic parameter partition near-field scanning testing device suitable for a non-flat composite material. The device comprises a sample fixing module used for fixing a non-flat composite material sample to be detected; the near-field probe group comprises a plurality of transmitting probes and receiving probes corresponding to the transmitting probes, the transmitting probes and the receiving probes are arranged on the upper side and the lower side of the detected sample respectively, and the near-field probe group is connected with the vector network analyzer; the multi-axis scanning platform moves the near-field probe group according to a preset path; the scanning control module is used for sending a motion instruction to the multi-axis scanning platform according to a partition plan and a scanning path set by a user; the data processing module is used for storing the reflection and transmission parameters acquired by the vector network analyzer and carrying out abnormal data detection; and the equivalent parameter inversion and reconstruction module is used for performing parameter inversion according to the acquired reflection and transmission parameters. The method is suitable for the electromagnetic parameter zoning near-field scanning test of the non-flat composite material.

Description

Technical Field

[0001] This invention relates to the field of electromagnetic parameter near-field scanning, and more specifically to an electromagnetic parameter partitioning near-field scanning test device suitable for non-flat composite materials. Background Technology

[0002] Currently, electromagnetic parameter testing methods mainly include the coaxial method, waveguide method, resonant cavity method, free space method, and near-field scanning method. For example, CN101713798A discloses a device for measuring the internal electric field distribution of composite materials. This device adopts a parallel plate waveguide structure, excites electromagnetic waves through a coaxial feed port, and sets a moving microwave probe above the sample surface. Under the control of a two-dimensional scanning platform, it performs planar scanning to collect the electric field response intensity at different locations inside the material.

[0003] However, existing technologies still have the following problems:

[0004] Existing devices rely on parallel plate waveguide cavity structures, requiring samples to meet requirements of planarity and regular thickness, thus limiting their applicability to regular flat plate materials. When dealing with non-flat composite materials with geometric features such as curvature, steps, slots, or thickness variations, this structure struggles to achieve effective bonding and scanning control, severely limiting the adaptability of the testing device.

[0005] Existing probe movement mechanisms are typically two-dimensional planar displacement platforms, which cannot achieve three-dimensional path following for irregular three-dimensional structures. For curved skins or cladding components commonly found in aerospace structures, this type of planar scanning strategy cannot cover the entire surface area, resulting in significant blind spots and incomplete data acquisition.

[0006] Existing devices use parallel plate waveguides as the excitation path, which means that such devices usually operate at mid-to-high frequencies (GHz level). However, the test conditions in the low-frequency band (such as 2MHz-30MHz) are difficult to meet the cutoff conditions of the waveguide mode, resulting in the inability to obtain stable signal excitation and reception in the low-frequency domain, and thus they are not suitable for low-frequency electromagnetic characteristic modeling.

[0007] Existing devices primarily focus on measuring electric field intensity distribution and lack mechanisms for partition mapping and equivalent parameter inversion. This makes it impossible to derive the equivalent permittivity and permeability of each region based on the collected data. The lack of an equivalent model renders the test results unsuitable for material parameter modeling, structural optimization design, or system-level electromagnetic analysis. Summary of the Invention

[0008] The purpose of this invention is to overcome the shortcomings of the prior art and provide an electromagnetic parameter partitioning near-field scanning test device suitable for non-flat composite materials. It adopts a multi-module collaborative approach to realize the entire process of sample partitioning, probe scanning, reflection and transmission parameter acquisition, data recording and equivalent parameter inversion, thereby improving the adaptability of the test device and the accuracy of the test results.

[0009] The present invention achieves the above objectives by adopting the following technical solution: The present invention provides an electromagnetic parameter partitioning near-field scanning testing device suitable for non-flat composite materials, comprising:

[0010] The sample fixing module is used to fix the non-flat composite material sample to be tested. It supports the mechanical fixing of curved, stepped, and concave-convex shaped samples and provides a reference surface for subsequent partition path generation.

[0011] The near-field probe group includes multiple transmitting probes and corresponding receiving probes, which are respectively arranged on the upper and lower sides of the sample under test. The near-field probe group is connected to a vector network analyzer.

[0012] The multi-axis scanning platform has three-axis motion control capabilities and is used to move the near-field probe group according to a preset path.

[0013] The scanning control module is used to send motion commands to the multi-axis scanning platform according to the user-defined zoning plan and scanning path, and to record the current position of the near-field probe group in real time.

[0014] The data processing module stores the reflection and transmission parameters collected by the vector network analyzer and introduces an anomaly identification mechanism based on probability sparsity detection to detect abnormal data.

[0015] The equivalent parameter inversion and reconstruction module is used to perform parameter inversion based on the collected reflection and transmission parameters, and to complete the extraction of equivalent electromagnetic parameters of each sub-region of the sample under test and the parameter synthesis of the overall structure.

[0016] Furthermore, anomaly detection specifically includes:

[0017] Construct probability density functions for the acquired magnitudes of reflection and transmission parameters using kernel density estimation:

[0018]

[0019] In the formula, n represents the number of valid S-parameter modulus samples collected from the historical scan path. Let K be the probability density function, h be the Gaussian kernel function, and s be the bandwidth parameter. i For historical data points, s = |S ij (f)| represents the sampled modulus value, i,j∈1,2, indicating the port number;

[0020] For the magnitudes of the currently acquired reflection and transmission parameters, calculate the estimated density values ​​under the probability density function;

[0021] Set a dynamic threshold and compare the calculated estimated density value with the dynamic threshold. If the estimated density value is less than the dynamic threshold, the currently collected reflection and transmission parameters are determined to be abnormal data, and a retest operation is automatically triggered.

[0022] Furthermore, the data processing module is also used for error correction, specifically including:

[0023] Transmitter and receiver probe correction:

[0024] Establish correction coefficients for the transmitting and receiving probes:

[0025]

[0026] In the formula, α(f) represents the correction coefficient, r0 represents the ideal position of the calibration point, and S meas (f,r0) represents the reflection and transmission parameters actually measured at the ideal position of the calibration point, S ref (f) represents the theoretical S-parameter distribution of the standard sample;

[0027] Position error correction:

[0028] The vertical position error is estimated by calculating the slope of the phase-frequency curve in the calibration data:

[0029]

[0030] In the formula, Δz represents the vertical position error, Δφ(f) is the frequency slope deviation between the measured phase and the theoretical phase, and c is the speed of light;

[0031] After measuring the vertical position error, adjustments are made using a multi-axis scanning platform or phase compensation is performed on the data.

[0032]

[0033] In the formula, S corrected (f) and S meas (f) S-parameter correction results and measured results at different frequency points at a certain location, respectively;

[0034] Environmental background error correction:

[0035] Differential filtering is performed on all measured reflection and transmission parameters:

[0036] S filtered (f,r)=S corrected (f,r)-Sbg (f)

[0037] In the formula, S bg (f) represents the background interference signal measured without a sample or with a free space reference sample.

[0038] Furthermore, the data processing module is also used for data preprocessing, specifically including:

[0039] After multiple acquisitions in each partition, the reflection and transmission parameter data at multiple frequency points are averaged, noise is removed, amplitude is normalized, and phase reference is corrected.

[0040] Furthermore, the equivalent parameter inversion and reconstruction module is specifically used for:

[0041] Raw data reading and frequency point processing:

[0042] Read the reflection and transmission parameter data collected from each partition, organize them uniformly according to frequency points, and record the complex reflection and transmission parameters of each partition at each frequency.

[0043] Local electromagnetic parameter inversion:

[0044] For each sub-region (i,j) with thickness d, the following parameter inversion calculation is performed:

[0045] Refractive index inversion:

[0046]

[0047] Wave impedance inversion:

[0048]

[0049] Equivalent parameter calculation:

[0050]

[0051] In the formula, d represents the thickness of each sub-region, Z0 is the free-space wave impedance, and S 11 (f) represents the reflection parameter data, S 21 (f) represents the transmission parameter data, ε i,j (f) represents the equivalent dielectric constant of each subregion, u i,j (f) represents the permeability of each subregion, and c is the speed of light. η represents the refractive index inversion result for each sub-region. i,j (f) represents the wave impedance inversion result;

[0052] Parameter consistency check and error assessment:

[0053] For each partition, construct an equivalent flat plate model, based on the inverted (ε) i,j ,μ i,j Calculate its theoretical reflection and transmission parameters, and compare the error with the measured values:

[0054]

[0055] If δ i,j (f) If multiple consecutive frequency points exceed the set threshold τ, then the region is marked as abnormal;

[0056] Overall equivalent parameter synthesis:

[0057] After completing the inversion of all zones, the equivalent electromagnetic parameters of the entire region are calculated using an area-weighted average method:

[0058]

[0059] In the formula, M represents the number of sub-regions divided along the vertical direction of the scanning area, N represents the number of sub-regions divided along the horizontal direction of the scanning area, and ε eq (f) represents the overall equivalent dielectric constant of the entire measurement region at frequency f, μ eq (f) represents the overall equivalent permeability of the entire region at frequency f;

[0060] Output:

[0061] ε for each partition i,j (f) and μ i,j (f) The output is a two-dimensional parameter distribution table, and ε is generated simultaneously. eq (f) and μ eq (f) Frequency response curve.

[0062] Furthermore, the equivalent parameter inversion and reconstruction module is also specifically used for:

[0063] Based on the error assessment results, the inversion process was optimized by adjusting the initial dielectric constant and the guessed permeability values ​​until the inversion results matched the actual test data.

[0064] Furthermore, the device also includes a report generation module, which presents the inversion results of each partition in the form of heat maps, three-dimensional surface maps, etc., and generates a complete test report. The report content includes the electromagnetic parameter inversion results of each partition, test environment description, error analysis, and model correction process.

[0065] The beneficial effects of this invention are as follows:

[0066] 1. Improved testing accuracy and spatial resolution

[0067] Traditional electromagnetic testing methods, especially for complex non-planar structures, are often affected by environmental factors, signal attenuation, and geometric mismatches. This invention employs near-field scanning technology, sampling zone by zone to accurately measure the electromagnetic response of each zone, significantly improving both spatial resolution and measurement accuracy. When dealing with complex composite material surfaces, it avoids errors caused by edge effects and sample defects found in traditional methods.

[0068] 2. Achieved adaptability to non-flat composite material structures.

[0069] This invention's device is adaptable to composite materials with non-flat structures such as curved surfaces, steps, and uneven surfaces, overcoming the problem that existing electromagnetic testing devices cannot effectively cover complex geometries. Through the flexible combination of a multi-axis scanning platform and a sample fixing module, it can ensure that the near-field probe accurately scans each zone on different surface features, exhibiting strong adaptability and enabling accurate measurement of the electromagnetic response of each zone.

[0070] 3. Overcame the problems of site and signal attenuation in low-frequency electromagnetic testing.

[0071] Traditional low-frequency electromagnetic parameter testing methods (such as the free-space method) typically require a large test area, and the test signal attenuates significantly, making it difficult to perform stable low-frequency measurements. This invention, through the design of multi-band test paths and precise signal optimization processing, enables stable testing in a smaller test space within the low-frequency range (e.g., 2MHz-30MHz), reducing the requirements for site size. Furthermore, by controlling background noise and using absorbing materials, the accuracy and stability of low-frequency testing are significantly improved.

[0072] 4. Real-time data feedback and correction of abnormal data

[0073] During testing, the system can collect and analyze the S-parameters of each partition in real time. If any data anomalies are detected (such as excessive signal attenuation or test error exceeding limits), the system will automatically provide data feedback and remind the operator to retest. Through this real-time feedback mechanism, the system can effectively reduce errors during testing, ensure the accuracy of electromagnetic response data for each partition, and improve the reliability of the overall testing process. Attached Figure Description

[0074] Figure 1 This is a structural diagram of an electromagnetic parameter partitioning near-field scanning test device suitable for non-flat composite materials provided in an embodiment of the present invention;

[0075] Figure 2 This is a flowchart of the device testing process provided in the embodiments of the present invention. Detailed Implementation

[0076] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings.

[0077] This invention provides a near-field scanning testing device for electromagnetic parameters of non-flat composite materials, such as... Figure 1 As shown, it specifically includes:

[0078] The sample fixing module is used to fix the non-flat composite material sample to be tested. It supports the mechanical fixing of curved, stepped, and concave-convex shaped samples and provides a reference surface for subsequent partition path generation.

[0079] The near-field probe group includes multiple transmitting probes and corresponding receiving probes, which are arranged on the upper and lower sides of the sample under test. By controlling the probe distance and alignment angle, a local near-field excitation and sensing area can be formed. The near-field probe group is connected to a vector network analyzer for outputting and acquiring high-frequency signals, such as S11 (reflection coefficient) and S21 (transmission coefficient).

[0080] The multi-axis scanning platform has three-axis (X, Y, Z) motion control capability. It can move the near-field probe group according to the preset path to ensure that the probe scans between each sub-region of the sample surface, maintaining a stable distance and contact angle.

[0081] The scanning control module, controlled by a microcontroller unit or host computer software, is responsible for sending motion commands to the multi-axis scanning platform according to the user-defined zoning plan and scanning path, and recording the current position of the near-field probe group in real time.

[0082] The data processing module stores the S-parameters collected by the vector network analyzer according to the partition number and spatial coordinates to form a "position-response" mapping table for subsequent inversion. It also introduces an anomaly identification mechanism based on probability sparsity detection to detect abnormal data.

[0083] The equivalent parameter inversion and reconstruction module is used to perform parameter inversion based on the acquired S-parameters, calculate the equivalent dielectric constant and permeability of each partition, and complete the extraction of equivalent electromagnetic parameters of each sub-region of the tested sample and the parameter synthesis of the overall structure. S-parameters are important parameters describing the electromagnetic properties of materials, including S11 (reflection coefficient) and S21 (transmission coefficient).

[0084] The invention will be further described below with reference to specific samples.

[0085] In aerospace vehicles such as aircraft, satellites, and missiles, composite materials are widely used in structural components such as skins, fuselage shells, radomes, and hatches. These structural components typically have complex geometries, including pronounced curved surfaces, helical slots, stepped variations, or embedded connecting strips, and their electromagnetic shielding performance directly affects the system's electromagnetic compatibility and anti-interference capabilities. Therefore, during the assembly or design phase, it is necessary to perform non-destructive, localized, and high-precision spatial measurements of the electromagnetic parameters of these structures.

[0086] Taking the electromagnetic parameter testing of aircraft canopy components as an example, combined with Figure 2 The complete testing procedure for the device of this invention is as follows:

[0087] Step 1: Sample installation and stable support;

[0088] The test chamber structure is placed on the platform provided by the sample fixation module, and its position and attitude are fixed using multi-point clamps. This sample fixation module supports stable support for non-flat, asymmetric material structures, ensuring that the sample does not shift during the entire scanning process and providing a geometric reference for subsequent path planning and data space mapping.

[0089] Step 2: Scan area setting and probe effective resolution calibration, as well as frequency band setting and scan path generation;

[0090] Scanning area settings and probe effective resolution calibration:

[0091] In the host computer software control interface, the geometric range of the area to be measured is set (e.g., a scanning area of ​​200mm × 200mm), and the partition size is determined according to the geometric characteristics of the sample (e.g., 20mm × 20mm). The scanning control module divides the target area into several rectangular partitions based on these parameters and calibrates the effective resolution of the probe for each partition. During calibration, the minimum sensing area of ​​the probe is automatically adjusted according to the surface characteristics of the sample to ensure that each partition can accurately acquire the required electromagnetic response data.

[0092] Frequency band settings and scan path generation:

[0093] First, based on testing requirements, multiple frequency bands (e.g., low-frequency band 2MHz-30MHz, high-frequency band 1GHz and above) are defined for testing range and sampling frequency. The scanning control module then sets the frequency bands accordingly and adjusts the probe's sampling accuracy and path planning based on the characteristics of different frequency bands, ensuring comprehensive electromagnetic parameter data is acquired within each band. Simultaneously, a scanning path point sequence is automatically generated to ensure that the scanning path for each frequency band covers all defined zones, meeting the testing requirements for different frequency bands.

[0094] Step 3: Initial probe positioning and dynamic bonding;

[0095] The scanning path sequence is sent to the multi-axis scanning platform via the command interface. The multi-axis scanning platform controls the near-field probe group to move along the X, Y, and Z axes to the starting scanning point. The dynamic height adjustment function of the Z-axis ensures that each probe always maintains a fixed near-field distance from the sample surface, adapting to curved surface structures and improving measurement consistency.

[0096] Step 4: Real-time S-parameter acquisition and abnormal data feedback;

[0097] During the scanning process, the near-field probe group excites a sweep frequency signal within a set frequency range (e.g., 2MHz-30MHz) in real time, and collects reflection (S11) and transmission (S21) parameters through a vector network analyzer. The complex S-parameters obtained from sampling at each frequency point are stored, and an anomaly identification mechanism based on probability sparsity detection is introduced to automatically determine whether the current data deviates from the reasonable physical range.

[0098] The specific process of anomaly data detection is as follows:

[0099] Step 4.1: Construct the probability distribution function of the modulus values ​​for the valid S-parameter data already collected in the historical scanning path. To improve the modeling ability for non-normal distributions, kernel density estimation (KDE) is used to establish the modulus probability density function. Where s = |S ij (f)| is the sampled modulus value (i,j∈1,2, representing the port number): During the scanning process, the near-field probe system excites a sweep frequency signal in the frequency range in real time (e.g., 2MHz-30MHz), and collects the reflection (S11) and transmission (S21) parameters through a vector network analyzer.

[0100]

[0101] In the formula, n represents the number of valid S-parameter modulus samples collected from the historical scan path. Let K be the probability density function, h be the Gaussian kernel function, and s be the bandwidth parameter. i For historical data points, i,j∈1,2, representing port numbers.

[0102] Step 4.2: Collect the S-parameter magnitudes of the current scan point. Calculate its probability density function Estimated density values

[0103] Step 4.3: Set the dynamic threshold τ. If the current sampled data falls into a low-probability sparse region, the point is determined to be an outlier, and a retest operation is automatically triggered.

[0104] This anomaly detection process utilizes historical multi-point distribution characteristics to construct a dynamic benchmark, avoiding reliance on fixed thresholds or linear rules, and can adapt to data differences under different frequency bands and structural types, thereby improving the adaptability and accuracy of anomaly identification.

[0105] Step 5: Standard sample error calibration;

[0106] After initial positioning, to ensure the testing accuracy and stability of the device, a standard sample calibration procedure was performed. The standard sample used was a homogeneous reference structure with known electromagnetic parameters, and its theoretical S-parameter distribution S... ref (f) can be obtained through electromagnetic simulation or theoretical calculation.

[0107] During the calibration process, the measured S-parameters S along the current probe path are collected. meas (f,r), and the following step-by-step strategy is used to identify and correct various errors:

[0108] Step 5.1, Probe response error correction;

[0109] Because the probe itself has non-ideal frequency response characteristics, its response typically exhibits amplitude compression and phase delay. The system performs complex domain ratio analysis on theoretical and measured values ​​at multiple frequency points to establish the probe correction coefficients at frequency f:

[0110]

[0111] In the formula, α(f) represents the probe correction coefficient, r0 represents the ideal position of the calibration point, and S meas (f,r0) represents the reflection and transmission parameters actually measured at the ideal position of the calibration point, S ref (f) represents the theoretical S-parameter distribution of the standard sample, where f represents the frequency;

[0112] This coefficient will subsequently be used for data correction at all scan points:

[0113] S corrected (f,r)=α(f)·S meas (f,r)

[0114] In the formula, S corrected (f,r) represents the S-parameter values ​​after probe frequency response error correction, i.e., the values ​​after complex domain correction of the amplitude and phase of the original position test results at the current position r and frequency f. meas (f,r) represents the original acquisition result of the measured S-parameters under the current probe path.

[0115] Step 5.2, Position Error Correction;

[0116] Even minute displacements between the probe and the sample can cause additional propagation phase delays, especially noticeable at high frequencies. The vertical positional error can be estimated by calculating the slope of the phase-frequency curve shift in the calibration data.

[0117]

[0118] In the formula, Δz represents the vertical position error, Δφ(f) is the frequency slope deviation between the measured phase and the theoretical phase, and c is the speed of light.

[0119] After measuring Δz, the data is adjusted or phase compensation is performed using a multi-axis scanning platform:

[0120]

[0121] Among them, S corrected (f) and S meas (f) S-parameter correction results and measured results at different frequency points at a certain location, respectively;

[0122] Step 5.3, Environmental background error correction;

[0123] Using the background interference signal measured without a sample or with a free-space reference sample as the baseline, differential filtering is performed on all measured S-parameters:

[0124] S filtered (f,r)=S corrected (f,r)-S bg (f)

[0125] In the formula, S bg (f) represents the background interference signal measured without a sample or with a free space reference sample.

[0126] Environmental background correction can effectively suppress spurious signals caused by environmental reflections and multipath interference, and is particularly suitable for background correction in low signal-to-noise ratio regions.

[0127] This step decomposes and models the errors of the probe, position, and environment, and uses three independent correction mechanisms—complex scaling, phase slope, and background difference—to compensate for various types of errors from different sources, ensuring that high-fidelity, physically consistent S-parameter data is obtained in subsequent scans.

[0128] If the standard sample error calibration result meets the index requirements, proceed to step 6; otherwise, perform model correction and result optimization, and return to step 4.

[0129] Step 6: Data preprocessing and averaging of multiple measurements;

[0130] After multiple acquisitions in each partition, the S-parameter data from multiple frequency points are averaged, noise is removed, amplitude normalization and phase reference correction are performed to obtain more stable and accurate electromagnetic response data.

[0131] Step 7: Electromagnetic parameter inversion and error assessment;

[0132] This step is based on the partition S-parameter data obtained by near-field scanning, calculates the equivalent permittivity and permeability of each partition, and completes the extraction of equivalent electromagnetic parameters of each sub-region and the parameter synthesis of the overall structure.

[0133] The inversion and error assessment process is detailed below:

[0134] Step 7.1: Raw data reading and frequency point organization;

[0135] Read the S data collected from each partition 11 (f) and S 21 (f) The data is organized uniformly according to frequency points, and the complex S-parameters of each partition at each frequency are recorded.

[0136] Step 7.2, Local electromagnetic parameter inversion;

[0137] For each sub-region (i,j) with thickness d, perform the following parameter inversion calculation:

[0138] Refractive index inversion:

[0139]

[0140] Wave impedance inversion:

[0141]

[0142] Where Z0 is the free space wave impedance.

[0143] Finally, the equivalent permittivity and permeability of each region are calculated using the following formula:

[0144]

[0145] In the formula, d represents the thickness of each sub-region, Z0 is the free-space wave impedance, and S 11 (f) represents the reflection parameter data, S 21 (f) represents the transmission parameter data, ε i,j (f) represents the equivalent dielectric constant of each subregion, u i,j (f) represents the permeability of each subregion, and c is the speed of light. η represents the refractive index inversion result for each sub-region. i,j (f) represents the wave impedance inversion result.

[0146] All inversion parameters are frequency-dependent complex numbers.

[0147] Step 7.3: Parameter consistency check and error assessment;

[0148] For each partition, the system constructs an equivalent flat plate model, based on the inverted (ε) i,j ,μ i,j Calculate its theoretical S-parameters and compare the error with the measured values:

[0149]

[0150] in and These are the measured S-parameters and the S-parameters obtained from equivalent modeling, respectively. If the inversion error δ i,j (f) If multiple consecutive frequency points exceed the set threshold τ, then the region is marked as abnormal.

[0151] Step 7.4, Error Exceedance Handling Mechanism;

[0152] For partitions that exceed the error limit, one of the following operations will be performed automatically:

[0153] 1. Remeasure the area;

[0154] 2. Perform smooth interpolation correction based on the parameters of adjacent partitions;

[0155] 3. Trigger local fine-tuning of deep learning models to enhance regional adaptability.

[0156] Step 7.5: Overall equivalent parameter synthesis;

[0157] After completing the inversion of all zones, the equivalent electromagnetic parameters of the entire region are calculated using an area-weighted average method:

[0158]

[0159] In the formula, M represents the number of sub-regions divided along the vertical direction (such as the y-axis or z-axis) of the scan area, N represents the number of sub-regions divided along the horizontal direction (such as the x-axis) of the scan area, and ε eq (f) represents the overall equivalent dielectric constant of the entire measurement region at frequency f, μ eq (f) represents the overall equivalent permeability of the entire region at frequency f;

[0160] Output:

[0161] ε for each partition i,j (f) and μ i,j (f) The output is a two-dimensional parameter distribution table, and ε is generated simultaneously. eq (f) and μ eq(f) Frequency response curve. All results can be exported as .xlsx, .mat, or database files for subsequent analysis.

[0162] Finally, determine whether the electromagnetic parameter inversion and error assessment results meet the requirements. If they do, proceed to step 8; otherwise, based on the error assessment results, adjust the initial dielectric constant and permeability guesses through the equivalent parameter inversion and reconstruction module to optimize the inversion model until the model results match the actual test data, and return to S4. This process may involve local optimization of electromagnetic parameters in different sub-regions within the frequency range.

[0163] Step 8: Visualize the results and generate a report;

[0164] The report generation module presents the inversion results of each partition in the form of heat maps, 3D surface plots, etc., and generates a complete test report. The report includes the electromagnetic parameter inversion results of each partition, test environment description, error analysis and model correction process to support subsequent simulation and design optimization.

[0165] Through this design, the testing process not only achieves multi-band coverage, real-time data feedback, error correction, and model optimization during the testing process, but also introduces the estimation of low-frequency parameters and comparative analysis of high-frequency test data, which greatly improves the accuracy and reliability of the test results.

[0166] In summary, this invention overcomes the technical problems of existing technologies, such as inability to adapt to complex non-planar structures, high difficulty in low-frequency testing, and poor data accuracy, by introducing innovative methods such as near-field scanning technology, multi-band test path planning, and real-time feedback and correction mechanisms. This significantly improves the accuracy and efficiency of electromagnetic parameter testing. Through these technical means, the device of this invention not only expands the applicability of electromagnetic testing but also significantly improves the intelligence and automation level of the testing process.

[0167] The above description is merely a preferred embodiment of the present invention. It should be understood that the present invention is not limited to the forms disclosed herein and should not be construed as excluding other embodiments. It can be used in various other combinations, modifications, and environments, and can be altered within the scope of the concept described herein through the above teachings or related technologies or knowledge. Modifications and variations made by those skilled in the art that do not depart from the spirit and scope of the present invention should be within the protection scope of the appended claims.

Claims

1. A near-field scanning testing device for electromagnetic parameters of non-flat composite materials, characterized in that, include: The sample fixing module is used to fix the non-flat composite material sample to be tested. It supports the mechanical fixing of curved, stepped, and concave-convex shaped samples and provides a reference surface for subsequent partition path generation. The near-field probe group includes multiple transmitting probes and corresponding receiving probes, which are respectively arranged on the upper and lower sides of the sample under test. The near-field probe group is connected to a vector network analyzer. The multi-axis scanning platform has three-axis motion control capabilities and is used to move the near-field probe group according to a preset path. The scanning control module is used to send motion commands to the multi-axis scanning platform according to the user-defined zoning plan and scanning path, and to record the current position of the near-field probe group in real time. The data processing module stores the reflection and transmission data collected by the vector network analyzer and introduces an anomaly detection mechanism based on probability sparsity detection to detect abnormal data. The equivalent parameter inversion and reconstruction module is used to perform parameter inversion based on the collected reflection and transmission data, calculate the equivalent dielectric constant and permeability of each partition, and complete the extraction of equivalent electromagnetic parameters of each sub-region of the sample under test and the parameter synthesis of the overall structure.

2. The electromagnetic parameter partitioning near-field scanning testing device for non-flat composite materials according to claim 1, characterized in that, Anomaly detection specifically includes: The probability density function of the acquired reflection and transmission modulus values ​​is constructed using kernel density estimation: In the formula, n represents the number of valid S-parameter modulus samples collected from the historical scan path. Let K be the probability density function, h be the Gaussian kernel function, and s be the bandwidth parameter. i For historical data points, s = |S ij (f)| represents the sampled modulus value, i,j∈1,2, indicating the port number; For the magnitudes of the currently acquired reflection and transmission parameters, calculate the estimated density values ​​under the probability density function; Set a dynamic threshold and compare the calculated estimated density value with the dynamic threshold. If the estimated density value is less than the dynamic threshold, the currently collected reflection and transmission parameters are determined to be abnormal data, and a retest operation is automatically triggered.

3. The electromagnetic parameter partitioning near-field scanning testing device for non-flat composite materials according to claim 1, characterized in that, The data processing module is also used for error correction, specifically including: Transmitter and receiver probe correction: Establish correction coefficients for the transmitting and receiving probes: In the formula, α(f) represents the correction coefficient, r0 represents the ideal position of the calibration point, and S meas (f,r0) represents the reflection and transmission parameters actually measured at the ideal position of the calibration point, S ref (f) represents the theoretical S-parameter distribution of the standard sample; Position error correction: The vertical position error is estimated by calculating the slope of the phase-frequency curve in the calibration data: In the formula, Δz represents the vertical position error, Δφ(f) is the frequency slope deviation between the measured phase and the theoretical phase, and c is the speed of light; After measuring the vertical position error, adjustments are made using a multi-axis scanning platform or phase compensation is performed on the data. In the formula, S corrected (f) and S meas (f) S-parameter correction results and measured results at different frequency points at a certain location, respectively; Environmental background error correction: Differential filtering is performed on all measured reflection and transmission parameters: S filtered (f,r)=S corrected (f,r)-S bg (f) In the formula, S bg (f) represents the background interference signal measured without a sample or with a free space reference sample.

4. The electromagnetic parameter partitioning near-field scanning testing device for non-flat composite materials according to claim 1, characterized in that, The data processing module is also used for data preprocessing, specifically including: After multiple acquisitions in each partition, the reflection and transmission parameter data at multiple frequency points are averaged, noise is removed, amplitude is normalized, and phase reference is corrected.

5. The electromagnetic parameter partitioning near-field scanning test device for non-flat composite materials according to claim 1, characterized in that, The equivalent parameter inversion and reconstruction module is specifically used for: Raw data reading and frequency point processing: Read the reflection and transmission parameter data collected from each partition, organize them uniformly according to frequency points, and record the complex reflection and transmission parameters of each partition at each frequency. Local electromagnetic parameter inversion: For each sub-region (i,j) with thickness d, the following parameter inversion calculation is performed: Refractive index inversion: Wave impedance inversion: Finally, the equivalent permittivity and permeability of each region are calculated using the following formula: In the formula, d represents the thickness of each sub-region, Z0 is the free-space wave impedance, and S 11 (f) represents the reflection parameter data, S 21 (f) represents the transmission parameter data, ε i,j (f) represents the equivalent dielectric constant of each subregion, u i,j (f) represents the permeability of each subregion, and c is the speed of light. η represents the refractive index inversion result for each sub-region. i,j (f) represents the wave impedance inversion result; Parameter consistency check and error assessment: For each partition, construct an equivalent flat plate model, based on the inverted (ε) i,j ,μ i,j Calculate its theoretical reflection and transmission parameters, and compare the error with the measured values: If δ i,j (f) If multiple consecutive frequency points exceed the set threshold τ, then the region is marked as abnormal; Overall equivalent parameter synthesis: After completing the inversion of all zones, the equivalent electromagnetic parameters of the entire region are calculated using an area-weighted average method: In the formula, M represents the number of sub-regions divided along the vertical direction of the scanning area, N represents the number of sub-regions divided along the horizontal direction of the scanning area, and ε eq (f) represents the overall equivalent dielectric constant of the entire measurement region at frequency f, μ eq (f) represents the overall equivalent permeability of the entire region at frequency f; Output: ε for each partition i,j (f) and μ i,j (f) The output is a two-dimensional parameter distribution table, and ε is generated simultaneously. eq (f) and μ eq (f) Frequency response curve.

6. The electromagnetic parameter partitioning near-field scanning testing device for non-flat composite materials according to claim 5, characterized in that, The equivalent parameter inversion and reconstruction module is also specifically used for: Based on the error assessment results, the inversion process was optimized by adjusting the initial dielectric constant and the guessed permeability values ​​until the inversion results matched the actual test data.

7. The electromagnetic parameter partitioning near-field scanning testing device for non-flat composite materials according to claim 1, characterized in that, The device also includes a report generation module, which presents the inversion results of each partition in the form of heat maps, three-dimensional surface plots, etc., and generates a complete test report. The report includes the electromagnetic parameter inversion results of each partition, test environment description, error analysis, and model correction process.

Images