A method for quickly detecting defects of a wave-absorbing coating based on incident angle modulation

Abstract

The application provides a kind of quick detection method of wave-absorbing coating defect based on incident angle modulation, belongs to the field of defect detection, comprising: initializing detection system;Multi-angle incident sweep frequency test is carried out on the surface of defect-free coating;Set the incident mode to grazing incidence, sweep frequency test is carried out on the surface of defective coating, and whether there is surface defect is determined by comparison and analysis with reference data;Set the incident mode to normal incidence, sweep frequency test is carried out on the surface of defective coating, and whether there is circular defect is determined;When it is determined that circular defect exists, multi-angle sweep frequency inversion is executed to quantify defect size, and actual physical size is obtained by using elliptical correction model correction;An adaptive calibration module based on deep learning is introduced to obtain the final defect size.The application can meet the needs of stealth aircraft and other aerospace equipment.

Description

Technical Field

[0001] This invention belongs to the field of defect detection, specifically relating to a rapid detection method for defects in absorbing coatings based on incident angle modulation. Background Technology

[0002] Stealth technology mainly encompasses two major directions: shape stealth and material stealth. As a crucial branch, material stealth technology utilizes multifunctional radar-absorbing coating systems on the aircraft surface to achieve efficient absorption and scattering modulation of radar waves, significantly enhancing the target's low detectability. Radar-absorbing coatings not only achieve radar stealth but also effectively suppress signal crosstalk in complex electromagnetic environments, ensuring the reliable operation of electronic equipment. However, in practical applications, radar-absorbing coatings face the risk of failure under thermal loads and mechanical stress, compromising structural integrity and consequently affecting the aircraft's stealth performance. Therefore, to ensure the stealth capability of aircraft and reduce maintenance costs, regular inspection and maintenance are necessary.

[0003] Current technologies for detecting defects in absorbing coatings can be mainly divided into the following three categories:

[0004] (1) Traditional microwave detection technology: Based on the interaction mechanism between electromagnetic waves and materials, the microwaves achieve defect detection through reflection, scattering, and transmission characteristics at the interface of the medium. The current mainstream reflective microwave detection technology establishes a quantitative mapping relationship between the defect area and characteristic parameters by accurately measuring the phase shift of the reflection coefficient at the coating-substrate interface. Experimental data show that under Ka-band conditions, the waveguide probe planar scanning technology can effectively identify circular hole-shaped debonding defects with a diameter of 1.6 mm, with a spatial resolution of sub-millimeter level. Traditional microwave testing has high accuracy, but it requires waveguide probes to scan point by point during testing, resulting in low defect detection efficiency. Moreover, the design of microwave transceiver modules in the high-frequency band is complex.

[0005] (2) Ultrasonic testing technology: This method uses the propagation characteristics of ultrasonic waves in materials to detect defects. When the coating debonds, the propagation of ultrasonic waves inside the coating is affected, which leads to abnormal received signals. C-scan imaging technology is commonly used. This method has good penetration and high sensitivity and is suitable for defect detection in various materials. However, it requires the use of a coupling agent and cannot detect curved structures. Contact measurement is prone to damaging the coating surface, and its efficiency is low in large models and on-site testing.

[0006] (3) Infrared thermal imaging technology: Detection is achieved by analyzing temperature changes caused by defects. Discontinuities within the object being measured, such as defects or structural differences, can affect the object's thermal diffusion characteristics, leading to abnormal surface temperature field distribution. Temperature resolution reaches 0.05℃. Existing technologies have high detection efficiency for surface and near-surface defects, but are significantly affected by ambient temperature and have poor adaptability to complex curved surfaces.

[0007] Therefore, existing rapid detection methods for defects in absorbing coatings have the following technical problems:

[0008] (1) Poor adaptability to different scenarios: Insufficient ability to detect curved surfaces; ultrasonic detection relies on planar coupling; infrared thermal imaging is affected by the deviation of the normal direction of the curved surface, resulting in a large positioning error. (2) Poor flexibility: Traditional methods have a single detection mode and rely on fixed detection parameters, making them unable to adapt to different defect types. (3) High risk of damage: The mechanical stress generated on the coating surface by contact detection methods (such as ultrasonic) is prone to causing secondary damage. (4) Target applications: Traditional defect detection methods are mainly focused on coating thickness measurement. Summary of the Invention

[0009] To address the aforementioned technical problems, this invention provides a rapid detection method for absorbing coating defects based on incident angle modulation. Utilizing the principle of radar cross section (RCS) reduction on the target surface caused by absorbing materials, the method detects the presence of defects by detecting abrupt changes in RCS caused by gaps or circular defects. It allows for continuous frequency sweeping with a user-defined step size and custom incident angle within a frequency range of 2–18 GHz, obtaining synchronized echo signal amplitude, phase, and polarization information. Furthermore, inverse Fourier transform converts the frequency domain S-parameters into a time domain signal, achieving one-dimensional imaging and significantly improving defect location accuracy. Simultaneously, based on multi-angle detection data, defect size can be assessed. This invention, by comparing the detected results with the initial defect-free result, is applicable to various complex shapes and curved surfaces, and can meet the needs of stealth aircraft and other aerospace equipment.

[0010] To achieve the above objectives, the present invention adopts the following technical solution:

[0011] A rapid detection method for defects in absorbing coatings based on incident angle modulation includes the following steps:

[0012] Step 1: Initialize the detection system, set the operating frequency range of the vector network analyzer, fix the sample to be tested on the three-dimensional scanning platform, and install the antenna on the adjustable antenna bracket.

[0013] Step 2: Perform multi-angle incident frequency sweep test on the defect-free coating surface to obtain echo signal data and perform inverse Fourier transform to obtain time domain data, which is used as reference data;

[0014] Step 3: Set the incident mode to grazing incidence and perform a frequency sweep test on the defective coating surface. By comparing and analyzing the data with the reference data, determine whether there are surface defects.

[0015] Step 4: Set the incident mode to vertical incidence and perform a frequency sweep test on the defective coating surface. By comparing and analyzing the data with the reference data, determine whether there are circular defects.

[0016] Step 5: Once the existence of the circular defect is confirmed, perform multi-angle frequency sweep inversion to quantify the defect size. Determine the range of the projected size of the circular defect on the incident plane using time-domain data at different incident angles, and use an ellipse correction model to correct and obtain the actual physical size of the circular defect; the time-domain data at different incident angles is the multi-angle time-domain data.

[0017] Step 6: Introduce a deep learning-based adaptive calibration module, construct a defect feature-size mapping model, and use a convolutional neural network to jointly analyze multi-angle time-domain data to obtain the final size of the circular defect.

[0018] Beneficial effects:

[0019] 1. This invention is based on the comparison between initial non-destructive data and subsequent test data, and judges the existence of defects by the difference in data, which can ignore the deviation caused by the curved surface; it uses multi-angle frequency sweep measurement to improve measurement accuracy; it transmits and receives echo signals through an antenna, without contact with the absorbing coating under test, which is a non-destructive testing method.

[0020] 2. This invention can identify defect types such as slits and circles through multi-angle testing. It also allows for customized testing frequencies based on a vector network analyzer, and the incident angle can be defined according to user needs. Furthermore, this invention is based on a handheld, miniaturized inspection system, making it flexible and convenient for various applications.

[0021] 3. This invention focuses on the identification, location, and size inversion of defects in microwave absorbing coatings. It employs novel detection methods and data processing technologies, utilizing a handheld detection device based on a vector network analyzer to rapidly scan the surface of the microwave absorbing coating from multiple angles, enabling defect type identification and size inversion. Traditional detection methods, however, can only qualitatively determine the existence of defects and are insufficient for detecting small gaps and circular defects. Attached Figure Description

[0022] Figure 1 This is a flowchart of a rapid detection method for absorbing coating defects based on incident angle modulation according to the present invention;

[0023] Figure 2 This is a schematic diagram of the detection system. Detailed Implementation

[0024] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention. Furthermore, the technical features involved in the various embodiments of this invention described below can be combined with each other as long as they do not conflict with each other.

[0025] like Figure 1 As shown, this invention provides a rapid detection method for defects in absorbing coatings based on incident angle modulation. By scanning the coating surface with a wideband wavelength at multiple incident angles, the echo signal intensity is inversely Fourier transformed to obtain a series of time-domain data. The presence of gaps or circular defects on the coating surface is quickly determined by the amplitude peak value, and the defect size is further deduced. Specifically, the method includes the following steps:

[0026] Step 1: Initialize the detection system and set the operating frequency range of the vector network analyzer according to the antenna frequency band range used.

[0027] like Figure 2 As shown, the detection system includes a vector network analyzer, a horn antenna (transmitting and receiving antennas), an antenna bracket, a three-dimensional scanning platform (for placing the sample to be tested or adjusting the antenna angle), data transmission cables (e.g., coaxial cables), and calibration components. The first port of the vector network analyzer is connected to the transmitting antenna via a coaxial cable, and the second port is connected to the receiving antenna via a coaxial cable. The sample to be tested (i.e., the target to be detected) is fixed on the three-dimensional scanning platform, and the antenna is mounted on an adjustable-angle antenna bracket.

[0028] Antennas typically use the 12-18 GHz frequency band, with an intermediate frequency bandwidth of 1 kHz. This bandwidth is an ideal starting point for general measurements, effectively balancing measurement accuracy (signal-to-noise ratio / dynamic range) and measurement speed. Simultaneously, the number of frequency points can be set to 801. More frequency points result in a higher upper limit for time-domain measurements, but also increase the scan time. The distance between the target and the antenna can be used to determine the location of the maximum peak in subsequent tests. It's important to note that the intermediate frequency bandwidth is not a fixed value; in practical applications, the bandwidth can be flexibly selected based on specific accuracy or speed requirements (narrower bandwidth results in higher accuracy but lower efficiency). Generally, the number of frequency points N needs to meet the following requirements:

[0029] ;

[0030] in, For frequency range, This represents the distance between the antenna and the target to be detected. The speed of light is generally taken as... By setting the above parameters, the detection system can be initialized.

[0031] Step 2: Perform multi-angle incident frequency sweep tests on the defect-free coating surface to obtain a series of echo signals, and substitute them into the inverse Fourier transform formula. To obtain the corresponding time-domain data This serves as the baseline data for subsequent tests. It refers to the angle of incidence being The time-domain data obtained at that time This is the index for the angle of incidence.

[0032] The formula for the inverse Fourier transform is as follows:

[0033] ;

[0034] in, Indicates the angle of incidence as Time-domain data, Index for the angle of incidence; Indicates the angle of incidence as Time-domain data; This represents the inverse Fourier transform operator; Represents the imaginary unit; Represents frequency variables; Represents a time variable.

[0035] For multi-angle incident angle selection, a uniform or non-uniform angle scanning strategy can be chosen according to actual needs.

[0036] For a uniform angle scanning strategy, a smaller step angle yields more reference data, but also results in a longer scanning time. Based on typical defect scattering patterns, an initial angle interval can be selected. (i.e., selection) , , , , , , To improve the accuracy of size inversion, a smaller interval, such as 5°, can be used for uniform scanning, but the corresponding scanning time will also increase.

[0037] For non-uniform angle scanning strategies, the angle-sensitive regions of the coating should be given special consideration. Typically, when the incident angle approaches the critical angle determined by the refractive index of the coating medium, the surface wave effect will significantly enhance the response sensitivity to small defects. At incidence, it exhibits the best detection efficiency for layered defects, but shows poor detection performance for crevice defects, which will help in subsequent differentiation of different defect types. Under a non-uniform strategy, six characteristic angles (i.e., based on fifth-order Legendre polynomial node distribution optimization) can typically be used. , , , , , This distribution is generated through zero-point mapping of orthogonal polynomials, covering key sensitive regions such as near-tangential incidence, Brewster's angle region, and perpendicular incidence, reducing the workload of full scanning while ensuring the completeness of defect response characteristics. In practice, to improve data reliability and minimize instability caused by the vector network analyzer, each incidence angle can be repeatedly measured and the average value taken as the benchmark for each angle.

[0038] Step 3: Set the incident mode to grazing incidence and perform a frequency sweep test on the defective coating surface.

[0039] The antenna is adjusted to a pre-selected characteristic grazing angle, which is typically chosen based on the critical angle calculated from the electromagnetic characteristic parameters of the absorbing coating. A typical value is... At this time, the electromagnetic wave forms a coupling effect with the coating surface, which increases the scattered field intensity of the micron-sized surface opening defect by 10-15 dB.

[0040] During testing, a continuous wave signal was emitted using a vector network analyzer to obtain frequency domain scattering parameters. Converted to time domain data via inverse Fourier transform By comparing with the baseline data obtained in step 2 (i.e., the time-domain data from step 2) In Take 6 feature angles , , , , , In Analyze the time-domain data and define if at the corresponding time A peak increase of more than 6 dB at the ns level indicates the presence of surface defects. The time window corresponding to the defect reflection peak, and the size of the sample under test in the incident direction. and the distance between the antenna and the sample under test Relevance can usually be inferred to satisfy [certain conditions]. In addition, to improve the stability of the determination, multiple measurements can be taken and the average value can be used to suppress false positives.

[0041] Step 4: Set the incident mode to vertical incidence and perform a frequency sweep test on the defective coating surface.

[0042] Adjust the antenna to At the incident angle, the slit-type defect has little effect on the echo signal, while the circular defect produces significant backscattering enhancement. During testing, the same frequency sweep parameters as the reference are maintained to obtain the frequency domain scattering parameters for this incident direction. Converted to time domain data via inverse Fourier transform By comparing with benchmark data (i.e., the time-domain data from step 2) In Take 6 feature angles , , , , , In )and Comparative analysis, defining if at the corresponding time A peak increase greater than 6 dB at the ns level indicates the presence of a circular defect; otherwise, the defect is not present. The time window corresponding to the defect reflection peak, and the size of the sample under test in the incident direction. and the distance between the antenna and the sample under test Relevance can usually be inferred to satisfy [certain conditions]. To improve the stability of the determination, measurements can be taken 3-10 times, and the average value can be used to suppress false positives.

[0043] By comparing with time domain data Comparative analysis, if the original time position ns The phenomenon of increased peak values, in If there is no response or a response of less than 2 dB at the same time location, it indicates that the defect has direction-selective scattering characteristics, consistent with a gap-type defect. To improve the stability of the determination, 3-10 measurements can be performed to suppress false positives.

[0044] Step 5: After confirming the existence of a circular defect in Step 4, multi-angle frequency sweep inversion needs to be performed to quantify the defect size.

[0045] Multi-angle frequency sweep tests can be performed by setting the incident angle. It is recommended that the incident angle value follow the principle of symmetrical scanning. , , , , , , Seven characteristic angles, this sequence covers the main lobe region of the defect scattering field ( ) and sidelobe sensitive areas ( When more precise detection results are required, the interval between adjacent angles can be reduced accordingly. A continuous wave signal is emitted using a vector network analyzer to obtain frequency domain scattering parameters, which are then converted into a series of time domain data using inverse Fourier transform. (i.e., multi-angle time-domain data) For each incident angle, the measurement can be repeated 3-10 times and the average value taken to improve the stability and reliability of the detection. Based on the change of incident angle, at different angles, compared with the reference data... The location of the peak increase exceeding 6dB after comparison (defined as the angle of incidence is) At that time, the time point at which a peak greater than 6dB appears compared to the baseline data may differ, resulting in different reflection peaks in the time domain. Therefore, the corresponding time position in step 4... Around ns, for each incident angle, find the specific time points in the time domain data where, compared to the baseline data at the same incident angle, a peak value greater than 6dB appears. At all the time points obtained. In this case, let the minimum echo time and corresponding angle be... and The maximum echo time and corresponding angle are and This allows us to determine the approximate projection size range of the circular defect on the incident plane. :

[0046] ;

[0047] ;

[0048] in, Indicates the maximum echo time. Indicates the minimum echo time. This represents the incident angle corresponding to the minimum echo time. This indicates the incident angle corresponding to the maximum echo time. This represents the average angle of incidence.

[0049] This dimension is the equivalent diameter of the defect within the inspection plane, and the actual physical size. It needs to be corrected using an elliptic correction model:

[0050] ;

[0051] in, This represents the approximate size range of the projected dimensions of a circular defect on the incident plane. This represents the incident angle corresponding to the minimum echo time. This indicates the incident angle corresponding to the maximum echo time. This represents the difference between the incident angle corresponding to the maximum echo time and the incident angle corresponding to the minimum echo time.

[0052] Step 6: Introduce a deep learning-based adaptive calibration module to achieve higher precision defect quantification by constructing a defect feature-size mapping model.

[0053] After obtaining the initial defect size range, a pre-trained convolutional neural network is used to process multi-angle temporal data. Perform joint analysis:

[0054] First, the time-domain signal amplitude sequences corresponding to different incident angles are arranged according to the angular dimension, that is, the multi-angle time-domain data. Compared with benchmark data Normalized difference processing is performed to form a two-dimensional time-angle feature matrix. :

[0055] ;

[0056] ;

[0057] in, Index of the angle of incidence; Here, T represents the number of sampling points in the time domain. Indicates the angle of incidence The maximum absolute value of the lower differential signal, This represents all time sampling points at that angle; This represents the value corresponding to the i-th row and j-th column of the two-dimensional time-angle characteristic matrix; Indicates the angle of incidence The coating containing defects Time-domain data at any given moment; Indicates the angle of incidence Defect-free coating Time-domain data at any given moment; Indicates the angle of incidence exist The differential signal value at time t.

[0058] Solving Maxwell's equations using the finite-difference time-domain (FDTD) method, and obtaining simulations using commercial electromagnetic simulation software FEKO or CST at different dielectric constants in the corresponding frequency bands ( ), defect morphology (circular / crack / compound defect), incident angle ( Frequency domain scattering field at a step size of 0.2° A time-domain training set is generated through Fourier transform. Data augmentation strategies can be introduced, such as adding Gaussian white noise with a signal-to-noise ratio of 30-50 dB to simulate interference in the actual measurement environment.

[0059] An end-to-end training method using a deep residual network (ResNet-34) is employed. An angle encoding layer is added at the input to map the incident angle into a 128-dimensional position encoding vector, which is then concatenated with the temporal signal before being input to the convolutional layer. The network structure includes: an initial convolutional layer (64 7×7 convolutional kernels, stride 2, 64 output channels), a max-pooling layer (3×3 pooling window, stride 2), and four sets of residual blocks (increasing channel counts of 64, 128, 256, and 512). A Monte Carlo dropout layer is added at the end to achieve joint angle-waveform feature extraction and defect size regression.

[0060] In practical applications, the pre-trained model is adapted to the current electromagnetic properties of the coating through transfer learning, based on the benchmark data from step 2. and the multi-angle time domain data in step 5 Construct a time-angle feature matrix and input it into a deep residual network. Add a probability output layer at the end of the network, and utilize the randomness of dropout to estimate the uncertainty of the model's predictions through 100 Monte Carlo forward propagations, and calculate the prediction variance. , obtain confidence level threshold For predictions with a confidence level below 90%, the angle sensitivity is calculated using a gradient-based activation map (Grad-CAM).

[0061] ;

[0062] in, It's an angle. The global sensitivity score indicates that the angle is more important for defect detection; This represents the predicted defect size output by the network. Representing the time-angle characteristic matrix in angle Time point The activation value at the location. Dynamically select the 3-5 optimal incident angles with the highest sensitivity to the current defect for supplementary scanning, and update the two-dimensional time-angle feature matrix. Until the error of two consecutive predictions is within the error threshold. This forms a closed-loop optimization system, controlling the final dimensional error within a certain range. level.

[0063] Those skilled in the art will readily understand that the above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A rapid detection method for defects in absorbing coatings based on incident angle modulation, characterized in that, Includes the following steps: Step 1: Initialize the detection system, set the operating frequency range of the vector network analyzer, fix the sample to be tested on the three-dimensional scanning platform, and install the antenna on the adjustable antenna bracket. Step 2: Perform multi-angle incident frequency sweep test on the defect-free coating surface to obtain echo signal data and perform inverse Fourier transform to obtain time domain data, which is used as reference data; Step 3: Set the incident mode to grazing incidence and perform a frequency sweep test on the defective coating surface. By comparing and analyzing the data with the reference data, determine whether there are surface defects. Step 4: Set the incident mode to vertical incidence and perform a frequency sweep test on the defective coating surface. By comparing and analyzing the data with the reference data, determine whether there are circular defects. Step 5: Once the existence of the circular defect is confirmed, perform multi-angle frequency sweep inversion to quantify the defect size. Determine the range of the projected size of the circular defect on the incident plane using time-domain data at different incident angles, and use an ellipse correction model to correct and obtain the actual physical size of the circular defect; the time-domain data at different incident angles is the multi-angle time-domain data. Step 6: Introduce a deep learning-based adaptive calibration module, construct a defect feature-size mapping model, and use a convolutional neural network to jointly analyze multi-angle time-domain data to obtain the final size of the circular defect.

2. The rapid detection method for absorbing coating defects based on incident angle modulation according to claim 1, characterized in that, In step 1, the first port of the vector network analyzer is connected to the transmitting antenna via a coaxial cable, and the second port of the vector network analyzer is connected to the receiving antenna via a coaxial cable.

3. The rapid detection method for absorbing coating defects based on incident angle modulation according to claim 1, characterized in that, In step 2, the multi-angle incident includes incident at a uniform angle or a non-uniform angle.

4. The rapid detection method for absorbing coating defects based on incident angle modulation according to claim 1, characterized in that, In step 3, the grazing incidence angle is selected based on the critical angle calculated from the electromagnetic characteristic parameters of the absorbing coating.

5. The rapid detection method for absorbing coating defects based on incident angle modulation according to claim 1, characterized in that, In step 4, the angle of vertical incidence is 90°.

6. The rapid detection method for absorbing coating defects based on incident angle modulation according to claim 1, characterized in that, In step 5, the value of the incident angle follows the principle of symmetrical scanning.

7. The rapid detection method for absorbing coating defects based on incident angle modulation according to claim 1, characterized in that, In step 6, the convolutional neural network is a deep residual network.

8. The rapid detection method for absorbing coating defects based on incident angle modulation according to claim 1, characterized in that, In step 6, the deep learning-based adaptive calibration module solves Maxwell's equations using the finite-difference time-domain method, obtains the frequency-domain scattering field under different dielectric constants, defect morphologies, and incident angles in the corresponding frequency band through simulation, and generates a time-domain training set through Fourier transform.

9. The rapid detection method for absorbing coating defects based on incident angle modulation according to claim 1, characterized in that, In step 6, the deep learning-based adaptive calibration module introduces a data augmentation strategy.

10. The rapid detection method for absorbing coating defects based on incident angle modulation according to claim 1, characterized in that, In step 6, the deep learning-based adaptive calibration module calculates the prediction variance and confidence level through 100 Monte Carlo forward propagation iterations. For prediction results with a confidence level below 90%, the gradient activation map is used to calculate the angle sensitivity. The 3-5 incident angles with the highest angle sensitivity are dynamically selected for supplementary scanning and updating until the prediction errors of two consecutive predictions are within the error threshold.

Images