A composite material ultrasonic defect depth detection and imaging method

Abstract

The application discloses a kind of composite material ultrasonic defect depth detection and imaging method, belong to nondestructive testing technical field, comprising: the A scanning signal of composite material test piece is collected by water immersion focusing probe;Signal is continuously wavelet transformed and dynamic threshold denoising;Particle swarm optimization wavelet transform modulus maximum (PSO-WTMM) model is constructed, to search optimal parameter with envelope entropy fusion peak penalty fitness function;Based on optimal parameter, extract upper and lower surface echo time difference, combine velocity to calculate defect depth;Point by point scanning constructs time difference and amplitude ratio matrix, and outputs C scanning image after interpolation enhancement.The application uses the above-mentioned composite material ultrasonic defect depth detection and imaging method, overcome the insufficient of traditional method parameter experience dependence, poor robustness, improve defect depth positioning precision and imaging quality, suitable for aviation composite material structure nondestructive testing and health monitoring.

Description

Technical Field

[0001] This invention relates to the field of nondestructive testing technology, and in particular to a method for detecting and imaging the burial depth of ultrasonic defects in composite materials. Background Technology

[0002] Composite materials, with their superior properties such as lightweight, high strength, fatigue resistance, and high designability, have become key structural materials in core fields such as aerospace (aircraft fuselages and wing components), rail transportation (high-speed train bodies), and new energy equipment (wind turbine blades). However, these materials are susceptible to process fluctuations and uneven resin distribution during manufacturing, and face assembly stress, complex loads, and environmental corrosion during service, leading to microscopic and macroscopic defects such as delamination, porosity, and fiber breakage. Among these, delamination damage, as the most significant failure mode of laminated structures, significantly reduces the material's stiffness, strength, and fatigue life, directly threatening the safe operation of equipment. Therefore, developing efficient and accurate internal defect detection and quantitative evaluation technologies is of great significance for ensuring the safety and reliability of composite material structures throughout their entire life cycle.

[0003] In the existing nondestructive testing (NDT) technology system, ultrasonic testing is widely used due to its high sensitivity to layered structural defects and moderate testing cost. Traditional contact ultrasonic testing relies on a coupling agent between the probe and the specimen to achieve sound energy transmission. However, uneven coupling pressure and coupling agent aging can easily lead to signal fluctuations, affecting the consistency and reliability of the test. Water immersion ultrasonic testing can achieve stable sound energy transmission through non-contact methods (such as water immersion coupling), effectively overcoming the shortcomings of contact coupling. It also has high sound energy transmission efficiency and can be adapted to the testing needs of complex structural components by adjusting the sound beam path. However, ultrasonic test signals are non-stationary and susceptible to interference from environmental noise and multi-interface reflections, making it difficult for traditional spectral analysis methods to effectively extract defect features.

[0004] The wavelet transform modulus maxima (WTMM) method, with its time-frequency localization properties, can accurately identify signal singularities (corresponding to defect reflection echoes), making it one of the core technologies for defect localization. However, the performance of this method is highly dependent on the selection of parameters such as the wavelet scale range, peak detection threshold, and minimum peak spacing. Existing technologies mostly set parameters empirically, lacking systematic optimization. This results in insufficient robustness when facing different materials and different defect types, large defects in depth measurement errors, and the need to improve imaging resolution and defect contour recognition capabilities.

[0005] To address the aforementioned issues, a composite material ultrasonic testing method is needed that can achieve adaptive parameter optimization and improve defect location accuracy and imaging quality, thereby resolving the core shortcomings of existing technologies such as parameter dependence on experience, large detection errors, and weak robustness. Summary of the Invention

[0006] The purpose of this invention is to provide a method for ultrasonic defect burial depth detection and imaging of composite materials. By adaptively optimizing the core parameters of wavelet transform modulus maxima (WTMM) using particle swarm optimization (PSO) algorithm, and combining the fitness function of envelope entropy and peak penalty mechanism, the method can achieve accurate positioning of defect depth and high-quality C-scan imaging, reduce detection error, and improve the robustness and engineering applicability of the method.

[0007] To achieve the above objectives, this invention provides a method for detecting and imaging the burial depth of ultrasonic defects in composite materials, comprising the following steps:

[0008] S1. Acquire A-scan ultrasonic signals of composite material specimens using an ultrasonic probe;

[0009] S2. Perform continuous wavelet transform on the A-scan ultrasound signal, implement dynamic threshold noise reduction, and extract the wavelet transform modulus maxima curve of the signal.

[0010] S3. Based on the wavelet transform modulus maxima curve, construct a particle swarm optimization wavelet transform modulus maxima model, and use the wavelet scale range, peak detection threshold and minimum peak spacing as optimization variables;

[0011] S4. Based on the fitness function of fusion envelope entropy and dynamic peak number penalty mechanism, the optimal parameter combination of optimization variables is searched through particle swarm optimization algorithm;

[0012] S5. Perform wavelet transform modulus maxima analysis based on the optimal parameter combination, extract the time difference between the upper surface echo and the defect or lower surface echo, and calculate the defect burial depth by combining the material sound velocity.

[0013] S6. Perform steps S2 to S5 on each scanning point within the detection area to construct the time difference matrix and amplitude ratio matrix, and improve the image resolution through interpolation to obtain the C-scan image of the defect.

[0014] Preferably, in step S1, the ultrasonic probe is a water immersion focusing probe, and the specimen is placed with the defect side facing down in the liquid immersion tank, and the ultrasonic probe is aligned with the defect area from the upper surface to collect signals.

[0015] Preferably, the signal acquisition steps specifically include:

[0016] The composite material specimen was placed horizontally in the liquid immersion tank with the pre-made defect side facing down, and the probe beam axis was aligned with the center area of ​​the defect.

[0017] Configure the center frequency of the probe and the sampling frequency of signal acquisition according to the detection requirements;

[0018] The control unit works in conjunction with the two-dimensional scanning frame to drive the water immersion focusing probe to scan the test area of ​​the specimen point by point.

[0019] Preferably, step S2 specifically includes:

[0020] S21. Using the Mexican hat wavelet as the mother wavelet, the time-domain expression of the Mexican hat wavelet is:

[0021] ;

[0022] in, Indicates time;

[0023] The frequency domain expression is:

[0024] ;

[0025] in, Indicates frequency;

[0026] Continuous wavelet transform was performed on the A-scan ultrasound signal to obtain the time-frequency domain wavelet coefficient matrix;

[0027] S22. The noise standard deviation is estimated based on the median absolute deviation (MAD) of wavelet coefficients. The calculation formula is as follows:

[0028] ;

[0029] in, The wavelet coefficient matrix, As a scale factor, The translation factor is... Represents the median function;

[0030] S23. Introducing a scale adjustment factor Construct an adaptive threshold that dynamically changes with scale. The calculation formula is:

[0031] ;

[0032] in, For signal length, The segmented weighted adjustment factor is calculated using the following formula:

[0033] ;

[0034] in, The scale-weighted sensitivity coefficient. To analyze the geometric mean of the scale range, , Indicates the smallest scale. Indicates the maximum scale;

[0035] S24. A soft thresholding function is used to filter and reduce noise in the wavelet coefficients. The expression is:

[0036] ;

[0037] in, It is a symbolic function;

[0038] S25. Reconstruct the time-domain signal using an energy-weighted method, with weighting coefficients... The calculation formula is:

[0039] ;

[0040] In the reconstructed signal, at the position corresponding to the significant peak of the original signal, the echo amplitude is adaptively enhanced by an enhancement factor, and then smoothed by a local window moving average to obtain the preprocessed signal;

[0041] S26. Perform continuous wavelet transform on the preprocessed signal again, extract the wavelet transform modulus maxima curve, and identify signal singularities.

[0042] Preferably, in step S3, constructing the Particle Swarm Optimized Wavelet Transform Modulus Maximum (PSO-WTMM) model specifically includes the following steps:

[0043] S31. Encode the wavelet scale range, peak detection threshold, and minimum peak spacing into particle position vectors. The parameter search boundary (unit: sampling points) is set according to the signal characteristics. Indicates the smallest scale. Indicates the maximum scale. Indicates the peak detection threshold. Indicates the minimum peak spacing;

[0044] S32. Initialize particle swarm parameters, and set the appropriate number of particles, number of iterations, inertia weight, learning factor and inertia weight decay coefficient according to optimization requirements.

[0045] S33. Design the fitness function :

[0046] ;

[0047] in, For the number of iterations, For the envelope entropy, This represents the deviation in the number of peak values. For the maximum penalty constant, This is a dynamic penalty coefficient. This represents the actual peak number of tests. This is the maximum value within the preset peak range;

[0048] S34. Iterative optimization based on the particle velocity and position update formula:

[0049] ;

[0050] in, For the first The particle reached the [number]th [particle]. The best position in the next iteration of history. For the entire population up to the [number]th The historical best position found in the next iteration and They represent particles respectively In the Velocity and position at the next iteration and They represent particles respectively In the Velocity and position at the next iteration Indicates inertia weight, and As a learning factor, Adjust the step size (cognitive component) of the particle as it flies toward its historical best position. Adjust the step size (social component) of the particle as it flies toward the global historical best position. and A random number between [0,1];

[0051] S35. When the maximum number of iterations is reached, output the globally optimal parameter combination.

[0052] Preferably, in step S4, the specific calculation steps of the fitness function include:

[0053] S41. Perform Hilbert transform on the wavelet transform modulus maxima curve corresponding to the current parameter combination to obtain the envelope of the wavelet transform modulus maxima curve;

[0054] S42. Calculate the Shannon entropy of the envelope and use this Shannon entropy as the envelope entropy. :

[0055] ;

[0056] in, For the envelope line The energy percentage of each sampling point This represents the total number of envelope sampling points;

[0057] S43, Number of actual peak detections on the detection envelope Set a reasonable range for the number of effective peak values. , To preset the lower limit of the peak value, To preset the upper limit of the peak value, calculate the deviation of the number of peak values. :

[0058] ;

[0059] S44. Constructing a dynamic penalty coefficient The calculation formula is:

[0060] ;

[0061] in, This represents the current iteration number. This represents the total number of iterations. and These are the initial penalty coefficient and the termination penalty coefficient, respectively.

[0062] S45, Combining envelope entropy Dynamic penalty coefficient and peak quantity deviation The fitness function value is obtained. .

[0063] Preferably, step S5 specifically includes:

[0064] S51. Extract the wavelet transform modulus maxima (WTMM) curve of the preprocessed signal based on the optimal parameter combination, identify the peak points of the modulus maxima of the echoes from the upper surface of the material and the echoes from defects or the lower surface, and record the time of the two peak points. and Calculate the time difference:

[0065] ;

[0066] S52. Calculate the burial depth of defects The calculation formula is:

[0067] ;

[0068] in, denoted as the ultrasonic wave propagation speed in the composite material.

[0069] Preferably, step S6 specifically includes:

[0070] S61. Define the detection area, set the scanning step size according to the detection resolution requirements, and drive the probe to scan the detection area point by point using a two-dimensional scanning frame. Repeat steps S2 to S4 to obtain the time difference of each scanning point. and amplitude ratio Wherein, the time difference is the time corresponding to the second peak of the modulus maximum curve minus the time corresponding to the first peak; the amplitude ratio is the amplitude corresponding to the second peak of the modulus maximum curve divided by the amplitude corresponding to the first peak.

[0071] S62. Constructing the time difference matrix And amplitude ratio matrix ;

[0072] S63. Using the natural neighborhood interpolation method for the time difference matrix And amplitude ratio matrix Perform resolution enhancement processing.

[0073] Therefore, the present invention employs the above-mentioned method for ultrasonic defect burial depth detection and imaging of composite materials, which has the following beneficial effects:

[0074] Adaptive parameter optimization and strong robustness: The PSO-WTMM model is constructed, and the wavelet scale range, peak detection threshold and minimum peak spacing are adaptively optimized through the particle swarm algorithm. This overcomes the shortcomings of the traditional WTMM method in that the parameters depend on experience. It can be adapted to detection scenarios of different materials and different defect types, and the generalization ability is significantly improved.

[0075] High defect location accuracy: Based on the fitness function that integrates envelope entropy and peak penalty mechanism, the physical rationality of peak detection is constrained while enhancing the significance of signal features, thereby reducing the measurement error of defect burial depth;

[0076] Excellent imaging quality: Based on dual-feature imaging of time difference ratio and amplitude ratio, combined with interpolation resolution enhancement technology, the difference between the defect area and the background is effectively magnified, avoiding artifact interference, and the location and outline of defects of different diameters and depths can be clearly presented.

[0077] Good detection stability: With wavelet denoising and signal enhancement processing, it can maintain stable detection performance even in complex scenarios such as edge regions and multiple defects.

[0078] Wide engineering applicability: Applicable to non-destructive testing of various layered composite materials, and can be extended to multiple fields such as aerospace and rail transportation, providing accurate and efficient technical support for health monitoring of composite material structures.

[0079] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0080] Figure 1 This is a flowchart of a method according to an embodiment of the present invention;

[0081] Figure 2 This is a flowchart of the PSO-WTMM model construction process according to an embodiment of the present invention;

[0082] Figure 3 These are comparison images of the ultrasonic signal before and after wavelet denoising in an embodiment of the present invention: (a) is the original signal, and (b) is the signal after denoising.

[0083] Figure 4 This is a graph showing the average magnitude of the modulus maximum under the initial parameters in an embodiment of the present invention.

[0084] Figure 5 The following are the WTMM curve envelope and peak detection diagrams of the present invention: (a) is a comparison of the modulus maxima and the Hilbert envelope, and (b) is the detection of singular points on the envelope.

[0085] Figure 6 The following is a diagram of the PSO algorithm parameter optimization iteration process in an embodiment of the present invention: (a) is the particle velocity change curve, (b) is the particle swarm diversity change curve, and (c) is the global optimal fitness convergence curve.

[0086] Figure 7 This is a graph showing the average magnitude of the modulus maximum under the optimal parameter combination in this embodiment of the invention.

[0087] Figure 8 The following are C-scan images of the elliptical flat-bottomed hole specimen according to an embodiment of the present invention: (a) is time-difference ratio imaging, and (b) is amplitude ratio imaging.

[0088] Figure 9 The following are abnormal point detection diagrams of the C-scan imaging of the elliptical flat-bottomed hole specimen according to an embodiment of the present invention: (a) is an abnormal point of time difference ratio, and (b) is an abnormal point of amplitude ratio.

[0089] Figure 10 The images shown are C-scan images of an elliptical flat-bottomed hole after image resolution enhancement by interpolation method according to an embodiment of the present invention: (a) is time difference ratio imaging, and (b) is amplitude ratio imaging.

[0090] Figure 11 This is a C-scan time difference ratio imaging diagram of a circular flat-bottomed hole specimen according to an embodiment of the present invention;

[0091] Figure 12 This is an image of the C-scan amplitude ratio of a circular flat-bottomed hole specimen according to an embodiment of the present invention. Detailed Implementation

[0092] The technical solution of the present invention will be further described below with reference to the accompanying drawings and embodiments.

[0093] Unless otherwise defined, the technical or scientific terms used in this invention shall have the ordinary meaning understood by one of ordinary skill in the art to which this invention pertains. The terms "first," "second," and similar terms used in this invention do not indicate any order, quantity, or importance, but are merely used to distinguish different components. Terms such as "comprising" or "including" mean that the element or object preceding the word encompasses the elements or objects listed following the word and their equivalents, without excluding other elements or objects. Terms such as "connected" or "linked" are not limited to physical or mechanical connections, but can include electrical connections, whether direct or indirect. Terms such as "upper," "lower," "left," and "right" are used only to indicate relative positional relationships; when the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0094] Example

[0095] In this embodiment, a T300 carbon fiber composite plate was selected as the sample plate, and two defect scenarios (single elliptical defect and multi-sized circular defect) were designed to verify the defect burial depth positioning accuracy and C-scan imaging effect of the proposed method.

[0096] First sample plate (elliptical defect): The dimensions are set to 200mm×200mm×5mm. One elliptical flat-bottomed hole defect is prefabricated with a major axis of 25mm, a minor axis of 5mm, a depth of 1.5mm (distance from the bottom of the hole to the lower surface of the specimen), and a remaining thickness of 3.5mm from the bottom of the hole to the test surface (upper surface).

[0097] The second sample plate (multi-specification circular defects): the size is set to 240mm×180mm×5mm, and 9 circular flat-bottomed hole defects are prefabricated in 3 rows×3 columns, row 1 (diameter 15mm), row 2 (diameter 10mm), row 3 (diameter 5mm), the defect depths of each row are 2mm, 1.5mm and 1mm respectively, and the corresponding remaining thicknesses are 4mm, 3.5mm and 3mm respectively.

[0098] like Figure 1 As shown in the figure, this embodiment provides a method for detecting and imaging the burial depth of ultrasonic defects in composite materials. The specific implementation steps are as follows:

[0099] S1: Detection system setup and calibration.

[0100] The water immersion ultrasonic testing system was built, and its core components include control software (developed based on LabVIEW), a signal generator (model JPR600C), a signal acquisition unit, a preamplifier, a host computer, a two-dimensional scanning frame, a liquid immersion tank, and a water immersion focusing probe (manufactured by Olympus, with a center frequency of 10MHz and a focal length of 5.08cm).

[0101] The system debugging and calibration steps are as follows:

[0102] Probe and sound velocity calibration: Using a 4mm thick T300 carbon fiber standard test block, the time difference between the upper surface echo (pr1) and the lower surface echo (pr2) was measured to infer the ultrasonic propagation velocity. The measurement was repeated 5 times and the average value was taken to obtain the calibrated sound velocity C = 2920.57m / s, ensuring that the sound velocity error is ≤0.5%.

[0103] Scanning frame accuracy calibration: The X and Y axis movement accuracy of the 2D scanning frame is verified by a laser rangefinder. The scanning step length is set to 1mm, and the position deviation is measured every 10mm. If the deviation exceeds ±0.05mm, adjustments are made to ensure accurate scanning position.

[0104] Signal system debugging: Adjust the signal generator transmission power, set the preamplifier gain to 30dB, the sampling frequency to 100MHz, and the probe distance from the upper surface of the sample plate to 45mm.

[0105] Sample plate placement: Place the sample plate horizontally in the immersion tank with the pre-made defect side facing down, ensuring that the center area of ​​the defect is directly aligned with the focal point of the water immersion focusing probe to avoid focusing deviation that could lead to signal attenuation.

[0106] S2: Signal acquisition and preprocessing.

[0107] S21, Raw signal acquisition.

[0108] The water immersion probe is driven by a two-dimensional scanning frame to scan the test area of ​​the sample plate point by point:

[0109] First sample plate (elliptical defect): A 40mm×40mm scanning area was defined, with a scanning step of 1mm, and the original A-scan ultrasonic signal of each scanning point was collected;

[0110] The second sample plate (multiple circular defects): the area is divided into 9 scanning units of 30mm×30mm, each corresponding to 9 circular defects. The scanning step size is 1mm. The original A-scan ultrasonic signal is collected unit by unit and stored as time domain signal data.

[0111] The original signal has non-stationary characteristics and contains environmental noise and multi-interface reflection interference, so it needs to be preprocessed to extract effective features.

[0112] S22, Signal preprocessing and WTMM extraction.

[0113] The original signal is processed using the Mexican hat wavelet, and the specific steps are as follows:

[0114] Continuous Wavelet Transform (CWT): Using the Mexh wavelet as the mother wavelet, a continuous wavelet transform is performed on the original A-scan signal to obtain the time-frequency domain wavelet coefficient matrix. ( As a scale factor, (Translation factor)

[0115] Noise Standard Deviation Estimation: Estimating Noise Standard Deviation Based on Median Absolute Deviation (MAD) of Wavelet Coefficients This method can robustly suppress the interference of outliers on noise estimation;

[0116] Adaptive threshold construction: Introducing a scale adjustment factor Construct a threshold that dynamically changes with scale. ;

[0117] Soft thresholding denoising: A soft thresholding function is used to select wavelet coefficients, avoiding signal abrupt distortion caused by hard thresholding. The effect before and after denoising is shown in the figure. Figure 3 As shown, high-frequency noise is significantly suppressed, and the signal peaks are more prominent.

[0118] Signal reconstruction and enhancement: The time-domain signal is reconstructed using an energy-weighted method. In the reconstructed signal, the significant peak positions of the original signal are determined by local maximum detection. An enhancement factor of 1.1-2.0 is used to adaptively enhance the echo amplitude. Then, the signal is smoothed by local window moving average to obtain the preprocessed signal.

[0119] WTMM Extraction: The preprocessed signal undergoes another continuous wavelet transform to extract the Wavelet Transform Modulus Maximum (WTMM) curve and identify signal singularities. This core process is integrated into the PSO-WTMM model construction, as follows: Figure 2 As shown.

[0120] S3: PSO-WTMM parameter optimization.

[0121] like Figure 2 As shown, a PSO-WTMM parameter optimization model is constructed to achieve adaptive optimization of wavelet scale range, peak detection threshold, and minimum peak spacing. The specific steps are as follows:

[0122] S31. Optimize parameter encoding and boundary settings.

[0123] The three core parameters are encoded into particle position vectors. :

[0124] Wavelet scale range: (Lower limit of scale) (Scale upper limit), peak detection threshold: (The proportion relative to the maximum amplitude of the signal);

[0125] Minimum peak spacing: (Unit: sampling points).

[0126] S32, Particle swarm parameter initialization.

[0127] Particle swarm optimization parameters: number of particles 30, number of iterations 20, inertia weight. =0.7, inertia weight decay coefficient 0.985 (after each iteration) = ×0.985), individual learning factor =1.8, social learning factor =1.6, , A uniformly random number between [0,1];

[0128] S33, Iterative optimization and optimal parameter output.

[0129] Iterative optimization based on the particle velocity-position update formula:

[0130] ;

[0131] in, For the first The particle reached the [number]th [particle]. The best position in the next iteration of history. For the entire population up to the [number]th The historical best position found in the next iteration and They represent particles respectively In the Velocity and position at the next iteration and They represent particles respectively In the Velocity and position at the next iteration Indicates inertia weight, and As a learning factor, Adjust the step size (cognitive component) of the particle as it flies toward its historical best position. Adjust the step size (social component) of the particle as it flies toward the global historical best position. and It is a random number between [0,1].

[0132] Iterative process as follows Figure 6 As shown: Figure 6 (a) in the figure shows the particle velocity change curve. After the 6th iteration, the velocity tends to stabilize, indicating that the algorithm has shifted from global exploration to local fine-tuning. Figure 6 (b) in the figure is the particle swarm diversity curve. The diversity is high in the early stage of iteration and tends to be stable in the later stage, reflecting the convergence of particles towards the optimal region. Figure 6(c) in the figure represents the global optimal fitness convergence curve. After the 6th iteration, the fitness value stabilizes, and the iteration terminates (in this embodiment, the convergence condition is met after the 6th iteration. Therefore, iterating to the preset maximum number of iterations (20 times) can fully guarantee that the algorithm converges to a stable optimal solution). Before outputting the optimal parameter combination, observe the average magnitude of the modulus maxima under the initial parameters (e.g., Figure 4 As shown in the figure, the curve fluctuates chaotically, and the peak detection accuracy is low. After adopting the optimal parameters (scale range 9:10, peak detection threshold 0.3 × maximum value, minimum peak spacing 59 sampling points), the average amplitude of the modulus maxima is shown in the figure. Figure 7 As shown, three significant peaks can be clearly identified, corresponding to pr1, pr2, and pr3 echoes respectively, and the peak positions are highly consistent with the actual echo times.

[0133] S4: Fitness function calculation.

[0134] Design a fitness function :

[0135] ;

[0136] Specific calculations:

[0137] Envelope Entropy Perform a Hilbert transform on the WTMM curve corresponding to the current parameter combination to obtain the envelope (e.g., Figure 5 As shown in (a), the correspondence between the WTMM curve and the Hilbert envelope is clearly presented. The Shannon entropy of the envelope is calculated, and this Shannon entropy is used as the envelope entropy. , The smaller the value, the more concentrated the signal characteristics;

[0138] Peak quantity deviation Set a reasonable range for the number of effective peak values. =[2,3], Detect the number of peak values ​​in the envelope. According to the formula Calculate the deviation:

[0139] ;

[0140] Dynamic penalty coefficient :

[0141] ;

[0142] in, This represents the current iteration number. This represents the total number of iterations. and The initial penalty coefficient and the final penalty coefficient are set to 0.01 and 100, respectively, to achieve dynamic adjustment of the penalty weight with coefficients ranging from 0.26 to 100.

[0143] Extreme situation punishment: when When =0, =1000, impose a severe penalty.

[0144] S5: Defect burial depth calculation.

[0145] Based on the optimal parameter combination, the WTMM curve is optimized to accurately extract the echo time difference and calculate the defect burial depth.

[0146] Peak point identification: A peak detection algorithm (combined with minimum peak spacing constraint) is used to identify peak points. Figure 7 In the optimized WTMM curve shown, the peak points of the modulus maxima of the upper surface echo and the defect / lower surface echo are identified, and the time of the two peak points is recorded. and Calculate the time difference ;

[0147] Burial depth calculation: according to the formula Calculate the burial depth of the defect ( =2920.57m / s is the ultrasonic propagation speed of T300 carbon fiber composite material.

[0148] Example result: A circular defect with a diameter of 15 mm and a depth of 2 mm in a circular sample plate, time difference =1.405μs, calculated burial depth 4.102mm, relative error 2.55%.

[0149] S6: Defect C-scan imaging.

[0150] A dual feature matrix is ​​constructed based on the optimized WTMM features, and the final C-scan image is obtained through interpolation enhancement. The specific steps are as follows:

[0151] Feature matrix construction: Traverse all scan points and construct the time difference matrix. And amplitude ratio matrix :

[0152] Time difference ;( The time corresponding to the second peak, (The time corresponding to the first peak);

[0153] Amplitude ratio ( The amplitude corresponding to the second peak. (The amplitude corresponding to the first peak);

[0154] Elliptical defect imaging: The C-scan imaging results of the first sample plate are as follows Figure 8 As shown, Figure 8 (a) in the image represents time-difference ratio imaging. Figure 8 (b) in the image shows amplitude ratio imaging, which can initially identify the location of elliptical defects; further anomaly detection (such as...) can be performed on the imaging results. Figure 9 As shown in the figure, the time difference ratio anomalies are concentrated inside the defect, while the amplitude ratio anomalies are distributed at the defect boundary, which verifies the accuracy of defect location.

[0155] Interpolation resolution enhancement: Natural neighborhood interpolation method is used for... and Interpolation processing is performed, increasing the image pixel density to four times the original density (equivalent scan step size 0.25mm). The enhanced imaging result is as follows. Figure 10 As shown, the defect edges are more clearly defined and the details are more prominent;

[0156] Multi-circular defect imaging: C-scan time-difference ratio imaging results of the second sample plate are as follows Figure 11 As shown, the amplitude ratio imaging is as follows: Figure 12 As shown, nine circular defects of different diameters and depths were accurately identified without artifact interference. The reflection area and signal intensity of defects of different diameters showed significant differences, which is consistent with the actual defect characteristics.

[0157] Therefore, the present invention adopts the above-mentioned method for ultrasonic defect burial depth detection and imaging of composite materials. Through precise system calibration, signal preprocessing, and PSO parameter optimization, the detection error is effectively reduced, and the robustness and engineering applicability of the method are improved. It can be widely used in non-destructive testing of composite material defects in aerospace, rail transportation and other fields.

[0158] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the technical solutions of the present invention, and these modifications or equivalent substitutions cannot cause the modified technical solutions to deviate from the spirit and scope of the technical solutions of the present invention.

Claims

1. A method for detecting and imaging the burial depth of ultrasonic defects in composite materials, characterized in that, Includes the following steps: S1. Acquire A-scan ultrasonic signals of composite material specimens using an ultrasonic probe; S2. Perform continuous wavelet transform on the A-scan ultrasound signal, implement dynamic threshold noise reduction, and extract the wavelet transform modulus maxima curve of the signal. S3. Based on the wavelet transform modulus maxima curve, construct a particle swarm optimization wavelet transform modulus maxima model, and use the wavelet scale range, peak detection threshold and minimum peak spacing as optimization variables; S4. Based on the fitness function of fusion envelope entropy and dynamic peak number penalty mechanism, the optimal parameter combination of optimization variables is searched through particle swarm optimization algorithm; S5. Perform wavelet transform modulus maxima analysis based on the optimal parameter combination, extract the time difference between the upper surface echo and the defect or lower surface echo, and calculate the defect burial depth by combining the material sound velocity. S6. Perform steps S2 to S5 on each scanning point within the detection area to construct the time difference matrix and amplitude ratio matrix, and improve the image resolution through interpolation to obtain the C-scan image of the defect. In step S4, the specific calculation steps of the fitness function include: S41. Perform Hilbert transform on the wavelet transform modulus maxima curve corresponding to the current parameter combination to obtain the envelope of the wavelet transform modulus maxima curve; S42. Calculate the Shannon entropy of the envelope and use this Shannon entropy as the envelope entropy. : ； in, For the envelope line The energy percentage of each sampling point This represents the total number of envelope sampling points; S43, Number of actual peak detections on the detection envelope Set a reasonable range for the number of effective peak values. , To preset the lower limit of the peak value, To preset the upper limit of the peak value, calculate the deviation of the number of peak values. : ； S44. Constructing a dynamic penalty coefficient The calculation formula is: ； in, For the number of iterations, This represents the total number of iterations. and These are the initial penalty coefficient and the termination penalty coefficient, respectively. S45, Combining envelope entropy Dynamic penalty coefficient and peak quantity deviation The fitness function value is obtained. .

2. The method for ultrasonic defect burial depth detection and imaging of composite materials according to claim 1, characterized in that, In step S1, the ultrasonic probe is a water immersion focusing probe. The specimen is placed with the defect side facing down in the liquid immersion tank, and the ultrasonic probe is aimed at the defect area from the upper surface to collect signals.

3. The method for ultrasonic defect burial depth detection and imaging of composite materials according to claim 2, characterized in that, The signal acquisition steps specifically include: The composite material specimen was placed horizontally in the liquid immersion tank with the pre-made defect side facing down, and the probe beam axis was aligned with the center area of ​​the defect. Configure the center frequency of the probe and the sampling frequency of signal acquisition according to the detection requirements; The control unit works in conjunction with the two-dimensional scanning frame to drive the water immersion focusing probe to scan the test area of ​​the specimen point by point.

4. The method for ultrasonic defect burial depth detection and imaging of composite materials according to claim 1, characterized in that, Step S2 specifically includes: S21. Using the Mexican hat wavelet as the mother wavelet, the time-domain expression of the Mexican hat wavelet is: ； in, Indicates time; The frequency domain expression is: ； in, Indicates frequency; Continuous wavelet transform was performed on the A-scan ultrasound signal to obtain the time-frequency domain wavelet coefficient matrix; S22. The median absolute deviation is used to estimate the noise standard deviation, and the calculation formula is as follows: ； in, The wavelet coefficient matrix, As a scale factor, The translation factor is... Represents the median function; S23. Introducing a scale adjustment factor Construct an adaptive threshold that dynamically changes with scale. The calculation formula is: ； in, For signal length, The segmented weighted adjustment factor is calculated using the following formula: ； in, The scale-weighted sensitivity coefficient. To analyze the geometric mean of the scale range, , Indicates the smallest scale. Indicates the maximum scale; S24. A soft thresholding function is used to filter and reduce noise in the wavelet coefficients. The expression is: ； in, It is a symbolic function; S25. Reconstruct the time-domain signal using an energy-weighted method, with weighting coefficients... The calculation formula is: ； In the reconstructed signal, at the position corresponding to the significant peak of the original signal, the echo amplitude is adaptively enhanced by an enhancement factor, and then smoothed by a local window moving average to obtain the preprocessed signal; S26. Perform continuous wavelet transform on the preprocessed signal again, extract the wavelet transform modulus maxima curve, and identify signal singularities.

5. The method for ultrasonic defect burial depth detection and imaging of composite materials according to claim 4, characterized in that, In step S3, the particle swarm optimization wavelet transform modulus maxima model is constructed, which specifically includes the following steps: S31. Encode the wavelet scale range, peak detection threshold, and minimum peak spacing into particle position vectors. The search boundary is set according to the signal characteristics, where, Indicates the smallest scale. Indicates the maximum scale. Indicates the peak detection threshold. Indicates the minimum peak spacing; S32. Initialize particle swarm parameters, and set the appropriate number of particles, number of iterations, inertia weight, learning factor and inertia weight decay coefficient according to optimization requirements. S33. Design the fitness function : ； in, For the number of iterations, For the envelope entropy, This represents the deviation in the number of peak values. For the maximum penalty constant, This is a dynamic penalty coefficient. This represents the actual peak number of tests. This is the maximum value within the preset peak range; S34. Iterative optimization based on the particle velocity and position update formula: ； in, For the first The particle reached the [number]th [particle]. The best position in the next iteration of history. For the entire population up to the [number]th The historical best position found in the next iteration and They represent particles respectively In the Velocity and position at the next iteration and They represent particles respectively In the Velocity and position at the next iteration Indicates inertia weight, and As a learning factor, Adjusting the step size of the particle's flight towards its historical best position. Adjust the step size of the particle's flight towards the global historical best position. and A random number between [0,1]; S35. When the maximum number of iterations is reached, output the globally optimal parameter combination.

6. The method for ultrasonic defect burial depth detection and imaging of composite materials according to claim 5, characterized in that, Step S5 specifically includes: S51. Based on the optimal parameter combination, extract the wavelet transform modulus maxima curve of the preprocessed signal, identify the modulus maxima peak points of the echoes from the upper surface of the material and the echoes from defects or the lower surface, and record the time of the two peak points. and Calculate the time difference: ； S52. Calculate the burial depth of defects The calculation formula is: ； in, denoted as the ultrasonic wave propagation speed in the composite material.

7. The method for ultrasonic defect burial depth detection and imaging of composite materials according to claim 6, characterized in that, Step S6 specifically includes: S61. Define the detection area, set the scanning step size according to the detection resolution requirements, and drive the probe to scan the detection area point by point using a two-dimensional scanning frame. Repeat steps S2 to S4 to obtain the time difference of each scanning point. and amplitude ratio Wherein, the time difference is the time corresponding to the second peak of the modulus maximum curve minus the time corresponding to the first peak; the amplitude ratio is the amplitude corresponding to the second peak of the modulus maximum curve divided by the amplitude corresponding to the first peak. S62. Constructing the time difference matrix And amplitude ratio matrix ; S63. Using the natural neighborhood interpolation method for the time difference matrix And amplitude ratio matrix Perform resolution enhancement processing.

Images