Solid plate defect detection method based on Lamb wave end face echoes

Abstract

The invention relates to the technical field of solid plate defect detection, in particular to a solid plate defect detection method based on Lamb wave end face echoes, which comprises the following steps: exciting and receiving wedge piezoelectric transducers are arranged on the same side of a solid plate, exciting and receiving a single-mode Lamb wave signal, and acquiring a multi-path signal comprising a direct wave, a side face echo and a defect reflection echo; by analyzing the difference between a direct wave and a side echo, a defect reflection signal is extracted, and by combining the time difference between the defect reflection signal and the side echo and the Lamb wave propagation speed, a geometric model is established, and the defect position is accurately calculated. By adjusting the position of a transducer, the time difference between a defect reflected wave and a direct wave is constant, a defect echo amplitude value is obtained under the equal propagation condition and compared with a standard test block amplitude value, a mapping relation between the amplitude value and the size is constructed, and accurate quantification of the defect size is achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of solid plate defect detection, and in particular to a solid plate defect detection method based on Lamb wave end face echo. Background Art

[0002] In fields such as aerospace, rail transit, energy equipment, and high-end manufacturing, solid plate structures serve as critical load-bearing components, and their integrity is directly related to the safety and reliability of the overall structure. However, in actual service, due to load fatigue, environmental corrosion, or manufacturing defects, microcracks, holes, or localized debonding may occur within or at the boundaries of solid plates. If not discovered and assessed in a timely manner, catastrophic failure may occur. Therefore, the development of highly sensitive and high-precision plate defect detection technology is of great engineering significance.

[0003] Lamb waves are currently being widely studied and applied in nondestructive testing of plate structures due to their long propagation distance, sensitivity to defects, and ability to achieve multimodal excitation. The propagation of Lamb waves in plates is disrupted by defects, resulting in recognizable reflections, scattering, and mode conversion phenomena, which in turn provide a basis for defect detection, location, and quantification.

[0004] Existing Lamb wave-based defect detection methods often employ a transmitter-receiver configuration, with each transmitter and receiver facing each other. They rely primarily on overall changes in signal amplitude or comparison thresholds for defect identification. However, these methods are less sensitive to defects in boundary areas or complex spatial locations, and struggle to effectively separate various echo components. This is particularly true in the presence of multiple reflections and interference waves, leading to missed detections or misjudgments. Furthermore, most methods lack a standardized mapping mechanism for quantifying defect size, relying solely on relative amplitude changes for estimation, making it impossible to accurately invert the actual size. Summary of the Invention

[0005] In order to solve the above problems, the present invention provides a solid plate defect detection method based on Lamb wave end face echo.

[0006] To achieve the above object, the technical solution adopted by the present invention is:

[0007] A solid plate defect detection method based on Lamb wave end face echo, comprising:

[0008] S1. On one side of a solid plate, a single-mode Lamb wave is excited by a fixed-angle angled piezoelectric transducer, and the Lamb wave signal is received by another fixed-angle angled piezoelectric transducer arranged on the same side as the first one, to obtain a direct wave signal and a side echo signal.

[0009] S2. Extracting a defect reflection echo signal caused by a defect through signals received between the direct wave signal and the side echo signal, and obtaining time information and amplitude information of the defect reflection echo;

[0010] S3. Calculate the defect position based on the time difference between the defect reflection echo time information and the side echo time information, combined with the pre-calibrated Lamb wave propagation velocity;

[0011] S4. By adjusting the transducer position to keep the time difference between the defect reflection echo and the direct wave at a fixed value, the defect reflection echo amplitude is synchronously obtained and compared with the reference echo amplitude of the preset size defect comparison block to obtain the quantitative size information of the defect.

[0012] Furthermore, the direct wave signal is obtained by the following steps:

[0013] An excitation and receiving wedge piezoelectric transducer is arranged in the defect-free area of ​​the solid plate. Based on the set excitation frequency and excitation angle, the excitation transducer is triggered to emit a single-mode Lamb wave signal.

[0014] The receiving transducer receives the Lamb wave signal that propagates directly along the plate surface between the excitation point and the receiving point, and extracts the first high-amplitude pulse signal that arrives in the waveform from the Lamb wave receiving signal as the direct wave signal;

[0015] The direct wave signal is subjected to time window constraint and amplitude filtering to determine the time domain characteristics of the direct wave.

[0016] Furthermore, the high-amplitude pulse signal is obtained by the following steps:

[0017] A time window is set in the Lamb wave received signal, all data points in the time window are scanned, the absolute values ​​of their amplitudes are calculated, and high-amplitude pulse signals whose amplitudes reach local maximum values ​​are identified.

[0018] Furthermore, it is characterized in that the side echo signal is obtained by the following steps:

[0019] In the Lamb wave received signal, based on the identified direct wave signal, a backward search is performed for a next-level high-amplitude pulse signal within a time window, and a reflection signal is extracted in combination with an expected propagation time interval corresponding to the distance from the plate end face as a side echo signal;

[0020] The side echo signal is subjected to time window constraint and amplitude filtering to determine the side echo time information.

[0021] Furthermore, the step S2 includes the following steps:

[0022] Based on the direct wave time information and the side echo time information, the target time interval between the two is determined, and the Lamb wave received signal is intercepted within the time interval to obtain the candidate echo signal segment;

[0023] Based on the candidate echo signal segments, performing amplitude peak extraction processing, identifying local extreme points above the background noise threshold, and obtaining several suspected defect reflection echo points;

[0024] Based on the time position and amplitude change characteristics of the suspected defect reflection echo point, envelope demodulation and time window filtering are performed to obtain preliminary characteristics of the defect reflection echo;

[0025] Based on the preliminary characteristics of the defect reflection echo, a joint time-frequency domain analysis is performed to extract the effective time domain center value and amplitude envelope peak of the echo signal as the final time information and amplitude information of the defect reflection echo.

[0026] Furthermore, the time-frequency domain joint analysis includes:

[0027] Based on the preliminary characteristic signal of the defect reflection echo, short-time Fourier transform processing is performed to obtain the frequency energy distribution of the signal in the time domain and obtain the time-frequency spectrum matrix;

[0028] Based on the time-frequency spectrum matrix, the time center value and the corresponding frequency component of the energy concentration area are extracted to obtain a preliminary time-frequency feature vector;

[0029] The preliminary characteristic signal of the defect reflection echo is further decomposed by wavelet packet to obtain the energy distribution of each frequency band at several scales and obtain the sub-band energy vector;

[0030] The time-frequency feature vector and the sub-band energy vector are concatenated in a time-aligned manner to generate a joint time-frequency feature vector;

[0031] Based on the concentration of the time component in the joint time-frequency eigenvector, the final time information of the defect reflection echo is extracted; based on the frequency component with the largest energy amplitude in the joint eigenvector, the final amplitude information of the defect reflection echo is extracted.

[0032] Furthermore, the step S3 includes the following steps:

[0033] Based on the time information of the defect reflection echo and the time information of the side echo, the time difference between the defect echo and the side echo is calculated;

[0034] Based on the propagation velocity of the Lamb wave in the solid plate pre-calibrated by the direct wave and the side echo in the defect-free area, combined with the time difference, the equivalent propagation distance from the defect to the transducer is calculated according to the propagation path relationship;

[0035] Based on the current position of the transducer on the surface of the solid plate, the known spatial relationship between the transducer and the plate boundary, and the equivalent propagation distance, a geometric solution model of the defect position is established, and the two-dimensional position coordinates of the defect in the solid plate coordinate system are obtained by solution.

[0036] Furthermore, the step of adjusting the position of the transducer so as to maintain the time difference between the defect reflection echo and the direct wave at a fixed value includes:

[0037] According to the time information of the defect reflection echo and the time information of the direct wave, the initial time difference between the two is calculated as a fixed reference time difference;

[0038] Keeping the excitation transducer in a fixed position, the receiving transducer is translated along the surface of the solid plate to collect Lamb wave receiving signals at different positions, and the defect reflection echo time and direct wave time corresponding to each position are extracted respectively;

[0039] Based on the defect reflection echo time and direct wave time corresponding to each position, the real-time time difference between the current defect reflection echo and the direct wave is calculated and compared with the reference time difference until the time difference between the defect reflection echo and the direct wave is stable and equal to the reference time difference, and the spatial position of the receiving transducer is synchronously recorded.

[0040] Furthermore, the quantitative size information of the defect is obtained by the following steps:

[0041] Based on the spatial position of the transducer, the Lamb wave receiving signal at the position is collected, and the defect reflection echo signal is extracted in the window corresponding to the fixed time difference;

[0042] Perform amplitude envelope analysis on the extracted defect reflection echo signal to obtain the peak amplitude of the echo signal as the reflection intensity characteristic value of the current defect;

[0043] On a standard test block with artificial reflectors of known sizes, equidistant propagation paths and the same excitation frequency conditions are set up. The echo signals of each reflector of known size are collected, and the corresponding envelope peak amplitude is extracted respectively to construct a corresponding relationship model between the reflection amplitude and the defect size.

[0044] The amplitude characteristic value of the defect reflection echo is input into the corresponding relationship model, matched with the corresponding defect size, and used as the quantitative size information of the detected defect.

[0045] Furthermore, the construction of the corresponding relationship model between the reflection amplitude and the defect size includes the following steps:

[0046] Several artificial reflectors of different sizes are preset on the standard comparison test block, corresponding to different reflective target areas;

[0047] For each artificial reflector position, arrange the transducer excitation and receiving configuration consistent with the actual detection path to stimulate Lamb waves and collect the corresponding received signals;

[0048] Extract the reflected echo signal corresponding to each artificial reflector, calculate its amplitude envelope peak, and establish a data pair containing the one-to-one correspondence between the defect size and the echo amplitude;

[0049] An interpolation fitting process is performed on the data pair to construct a continuous function between the reflected echo amplitude and the defect size as a corresponding relationship model.

[0050] The beneficial effects of the present invention are as follows: by arranging the excitation and receiving wedge piezoelectric transducers on the same side of a solid plate, the present invention excites and receives a single-mode Lamb wave signal propagating within the plate, obtaining multiple propagation path signals including direct waves, side echoes, and defect reflection echoes, significantly improving the ability to identify edge areas and local defects, and overcoming the limitation of the existing opposite arrangement method that is prone to missed detection in boundary areas. Based on the echo sequence difference between the direct wave and the side echo, the reflection signal caused by the defect is accurately extracted, and combined with the time difference between the defect echo and the side echo and the pre-calibrated Lamb wave propagation speed, a high-precision geometric model is constructed to achieve quantitative calculation of the spatial position of the defect, effectively improving the positioning accuracy, and solving the problem of difficult echo separation and difficult position determination in a multi-interference environment. Furthermore, by adjusting the position of the transducer to keep the time difference between the defect reflection wave and the direct wave constant, the defect echo amplitude is obtained under equal propagation conditions. The amplitude is mapped and compared with the reference amplitude of the standard test block, and a functional relationship between the amplitude and size is established to achieve quantitative inversion of the actual size of the defect, overcoming the uncertainty problem of traditional methods that estimate size based solely on amplitude changes. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] Figure 1 This is a flow chart of the steps of a solid plate defect detection method based on Lamb wave end face echo in the present invention.

[0052] Figure 2 It is a flow chart of the steps for obtaining quantitative size information of defects in the present invention. DETAILED DESCRIPTION

[0053] See also Figure 1-2 As shown, the present invention relates to a solid plate defect detection method based on Lamb wave end face echo, comprising:

[0054] S1. On one side of a solid plate, a single-mode Lamb wave is excited by a fixed-angle angled piezoelectric transducer, and the Lamb wave signal is received by another fixed-angle angled piezoelectric transducer arranged on the same side as the first one, to obtain a direct wave signal and a side echo signal.

[0055] S2. Extracting a defect reflection echo signal caused by a defect through signals received between the direct wave signal and the side echo signal, and obtaining time information and amplitude information of the defect reflection echo;

[0056] S3. Calculate the defect position based on the time difference between the defect reflection echo time information and the side echo time information, combined with the pre-calibrated Lamb wave propagation velocity;

[0057] S4. By adjusting the transducer position to keep the time difference between the defect reflection echo and the direct wave at a fixed value, the defect reflection echo amplitude is synchronously obtained and compared with the reference echo amplitude of the preset size defect comparison block to obtain the quantitative size information of the defect.

[0058] It should be noted that Lamb waves in solid plates are elastic waves that propagate within a plate-like structure. They are formed by multiple reflections between the two free surfaces of the plate and are characterized by low energy attenuation and long propagation distance. When a Lamb wave encounters a defect or interface in the material, some energy scattering and attenuation occur. By detecting changes in the propagation velocity and amplitude of the wave, the location, size, and type of the defect within the material can be inferred. Modal conversion: Lamb waves exhibit different propagation properties at different frequencies and modes, responding differently to defects within the material. By analyzing the waveforms and propagation characteristics in different modes, defect information within the material can be identified. Therefore, this technique has found widespread application in nondestructive testing. In some embodiments, the excitation transducer and receiving transducer are arranged on the same side of the solid plate and coupled to the plate surface via an angled wedge structure, creating a fixed excitation angle. The excitation transducer excites a single-mode Lamb wave signal. After propagation within the plate, the receiving transducer receives a signal consisting of the direct wave, side echoes reflected from the plate end faces, and possible reflections from defects within the plate. This same-side arrangement enables clear separation of echo signals corresponding to different propagation paths within a single receiving signal channel, significantly improving detection capabilities for structural edge areas and local defects within panels. This overcomes the problem of missed detection in edge areas often encountered with existing opposite-side arrangements. Time-domain analysis of the received signal first extracts the temporal positions of the direct and side echoes. Then, within the time window between the two, the intermediate echo signal, formed by the defect reflection, is identified. This signal exhibits characteristics such as a sudden amplitude change and concentrated energy, and is confirmed as a defect echo by comparison with the background waveform. Furthermore, by measuring the time difference between the defect echo and the side echo and combining it with the propagation velocity of Lamb waves in the material, an equivalent propagation path model is established. The spatial position of the defect in the panel coordinate system is inferred, enabling precise defect location. Once the defect is located, the time difference between the defect echo and the direct echo is adjusted to a set constant value by translating the receiving transducer to quantitatively assess the defect size. This ensures that the defect reflection signal is captured under the same propagation path conditions. Under constant time difference conditions, the amplitude characteristics of the defect echo are extracted and compared with the echo amplitude corresponding to a reflector of known size in a preset standard test block. Through the functional mapping relationship between amplitude and size, the defect size can be inverted and quantified. Compared with the existing technology, this embodiment has significant innovations.First, the same-side arrangement effectively improves the detectability of edge defects and enhances the distinguishability of multi-path propagation signals; second, by introducing side echoes as stable reference points, a defect positioning method based on time difference is established, which improves the stability and accuracy of positioning; finally, through amplitude acquisition and standard amplitude comparison mechanism under constant propagation conditions, standardized inversion of defect size is achieved, overcoming the defects of traditional methods that are strongly dependent on the absolute value of amplitude and inaccurate size estimation, reflecting the comprehensive technical advantages of the present invention in defect detection integrity, spatial positioning capability and size quantification accuracy.

[0059] Furthermore, the direct wave signal is obtained by the following steps:

[0060] An excitation and receiving wedge piezoelectric transducer is arranged in the defect-free area of ​​the solid plate. Based on the set excitation frequency and excitation angle, the excitation transducer is triggered to emit a single-mode Lamb wave signal.

[0061] The receiving transducer receives the Lamb wave signal that propagates directly along the plate surface between the excitation point and the receiving point, and extracts the first high-amplitude pulse signal that arrives in the waveform from the Lamb wave receiving signal as the direct wave signal;

[0062] The direct wave signal is subjected to time window constraint and amplitude filtering to determine the time domain characteristics of the direct wave.

[0063] In some embodiments, a continuous section of the plate surface is first selected in a defect-free area of ​​the solid plate as a signal reference acquisition area, and the excitation and receiving oblique wedge piezoelectric transducers are arranged. The excitation frequency and excitation angle of the transducer are pre-set according to the dispersion characteristics of the target Lamb wave mode and kept stable. After the excitation transducer emits a single-mode Lamb wave signal, the signal directly reaches the receiving transducer along the propagation path within the plate. The propagation path does not undergo any interface reflection or scattering process, so the first set of waveforms received contains the purest direct wave information. In the process of analyzing and processing the received signal, the system performs amplitude absolute value conversion on the complete waveform data and identifies the position of the first high-amplitude pulse, that is, the first energy concentration peak. This signal usually has a prominent peak feature and is located in the earliest effective band after the excitation is triggered. Extracting it as a direct wave signal will help to establish an echo reference benchmark later. In order to enhance the stability and repeatability of the direct wave signal, a limited time window is further set within its time neighborhood, and amplitude filtering is performed on the waveform within the window to suppress high-frequency noise and low-amplitude interference items, extract its main wave packet structure, and form a direct wave reference template with clear time domain characteristics and clear waveform boundaries. Different from the existing technology, it uses the local standard path in the defect-free area to extract the real propagation echo to ensure that the extracted direct wave signal has structural symmetry and propagation consistency, avoiding the misjudgment problem caused by multipath superposition and signal aliasing. At the same time, the signal quality is enhanced through the time window and filtering mechanism, so that the subsequent time and amplitude comparison with the defect echo and side echo has an accurate reference standard, which improves the robustness and measurement accuracy of the overall defect detection process.

[0064] Furthermore, the high-amplitude pulse signal is obtained by the following steps:

[0065] A time window is set in the Lamb wave received signal, all data points in the time window are scanned, the absolute values ​​of their amplitudes are calculated, and high-amplitude pulse signals whose amplitudes reach local maximum values ​​are identified.

[0066] It should be noted that to extract valid high-amplitude pulse signals from Lamb wave received signals, a physically meaningful time window is first set within the original received signal. This time window is typically determined based on a priori propagation path information to cover the expected valid echo region while excluding interference signals caused by excitation delay, initial coupling noise, and multiple reflections at the far end. Within the selected time window, the absolute value of the amplitude is calculated for each discrete sampling point of the received signal, and an envelope curve is formed to highlight the waveform energy distribution characteristics. The system then performs a local extreme value scan on this amplitude curve, determining which points in the continuous sampling point meet the extreme value condition: that is, the amplitude of the current sampling point is greater than that of its adjacent previous and subsequent points and is also above a set noise threshold. Pulse signals that meet these conditions are extracted as high-amplitude pulse signals and are considered key signal units that may contain defect information or reference information. This processing method differs from traditional methods that filter waveforms based on overall amplitude changes or fixed thresholds. Instead, it uses a dynamic analysis method based on the signal's local extreme value characteristics, which can accurately extract high-energy pulse signals reflecting structural response characteristics from complex waveform backgrounds. By limiting the time window and introducing a local extreme value judgment mechanism, the extraction results are made more stable and robust, establishing a reliable signal foundation for subsequent timing discrimination, defect identification, and amplitude comparison. This method improves defect identification accuracy while reducing the risk of false triggering due to random noise.

[0067] Furthermore, it is characterized in that the side echo signal is obtained by the following steps:

[0068] In the Lamb wave received signal, based on the identified direct wave signal, a backward search is performed for a next-level high-amplitude pulse signal within a time window, and a reflection signal is extracted in combination with an expected propagation time interval corresponding to the distance from the plate end face as a side echo signal;

[0069] The side echo signal is subjected to time window constraint and amplitude filtering to determine the side echo time information.

[0070] In this embodiment, to accurately extract side echo signals from Lamb wave received signals, a target time window is first set backward from the identified direct wave signal's time position as a reference. The start and end ranges of this time window are calculated based on the physical distance between the receiving transducer and the panel end face and the pre-calibrated Lamb wave propagation velocity. This ensures that the window covers the theoretical arrival time interval of the end face reflection wave while eliminating any interference segments that may overlap with defect reflection waves. Within this target time window, the amplitude curve of the received signal is subjected to local extremum analysis, searching for high-amplitude pulse signals with prominent amplitudes. Because Lamb waves, after reflecting from the panel end face, return along their original path in a specific pattern, the reflected wave's energy is concentrated and its position is stable. Therefore, the extracted secondary maximum signal exhibits distinct waveform characteristics. This pulse signal is then used as a candidate side echo signal, and its neighboring waveform is further subjected to time windowing and amplitude filtering to reduce background interference while preserving the main waveform energy. Ultimately, the peak time of the waveform envelope is extracted to determine the effective time information of the side echo. This method effectively improves the accuracy of end-face reflection signal extraction by constructing a time reference system based on direct waves, combined with a time prediction mechanism based on theoretical propagation path calculations. It avoids the identification errors caused by echo overlap and unclear timing in traditional methods. Particularly in inspection environments with complex boundary conditions or significant multi-reflection interference, this solution still achieves highly robust side echo extraction, providing a stable reference time scale for the time location and difference analysis of defect reflection waves.

[0071] Furthermore, the step S2 includes the following steps:

[0072] Based on the direct wave time information and the side echo time information, the target time interval between the two is determined, and the Lamb wave received signal is intercepted within the time interval to obtain the candidate echo signal segment;

[0073] Based on the candidate echo signal segments, performing amplitude peak extraction processing, identifying local extreme points above the background noise threshold, and obtaining several suspected defect reflection echo points;

[0074] Based on the time position and amplitude change characteristics of the suspected defect reflection echo point, envelope demodulation and time window filtering are performed to obtain preliminary characteristics of the defect reflection echo;

[0075] Based on the preliminary characteristics of the defect reflection echo, a joint time-frequency domain analysis is performed to extract the effective time domain center value and amplitude envelope peak of the echo signal as the final time information and amplitude information of the defect reflection echo.

[0076] In this embodiment, to accurately extract reflected echo signals caused by structural defects, a target time interval is first determined based on the time information of the direct wave and side echoes obtained in the previous identification. This time interval theoretically corresponds to the possible occurrence range of defect reflection waves and has physical propagation path significance, helping to isolate the true and valid defect signal from the complex background. Subsequently, the complete Lamb wave received signal is intercepted within this time interval to form a candidate echo signal segment, providing an analytical basis for subsequent feature identification. Within this candidate signal segment, the system performs amplitude peak extraction, scanning all sampling points and identifying local extreme points with amplitudes above a set background noise threshold. This process ensures that only signal structures with significant energy changes are retained, preliminarily screening several reflection wave locations that may be related to defects. Next, for each suspected echo point, a local time window is constructed to extract the waveform signal of its neighborhood. Envelope demodulation is performed to extract the main peak feature from the amplitude envelope. To enhance signal reliability and anti-interference capabilities, time window filtering technology is further incorporated to suppress secondary noise interference and output preliminary defect reflection echo features with time-domain focusing characteristics. On the basis of obtaining preliminary features, the system performs joint time-frequency analysis on these candidate signals. Specifically, it includes: first, obtaining the frequency energy distribution of the signal in different time periods through short-time Fourier transform; secondly, analyzing the frequency band energy changes at different scales through wavelet packet decomposition. Finally, the two features are aligned and fused on the time axis to extract the time center value where the signal energy is most concentrated and the corresponding envelope main peak amplitude. This information serves as the final time information and amplitude information of the defect reflection echo, providing high-precision input for subsequent position inversion and size estimation. Compared with the traditional means of relying solely on amplitude changes or fixed threshold detection, this embodiment introduces time reference constraints, waveform structure analysis, frequency energy deconstruction and feature fusion mechanisms at multiple key processing nodes, so that defect echo identification has stronger stability and selectivity. Especially in scenarios with severe multipath interference or low signal-to-noise ratio, a high defect recognition rate can still be maintained.

[0077] Furthermore, the time-frequency domain joint analysis includes:

[0078] Based on the preliminary characteristic signal of the defect reflection echo, short-time Fourier transform processing is performed to obtain the frequency energy distribution of the signal in the time domain and obtain the time-frequency spectrum matrix;

[0079] Based on the time-frequency spectrum matrix, the time center value and the corresponding frequency component of the energy concentration area are extracted to obtain a preliminary time-frequency feature vector;

[0080] The preliminary characteristic signal of the defect reflection echo is further decomposed by wavelet packet to obtain the energy distribution of each frequency band at several scales and obtain the sub-band energy vector;

[0081] The time-frequency feature vector and the sub-band energy vector are concatenated in a time-aligned manner to generate a joint time-frequency feature vector;

[0082] Based on the concentration of the time component in the joint time-frequency eigenvector, the final time information of the defect reflection echo is extracted; based on the frequency component with the largest energy amplitude in the joint eigenvector, the final amplitude information of the defect reflection echo is extracted.

[0083] In this embodiment, to further enhance the accuracy and robustness of identifying defect reflection echo characteristics, a joint time-frequency domain analysis process is introduced based on the extracted preliminary defect reflection echo characteristic signal. First, a short-time Fourier transform is performed on the preliminary signal. By setting an appropriate sliding window function and time step, the signal is subjected to local frequency analysis to generate a complete time-frequency spectrum matrix. This matrix reflects the frequency energy distribution of the signal over different time periods and can intuitively demonstrate the two-dimensional variation of signal energy over time and frequency. After obtaining the time-frequency spectrum matrix, the region with the most concentrated energy is extracted. Its temporal center position and the corresponding dominant frequency component are analyzed to construct a preliminary time-frequency feature vector reflecting the local frequency characteristics of the signal. This vector represents the time-frequency energy core of the defect reflection wave and facilitates subsequent feature fusion to enhance discrimination. To improve the ability to capture details of non-stationary or multi-scale signals, wavelet packet decomposition is performed on the same preliminary feature signal to extract its energy distribution at multiple frequency bands and scales. The energy vector is reconstructed by decomposing the coefficients to obtain the signal's subband energy vector, reflecting its local frequency domain variation characteristics. Subsequently, the time-frequency feature vectors obtained by short-time Fourier analysis and the subband energy vectors obtained by wavelet packet decomposition are aligned along the time axis and spliced ​​together to construct a data structure of uniform length and time consistency, generating a joint time-frequency feature vector. This vector comprehensively retains the global frequency evolution information and local scale energy variation characteristics, improving the specificity and stability of defect wave identification. In the joint feature vector, analyzing the concentration of the time component can obtain the final time information of the defect reflection wave, that is, the time center value where the signal energy is most concentrated; at the same time, the position with the largest energy amplitude is extracted from the frequency dimension to obtain the final amplitude information as a significant characteristic indicator of the defect reflection wave.

[0084] Furthermore, the step S3 includes the following steps:

[0085] Based on the time information of the defect reflection echo and the time information of the side echo, the time difference between the defect echo and the side echo is calculated;

[0086] Based on the propagation velocity of the Lamb wave in the solid plate pre-calibrated by the direct wave and the side echo in the defect-free area, combined with the time difference, the equivalent propagation distance from the defect to the transducer is calculated according to the propagation path relationship;

[0087] Based on the current position of the transducer on the surface of the solid plate, the known spatial relationship between the transducer and the plate boundary, and the equivalent propagation distance, a geometric solution model of the defect position is established, and the two-dimensional position coordinates of the defect in the solid plate coordinate system are obtained by solution.

[0088] In this embodiment, to spatially locate a defect on a solid plate surface, the propagation time difference between the extracted defect reflection echo and the side echo is first calculated. This time difference reflects the difference between the path length from the defect location to the receiving transducer and the path length from the end face reflection point to the receiving transducer, and has direct geometric propagation significance. Subsequently, the Lamb wave propagation velocity, previously obtained in the system through calibration of direct waves and side echoes in a defect-free area, is called and multiplied by the above time difference to obtain the equivalent propagation distance corresponding to the defect echo. Furthermore, based on the known spatial position of the transducer on the solid plate surface, the position of the solid plate end face, the distance from the excitation transducer to the end face, and the angle between the receiving transducer and the possible defect path direction, a solution model is constructed that includes propagation path constraints and boundary geometric relationships. In this model, the defect location is represented as a unique solution in a set of points with a fixed distance from the receiving transducer and that satisfy the propagation path symmetry constraints. By performing algebraic analysis on the above-mentioned propagation distance, angle constraint and boundary position parameters, the two-dimensional position coordinates of the defect in the solid plate coordinate system are finally solved, and the output is a point result with actual spatial reference significance. Compared with the existing method of inferring the defect position based on the overall waveform offset, this embodiment has a clearer physical basis and a quantifiable propagation model. It does not rely on empirical regression or inversion training models, and avoids excessive sensitivity to factors such as environmental noise and structural complexity. Especially in scenarios with multipath interference or signal delay overlap, this method can still establish a time difference framework through a stable side echo reference, effectively achieving high-precision solution of the defect position.

[0089] Furthermore, the step of adjusting the position of the transducer so as to maintain the time difference between the defect reflection echo and the direct wave at a fixed value includes:

[0090] According to the time information of the defect reflection echo and the time information of the direct wave, the initial time difference between the two is calculated as a fixed reference time difference;

[0091] Keeping the excitation transducer in a fixed position, the receiving transducer is translated along the surface of the solid plate to collect Lamb wave receiving signals at different positions, and the defect reflection echo time and direct wave time corresponding to each position are extracted respectively;

[0092] Based on the defect reflection echo time and direct wave time corresponding to each position, the real-time time difference between the current defect reflection echo and the direct wave is calculated and compared with the reference time difference until the time difference between the defect reflection echo and the direct wave is stable and equal to the reference time difference, and the spatial position of the receiving transducer is synchronously recorded.

[0093] In this embodiment, the time difference between the extracted defect reflection echo and direct wave time information is first calculated, serving as a fixed reference time difference. This reference time difference represents the relative propagation delay between the defect wave and the direct wave under a specific spatial layout and serves as the target parameter for subsequent matching adjustments. Subsequently, while maintaining the position of the excitation transducer, the receiving transducer is translated along the surface of the solid plate at regular steps. After each position adjustment, the system collects the complete Lamb wave received signal at the current position and re-extracts the direct wave arrival time and the defect reflection wave arrival time under the current conditions, calculating the real-time time difference between the two. By sequentially comparing this real-time time difference with a preset reference time difference, the system determines whether the current transducer position meets the time difference constant condition. When the real-time time difference and the reference value are consistently consistent within a tolerance range, the receiving position is considered to meet the propagation path equivalence requirement. The coordinates of this position are then recorded, and the defect echo signal corresponding to this position is used as the amplitude sampling result under standard conditions for subsequent quantitative comparison with a standard defect amplitude model. This implementation not only establishes an amplitude acquisition mechanism under time-difference control but also physically reconstructs equivalent propagation path conditions through spatial displacement, avoiding amplitude drift caused by inconsistent excitation paths. Unlike traditional methods that directly sample fixed-point waveforms for amplitude comparison while ignoring the limitations of propagation geometry, the introduction of a spatial adjustment strategy using time-difference matching control ensures that defect reflection signals are acquired under highly consistent propagation conditions, thereby improving the accuracy and repeatability of quantitative defect size assessment.

[0094] Furthermore, the quantitative size information of the defect is obtained by the following steps:

[0095] Based on the spatial position of the transducer, the Lamb wave receiving signal at the position is collected, and the defect reflection echo signal is extracted in the window corresponding to the fixed time difference;

[0096] Perform amplitude envelope analysis on the extracted defect reflection echo signal to obtain the peak amplitude of the echo signal as the reflection intensity characteristic value of the current defect;

[0097] On a standard test block with artificial reflectors of known sizes, equidistant propagation paths and the same excitation frequency conditions are set up. The echo signals of each reflector of known size are collected, and the corresponding envelope peak amplitude is extracted respectively to construct a corresponding relationship model between the reflection amplitude and the defect size.

[0098] The amplitude characteristic value of the defect reflection echo is input into the corresponding relationship model, matched with the corresponding defect size, and used as the quantitative size information of the detected defect.

[0099] In this embodiment, the Lamb wave received signal at the receiving transducer's spatial position, recorded under the aforementioned constant time difference conditions, is first acquired. Based on the reference time corresponding to the direct wave, the system intercepts a window interval defined by the fixed time difference and extracts the defect reflection echo signal from it. To enhance the amplitude discriminability of the echo, the extracted signal undergoes amplitude envelope processing to obtain its main envelope curve. The envelope peak is used as the intensity characteristic value of the current defect reflection for subsequent amplitude-size mapping analysis. To establish a physically meaningful amplitude-size relationship, echo calibration is performed using a standard comparison test block during the experimental preparation phase. This standard test block is pre-installed with multiple artificial reflectors of varying sizes to simulate typical defect types. Echoes are acquired from each reflector under transducer placement and excitation frequency consistent with the actual inspection path. Envelope analysis is also performed on each of the acquired standard echo signals, extracting the amplitude peak corresponding to each known defect. A set of "known size-envelope amplitude" data pairs is constructed and fitted using interpolation or regression algorithms to form a continuous mapping model. The defect reflection echo amplitude characteristic value obtained in the actual detection is used as the model input, and the corresponding defect size is obtained through function matching to achieve quantitative inversion of the defect geometric characteristics. This method avoids the problem that the traditional amplitude comparison process relies on empirical judgment or a single threshold strategy, and is instead based on physical equivalent propagation conditions and standard reference curves, thereby improving the accuracy and stability of size identification. This embodiment ensures the consistency of the propagation path by keeping the time difference constant, thereby achieving normalized acquisition of the echo amplitude, and establishing an accurate reflection amplitude-size mapping model in combination with the standard test block, thereby breaking through the technical bottleneck of the traditional method that the defect amplitude is easily affected by factors such as path changes and coupling instability, resulting in large size estimation errors, and significantly improving the engineering applicability and measurement credibility of the defect size assessment capability based on Lamb waves.

[0100] Furthermore, the construction of the corresponding relationship model between the reflection amplitude and the defect size includes the following steps:

[0101] Several artificial reflectors of different sizes are preset on the standard comparison test block, corresponding to different reflective target areas;

[0102] For each artificial reflector position, arrange the transducer excitation and receiving configuration consistent with the actual detection path to stimulate Lamb waves and collect the corresponding received signals;

[0103] Extract the reflected echo signal corresponding to each artificial reflector, calculate its amplitude envelope peak, and establish a data pair containing the one-to-one correspondence between the defect size and the echo amplitude;

[0104] An interpolation fitting process is performed on the data pair to construct a continuous function between the reflected echo amplitude and the defect size as a corresponding relationship model.

[0105] In this embodiment, to achieve a quantitative mapping between defect echo amplitude and actual geometric dimensions, a standard comparison test block, made of the same material as the solid plate being tested, was first used as the modeling foundation. Multiple artificial reflectors of known dimensions were then pre-set at different locations. These typically consisted of blind holes or etched grooves of varying diameters, corresponding to the reflective target areas simulating defects of varying scales. Each artificial reflector constituted a standard reflection unit, simulating the reflection behavior of Lamb waves under conditions of varying physical scales. Subsequently, for each artificial reflector, a transducer arrangement and excitation parameters consistent with the actual testing process were employed, ensuring that the excitation frequency, excitation angle, and propagation path remained consistent with subsequent field measurements. While maintaining stable propagation conditions, Lamb waves were excited and the reflected echo signals generated at each artificial reflector were collected. The received echo signals were subjected to amplitude envelope processing, and their main envelope peaks were extracted as representative amplitude characteristics for the reflector. Ultimately, a set of data pairs consisting of defect size and corresponding reflected echo amplitudes was obtained, constructing a sample set of reflection responses for multi-scale sampling. In order to establish a universal continuous function model, interpolation or fitting processing is performed on the obtained discrete data pairs, including the use of polynomial fitting, spline interpolation or exponential function fitting, to obtain a continuous mapping relationship between the reflection amplitude and the defect size. This function model can be used as the input basis for the subsequent measured defect amplitude characteristic values ​​to achieve rapid inversion and quantitative output of the actual defect size. Unlike existing methods that rely on empirical estimation or relative amplitude change trends to determine the size of defects, this embodiment constructs a physical quantitative mapping mechanism based on standard test blocks to extract amplitude responses under strictly unified propagation paths and excitation conditions, thereby ensuring the accuracy, repeatability and physical interpretability of the model. Especially when faced with the need to identify multi-size defects in engineering scenarios, this model provides a highly self-consistent and verifiable echo-size correlation structure.

[0106] The above embodiments are merely descriptions of preferred embodiments of the present invention and are not intended to limit the scope of the present invention. Without departing from the design spirit of the present invention, various modifications and improvements made to the technical solutions of the present invention by ordinary engineering technicians in this field should fall within the scope of protection determined by the claims of the present invention.

Claims

1. A solid plate defect detection method based on Lamb wave end face echo, characterized in that: include: S1. On one side of a solid plate, a single-mode Lamb wave is excited by a fixed-angle angled piezoelectric transducer, and the Lamb wave signal is received by another fixed-angle angled piezoelectric transducer arranged on the same side as the first one, to obtain a direct wave signal and a side echo signal. S2. Extracting a defect reflection echo signal caused by a defect through signals received between the direct wave signal and the side echo signal, and obtaining time information and amplitude information of the defect reflection echo; S3. Calculate the defect position based on the time difference between the defect reflection echo time information and the side echo time information, combined with the pre-calibrated Lamb wave propagation velocity; S4. By adjusting the transducer position to keep the time difference between the defect reflection echo and the direct wave at a fixed value, the defect reflection echo amplitude is synchronously obtained and compared with the reference echo amplitude of the preset size defect comparison block to obtain the quantitative size information of the defect.

2. The method for detecting defects in a solid plate based on Lamb wave end face echo according to claim 1, characterized in that: The direct wave signal is obtained by the following steps: An excitation and receiving wedge piezoelectric transducer is arranged in the defect-free area of ​​the solid plate. Based on the set excitation frequency and excitation angle, the excitation transducer is triggered to emit a single-mode Lamb wave signal. The receiving transducer receives the Lamb wave signal that propagates directly along the plate surface between the excitation point and the receiving point, and extracts the first high-amplitude pulse signal that arrives in the waveform from the Lamb wave receiving signal as the direct wave signal; The direct wave signal is subjected to time window constraint and amplitude filtering to determine the time domain characteristics of the direct wave.

3. The method for detecting defects in a solid plate based on Lamb wave end face echo according to claim 2, characterized in that: The high-amplitude pulse signal is obtained by the following steps: A time window is set in the Lamb wave received signal, all data points in the time window are scanned, the absolute values ​​of their amplitudes are calculated, and high-amplitude pulse signals whose amplitudes reach local maximum values ​​are identified.

4. The method for detecting defects in a solid plate based on Lamb wave end face echo according to claim 3, characterized in that: The side echo signal is obtained by the following steps: In the Lamb wave received signal, based on the identified direct wave signal, a backward search is performed for a next-level high-amplitude pulse signal within a time window, and a reflection signal is extracted in combination with an expected propagation time interval corresponding to the distance from the plate end face as a side echo signal; The side echo signal is subjected to time window constraint and amplitude filtering to determine the side echo time information.

5. The method for detecting defects in a solid plate based on Lamb wave end face echo according to claim 1, characterized in that: The S2 comprises the following steps: Based on the direct wave time information and the side echo time information, the target time interval between the two is determined, and the Lamb wave received signal is intercepted within the time interval to obtain the candidate echo signal segment; Based on the candidate echo signal segments, performing amplitude peak extraction processing, identifying local extreme points above the background noise threshold, and obtaining several suspected defect reflection echo points; Based on the time position and amplitude change characteristics of the suspected defect reflection echo point, envelope demodulation and time window filtering are performed to obtain preliminary characteristics of the defect reflection echo; Based on the preliminary characteristics of the defect reflection echo, a joint time-frequency domain analysis is performed to extract the effective time domain center value and amplitude envelope peak of the echo signal as the final time information and amplitude information of the defect reflection echo.

6. The method for detecting defects in a solid plate based on Lamb wave end face echo according to claim 5, characterized in that: The time-frequency domain joint analysis includes: Based on the preliminary characteristic signal of the defect reflection echo, short-time Fourier transform processing is performed to obtain the frequency energy distribution of the signal in the time domain and obtain the time-frequency spectrum matrix; Based on the time-frequency spectrum matrix, the time center value and the corresponding frequency component of the energy concentration area are extracted to obtain a preliminary time-frequency feature vector; The preliminary characteristic signal of the defect reflection echo is further decomposed by wavelet packet to obtain the energy distribution of each frequency band at several scales and obtain the sub-band energy vector; The time-frequency feature vector and the sub-band energy vector are concatenated in a time-aligned manner to generate a joint time-frequency feature vector; Based on the concentration of the time component in the joint time-frequency eigenvector, the final time information of the defect reflection echo is extracted; based on the frequency component with the largest energy amplitude in the joint eigenvector, the final amplitude information of the defect reflection echo is extracted.

7. The method for detecting defects in a solid plate based on Lamb wave end face echo according to claim 6, characterized in that: The S3 includes the following steps: Based on the time information of the defect reflection echo and the time information of the side echo, the time difference between the defect echo and the side echo is calculated; Based on the propagation velocity of the Lamb wave in the solid plate pre-calibrated by the direct wave and the side echo in the defect-free area, combined with the time difference, the equivalent propagation distance from the defect to the transducer is calculated according to the propagation path relationship; Based on the current position of the transducer on the surface of the solid plate, the known spatial relationship between the transducer and the plate boundary, and the equivalent propagation distance, a geometric solution model of the defect position is established, and the two-dimensional position coordinates of the defect in the solid plate coordinate system are obtained by solution.

8. The method for detecting defects in a solid plate based on Lamb wave end face echo according to claim 7, characterized in that: The step of adjusting the transducer position so as to maintain the time difference between the defect reflection echo and the direct wave at a fixed value comprises: According to the time information of the defect reflection echo and the time information of the direct wave, the initial time difference between the two is calculated as a fixed reference time difference; Keeping the excitation transducer in a fixed position, the receiving transducer is translated along the surface of the solid plate to collect Lamb wave receiving signals at different positions, and the defect reflection echo time and direct wave time corresponding to each position are extracted respectively; Based on the defect reflection echo time and direct wave time corresponding to each position, the real-time time difference between the current defect reflection echo and the direct wave is calculated and compared with the reference time difference until the time difference between the defect reflection echo and the direct wave is stable and equal to the reference time difference, and the spatial position of the receiving transducer is synchronously recorded.

9. The method for detecting defects in a solid plate based on Lamb wave end face echo according to claim 8, characterized in that: The quantitative size information of the defect is obtained by the following steps: Based on the spatial position of the transducer, the Lamb wave receiving signal at the position is collected, and the defect reflection echo signal is extracted in the window corresponding to the fixed time difference; Perform amplitude envelope analysis on the extracted defect reflection echo signal to obtain the peak amplitude of the echo signal as the reflection intensity characteristic value of the current defect; On a standard test block with artificial reflectors of known sizes, equidistant propagation paths and the same excitation frequency conditions are set up. The echo signals of each reflector of known size are collected, and the corresponding envelope peak amplitude is extracted respectively to construct a corresponding relationship model between the reflection amplitude and the defect size. The amplitude characteristic value of the defect reflection echo is input into the corresponding relationship model, matched with the corresponding defect size, and used as the quantitative size information of the detected defect.

10. The method for detecting defects in a solid plate based on Lamb wave end face echo according to claim 9, characterized in that: The construction of the corresponding relationship model between the reflection amplitude and the defect size includes the following steps: Several artificial reflectors of different sizes are preset on the standard comparison test block, corresponding to different reflective target areas; For each artificial reflector position, arrange the transducer excitation and receiving configuration consistent with the actual detection path to excite Lamb waves and collect the corresponding received signals; Extract the reflected echo signal corresponding to each artificial reflector, calculate its amplitude envelope peak, and establish a data pair containing the one-to-one correspondence between the defect size and the echo amplitude; An interpolation fitting process is performed on the data pair to construct a continuous function between the reflected echo amplitude and the defect size as a corresponding relationship model.

Images