Post-welding welding spot structure integrity detection method and system

Abstract

The invention relates to the technical field of nondestructive testing, in particular to a postwelding welding spot structure integrity detection method and system.The method comprises the steps that acoustic emission signals of a welding spot in a specified time period are obtained and preprocessed, and preprocessed acoustic emission signals are obtained; the acoustic emission signal is controlled transient mechanical excitation applied to the position of a welding spot structure to which a welding spot which is welded and cooled and has a distance not greater than a first preset distance from the welding spot belongs, so that the welding spot generates an elastic response and is spread in the welding spot structure; the acoustic emission sensor is arranged at a position which is not greater than a second preset distance away from the welding spot; the specified time period is a preset time period from the moment when the controlled transient mechanical excitation is applied; acquiring characteristic parameters in the preprocessed acoustic emission signal based on the preprocessed acoustic emission signal; and judging the integrity of the welding spot structure based on the characteristic parameters in the preprocessed acoustic emission signal to obtain a judgment result.

Description

Technical Field

[0001] This application relates to the field of nondestructive testing technology, and in particular to a method and system for detecting the structural integrity of weld joints after welding. Background Technology

[0002] After welding, structural defects such as microcracks, incomplete welds, and weld detachment may occur within the weld joint due to factors such as cooling shrinkage, material mismatch, and process fluctuations. These defects are often difficult to identify through visual inspection or conventional electrical performance testing in the static state after welding. Therefore, the inspection of the structural integrity of weld joints after welding has gradually become a key technical requirement in high-reliability manufacturing scenarios. Existing methods for inspecting weld joint integrity mostly rely on visual inspection, X-ray inspection, or electrical performance testing. Among these, visual inspection can only identify surface defects and is difficult to reflect the internal structural state of the weld joint; X-ray inspection equipment is costly and has low inspection efficiency, and is not suitable for online or batch rapid inspection; while electrical performance testing often only reveals anomalies when the weld joint has already shown obvious failure, lacking sensitivity to early structural degradation. In addition, some methods based on passive acoustic emission monitoring usually rely on acoustic emission events that are naturally generated during the service of the weld joint. These signals are characterized by strong randomness and uncontrollable triggering conditions, resulting in uncertain timing and low signal-to-noise ratio of the acquired acoustic emission signals, making the judgment of weld joint structural integrity unreliable. Summary of the Invention

[0003] (a) Technical problems to be solved

[0004] In view of the above-mentioned shortcomings and deficiencies of the prior art, this application provides a method and system for detecting the structural integrity of weld joints after welding, which solves the technical problems in the prior art of making it difficult to obtain stable and effective detection information of weld joints under non-destructive conditions and making it difficult to reliably judge the structural integrity of weld joints.

[0005] (II) Technical Solution

[0006] To achieve the above objectives, the main technical solutions adopted in this application include:

[0007] In a first aspect, embodiments of this application provide a method for detecting the structural integrity of a weld joint after welding, comprising: S1, acquiring acoustic emission signals of the weld joint within a specified time period, and preprocessing the acoustic emission signals to obtain preprocessed acoustic emission signals; wherein, the acoustic emission signals are signals acquired by at least one acoustic emission sensor located at a position no greater than a second preset distance from the weld joint after welding is completed and cooled, where a controlled transient mechanical excitation is applied to the weld joint structure to which the weld joint belongs, causing the weld joint to generate an elastic response and propagate within the weld joint structure to which the weld joint belongs; the specified time period is a preset time period starting from the moment the controlled transient mechanical excitation is applied;

[0008] S2. Based on the preprocessed acoustic emission signal, obtain the characteristic parameters in the preprocessed acoustic emission signal; S3. Based on the characteristic parameters in the preprocessed acoustic emission signal, judge the integrity of the solder joint structure to which the solder joint belongs, and obtain the judgment result.

[0009] Preferably, in some embodiments of this application, the preprocessing of the acoustic emission signal to obtain a preprocessed acoustic emission signal includes: performing bandpass filtering on the acoustic emission signal collected by each acoustic emission sensor to obtain a filtered acoustic emission signal; performing denoising on the filtered acoustic emission signal to obtain a denoised acoustic emission signal, wherein the denoising includes threshold denoising, wavelet denoising, or adaptive denoising based on the statistical characteristics of background noise; and performing amplitude normalization or energy normalization on the denoised acoustic emission signal to obtain a normalized acoustic emission signal corresponding to the acoustic emission sensor; wherein the preprocessed acoustic emission signal includes: normalized acoustic emission signals corresponding to all acoustic emission sensors respectively.

[0010] Preferably, in some embodiments of this application, S2 specifically includes: for each acoustic emission sensor, obtaining characteristic parameters in the normalized acoustic emission signal corresponding to that acoustic emission sensor; wherein, the characteristic parameters in the preprocessed acoustic emission signal include characteristic parameters in the normalized acoustic emission signal corresponding to each acoustic emission sensor; the characteristic parameters include the normalized: cumulative energy of the acoustic emission signal, maximum instantaneous amplitude of the acoustic emission signal, root mean square amplitude of the acoustic emission signal, dominant frequency of the acoustic emission signal, spectral centroid of the acoustic emission signal, frequency band energy ratio of the acoustic emission signal within a preset frequency band, rise time of the acoustic emission signal, duration of the acoustic emission signal, energy decay rate of the acoustic emission signal, first occurrence time of high-frequency components in the acoustic emission signal, duration of high-frequency components in the acoustic emission signal, time-frequency energy concentration of the acoustic emission signal, and response stability coefficient of the acoustic emission signal.

[0011] Preferably, in some embodiments of this application, S3 specifically includes: when the acoustic emission signal is acquired by an acoustic emission sensor located at a position no greater than a second preset distance from the solder joint, based on the characteristic parameters in the normalized acoustic emission signal corresponding to the acoustic emission sensor, a first method is used to determine the integrity of the solder joint structure to which the solder joint belongs, and a determination result is obtained. Specifically, this includes: for each pre-specified first characteristic parameter in the characteristic parameters in the normalized acoustic emission signal corresponding to the acoustic emission sensor, comparing the pre-specified first characteristic parameter with its corresponding preset reference interval; when the pre-specified first characteristic parameter falls into its corresponding preset reference interval, determining that the solder joint structure to which the solder joint belongs is complete, and taking the integrity of the solder joint structure to which the solder joint belongs as the determination result; when at least one pre-specified first characteristic parameter does not fall into its corresponding preset reference interval, determining that the solder joint structure to which the solder joint belongs has a structural anomaly, and taking the structural anomaly of the solder joint to which the solder joint belongs as the determination result; the pre-specified first characteristic parameters include: the cumulative energy of the acoustic emission signal, the maximum instantaneous amplitude of the acoustic emission signal, and the root mean square amplitude of the acoustic emission signal.

[0012] Preferably, in some embodiments of this application, S3 specifically includes: when the acoustic emission signal is acquired by an acoustic emission sensor located at a distance of no more than a second preset distance from the solder joint, based on the feature parameters in the normalized acoustic emission signal corresponding to the acoustic emission sensor, the integrity of the solder joint structure to which the solder joint belongs is judged in a second manner to obtain a judgment result, specifically including: judging whether each pre-specified second feature parameter in the feature parameters in the normalized acoustic emission signal corresponding to the acoustic emission sensor is greater than its corresponding preset minimum threshold, and when each pre-specified second feature parameter is greater than or equal to its corresponding preset minimum threshold, a comprehensive score is obtained using formula (1);

[0013] The second characteristic parameter includes: the cumulative energy of the acoustic emission signal, the frequency band energy ratio of the acoustic emission signal in the preset frequency band, the energy attenuation rate of the acoustic emission signal, and the response stability coefficient of the acoustic emission signal; wherein, the formula (1) is: C=w1·E+w2·F+w3·D+w4·S;

[0014] C represents the comprehensive score; w1 is the first preset weight, w2 is the second preset weight, w3 is the third preset weight, and w4 is the fourth preset weight; where 1 = w1 + w2 + w3 + w4; E is the cumulative energy of the acoustic emission signal in the normalized acoustic emission signal, F is the proportion of the frequency band energy of the acoustic emission signal in the normalized acoustic emission signal within the preset frequency band, D is the energy attenuation rate of the acoustic emission signal in the normalized acoustic emission signal, and S is the response stability coefficient of the acoustic emission signal in the normalized acoustic emission signal; the comprehensive score C is compared with the preset comprehensive score threshold C. th A comparison is made, and when the comprehensive score C is greater than or equal to the preset comprehensive score threshold C... th In the case where the solder joint structure is determined to be complete, the completeness of the solder joint structure is taken as the judgment result; when the comprehensive score C is less than the preset comprehensive score threshold C th In such cases, it is determined that the structure of the weld point to which the weld point belongs has a structural abnormality, and the structural abnormality of the weld point to which the weld point belongs is taken as the judgment result.

[0015] Preferably, in some embodiments of this application, step S3 specifically includes: when the acoustic emission signal is acquired by an acoustic emission sensor located at a distance no greater than a second preset distance from the solder joint, based on the characteristic parameters in the normalized acoustic emission signal corresponding to the acoustic emission sensor, a third method is used to determine the integrity of the solder joint structure to which the solder joint belongs, and a determination result is obtained. Specifically, this includes: obtaining a characteristic parameter vector based on a pre-specified third characteristic parameter in the characteristic parameters in the normalized acoustic emission signal corresponding to the acoustic emission sensor; the third characteristic parameter includes: the cumulative energy of the acoustic emission signal, the maximum instantaneous amplitude of the acoustic emission signal, the root mean square amplitude of the acoustic emission signal, the dominant frequency of the acoustic emission signal, the spectral centroid of the acoustic emission signal, and the acoustic emission signal at a preset frequency. The system selects at least one of the following: frequency band energy percentage within the band, rise time of the acoustic emission signal, duration of the acoustic emission signal, energy decay rate of the acoustic emission signal, first appearance time of high-frequency components in the acoustic emission signal, duration of high-frequency components in the acoustic emission signal, time-frequency energy concentration of the acoustic emission signal, and response stability coefficient of the acoustic emission signal. The feature parameter vector is input into a pre-trained machine learning model to obtain a judgment result. The machine learning model is pre-trained using feature parameter vectors from historical time periods. Each feature parameter vector from a historical time period is labeled to indicate whether the solder joint structure is intact or defective. The machine learning model is a support vector machine, random forest model, or neural network model.

[0016] Preferably, in some embodiments of this application, the method further includes: S4, when the acoustic emission signal is acquired by multiple acoustic emission sensors located at a distance of no more than a second preset distance from the solder joint, based on the characteristic parameters in the normalized acoustic emission signal corresponding to each acoustic emission sensor, the integrity of the solder joint structure to which the solder joint belongs is judged using a first method, a second method, or a third method to obtain a judgment result corresponding to the acoustic emission sensor; based on the judgment results corresponding to each acoustic emission sensor, a final judgment result is obtained.

[0017] Preferably, in some embodiments of this application, the final judgment result is obtained based on the judgment results corresponding to each acoustic emission sensor, specifically including: obtaining a weighted total score using formula (2) based on the judgment results corresponding to each acoustic emission sensor; formula (2) is: ;in, For weighted total score; Let be the confidence weight corresponding to the i-th acoustic emission sensor; This represents the judgment result corresponding to the i-th acoustic emission sensor; For indicator functions, when A score of 1 indicates a complete solder joint structure, otherwise a score of 0 is used; the weighted total score is calculated as follows. With the preset score threshold T th Comparison, when the weighted total score Greater than or equal to the preset score threshold T th In the case of a solder joint, the solder joint structure to which the solder joint belongs is determined to be complete, and the completeness of the solder joint structure to which the solder joint belongs is taken as the final judgment result; when the weighted total score Less than the preset score threshold T th In such cases, it is determined that the weld joint structure to which the weld joint belongs has a structural abnormality, and the structural abnormality of the weld joint to which the weld joint belongs is taken as the final judgment result.

[0018] Preferably, in some embodiments of this application, the confidence weight corresponding to the i-th acoustic emission sensor is... It is calculated using formula (3); formula (3) is:

[0019] ;

[0020] in, Let be the signal-to-noise ratio of the acoustic emission signal collected by the i-th acoustic emission sensor, where the signal-to-noise ratio is a dimensionless quantity; The maximum value among the signal-to-noise ratios of each acoustic emission sensor involved in the judgment; Let be the accuracy rate of the i-th acoustic emission sensor in the historical weld joint structural integrity detection, where the accuracy rate is a dimensionless quantity. The maximum value among the judgment accuracy rates of each acoustic emission sensor involved in the judgment; The actual spatial distance between the i-th acoustic emission sensor and the solder joint; The second preset distance, and satisfying 0≤ < ;

[0021] N represents the number of acoustic emission sensors involved in determining the structural integrity of the weld joint. This is the first adjustment coefficient; This is the second adjustment coefficient; This is the third adjustment coefficient.

[0022] On the other hand, this application also provides a post-weld weld joint structural integrity detection system, comprising: a mechanical excitation device for applying controlled transient mechanical excitation to a location of the weld joint structure to which the weld joint belongs, at a distance not greater than a first preset distance after welding is completed and cooled; and at least one acoustic emission sensor, disposed at a location not greater than a second preset distance from the weld joint, for acquiring acoustic emission signals of the weld joint within a specified time period, and preprocessing the acoustic emission signals to obtain preprocessed acoustic emission signals; wherein the acoustic emission signal is: at the location of the weld joint structure to which the weld joint belongs, at a distance not greater than the first preset distance after welding is completed and cooled. A controlled transient mechanical excitation is applied to cause the solder joint to generate an elastic response that propagates within the solder joint structure to which the solder joint belongs, and is acquired by the acoustic emission sensor; the specified time period is a preset time period starting from the moment the controlled transient mechanical excitation is applied; a signal processing module is used to acquire characteristic parameters in the preprocessed acoustic emission signal based on the preprocessed acoustic emission signal; a judgment module is used to judge the integrity of the solder joint structure to which the solder joint belongs based on the characteristic parameters, and obtain a judgment result; the mechanical excitation device is a spring-mass impact device, a piezoelectric actuator, or an electromagnetic exciter; the system also includes a display device for displaying the characteristic parameters and the judgment result.

[0023] (III) Beneficial Effects

[0024] This application provides a method for detecting the structural integrity of weld joints after welding. After welding is completed and cooled, a controlled transient mechanical excitation is applied to the weld joint structure, causing an elastic response that propagates within the weld joint structure. This actively introduces a mechanical disturbance directly related to the structural state of the weld joint. Compared to detection methods that rely on spontaneous acoustic emission signals in a static state, this method can stably excite acoustic emission signals from the weld joint under controlled conditions, allowing the influence of internal structural differences on the elastic response characteristics to be reflected, providing an effective signal basis for subsequent detection. By limiting the acoustic emission signal to signals collected within a preset time period from the moment the controlled transient mechanical excitation is applied, interference from environmental noise and non-excitation-related signals can be effectively suppressed. This ensures that the collected acoustic emission signals form a clear correspondence with the excitation event in the time dimension, thereby improving the correlation between the acoustic emission signal and the structural state of the weld joint, and enhancing the stability and consistency of the detection results. Furthermore, by applying the controlled transient mechanical excitation at a location no greater than a first preset distance from the solder joint, and collecting the acoustic emission signal by at least one acoustic emission sensor positioned no greater than a second preset distance from the solder joint, it is possible to ensure that the excitation energy effectively acts on the solder joint area. Simultaneously, it reduces signal attenuation and distortion during propagation within the solder joint structure, making the collected acoustic emission signal more reflective of the true elastic response characteristics of the solder joint structure. Based on this, the collected acoustic emission signal is preprocessed to obtain its characteristic parameters. Then, the integrity of the solder joint structure is judged based on these characteristic parameters. This transforms the elastic response characteristics exhibited by the solder joint under controlled excitation into quantitative information that can be used for analysis and judgment, thereby achieving non-destructive testing of the structural integrity of the solder joint after welding. Thus, the judgment of the structural integrity of the solder joint no longer relies on a single appearance or electrical performance indicator, but is based on the mechanical response characteristics of the solder joint structure itself, improving the reliability of post-weld solder joint structural integrity testing. Attached Figure Description

[0025] Figure 1 This is a flowchart illustrating a method for detecting the structural integrity of weld joints after welding, according to an embodiment of this application.

[0026] Figure 2 This is a schematic diagram of a post-weld weld joint structural integrity detection system according to an embodiment of this application. Detailed Implementation

[0027] To better explain and facilitate understanding of this application, the following detailed description of the application is provided in conjunction with the accompanying drawings and specific embodiments.

[0028] In existing technologies, post-weld weld joint structural integrity testing methods can be mainly categorized as follows: The first category is traditional testing schemes based on appearance or electrical performance. These schemes directly measure the surface morphology, macroscopic dimensions, or electrical performance indicators of the weld joint and compare the results with preset standards or historical data to determine the presence of defects. However, this scheme only reflects the surface state or overall conductivity of the weld joint. For weld joints with microcracks, metallurgical defects, or high residual stress that have not significantly changed macroscopic morphology, its judgment capability is limited, and it is difficult to obtain repeatable and reliable structural integrity information in the static state after welding and cooling. The second category is testing schemes based on passive acoustic emission signal monitoring. This scheme utilizes the transient acoustic emission signals spontaneously generated by the weld joint during service or loading to determine the structural state of the weld joint through feature extraction and analysis. Although it can reflect internal defects under certain conditions, the weld joint usually does not generate sufficiently strong spontaneous acoustic emission events in the static state after cooling, resulting in high randomness and low signal-to-noise ratio in the collected signals, making it difficult to form a stable and reliable judgment basis.

[0029] To overcome the above problems, the post-weld joint structural integrity detection method provided in this application applies a controlled transient mechanical excitation to the structural location of the weld joint after welding and cooling, actively stimulating the elastic response of the weld joint, and collecting acoustic emission signals near the weld joint within a preset time period. Simultaneously, the collected acoustic emission signals are subjected to bandpass filtering, threshold or wavelet denoising, and amplitude / energy normalization to obtain a stable and consistent signal data basis, from which feature parameters reflecting the structural state of the weld joint are extracted. Through the above processing, the method of this application can transform the elastic response of the weld joint under controlled excitation into quantitative feature information that can be used to judge the structural integrity of the weld joint, thereby achieving non-destructive, stable, consistent, and repeatable integrity detection of post-weld joints. This method not only improves detection accuracy and reliability but is also applicable to industrial detection scenarios under different weld joint arrangements and sensor configurations, achieving high versatility and operability in post-weld joint integrity detection. To better understand the above technical solution, exemplary embodiments of this application will be described in more detail below with reference to the accompanying drawings. Figure 1 This is a flowchart illustrating a method for detecting the structural integrity of weld joints after welding, according to an embodiment of this application. Figure 1 As shown, the method for detecting the structural integrity of weld joints after welding includes:

[0030] S1. Acquire the acoustic emission signal of the solder joint within a specified time period, and preprocess the acoustic emission signal to obtain a preprocessed acoustic emission signal; wherein, the acoustic emission signal is a signal acquired by at least one acoustic emission sensor located at a position no greater than a second preset distance from the solder joint after welding is completed and cooled, where a controlled transient mechanical excitation is applied to the solder joint structure to which the solder joint belongs, causing the solder joint to generate an elastic response and propagate within the solder joint structure to which the solder joint belongs; the specified time period is a preset time period starting from the moment the controlled transient mechanical excitation is applied;

[0031] For example, taking a single tab weld in a power battery module as an example, this weld is formed by laser welding. After welding is completed and it cools naturally, its structural integrity needs to be non-destructively tested. First, at a distance of no more than a first preset distance from the weld, a controlled transient mechanical excitation device applies a transient mechanical excitation to the weld structure, causing the weld to generate an instantaneous elastic response without plastic deformation. This elastic response propagates in the structure of the weld in the form of an acoustic emission signal, which is simultaneously collected by multiple acoustic emission sensors arranged at a distance of no more than a second preset distance from the weld. During signal acquisition, only acoustic emission signals within a preset time period from the moment the controlled transient mechanical excitation is applied are collected, thereby ensuring that the collected signals correspond one-to-one with the excitation event and avoiding interference from irrelevant signals such as environmental vibration and equipment operating noise. Subsequently, the acoustic emission signals acquired by each acoustic emission sensor are bandpass filtered to retain the effective frequency band corresponding to the elastic response of the solder joint; then, the filtered signals are subjected to threshold denoising or wavelet denoising to reduce the influence of background noise on signal details; finally, the denoised signals are normalized by amplitude or energy to obtain the normalized acoustic emission signals corresponding to each acoustic emission sensor.

[0032] The process of preprocessing the acoustic emission signals to obtain preprocessed acoustic emission signals includes: performing bandpass filtering on the acoustic emission signals collected by each acoustic emission sensor to obtain filtered acoustic emission signals; performing denoising on the filtered acoustic emission signals to obtain denoised acoustic emission signals, wherein the denoising process includes threshold denoising, wavelet denoising, or adaptive denoising based on the statistical characteristics of background noise; and performing amplitude normalization or energy normalization on the denoised acoustic emission signals to obtain normalized acoustic emission signals corresponding to the acoustic emission sensor; wherein the preprocessed acoustic emission signals include: normalized acoustic emission signals corresponding to all acoustic emission sensors respectively.

[0033] S2. Based on the preprocessed acoustic emission signal, obtain the feature parameters in the preprocessed acoustic emission signal; S2 specifically includes: for each acoustic emission sensor, obtain the feature parameters in the normalized acoustic emission signal corresponding to that acoustic emission sensor; wherein, the feature parameters in the preprocessed acoustic emission signal include the feature parameters in the normalized acoustic emission signal corresponding to each acoustic emission sensor.

[0034] The characteristic parameters include, after normalization: the cumulative energy of the acoustic emission signal, the maximum instantaneous amplitude of the acoustic emission signal, the root mean square amplitude of the acoustic emission signal, the dominant frequency of the acoustic emission signal, the spectral centroid of the acoustic emission signal, the frequency band energy ratio of the acoustic emission signal within the preset frequency band, the rise time of the acoustic emission signal, the duration of the acoustic emission signal, the energy decay rate of the acoustic emission signal, the first occurrence time of the high-frequency components in the acoustic emission signal, the duration of the high-frequency components in the acoustic emission signal, the time-frequency energy concentration of the acoustic emission signal, and the response stability coefficient of the acoustic emission signal.

[0035] S3. Based on the characteristic parameters in the preprocessed acoustic emission signal, the integrity of the solder joint structure to which the solder joint belongs is judged, and the judgment result is obtained.

[0036] For example, for small solder joints in battery modules or automotive parts, the applied transient mechanical excitation can be achieved using a miniature impactor or piezoelectric vibrator, causing the solder joint to undergo minute elastic deformation and generate propagating acoustic emission waves. These waves propagate along the solder joint and its associated structure, and are collected by sensors placed near the solder joint, thus obtaining acoustic emission signals closely related to the structural state of the solder joint. By limiting signal acquisition to a preset time period after the excitation is applied, interference from environmental noise or unrelated mechanical vibrations can be effectively eliminated, giving the acquired signals clear temporal correspondence and high correlation. In the signal preprocessing stage, the raw signals collected by each acoustic emission sensor are first bandpass filtered to remove unwanted low-frequency or high-frequency interference components during the propagation of the solder joint structure. Subsequently, the filtered signals are further processed using threshold denoising, wavelet denoising, or adaptive denoising methods based on the statistical characteristics of background noise to enhance the signal-to-noise ratio, enabling the effective capture of features generated by minute defects in the solder joint response. Finally, through amplitude or energy normalization, the signals from each sensor are unified to the same standard scale, ensuring the comparability of signals collected by different sensors in subsequent feature extraction and comparison. For example, in weld joints with slight microcracks or poor local bonding, the local response energy of the crack region in the normalized acoustic emission signal often exhibits abnormal cumulative energy curves and peak amplitude characteristics, which may be masked by noise in the unprocessed raw signal.

[0037] In the feature parameter extraction stage, the method of this application performs a detailed analysis of the normalized signal corresponding to each acoustic emission sensor to obtain multi-dimensional feature parameters, including cumulative energy, maximum instantaneous amplitude, root mean square amplitude, dominant frequency, spectral centroid, energy proportion of preset frequency band, rise time, duration, energy decay rate, first appearance time and duration of high-frequency components, time-frequency energy concentration, and response stability coefficient. By comprehensively analyzing these feature parameters, the response modes of normal and defective internal structures of the solder joint can be clearly distinguished. For example, in solder joints with microcracks, high-frequency components typically appear earlier and last longer, while the response stability coefficient is lower, indicating that the elastic response of the solder joint under mechanical excitation is uneven. Conversely, the acoustic emission signal feature parameters of intact solder joints show high consistency across sensors, with high time-frequency energy concentration, reflecting good overall stiffness and integrity of the solder joint structure.

[0038] Based on the above multidimensional feature parameters, the method of this application can reliably determine the integrity of the structure to which the weld point belongs, and realize non-destructive, quantitative and repeatable detection of the weld point after welding.

[0039] In some embodiments of this application, taking a single weld point in a large welded structure as an example, multiple acoustic emission sensors, such as a first acoustic emission sensor, a second acoustic emission sensor, and a third acoustic emission sensor, are set at positions no greater than a second preset distance from the weld point along different directions and propagation paths around the weld point. Each acoustic emission sensor is used to collect the acoustic emission signal generated by the weld point under controlled transient mechanical excitation and propagating in the weld point structure. When only one acoustic emission sensor is set, the collected acoustic emission signal mainly reflects the elastic response characteristics of the weld point along the direction of the sensor, which is easily affected by the geometry of the weld point, structural boundary conditions, and differences in local propagation paths. By setting at least one acoustic emission sensor and simultaneously collecting the elastic response of the weld point in multiple directions, acoustic emission response information of the weld point on different propagation paths can be obtained, thereby avoiding misjudgment caused by anomalies in a single propagation path.

[0040] In the method of this application, the same preprocessing procedure is applied to the acoustic emission signals collected by each acoustic emission sensor, including bandpass filtering, noise reduction, and amplitude or energy normalization, so that the acoustic emission signals corresponding to different acoustic emission sensors are comparable in scale and energy level. Based on this, characteristic parameters are extracted from the normalized acoustic emission signals corresponding to each acoustic emission sensor, including cumulative energy, maximum instantaneous amplitude, root mean square amplitude, dominant frequency, spectral centroid, frequency band energy ratio, rise time, duration, energy decay rate, first appearance time of high-frequency components, duration of high-frequency components, time-frequency energy concentration, and response stability coefficient. For example, when the solder joint structure is intact, the acoustic emission signals collected by different acoustic emission sensors show high consistency in the above characteristic parameters. The cumulative energy change trends of each sensor are similar, the first appearance time of high-frequency components is basically synchronized, the energy decay rate difference is small, and the response stability coefficients are all at a high level, indicating that the elastic response of the solder joint in all propagation directions is uniform and continuous. Conversely, when microcracks or poor local bonding exist within the solder joint, the defects directionally affect the propagation path and energy distribution of acoustic emission waves. Consequently, the characteristic parameters corresponding to different acoustic emission sensors exhibit significant differences. For example, high-frequency components in the acoustic emission signals of some sensors may appear earlier or last longer, resulting in uneven cumulative energy distribution and abrupt changes in energy decay rates. This leads to a decrease in the consistency of response stability coefficients among the sensors. Based on the consistency or differences in the characteristic parameters of multiple acoustic emission sensors, the method in this application can cross-verify the structural state of the solder joint from multiple propagation paths. This allows the determination of the solder joint's structural integrity to no longer rely on the local information of a single sensor, but rather on the comprehensive analysis results of multi-source acoustic emission signals, thereby improving the reliability, anti-interference ability, and judgment stability of post-weld solder joint structural integrity detection.

[0041] In some embodiments of this application, S3 specifically includes: when the acoustic emission signal is acquired by an acoustic emission sensor located at a distance of no more than a second preset distance from the solder joint, based on the characteristic parameters in the normalized acoustic emission signal corresponding to the acoustic emission sensor, judging the integrity of the solder joint structure to which the solder joint belongs in a first manner to obtain a judgment result, specifically including: for each pre-specified first characteristic parameter in the characteristic parameters in the normalized acoustic emission signal corresponding to the acoustic emission sensor, comparing the pre-specified first characteristic parameter with its corresponding preset reference interval;

[0042] When the pre-specified first feature parameter falls into its corresponding preset reference range, the solder joint structure to which the solder joint belongs is determined to be complete, and the completeness of the solder joint structure to which the solder joint belongs is taken as the judgment result;

[0043] When at least one pre-specified first characteristic parameter does not fall into its corresponding preset reference range, it is determined that the solder joint structure to which the solder joint belongs has a structural abnormality, and the structural abnormality of the solder joint to which the solder joint belongs is taken as the judgment result.

[0044] The pre-specified first characteristic parameter includes: the cumulative energy of the acoustic emission signal, the maximum instantaneous amplitude of the acoustic emission signal, and the root mean square amplitude of the acoustic emission signal.

[0045] For example, when the acoustic emission signal is acquired by an acoustic emission sensor located at a distance no greater than a second preset distance from the solder joint, the integrity judgment of the solder joint structure to which the solder joint belongs adopts the first method. Taking a single solder joint after welding and cooling as an example, after applying a controlled transient mechanical excitation to the solder joint structure to which the solder joint belongs, the acoustic emission sensor acquires the corresponding normalized acoustic emission signal within the preset time period. In this embodiment, a pre-specified first characteristic parameter is first extracted from the normalized acoustic emission signal corresponding to the acoustic emission sensor, including the cumulative energy of the acoustic emission signal, the maximum instantaneous amplitude of the acoustic emission signal, and the root mean square amplitude of the acoustic emission signal. The above-mentioned first characteristic parameters are used to characterize the overall energy release level, instantaneous response intensity, and average energy characteristics of the solder joint under controlled excitation, respectively, and can reflect the elastic response capability of the solder joint structure to mechanical excitation from different perspectives. Subsequently, for each of the first characteristic parameters, it is compared with a pre-established preset reference interval corresponding to that characteristic parameter. The preset reference interval can be determined based on historical detection data or calibration experimental data of structurally intact weld joints under the same excitation conditions. It describes the reasonable range of values ​​for this characteristic parameter when the weld joint is structurally intact. For example, when the weld joint is structurally intact, under controlled transient mechanical excitation, the cumulative energy of its acoustic emission signal is usually within a stable range, neither significantly lower due to rapid energy dissipation nor abnormally higher due to local structural anomalies. Simultaneously, its maximum instantaneous amplitude and root mean square amplitude will also fall within the preset reference interval corresponding to the intact weld joint, indicating that the weld joint has uniform and continuous elastic characteristics throughout the response process. In this case, when the cumulative energy, maximum instantaneous amplitude, and root mean square amplitude of the acoustic emission signal all fall within their respective preset reference intervals, the weld joint structure is determined to be intact, and the integrity of the weld joint structure is taken as the judgment result. Conversely, when there are structural anomalies such as incomplete welding, incomplete fusion, or microcracks inside the weld joint, its acoustic emission response under controlled transient mechanical excitation will change significantly. For example, local structural discontinuities can lead to abnormal acoustic emission energy distribution, causing the accumulated energy to deviate from the normal range; sudden changes in structural stiffness can cause abnormal increases or decreases in the maximum instantaneous amplitude; and non-uniform response processes can cause the root mean square amplitude to deviate from the preset reference range. In such cases, as long as at least one of the pre-specified first characteristic parameters does not fall within its corresponding preset reference range, it can be determined that the weld joint structure to which the weld joint belongs has a structural anomaly, and the structural anomaly of the weld joint to which the weld joint belongs is taken as the judgment result.

[0046] As can be seen from the above embodiments, this application, with only one acoustic emission sensor, can effectively determine the structural integrity of weld joints based on the comparison between the cumulative energy, maximum instantaneous amplitude, and root mean square amplitude of the acoustic emission signal and a preset reference interval. This determination method does not rely on complex models or multi-sensor collaboration; the judgment logic is clear and simple to implement, making it suitable for detection scenarios with limited space or sensor placement conditions. Furthermore, by introducing cross-constraints of multiple first characteristic parameters, misjudgments caused by fluctuations in a single characteristic parameter can be avoided, thereby improving the reliability and stability of post-weld weld joint structural integrity detection.

[0047] In some embodiments of this application, step S3 specifically includes: when the acoustic emission signal is acquired by an acoustic emission sensor located at a position no greater than a second preset distance from the solder joint, based on the characteristic parameters in the normalized acoustic emission signal corresponding to the acoustic emission sensor, judging the integrity of the solder joint structure to which the solder joint belongs using a second method to obtain a judgment result, specifically including:

[0048] Determine whether each pre-specified second feature parameter in the normalized acoustic emission signal corresponding to the acoustic emission sensor is greater than its corresponding preset minimum threshold, and if each pre-specified second feature parameter is greater than or equal to its corresponding preset minimum threshold, use formula (1) to obtain the comprehensive score.

[0049] The second characteristic parameter includes: the cumulative energy of the acoustic emission signal, the frequency band energy ratio of the acoustic emission signal in the preset frequency band, the energy attenuation rate of the acoustic emission signal, and the response stability coefficient of the acoustic emission signal; wherein, the formula (1) is: C=w1·E+w2·F+w3·D+w4·S;

[0050] C represents the overall score; w1 represents the first preset weight, w2 represents the second preset weight, w3 represents the third preset weight, and w4 represents the fourth preset weight; where 1 = w1 + w2 + w3 + w4; E represents the cumulative energy of the acoustic emission signal in the normalized acoustic emission signal, F represents the proportion of the frequency band energy of the acoustic emission signal in the preset frequency band in the normalized acoustic emission signal, D represents the energy attenuation rate of the acoustic emission signal in the normalized acoustic emission signal, and S represents the response stability coefficient of the acoustic emission signal in the normalized acoustic emission signal.

[0051] Compare the overall score C with the preset overall score threshold C th A comparison is made, and when the comprehensive score C is greater than or equal to the preset comprehensive score threshold C... th In the case of a solder joint, the solder joint structure to which the solder joint belongs is determined to be complete, and the completeness of the solder joint structure to which the solder joint belongs is taken as the judgment result;

[0052] When the comprehensive score C is less than the preset comprehensive score threshold C th In such cases, it is determined that the structure of the weld point to which the weld point belongs has a structural abnormality, and the structural abnormality of the weld point to which the weld point belongs is taken as the judgment result.

[0053] Specifically, after transient mechanical excitation of a weld point, an acoustic emission sensor is placed at a distance no greater than a second preset distance from the weld point to collect the acoustic emission signal of the elastic response of the weld point structure. Since the sensor is close to the weld point, the collected acoustic emission signal can more completely reflect the energy release characteristics, spectral distribution characteristics, attenuation and stability characteristics of the weld point and its surrounding base material under excited state. In this embodiment, the collected acoustic emission signal is first normalized to eliminate the influence of different excitation amplitudes and different installation conditions on the signal amplitude and absolute energy value. Subsequently, four second characteristic parameters are extracted from the normalized acoustic emission signal: cumulative energy, frequency band energy ratio within the preset frequency band, energy attenuation rate, and response stability coefficient. The second characteristic parameters are then judged to be greater than the corresponding preset minimum threshold. Only when all the second characteristic parameters reach the minimum effective discrimination condition is the weld point structure further evaluated using formula (1) to avoid misjudgment caused by invalid signals or low signal-to-noise ratio signals. In the above detection process, the cumulative energy E of the acoustic emission signal can reflect the overall energy response level of the weld joint under transient mechanical excitation. For example, for weld joints with intact structures and good metallurgical bonding, their internal continuity is high, the elastic wave propagation path is continuous, and the energy release of the acoustic emission signal is more complete, resulting in a higher normalized cumulative energy E. However, for weld joints with incomplete welds, incomplete penetration, or microcracks, the elastic wave is scattered and dissipated at the defects, leading to a significant decrease in cumulative energy. The frequency band energy proportion F of the acoustic emission signal within a preset frequency band is used to characterize the response characteristics of the weld joint structure to specific frequency components. For example, in an intact weld joint, the inherent characteristics of the structure cause the energy of the acoustic emission signal to be mainly concentrated in a preset frequency band that matches the modal response of the weld joint structure, resulting in a higher proportion of energy in this frequency band. However, when there are structural anomalies inside the weld joint, the spectral distribution of the acoustic emission signal is often more dispersed, with an increased proportion of energy in high-frequency or non-characteristic frequency bands, leading to a decrease in the proportion of energy within the preset frequency band. The energy decay rate D of the acoustic emission signal reflects the energy decay characteristics of the weld joint structure over time after being subjected to transient excitation. For example, in weld joints with intact structures and strong interface bonding, the acoustic emission signal attenuates relatively slowly during propagation within the structure, with the attenuation rate within a reasonable range. However, when cracks, pores, or poorly bonded areas exist in the weld joint, the acoustic emission energy dissipates rapidly during propagation, resulting in a significantly increased energy attenuation rate, thus reducing the normalized value of D. The response stability coefficient S of the acoustic emission signal is used to measure the waveform consistency and time stability of the acoustic emission signal throughout the entire response process. For example, in weld joints with intact structures, due to good material continuity and stable structural boundaries, the amplitude and spectral changes of the acoustic emission signal are small during multiple or continuous responses, resulting in a high stability coefficient. However, in weld joints with structural anomalies, local defects lead to increased random fluctuations during the response process, significantly reducing the stability coefficient.This application embodiment uses a comprehensive scoring method to evaluate the weld joint structure, enabling characteristic parameters with different physical meanings to participate in structural integrity judgment under a unified scale. For example, in a certain detection instance, even if the cumulative energy E of the weld joint is high, if its frequency band energy ratio F or response stability coefficient S is significantly low, the comprehensive score C after weighted summation may still be lower than the preset comprehensive score threshold C. th This allows for the determination of a structural anomaly at the weld joint. This method avoids the problem of a single characteristic parameter being "accidentally high" and masking potential structural defects.

[0054] By employing the judgment strategy described in the second approach, this embodiment no longer relies on a single acoustic emission feature to determine the structural integrity of the solder joint. Instead, it constructs a weighted comprehensive scoring model based on multiple second feature parameters with clear physical meanings to jointly evaluate the energy response characteristics, spectral concentration, energy attenuation behavior, and response stability of the solder joint. This multi-feature collaborative discrimination mechanism allows the detection results to simultaneously reflect both the overall continuity and local anomalies of the solder joint structure. Furthermore, comprehensive scoring is only performed when each second feature parameter meets a preset minimum threshold, effectively avoiding interference from low-quality acoustic emission signals, environmental noise, or ineffective excitation signals on the discrimination results, thus improving the reliability of solder joint structural integrity detection.

[0055] In some embodiments of this application, S3 specifically includes: when the acoustic emission signal is acquired by an acoustic emission sensor located at a position no greater than a second preset distance from the solder joint, based on the characteristic parameters in the normalized acoustic emission signal corresponding to the acoustic emission sensor, a third method is used to determine the integrity of the solder joint structure to which the solder joint belongs, and a determination result is obtained. Specifically, this includes: obtaining a characteristic parameter vector based on a pre-specified third characteristic parameter in the characteristic parameters in the normalized acoustic emission signal corresponding to the acoustic emission sensor; the third characteristic parameter includes at least one of the following: the cumulative energy of the acoustic emission signal, the maximum instantaneous amplitude of the acoustic emission signal, the root mean square amplitude of the acoustic emission signal, the dominant frequency of the acoustic emission signal, the spectral centroid of the acoustic emission signal, the frequency band energy ratio of the acoustic emission signal within a preset frequency band, the rise time of the acoustic emission signal, the duration of the acoustic emission signal, the energy decay rate of the acoustic emission signal, the first occurrence time of the high-frequency components in the acoustic emission signal, the duration of the high-frequency components in the acoustic emission signal, the time-frequency energy concentration of the acoustic emission signal, and the response stability coefficient of the acoustic emission signal.

[0056] The feature parameter vector is input into a pre-trained machine learning model to obtain a judgment result;

[0057] The machine learning model is trained by using feature parameter vectors from historical time periods. Each feature parameter vector from a historical time period is labeled to indicate whether the solder joint structure is intact or defective. The machine learning model is a support vector machine, a random forest model, or a neural network model.

[0058] For example, after a welded component is completed and cooled, a controlled transient mechanical excitation is applied to the target weld point. Only one acoustic emission sensor is placed at a distance no greater than a second preset distance from the weld point to collect the acoustic emission signal generated by the weld point structure under excitation. Because the sensor's placement is fixed and the distance is controlled, the collected acoustic emission signal exhibits good consistency in spatial attenuation and propagation path, making it suitable for constructing a feature vector-based discrimination model. In this embodiment, after preprocessing and normalizing the collected acoustic emission signal, multiple third feature parameters with different physical meanings and time-frequency characteristics are extracted from the normalized acoustic emission signal. These extracted third feature parameters are then combined into a feature parameter vector in a preset order. Subsequently, this feature parameter vector is input into a pre-trained machine learning model, which outputs a judgment result regarding the integrity of the weld point structure. In the above embodiment, the constructed feature parameter vector includes not only feature parameters reflecting energy and amplitude levels but also feature parameters reflecting time-domain, frequency-domain, and time-frequency joint characteristics. For example, the cumulative energy, maximum instantaneous amplitude, and root mean square amplitude of the acoustic emission signal are used to characterize the overall response intensity and local peak characteristics of the solder joint under transient excitation; the dominant frequency and spectral centroid are used to characterize the energy concentration trend and structural response characteristic frequencies of the acoustic emission signal in the frequency domain. Furthermore, the frequency band energy ratio within a preset frequency band, the first occurrence time of high-frequency components, and the duration of high-frequency components are used to reflect the spectral variation characteristics of the acoustic emission signal related to structural discontinuities and interface scattering; the rise time and duration are used to characterize the occurrence speed of the acoustic emission event and the timescale of the energy release process; while the energy decay rate, time-frequency energy concentration, and response stability coefficient are used to comprehensively reflect the attenuation characteristics of the solder joint structure to elastic wave propagation, the degree of energy distribution concentration, and the consistency of the response process. By combining at least one of the above third characteristic parameters into a characteristic parameter vector, this characteristic parameter vector can describe the state of the solder joint structure from multiple dimensions, rather than being limited to a single or a few acoustic emission features.

[0059] During the training phase of the machine learning model, a large amount of solder joint detection data from a historical time period is pre-selected as training samples. For each historical solder joint sample, acoustic emission signals are acquired under the same or similar detection conditions, and the corresponding third feature parameter vector is extracted. Simultaneously, based on subsequent destructive testing, microscopic inspection, or reliability testing results, each feature parameter vector is labeled as either "solder joint structure intact" or "solder joint structure defective." For example, when training a support vector machine model, training the labeled feature parameter vectors enables the model to learn the optimal separating hyperplane to distinguish between "intact solder joints" and "defective solder joints" in a high-dimensional feature space. When using a random forest model, multiple decision trees are used to discriminate different combinations of feature parameters, reducing the random influence of a single feature on the judgment result. When using a neural network model, multi-layer nonlinear mapping is used to learn the complex relationships between different feature parameters, thereby improving the ability to identify complex defect patterns. For example, in some solder joints, even if the cumulative energy and maximum amplitude of the acoustic emission signal are within the normal range, the presence of tiny unbonded areas within the joint causes the high-frequency components to appear earlier, the spectral center of gravity to shift, and the energy decay rate to increase abnormally. For discrimination methods based on a single threshold or linear weighting rules, such solder joints are easily misjudged as structurally intact. However, by inputting the aforementioned feature parameters into a machine learning model, the model can identify the difference between this solder joint and a "complete solder joint" in the feature space based on feature combination patterns learned from historical samples, thus providing a judgment of structural abnormality.

[0060] By employing a third approach to construct a feature parameter vector from multiple third feature parameters and introducing a pre-trained machine learning model to judge the structural integrity of the weld joint, the technical solution of this application can fully exploit the multi-dimensional information of the acoustic emission signal in the time domain, frequency domain, and time-frequency domain, achieving a comprehensive characterization of the weld joint's structural state. Compared with discrimination methods based on fixed thresholds or simple weighted rules, this method no longer relies on manually set discrimination boundaries, but automatically forms discrimination relationships between features through learning from historical samples, thus better adapting to the changes in the acoustic emission characteristics of weld joints under different welding processes, materials, and structural forms. Furthermore, by setting an acoustic emission sensor only at a location no greater than a second preset distance from the weld joint, combined with the discrimination method of feature vectors and machine learning models, the complexity of sensor placement and system cost are reduced while ensuring detection accuracy. This third approach is particularly suitable for detecting weld joints with complex structural state change patterns and diverse defect types, effectively improving the ability to identify early hidden defects and atypical structural anomalies.

[0061] In some embodiments of this application, the method further includes: S4, when the acoustic emission signal is acquired by multiple acoustic emission sensors located at a distance of no more than a second preset distance from the solder joint, based on the characteristic parameters in the normalized acoustic emission signal corresponding to each acoustic emission sensor, the integrity of the solder joint structure to which the solder joint belongs is judged using a first method, a second method, or a third method to obtain a judgment result corresponding to the acoustic emission sensor; based on the judgment results corresponding to each acoustic emission sensor, a final judgment result is obtained.

[0062] For example, in some embodiments of this application, to further improve the reliability of post-weld joint structural integrity detection, multiple acoustic emission sensors are set at a distance no greater than a second preset distance from the weld joint to collect acoustic emission signals generated by the weld joint structure under controlled transient mechanical excitation from different spatial locations. Since the propagation path, propagation direction, and local structural response characteristics of the acoustic emission signal in the weld joint structure differ, the acoustic emission signals collected by different acoustic emission sensors may differ in amplitude, spectral distribution, and time-frequency characteristics. In this case, the acoustic emission signal collected by each acoustic emission sensor is preprocessed, normalized, and its feature parameters are extracted. Based on the feature parameters in the normalized acoustic emission signal corresponding to each acoustic emission sensor, the integrity of the weld joint structure to which the weld joint belongs is judged using a first, second, or third method, thereby obtaining a judgment result corresponding to each acoustic emission sensor.

[0063] For example, in one embodiment, for a first acoustic emission sensor located near the solder joint, a first method is used to compare the cumulative energy, maximum instantaneous amplitude, and root mean square amplitude of the acoustic emission signal with a preset reference interval to obtain the judgment result corresponding to that acoustic emission sensor. For a second acoustic emission sensor located on the other side of the solder joint, a second method is used to calculate a comprehensive score based on a comprehensive scoring formula and compare it with a preset comprehensive scoring threshold to obtain the judgment result. For a third acoustic emission sensor located at the far end of the solder joint structure, a third method is used, inputting the extracted feature parameter vector into a pre-trained machine learning model to obtain the judgment result. By allowing different discrimination methods to be selected for different acoustic emission sensors, this method can flexibly adjust the judgment strategy according to sensor placement conditions, signal quality, and application scenarios, thereby improving the adaptability of the overall detection scheme. After obtaining the judgment results corresponding to each acoustic emission sensor, the final judgment result is obtained based on the judgment results. By fusing the judgment results of multiple acoustic emission sensors in the above manner, the final judgment result no longer depends on the acoustic emission signal characteristics under a single sensor or a single propagation path, but is based on the comprehensive judgment of multiple spatial sampling points, thereby effectively suppressing the risk of misjudgment caused by accidental noise, local structural differences or individual sensor installation deviations.

[0064] In this embodiment, by setting multiple acoustic emission sensors at a distance no greater than a second preset distance from the solder joint, and using a first, second, or third method to judge based on the characteristic parameters in the normalized acoustic emission signal corresponding to each acoustic emission sensor, the technical solution of this application can sample the elastic response of the solder joint structure from multiple spatial locations, realizing a multi-angle characterization of the solder joint structural integrity. Based on this, by fusing the judgment results corresponding to each acoustic emission sensor, a final judgment result is obtained, making the detection conclusion more consistent and stable. Compared with detection methods relying solely on a single acoustic emission sensor, this multi-sensor judgment mechanism can effectively reduce the probability of misjudgment introduced by differences in propagation paths, local structural non-uniformity, and occasional noise interference, significantly improving the reliability and repeatability of the post-soldering solder joint structural integrity detection results.

[0065] In this embodiment, the final judgment result is obtained based on the judgment results corresponding to each acoustic emission sensor. Specifically, this includes obtaining the weighted total score using formula (2) based on the judgment results corresponding to each acoustic emission sensor. Formula (2) is as follows: ;in, For weighted total score; Let be the confidence weight corresponding to the i-th acoustic emission sensor; This represents the judgment result corresponding to the i-th acoustic emission sensor; For indicator functions, when A score of 1 indicates a complete solder joint structure, otherwise a score of 0 is used; the weighted total score is calculated as follows. With the preset score threshold T th Comparison, when the weighted total score Greater than or equal to the preset score threshold T th In the case of a solder joint, the solder joint structure to which the solder joint belongs is determined to be complete, and the completeness of the solder joint structure to which the solder joint belongs is taken as the final judgment result; when the weighted total score Less than the preset score threshold T th In such cases, it is determined that the weld joint structure to which the weld joint belongs has a structural abnormality, and the structural abnormality of the weld joint to which the weld joint belongs is taken as the final judgment result.

[0066] For example, when the solder joint structure to which the solder joint belongs is subjected to controlled transient mechanical excitation, and multiple acoustic emission sensors located at a distance of no more than a second preset distance from the solder joint collect acoustic emission signals, the reliability of different acoustic emission sensors in judging the integrity of the solder joint structure is not completely the same due to differences in the spatial positional relationship between each acoustic emission sensor and the solder joint, the propagation path length, and the signal attenuation conditions. Based on the above objective situation, the embodiments of this application do not simply perform equal weighting on the judgment results of each acoustic emission sensor, but pre-set a confidence weight corresponding to its detection reliability for each acoustic emission sensor, and calculate the weighted total score using formula (2) based on the judgment results corresponding to each acoustic emission sensor, thereby obtaining the final judgment result. For example, in a certain embodiment, for the same solder joint structure, three acoustic emission sensors are set as the first acoustic emission sensor, the second acoustic emission sensor, and the third acoustic emission sensor. According to historical detection data, sensor installation position, and signal stability, confidence weights c1=0.4, c2=0.35, and c3=0.25 are set for the three, and the sum of the weights is 1. After inspecting the solder joint structure, the following judgment results were obtained: the first acoustic emission sensor determined that the solder joint structure was intact; the second acoustic emission sensor determined that the solder joint structure was intact; and the third acoustic emission sensor determined that the solder joint structure had an anomaly. In the above embodiment, a preset score threshold T was pre-set. h =0.6. When the weighted total score With the preset score threshold T h When making comparisons, because =0.75≥T h The system determines that the solder joint structure to which the solder joint belongs is complete, and takes the integrity of the solder joint structure to which the solder joint belongs as the final judgment result. Conversely, in another embodiment, if only the first acoustic emission sensor determines that the solder joint structure is complete, while the second and third acoustic emission sensors both determine that the solder joint structure is abnormal, then the corresponding weighted total score is: =0.4×1+0.35×0+0.25×0=0.4; when <T hWhen a structural anomaly is determined to exist in the weld joint structure to which the weld joint belongs, this structural anomaly is taken as the final judgment result. By employing a weighted total score decision method that incorporates confidence weights and indicator functions based on the judgment results corresponding to each acoustic emission sensor, the technical solution of this application can achieve quantitative fusion judgment of weld joint structural integrity in multi-acoustic emission sensor detection scenarios. This method not only considers the consistency of judgment results from different acoustic emission sensors but also fully reflects the differences in spatial location, signal quality, and detection reliability among the acoustic emission sensors. Compared with simple equal-weighted voting or single-sensor judgment methods, this weighted decision mechanism can effectively reduce the interference of occasional noise or local anomalies on the final judgment result, improving the stability, repeatability, and engineering reliability of the post-weld weld joint structural integrity detection results.

[0067] Specifically, the confidence weight corresponding to the i-th acoustic emission sensor It is calculated using formula (3); formula (3) is:

[0068] ;

[0069] in, The signal-to-noise ratio (SNR) of the acoustic emission signal acquired by the i-th acoustic emission sensor is a dimensionless quantity. The higher the SNR, the more prominent the effective acoustic emission signal is relative to the background noise, and the less the extracted feature parameters are affected by noise. Therefore, the reliability of the signal-to-noise ratio in judging the integrity of the weld joint structure is higher. The maximum value among the signal-to-noise ratios of each acoustic emission sensor involved in the judgment; Let be the accuracy rate of the i-th acoustic emission sensor in the historical weld joint structural integrity detection, where the accuracy rate is dimensionless; historical accuracy rate. This reflects the stability and reliability of the acoustic emission sensor in the long-term detection of weld joint structural integrity. This parameter is obtained by statistically analyzing the detection results with integrity / abnormal labels over a historical period, so that the confidence weight can dynamically reflect the long-term discrimination capability of the sensor as the actual detection performance changes. The maximum value among the judgment accuracy rates of each acoustic emission sensor involved in the judgment; The actual spatial distance between the i-th acoustic emission sensor and the solder joint; The second preset distance, and satisfying 0≤ < N represents the number of acoustic emission sensors involved in determining the structural integrity of the weld joint. This is the first adjustment coefficient; This is the second adjustment coefficient; This is the third adjustment coefficient. In this embodiment, it can be appropriately increased in a strong noise environment. This enhances the impact of signal-to-noise ratio factors; in scenarios with mass production and sufficient historical data, it can increase... This highlights the importance of historical accuracy; in scenarios with complex structures or significantly different propagation paths, it can increase... This is to enhance the ability of spatial distance factors to differentiate confidence weights.

[0070] For example, in a specific embodiment, for the post-weld joint structural integrity detection scenario in a general industrial production environment, the signal-to-noise ratio of the acoustic emission signal, the historical judgment accuracy of the acoustic emission sensor, and the spatial distance between the acoustic emission sensor and the weld joint all have a significant impact on the detection results, and their importance is similar. Therefore, in this embodiment, the first adjustment coefficient is... Second adjustment coefficient and the third adjustment coefficient All are set to 1, that is =1, =1, =1. Under the above value method, the three factors in formula (3) participate in the calculation of confidence weight in a linear and equal weight manner, so that the confidence weight can comprehensively reflect the comprehensive influence of acoustic emission signal quality, historical detection stability and spatial propagation conditions, and is suitable for conventional weld point structural integrity detection scenarios with moderate noise levels and relatively uniform sensor arrangement.

[0071] In another embodiment, strong environmental mechanical vibration or electromagnetic interference exists at the detection site, resulting in a high level of background noise in the acoustic emission signal. In this case, the signal-to-noise ratio (SNR) has a more critical impact on the reliability of feature parameter extraction. Therefore, in this embodiment, the first adjustment coefficient... Set to a value greater than 1, for example =2, while the second adjustment coefficient and the third adjustment coefficient Still set to 1, that is =1, =1. Through the above settings, the influence of the signal-to-noise ratio term in the weight calculation of formula (3) is amplified, so that the acoustic emission sensor that collects the high signal-to-noise ratio acoustic emission signal occupies a higher weight in the weighted fusion decision, thereby effectively suppressing the adverse effect of the low signal-to-noise ratio acoustic emission sensor on the final judgment result and improving the stability and reliability of the overall detection result in a strong noise environment.

[0072] In this embodiment, formula (3) is not a simple mathematical weighting, but rather an objective quantification of the reliability of multiple acoustic emission sensors by jointly modeling the acoustic emission signal quality, historical judgment reliability and spatial propagation conditions, thereby improving the stability, consistency and engineering reliability of the weld point structural integrity detection results after welding.

[0073] In this embodiment, the cumulative energy of the acoustic emission signal refers to the total energy obtained by integrating the square of the normalized acoustic emission signal amplitude over time within the specified time period. It is used to characterize the ability of the weld joint structure to release elastic response energy as a whole under controlled transient mechanical excitation.

[0074] The maximum instantaneous amplitude of the acoustic emission signal refers to the maximum amplitude reached by the normalized acoustic emission signal within the specified time period, reflecting the peak response intensity of the weld joint structure under transient excitation. The root mean square amplitude of the acoustic emission signal is the square root of the average of the squared amplitudes of the normalized acoustic emission signal within the specified time period, used to characterize the overall energy level and sustained response intensity of the acoustic emission signal. The dominant frequency of the acoustic emission signal refers to the frequency component with the largest energy proportion in the spectrum after spectral analysis of the normalized acoustic emission signal, used to reflect the main vibration mode characteristics of the weld joint structure. The spectral centroid of the acoustic emission signal refers to the average frequency position calculated by weighting each frequency component according to its energy in the spectrum of the normalized acoustic emission signal, used to reflect the overall distribution trend of acoustic emission energy in the frequency domain. The frequency band energy proportion of the acoustic emission signal within the preset frequency band refers to the ratio of the energy of the normalized acoustic emission signal within the preset frequency band to the total energy of the acoustic emission signal, used to characterize the contribution of the acoustic emission response to the overall response within a specific frequency range. The rise time of the acoustic emission signal refers to the time interval from when the normalized acoustic emission signal first exceeds a preset trigger threshold to when it reaches its maximum instantaneous amplitude, reflecting the response speed of the weld joint structure to transient excitation. The duration of the acoustic emission signal refers to the time from when the normalized acoustic emission signal first exceeds the preset trigger threshold until the signal amplitude decays below the trigger threshold, characterizing the sustained characteristics of the weld joint structure's vibration response. The energy decay rate of the acoustic emission signal refers to the rate at which the energy of the normalized acoustic emission signal decays over time within the specified time period, reflecting the energy dissipation characteristics within the weld joint structure. The first occurrence time of the high-frequency components in the acoustic emission signal refers to the time point at which frequency components above a preset high-frequency threshold first appear during time-frequency analysis of the normalized acoustic emission signal, reflecting the excitation timing characteristics of the high-frequency response in the weld joint structure. The duration of the high-frequency components in the acoustic emission signal refers to the length of time during which frequency components above a preset high-frequency threshold continuously exist in the normalized acoustic emission signal, characterizing the stability of the high-frequency response. The time-frequency energy concentration of an acoustic emission signal refers to the degree of concentration of the normalized acoustic emission signal energy in the main time-frequency region within the time-frequency domain, reflecting whether the acoustic emission energy exhibits a localized concentrated distribution characteristic. The response stability coefficient of an acoustic emission signal is a dimensionless parameter determined based on the degree of change of the characteristic parameters between adjacent sub-time windows after a controlled transient mechanical excitation is applied to the weld joint structure. This is achieved by dividing the normalized acoustic emission signal acquired by the same acoustic emission sensor into multiple consecutive sub-time windows within the specified time period, calculating at least one time-domain characteristic parameter and / or frequency-domain characteristic parameter of the acoustic emission signal within each sub-time window, and defining the dimensionless parameter.The degree of change is used to characterize whether the amplitude variation, energy distribution, and frequency composition of the acoustic emission signal exhibit continuous, smooth, and consistent characteristics during its time evolution. In one embodiment, the response stability coefficient is determined based on the degree of difference of the characteristic parameters in adjacent sub-time windows. When the degree of difference is small, the response stability coefficient indicates that the acoustic emission signal has high response stability; when the degree of difference is large, the response stability coefficient indicates that the acoustic emission signal has low response stability. The degree of change includes at least one of the absolute change, relative change, or statistical dispersion of the characteristic parameters between adjacent sub-time windows.

[0075] On the other hand, see Figure 2 This application also provides a post-weld weld joint structural integrity detection system, comprising: a mechanical excitation device for applying controlled transient mechanical excitation to a location of the weld joint structure to which the weld joint belongs, at a distance not greater than a first preset distance after welding is completed and cooled; and at least one acoustic emission sensor, disposed at a location not greater than a second preset distance from the weld joint, for acquiring acoustic emission signals of the weld joint within a specified time period and preprocessing the acoustic emission signals to obtain preprocessed acoustic emission signals; wherein the acoustic emission signal is: the signal obtained by applying controlled transient mechanical excitation to a location of the weld joint structure to which the weld joint belongs, at a distance not greater than the first preset distance after welding is completed and cooled. A controlled transient mechanical excitation causes the solder joint to produce an elastic response that propagates within the solder joint structure to which the solder joint belongs, and is acquired by the acoustic emission sensor; the specified time period is a preset time period from the moment the controlled transient mechanical excitation is applied; a signal processing module is used to acquire characteristic parameters in the preprocessed acoustic emission signal based on the preprocessed acoustic emission signal; a judgment module is used to judge the integrity of the solder joint structure to which the solder joint belongs based on the characteristic parameters, and obtain a judgment result; the mechanical excitation device is a spring-mass impact device, a piezoelectric actuator, or an electromagnetic exciter; the system also includes a display device for displaying the characteristic parameters and the judgment result.

[0076] For example, the post-weld joint structural integrity detection system provided in this application includes a mechanical excitation device, at least one acoustic emission sensor, a signal processing module, a judgment module, and a display device. The mechanical excitation device applies controlled transient mechanical excitation to the weld joint structure after welding and cooling. This excitation induces an elastic response in the weld joint and propagates within the weld joint structure, thereby generating characteristic acoustic emission signals from potential defects. The mechanical excitation device can be a spring-mass impact device, a piezoelectric actuator, or an electromagnetic exciter. These devices are characterized by high controllability, fast response, and good repeatability, ensuring consistent excitation conditions each time, thus guaranteeing the reliability and repeatability of the detection results. The acoustic emission sensor is positioned at a distance no greater than a second preset distance from the weld joint to collect the acoustic emission signal generated by the weld joint after mechanical excitation. The preprocessing module performs noise suppression, filtering, and amplification on the collected signal to obtain a high-quality preprocessed signal. The signal processing module further extracts characteristic parameters from the preprocessed signal, such as instantaneous amplitude, frequency components, energy distribution, and envelope characteristics. These parameters can reflect the presence of internal defects, microcracks, or poor bonding within the weld joint. In the judgment module, the integrity of the weld joint structure can be assessed by analyzing the extracted feature parameters, resulting in a judgment result. Compared to traditional manual visual inspection or simple static testing, this method can more sensitively detect minute defects and potential structural anomalies, improving the safety and reliability of weld joints. Simultaneously, the display device intuitively presents the feature parameters and judgment results to the operator, facilitating rapid assessment of weld joint quality and enabling the recording and traceability of welding quality.

[0077] Although embodiments of this application have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting this application. Those skilled in the art can make modifications, alterations, substitutions and variations to the above embodiments within the scope of this application.

Claims

1. A method for detecting the structural integrity of weld points after welding, characterized in that, include: S1. Acquire the acoustic emission signal of the solder joint within a specified time period, and preprocess the acoustic emission signal to obtain the preprocessed acoustic emission signal; The acoustic emission signal is a signal obtained by applying a controlled transient mechanical excitation at a position of the weld joint structure to which the weld joint belongs, at a distance of no more than a first preset distance from the weld joint after the welding is completed and cooled, so that the weld joint produces an elastic response and propagates in the weld joint structure to which the weld joint belongs, and is collected by at least one acoustic emission sensor located at a position of no more than a second preset distance from the weld joint. The specified time period is a preset time period starting from the moment the controlled transient mechanical excitation is applied; S2. Based on the preprocessed acoustic emission signal, obtain the characteristic parameters in the preprocessed acoustic emission signal; S3. Based on the characteristic parameters in the preprocessed acoustic emission signal, the integrity of the solder joint structure to which the solder joint belongs is judged, and the judgment result is obtained.

2. The method for detecting the structural integrity of weld points after welding according to claim 1, characterized in that, in, The acoustic emission signal is preprocessed to obtain a preprocessed acoustic emission signal, including: The acoustic emission signal collected by each acoustic emission sensor is bandpass filtered to obtain the filtered acoustic emission signal. The filtered acoustic emission signal is denoised to obtain a denoised acoustic emission signal. The denoising process includes threshold denoising, wavelet denoising, or adaptive denoising based on the statistical characteristics of background noise. The amplitude or energy of the denoised acoustic emission signal is normalized to obtain the normalized acoustic emission signal corresponding to the acoustic emission sensor. The preprocessed acoustic emission signals include: normalized acoustic emission signals corresponding to all acoustic emission sensors.

3. The method for detecting the structural integrity of weld points after welding according to claim 2, characterized in that, S2 specifically includes: For each acoustic emission sensor, the characteristic parameters in the normalized acoustic emission signal corresponding to that acoustic emission sensor are obtained. The characteristic parameters in the preprocessed acoustic emission signal include the characteristic parameters in the normalized acoustic emission signal corresponding to each acoustic emission sensor. The characteristic parameters include, after normalization: the cumulative energy of the acoustic emission signal, the maximum instantaneous amplitude of the acoustic emission signal, the root mean square amplitude of the acoustic emission signal, the dominant frequency of the acoustic emission signal, the spectral centroid of the acoustic emission signal, the frequency band energy ratio of the acoustic emission signal within the preset frequency band, the rise time of the acoustic emission signal, the duration of the acoustic emission signal, the energy decay rate of the acoustic emission signal, the first occurrence time of the high-frequency components in the acoustic emission signal, the duration of the high-frequency components in the acoustic emission signal, the time-frequency energy concentration of the acoustic emission signal, and the response stability coefficient of the acoustic emission signal.

4. The method for detecting the structural integrity of weld points after welding according to claim 3, characterized in that, S3 specifically includes: When the acoustic emission signal is acquired by an acoustic emission sensor located at a distance of no more than a second preset distance from the solder joint, the integrity of the solder joint structure to which the solder joint belongs is judged based on the characteristic parameters in the normalized acoustic emission signal corresponding to the acoustic emission sensor, using a first method, to obtain a judgment result, specifically including: For each pre-specified first characteristic parameter in the normalized acoustic emission signal corresponding to the acoustic emission sensor, the pre-specified first characteristic parameter is compared with its corresponding preset reference interval. When the pre-specified first feature parameter falls into its corresponding preset reference range, the solder joint structure to which the solder joint belongs is determined to be complete, and the completeness of the solder joint structure to which the solder joint belongs is taken as the judgment result; When at least one pre-specified first characteristic parameter does not fall into its corresponding preset reference range, it is determined that the solder joint structure to which the solder joint belongs has a structural abnormality, and the structural abnormality of the solder joint to which the solder joint belongs is taken as the judgment result. The pre-specified first characteristic parameter includes: the cumulative energy of the acoustic emission signal, the maximum instantaneous amplitude of the acoustic emission signal, and the root mean square amplitude of the acoustic emission signal.

5. The method for detecting the structural integrity of weld points after welding according to claim 3, characterized in that, S3 specifically includes: When the acoustic emission signal is acquired by an acoustic emission sensor located at a distance of no more than a second preset distance from the solder joint, the integrity of the solder joint structure to which the solder joint belongs is judged using a second method based on the characteristic parameters in the normalized acoustic emission signal corresponding to the acoustic emission sensor, and the judgment result is obtained, specifically including: Determine whether each pre-specified second feature parameter in the normalized acoustic emission signal corresponding to the acoustic emission sensor is greater than its corresponding preset minimum threshold, and if each pre-specified second feature parameter is greater than or equal to its corresponding preset minimum threshold, use formula (1) to obtain the comprehensive score. The second characteristic parameters include: the cumulative energy of the acoustic emission signal, the proportion of the frequency band energy of the acoustic emission signal within the preset frequency band, the energy attenuation rate of the acoustic emission signal, and the response stability coefficient of the acoustic emission signal; Wherein, the formula (1) is: C=w1·E+w2·F+w3·D+w4·S; C represents the overall score; w1 represents the first preset weight, w2 represents the second preset weight, w3 represents the third preset weight, and w4 represents the fourth preset weight; where 1 = w1 + w2 + w3 + w4; E represents the cumulative energy of the acoustic emission signal in the normalized acoustic emission signal, F represents the proportion of the frequency band energy of the acoustic emission signal in the preset frequency band in the normalized acoustic emission signal, D represents the energy attenuation rate of the acoustic emission signal in the normalized acoustic emission signal, and S represents the response stability coefficient of the acoustic emission signal in the normalized acoustic emission signal. Compare the overall score C with the preset overall score threshold C th A comparison is made, and when the comprehensive score C is greater than or equal to the preset comprehensive score threshold C... th In the case of a solder joint, the solder joint structure to which the solder joint belongs is determined to be complete, and the completeness of the solder joint structure to which the solder joint belongs is taken as the judgment result; When the comprehensive score C is less than the preset comprehensive score threshold C th In such cases, it is determined that the structure of the weld point to which the weld point belongs has a structural abnormality, and the structural abnormality of the weld point to which the weld point belongs is taken as the judgment result.

6. The method for detecting the structural integrity of weld points after welding according to claim 3, characterized in that, S3 specifically includes: When the acoustic emission signal is acquired by an acoustic emission sensor located at a distance no greater than a second preset distance from the solder joint, the integrity of the solder joint structure to which the solder joint belongs is judged using a third method based on the characteristic parameters in the normalized acoustic emission signal corresponding to the acoustic emission sensor, and the judgment result is obtained, specifically including: Based on the third pre-specified feature parameter in the feature parameters of the normalized acoustic emission signal corresponding to the acoustic emission sensor, the feature parameter vector is obtained. The third characteristic parameter includes at least one of the following: cumulative energy of the acoustic emission signal, maximum instantaneous amplitude of the acoustic emission signal, root mean square amplitude of the acoustic emission signal, dominant frequency of the acoustic emission signal, spectral centroid of the acoustic emission signal, frequency band energy ratio of the acoustic emission signal within a preset frequency band, rise time of the acoustic emission signal, duration of the acoustic emission signal, energy decay rate of the acoustic emission signal, first occurrence time of high-frequency components in the acoustic emission signal, duration of high-frequency components in the acoustic emission signal, time-frequency energy concentration of the acoustic emission signal, and response stability coefficient of the acoustic emission signal. The feature parameter vector is input into a pre-trained machine learning model to obtain a judgment result; The machine learning model is trained by using feature parameter vectors from historical time periods. Each feature parameter vector from a historical time period is labeled to indicate whether the solder joint structure is intact or defective. The machine learning models mentioned are support vector machines, random forest models, and neural network models.

7. The method for detecting the structural integrity of weld points after welding according to any one of claims 4-6, characterized in that, The method further includes: S4. When the acoustic emission signal is acquired by multiple acoustic emission sensors located at a distance no greater than a second preset distance from the solder joint. Based on the characteristic parameters in the normalized acoustic emission signal corresponding to each acoustic emission sensor, the integrity of the solder joint structure to which the solder joint belongs is judged by the first method, the second method, or the third method, and the judgment result corresponding to the acoustic emission sensor is obtained. The final judgment result is obtained based on the judgment results corresponding to each acoustic emission sensor.

8. The method for detecting the structural integrity of weld points after welding according to claim 7, characterized in that, Based on the judgment results corresponding to each acoustic emission sensor, the final judgment result is obtained, specifically including: Based on the judgment results corresponding to each acoustic emission sensor, the weighted total score is obtained using formula (2); Formula (2) is: ; in, For weighted total score; Let be the confidence weight corresponding to the i-th acoustic emission sensor; This represents the judgment result corresponding to the i-th acoustic emission sensor; For indicator functions, when A value of 1 indicates that the solder joint structure is complete, otherwise a value of 0 is used. The weighted total score With the preset score threshold T th Comparison, when the weighted total score Greater than or equal to the preset score threshold T th In the case of a solder joint, the solder joint structure to which the solder joint belongs is determined to be complete, and the completeness of the solder joint structure to which the solder joint belongs is taken as the final judgment result; When the weighted total score Less than the preset score threshold T th In such cases, it is determined that the weld joint structure to which the weld joint belongs has a structural abnormality, and the structural abnormality of the weld joint to which the weld joint belongs is taken as the final judgment result.

9. The method for detecting the structural integrity of weld points after welding according to claim 8, characterized in that, in, The confidence weight corresponding to the i-th acoustic emission sensor It is calculated using formula (3); The formula (3) is: ； in, Let be the signal-to-noise ratio of the acoustic emission signal collected by the i-th acoustic emission sensor, where the signal-to-noise ratio is a dimensionless quantity; The maximum value among the signal-to-noise ratios of each acoustic emission sensor involved in the judgment; Let be the accuracy rate of the i-th acoustic emission sensor in the historical weld joint structural integrity detection, where the accuracy rate is a dimensionless quantity. The maximum value among the judgment accuracy rates of each acoustic emission sensor involved in the judgment; The actual spatial distance between the i-th acoustic emission sensor and the solder joint; The second preset distance, and satisfying 0≤ < ; N represents the number of acoustic emission sensors involved in determining the structural integrity of the weld joint. This is the first adjustment coefficient; This is the second adjustment coefficient; This is the third adjustment coefficient.

10. A post-weld weld joint structural integrity detection system, characterized in that, include: A mechanical excitation device is used to apply controlled transient mechanical excitation to the position of the weld point structure to which the weld point belongs, which is no more than a first preset distance from the weld point after welding is completed and cooled. At least one acoustic emission sensor is set at a position no greater than a second preset distance from the solder joint, for acquiring the acoustic emission signal of the solder joint within a specified time period, and preprocessing the acoustic emission signal to obtain a preprocessed acoustic emission signal; The acoustic emission signal is: a controlled transient mechanical excitation is applied at the position of the weld point structure to which the weld point belongs, at a distance of no more than a first preset distance from the weld point after welding is completed and cooled, so that the weld point generates an elastic response and propagates in the weld point structure to which the weld point belongs, and is collected by the acoustic emission sensor; The specified time period is a preset time period starting from the moment the controlled transient mechanical excitation is applied; The signal processing module is used to obtain characteristic parameters in the preprocessed acoustic emission signal based on the preprocessed acoustic emission signal; The determination module is used to determine the integrity of the solder joint structure to which the solder joint belongs based on the feature parameters, and obtain the determination result; The mechanical excitation device is a spring-mass impact device, a piezoelectric actuator, or an electromagnetic exciter. The system further includes a display device for displaying the feature parameters and the judgment result.

Images