Acoustic emission signal detection method for microscopic cracks in aluminum profile welding joint area

Abstract

The invention relates to the technical field of aluminum profile processing, in particular to an acoustic emission signal detection method for microscopic cracks in an aluminum profile welding joint area, which comprises the following steps: S1, welding joint preparation: cleaning an aluminum profile to be detected; s2, acoustic emission sensor arrangement: comprehensive monitoring and signal acquisition are carried out on a welding joint; s3, signal acquisition and processing: starting an acoustic emission sensor to acquire an acoustic emission signal; s4, multi-modal data fusion: fusing the acoustic emission signal and the physical state data into comprehensive data; s5, self-adaptive crack propagation prediction: identifying potential microscopic crack characteristics, and predicting future development of microscopic cracks; and S6, evaluation and report generation: generating an evaluation report. According to the method, the microscopic crack characteristics and the development trend of the welding joint area can be accurately identified and evaluated, the accuracy and the real-time performance of detection are remarkably improved, and effective signals and background noise can be efficiently separated.

Description

Technical Field

[0001] The invention relates to the technical field of aluminum profile processing, and in particular to an acoustic emission signal detection method for micro cracks in an aluminum profile welding joint area. Background Art

[0002] Aluminum profiles are widely used in construction, transportation, aerospace and other fields due to their excellent strength, corrosion resistance and light weight. However, welding, as the main process for connecting aluminum profiles, is very likely to produce micro cracks in the welding joint area. These micro cracks gradually expand under the changes in structural stress or environmental conditions, which may lead to a decrease in the strength of the welding joint and even cause structural failure. Early detection and accurate evaluation of micro cracks in the welding joint area of ​​aluminum profiles has become an important part of ensuring structural safety and service life.

[0003] Existing micro-crack detection technology mainly relies on acoustic emission signal monitoring, but it faces some shortcomings in the detection process. First, acoustic emission signals are easily affected by environmental noise and various interference factors, resulting in insufficient signal accuracy and reliability. Second, traditional data analysis methods usually rely on only a single signal source and fail to fully consider multimodal data fusion of physical states such as temperature, strain and vibration, which limits the accuracy and comprehensiveness of crack feature identification. In addition, for the prediction of crack extension, existing technologies find it difficult to adaptively adjust model parameters in complex environments, resulting in low accuracy and adaptability of prediction results, and unable to effectively provide accurate risk warnings and maintenance guidance.

[0004] To overcome the above-mentioned shortcomings, the present invention provides an improved acoustic emission signal detection method for micro cracks in the welding joint area of ​​aluminum profiles, which improves the adaptability and accuracy of the prediction, and ultimately generates a crack distribution map and maintenance recommendations, providing a scientific basis for structural health monitoring and maintenance decisions. Summary of the invention

[0005] The invention provides an acoustic emission signal detection method for micro cracks in a welding joint area of ​​an aluminum profile.

[0006] The acoustic emission signal detection method for micro cracks in the welding joint area of ​​aluminum profiles comprises the following steps:

[0007] S1, welding joint preparation: clean the aluminum profile to be tested, and use ultrasonic cleaning method to remove surface impurities and oxide layer;

[0008] S2, acoustic emission sensor arrangement: acoustic emission sensors are arranged in and around the welding joint area to comprehensively monitor and collect signals of the welding joint;

[0009] S3, signal acquisition and processing: after welding is completed, the acoustic emission sensor is started to collect acoustic emission signals, and the collected acoustic emission signals are denoised;

[0010] S4, multimodal data fusion: after signal processing, the multimodal data fusion algorithm is used to fuse the acoustic emission signal with the physical state data (temperature, strain and vibration data) into comprehensive data;

[0011] S5, Adaptive Crack Growth Prediction: Based on the integrated data, potential micro crack characteristics are identified and the future development of micro cracks is predicted through the adaptive crack growth prediction model, including:

[0012] S51, micro crack feature recognition: extract micro crack features from acoustic emission signals based on the fused comprehensive data, including crack location, length, expansion rate, morphology and width;

[0013] S52, crack location and severity determination: using the identified crack features, determine the location and severity of the crack;

[0014] S53, Adaptive Crack Extension Prediction: Through the adaptive crack extension prediction model, combined with the current environment and stress state, the identified micro crack characteristics are input to predict the future development of the crack;

[0015] S6, evaluation and report generation: Based on the identified micro crack characteristics and the predicted results of crack propagation, an evaluation report is generated. The report content includes the distribution and severity of the cracks and corresponding maintenance recommendations.

[0016] Optionally, the acoustic emission sensor arrangement in S2 includes:

[0017] S21, select sensor type: select acoustic emission sensor according to the material characteristics of the welding joint and monitoring requirements;

[0018] S22, determine the layout position: evenly install the acoustic emission sensors around the welding joint;

[0019] S23, sensor installation: Use adhesive or mechanical fixing method to install the acoustic emission sensor at the predetermined location.

[0020] Optionally, the signal acquisition and processing in S3 includes:

[0021] S31, signal acquisition: start the arranged acoustic emission sensors to collect the acoustic emission signals of the welding joint area in real time;

[0022] S32, denoising: decomposing the acoustic emission signal by wavelet transform to extract different frequency components, and applying soft threshold method to denoise the wavelet coefficients;

[0023] S33, signal reconstruction: reconstructing the denoised wavelet coefficients into an acoustic emission signal through inverse wavelet transform.

[0024] Optionally, the multimodal data fusion in S4 includes:

[0025] S41, data acquisition and preprocessing: real-time acquisition of physical state data, including temperature, strain and vibration data, and preprocessing of the denoised acoustic emission signal and the acquired physical state data, including normalization and time alignment;

[0026] S42, multimodal data fusion: combining the preprocessed acoustic emission signals with the physical state data to form comprehensive data.

[0027] Optionally, the data collection and preprocessing in S41 includes:

[0028] S411, physical state data collection: using temperature sensors, strain sensors, and vibration sensors to collect temperature, strain, and vibration in the welding process in real time;

[0029] S412, data preprocessing: preprocessing the denoised acoustic emission signal and the collected physical state data, specifically including:

[0030] Normalization: standardize different data sources to the same range;

[0031] Time alignment: The acoustic emission signal and the physical state data are aligned in time by linear interpolation.

[0032] Optionally, the multimodal data fusion in S42 adopts a weighted fusion algorithm to combine the preprocessed acoustic emission signal with the physical state data to form a comprehensive data D fused .

[0033] Optionally, the micro crack feature identification in S51 includes:

[0034] S511, Crack location identification: Use the peak detection algorithm to identify abnormal peaks in the acoustic emission signal and determine the location of the potential crack P crack ;

[0035] S512, Crack length estimation: By calculating the time interval between the crack characteristic peaks and combining the wave velocity c, the crack length L is estimated. crack ;

[0036] S513, expansion rate calculation: Calculate the crack expansion rate R according to the change of the crack in the time series crack ;

[0037] S514, Crack morphology analysis: Use wavelet transform to analyze the morphological characteristics of crack signals and obtain the frequency distribution F of the signal. shape (t);

[0038] S515, Crack Width Estimation: By analyzing the amplitude of the acoustic emission signal and combining the spectrum characteristics, the crack width W is estimated. crack .

[0039] Optionally, the determination of the crack location and severity in S52 includes:

[0040] S521, Crack location location: Based on the result of crack location identification P crack , determine the location of the crack;

[0041] S522, Crack severity assessment: Use the length, growth rate and amplitude of the acoustic emission signal to assess the severity of the crack and calculate the crack severity score S severity ;

[0042] S523, Severity classification: The severity of the crack is graded S severity The upper limit of the preset severity threshold Severity lower threshold Compare and classify the severity of cracks. When When , indicating a severe crack.

[0043] Optionally, the adaptive crack growth prediction model in S53 is constructed based on a Gaussian process regression model, and the adaptive crack growth prediction model includes:

[0044] S531, define a multi-input feature space: considering that crack extension is affected by multiple physical states, a multi-input feature space is introduced, including crack length L, extension rate R, temperature T, strain S and vibration V, which constitute an input vector X;

[0045] S532, Adaptive multimodal covariance function design: Considering the influence of different characteristics on crack propagation, define the combined covariance function k(x,x′);

[0046] S533, multimodal prediction mean and covariance calculation: based on the adaptive multimodal covariance function, the prediction mean and covariance matrix are calculated, specifically including:

[0047] Covariance matrix calculation: Let the observed data X = [X1, X2, ..., X n ], the corresponding target value is Y=[Y1,Y2,...,Y n], calculate the covariance matrix K;

[0048] Predicted mean: For a new input X * , predicted mean That is, the estimated value of the predicted crack characteristics at a future point in time;

[0049] Prediction covariance: prediction covariance matrix cov(f * ), which is used to estimate the variance of the prediction results and represents the uncertainty of the prediction results;

[0050] S534, Uncertainty measurement and dynamic adjustment: Use the variance information in the prediction covariance to evaluate the uncertainty of the prediction results and calculate the prediction variance σ f* , when the prediction variance σ f* When the preset threshold is exceeded, the length scale l is increased.

[0051] Optionally, the evaluation and report generation in S6 includes:

[0052] S61, crack distribution map generation: based on the location and distribution characteristics of the cracks, a crack distribution map of the weld joint area is generated to show the spatial distribution of the cracks;

[0053] S62, maintenance recommendations: Based on the severity of the crack and the predicted expansion trend, maintenance recommendations are made, including inspection frequency, repair or replacement plans;

[0054] S63, report generation and export: Crack distribution map generation and maintenance recommendations are provided, and an assessment report is generated. The report content includes the distribution of cracks, severity level and maintenance recommendations.

[0055] Beneficial effects of the present invention:

[0056] The present invention can accurately identify and evaluate the micro crack characteristics and development trends of the welding joint area by systematically cleaning the welding joint, reasonably arranging the acoustic emission sensors, collecting and processing the acoustic emission signals in real time, and combining the multimodal data fusion and adaptive crack propagation prediction method. It not only significantly improves the accuracy and real-time performance of the detection, but also can efficiently separate the effective signal from the background noise, ensure the integrity and reliability of the data, and thus enhance the health monitoring capability of the welding structure.

[0057] The present invention introduces multimodal data fusion technology to comprehensively analyze acoustic emission signals with physical state data such as temperature, strain and vibration, so that the identification of crack characteristics is more comprehensive and the analysis depth is further improved. An adaptive crack growth prediction model is adopted to reasonably set the multimodal covariance function and the dynamic adjustment mechanism of uncertainty measurement, so that a high prediction accuracy can be maintained under environmental changes. When the uncertainty is high, the model reduces the sensitivity to noise through dynamic adjustment, thereby improving the adaptability and robustness of crack growth prediction, and providing a scientific basis for the long-term reliability of engineering structures.

[0058] The present invention generates an assessment report based on crack distribution and expansion trends. The report content includes a spatial distribution map of cracks, severity levels, and targeted maintenance suggestions, which helps engineering personnel quickly identify high-risk areas and take corresponding measures. This not only improves the efficiency and accuracy of detection and assessment, but also greatly reduces the risk of structural failure, saves maintenance costs, and provides effective protection for the long-term safe use of welded joints. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings in the following description are only for the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0060] Figure 1 A schematic diagram of a detection method flow chart of an embodiment of the present invention;

[0061] Figure 2 Schematic diagram of adaptive crack extension prediction according to an embodiment of the present invention. DETAILED DESCRIPTION

[0062] The present invention is described in detail below in conjunction with the accompanying drawings and specific embodiments. At the same time, it is explained here that in order to make the embodiments more detailed, the following embodiments are the best and preferred embodiments, and those skilled in the art may also adopt other alternatives to implement some known technologies; and the accompanying drawings are only for more specific description of the embodiments, and are not intended to specifically limit the present invention.

[0063] It should be noted that the references to "one embodiment", "an embodiment", "an exemplary embodiment", "some embodiments" and the like in the specification indicate that the embodiments described may include specific features, structures or characteristics, but not every embodiment may include the specific features, structures or characteristics. In addition, when a specific feature, structure or characteristic is described in conjunction with an embodiment, it should be within the knowledge of a person skilled in the art to implement such feature, structure or characteristic in conjunction with other embodiments (whether or not explicitly described).

[0064] In general, a term can be understood, at least in part, from its use in context. For example, depending, at least in part, on the context, the term "one or more" as used herein can be used to describe any feature, structure, or characteristic in the singular sense, or can be used to describe a combination of features, structures, or characteristics in the plural sense. Additionally, the term "based on" can be understood as not necessarily intended to convey an exclusive set of factors, but can instead, depending, at least in part, on the context, allow for the presence of other factors that are not necessarily explicitly described.

[0065] like Figure 1-Figure 2 As shown, the acoustic emission signal detection method for micro cracks in the welding joint area of ​​aluminum profiles includes the following steps:

[0066] S1, welding joint preparation: clean the aluminum profile to be tested, and use ultrasonic cleaning method to remove surface impurities and oxide layer;

[0067] S2, acoustic emission sensor arrangement: acoustic emission sensors are arranged in and around the welding joint area to comprehensively monitor and collect signals of the welding joint;

[0068] S3, signal acquisition and processing: after welding is completed, the acoustic emission sensor is started to collect acoustic emission signals, and the collected acoustic emission signals are denoised;

[0069] S4, multimodal data fusion: after signal processing, the multimodal data fusion algorithm is used to fuse the acoustic emission signal with the physical state data (temperature, strain and vibration data) into comprehensive data;

[0070] S5, Adaptive Crack Growth Prediction: Based on the integrated data after fusion, potential micro crack characteristics are identified and the future development of micro cracks is predicted through the adaptive crack growth prediction model, including:

[0071] S51, micro crack feature recognition: extract micro crack features from acoustic emission signals based on the fused comprehensive data, including crack location, length, expansion rate, morphology and width;

[0072] S52, crack location and severity determination: using the identified crack features, determine the location and severity of the crack;

[0073] S53, Adaptive Crack Extension Prediction: Through the adaptive crack extension prediction model, combined with the current environment and stress state, the identified micro crack characteristics are input to predict the future development of the crack;

[0074] S6, evaluation and report generation: based on the identified micro crack characteristics and the predicted results of crack extension, an evaluation report is generated, which includes the distribution and severity of the cracks and corresponding maintenance recommendations;

[0075] Through the above content, systematic cleaning of welding joints, arrangement of sensors, collection and processing of acoustic emission signals, combined with multimodal data fusion and adaptive crack propagation prediction, it is possible to accurately identify and evaluate micro crack characteristics and their development trends, which not only improves the accuracy and timeliness of detection, but also helps to effectively prevent structural failures and reduce maintenance costs, providing a scientific basis for the long-term reliability of aluminum profile welding joints.

[0076] The acoustic emission sensor arrangement in S2 includes:

[0077] S21, select sensor type: select acoustic emission sensor according to the material characteristics of the welding joint and monitoring requirements;

[0078] S22, determine the layout position: evenly install the acoustic emission sensors around the welding joint;

[0079] S23, sensor installation: using adhesive or mechanical fixing method, install the acoustic emission sensor at the predetermined position;

[0080] Through the above content, selecting the appropriate sensor type, determining a uniform installation location, and using a stable installation method can effectively improve the monitoring accuracy and signal acquisition integrity of the weld joint area, which not only ensures a timely response to potential crack formation and expansion, but also enhances the reliability of the data.

[0081] Signal acquisition and processing in S3 include:

[0082] S31, signal acquisition: start the arranged acoustic emission sensors to collect the acoustic emission signals of the welding joint area in real time;

[0083] S32, denoising: decompose the acoustic emission signal by wavelet transform to extract different frequency components, and apply soft threshold method to denoise the wavelet coefficients, which is expressed as:

[0084]

[0085] Where W(a,b) is the wavelet coefficient, f(t) is the acoustic emission signal, ψ is the wavelet mother function, a and b are the scale and translation parameters respectively;

[0086]

[0087] Among them, λ is the set threshold, which is a multiple of the signal noise standard deviation;

[0088] S33, signal reconstruction: The denoised wavelet coefficients are reconstructed into acoustic emission signals through inverse wavelet transform, which is expressed as:

[0089]

[0090] Among them, f(t) is the reconstructed original signal, C ψ is a normalization constant that depends on the selected wavelet mother function, W(a,b) is the wavelet coefficient, ψ * is the conjugate of the wavelet mother function (in the real case, ψ * =ψ), a and b are scale and translation parameters respectively;

[0091] Through the above content, the clarity and accuracy of the signal can be improved. The wavelet transform and soft threshold denoising methods can be used to effectively separate the useful signal from the background noise, thereby retaining the key information of the micro crack characteristics. It not only ensures high-quality signal data to provide a basis for subsequent analysis, but also improves the sensitivity and reliability of detection, providing strong support for the timely identification of potential defects and structural health monitoring.

[0092] Multimodal data fusion in S4 includes:

[0093] S41, data acquisition and preprocessing: real-time acquisition of physical state data, including temperature, strain and vibration data, and preprocessing of the denoised acoustic emission signal and the acquired physical state data, including normalization and time alignment;

[0094] S42, multimodal data fusion: combining the preprocessed acoustic emission signals with the physical state data to form comprehensive data;

[0095] Through the above content, more comprehensive and accurate monitoring of the health status of welded joints can be provided, which not only enhances the ability to identify potential crack characteristics, but also improves the reliability of data and the depth of analysis, thereby providing strong support for structural health monitoring. It can more effectively reflect the performance of welded joints under actual working conditions, help to detect problems in time and take preventive measures, and ultimately improve the safety and reliability of welded structures.

[0096] Data collection and preprocessing in S41 include:

[0097] S411, physical state data collection: using temperature sensors, strain sensors, and vibration sensors to collect temperature, strain, and vibration in the welding process in real time;

[0098] S412, data preprocessing: preprocessing the denoised acoustic emission signal and the collected physical state data, specifically including:

[0099] Normalization: Standardize different data sources to the same range to eliminate scale differences, expressed as:

[0100]

[0101] Among them, D is the original data, D min and D max are the minimum and maximum values ​​of the data set respectively;

[0102] Time alignment: The acoustic emission signal and the physical state data are aligned in time by linear interpolation to ensure that the corresponding data are obtained at the same time point, which is expressed as:

[0103]

[0104] Wherein, D1(t1) and D2(t2) are the original data at time points t1 and t2;

[0105] Through the above content, the accuracy and reliability of the data are improved, and the consistency between various types of data is ensured, thereby enhancing the comprehensive monitoring capability of the health status of welded joints. It can not only identify potential structural defects in a timely manner, but also provide an important basis for effective maintenance decisions, thereby improving the safety and reliability of welded structures.

[0106] The multimodal data fusion in S42 uses a weighted fusion algorithm to combine the preprocessed acoustic emission signals with the physical state data to form a comprehensive data D fused , expressed as:

[0107] D fused =w1·D AE +w2·D temp +w3·D strain +w4·D vibration ;

[0108] Among them, D AE is the denoised acoustic emission signal, D temp , D strain , D vibration are the normalized temperature, strain and vibration data, respectively, w1, w2, w3, w4 are the weights of each data source;

[0109] Through the above content, the comprehensiveness and accuracy of the data are improved, and the information of each data source can be fully utilized, making the monitoring of the health status of the welded joint more comprehensive and detailed. In addition, the weighted fusion method ensures that the importance of each data source is reasonably reflected, which not only improves the ability to identify potential defects, but also provides an important basis for maintenance decisions, ultimately enhancing the safety and reliability of the welded structure.

[0110] Micro crack feature recognition in S51 includes:

[0111] S511, Crack location identification: Use the peak detection algorithm to identify abnormal peaks in the acoustic emission signal and determine the location of the potential crack P crack , expressed as:

[0112]

[0113] Among them, P i is the position of the i-th peak, A i is the corresponding peak amplitude, N is the number of detected peaks;

[0114] S512, Crack length estimation: By calculating the time interval between the crack characteristic peaks and combining the wave velocity c, the crack length L is estimated. crack , expressed as:

[0115] L crack = c·Δt;

[0116] Where Δt is the time interval between peaks;

[0117] S513, expansion rate calculation: Calculate the crack expansion rate R according to the change of the crack in the time series crack , expressed as:

[0118]

[0119] Among them, L crack (t1) and L crack (t2) are the crack lengths at different time points;

[0120] S514, Crack morphology analysis: Use wavelet transform to analyze the morphological characteristics of crack signals and obtain the frequency distribution F of the signal. shape (t), expressed as:

[0121]

[0122] in, represents the wavelet transform operation;

[0123] S515, Crack Width Estimation: By analyzing the amplitude of the acoustic emission signal and combining the spectrum characteristics, the crack width W is estimated. crack , expressed as:

[0124]

[0125] Among them, A max is the maximum amplitude of the acoustic emission signal, f peak is the peak value of the corresponding frequency;

[0126] Through the above content, the monitoring capability of the health status of welded joints is enhanced, which not only improves the accuracy of identifying potential cracks, but also enables real-time evaluation of the severity and development trend of cracks, providing a reliable basis for maintenance decisions, and timely discovering structural hidden dangers and reducing the risk of failure.

[0127] Crack location and severity determination in S52 includes:

[0128] S521, Crack location location: Based on the result of crack location identification P crack , determine the location of the crack;

[0129] S522, Crack severity assessment: Use the length, growth rate and amplitude of the acoustic emission signal to assess the severity of the crack and calculate the crack severity score S severity , expressed as:

[0130] S severity =k1·L crack +k2·R crack +k3·A max ;

[0131] Among them, S severity is the crack severity score, k1, k2, k3 are the importance weights of each feature, L crack is the crack length, R crack is the crack growth rate, A max is the maximum amplitude of the acoustic emission signal;

[0132] S523, Severity classification: The severity of the crack is graded S severity The upper limit of the preset severity threshold Severity lower threshold Compare and classify the severity of cracks. When When , it indicates a serious crack;

[0133] Severity threshold upper limit Severity lower threshold It is set according to the material properties and crack propagation characteristics and is expressed as:

[0134]

[0135] Among them, K IC is the fracture toughness of the material, σ max is the maximum stress in the application environment, L max is the maximum allowable crack length, R max is the maximum allowable value of crack growth rate, k, α and β are adjustment parameters, which can be optimized based on experiments or historical data;

[0136]

[0137] Among them, σ fatigue is the fatigue limit of the material, L min is the minimum detectable crack length, R min is the minimum detectable value of the crack growth rate, m, γ and δ are adjustment parameters used to optimize according to the safety standards of different materials and structures;

[0138] Through the above content, the specific location of the crack and its potential threat to the structure can be accurately identified, ensuring a scientific assessment of the crack risk, enabling the monitoring system to distinguish between minor, moderate and severe cracks, and thus making targeted maintenance recommendations, improving the accuracy and reliability of monitoring, and timely discovering and responding to potential structural risks, thereby ensuring the safety and service life of the welded structure.

[0139] The adaptive crack growth prediction model in S53 is built based on the Gaussian process regression model. The adaptive crack growth prediction model includes:

[0140] S531, define multi-input feature space: Considering that crack extension is affected by multiple physical states, a multi-input feature space is introduced, including crack length L, extension rate R, temperature T, strain S and vibration V, which constitute the input vector X, expressed as:

[0141] X = [L, R, T, S, V];

[0142] S532, Adaptive multimodal covariance function design: Considering the influence of different characteristics on crack propagation, the combined covariance function k(x,x′) is defined, which is expressed as:

[0143] k(x,x′)=k RBF (L,L′)+k Matern (R,R′)+k Periodic (T,T′)

[0144] k Linear (S,S′)·kRationalQuadratic (V, V′);

[0145]

[0146] Among them, k(x,x′) is the covariance function, which is defined as a combination of multiple kernel functions. L′, R′, T′, S′, and V′ are the crack length, crack growth rate, temperature value, strain value, and vibration value of another point, respectively. RBF (L,L′) is the RBF kernel function, which is used for the crack length L, σ f is the signal amplitude of the RBF kernel, indicating the overall scale of the kernel function, l L is the length scale parameter of the RBF core, controlling the influence range of the crack length L, k Matern (R, R′) is the Matern kernel function used for the crack growth rate R, σ m is the amplitude parameter of the Matern kernel, l R is the length scale parameter of the Matern kernel, controlling the influence range of the crack growth rate R, k Periodic (T,T′) is the periodic kernel function for temperature T, σ p is the signal amplitude of the periodic kernel, p is the period parameter, indicating the period of temperature change, l T is the length scale parameter of temperature, controlling the influence range of temperature T, k Linear (S,S′) is the linear kernel function for strain S, σ l is the bias term of the linear kernel, w5 is the proportional factor, which controls the strength of the linear relationship of strain S, and k RationalQuadratic (V,V′) is a rational quadratic kernel function used to oscillate V, σ r is the amplitude parameter of the rational quadratic kernel, w6 is the shape parameter, controlling the smoothness of the vibration V, l V is the length scale parameter, which controls the influence range of the vibration V;

[0147] S533, multimodal prediction mean and covariance calculation: based on the adaptive multimodal covariance function, the prediction mean and covariance matrix are calculated, specifically including:

[0148] Covariance matrix calculation: Let the observed data X = [X1, X2, ..., X n ], the corresponding target value is Y=[Y1,Y2,...,Y n ], calculate the covariance matrix K, expressed as:

[0149]

[0150] Among them, K(X,X) is the covariance matrix, which is used to calculate the covariance between the observed data X and itself. is the noise variance, which is used to control the size of the observation noise, and I is the unit matrix, which is used together with the noise variance to increase the diagonal noise;

[0151] Predicted mean: For a new input X * , predicted mean That is, the estimated value of the predicted crack characteristics at a future time point, expressed as:

[0152]

[0153] in, is the predicted mean, which is the estimated value of the future development of the crack, K(X * ,X) is the new input X * and the covariance matrix between the observed data X, K(X,X) -1 is the inverse matrix of the observed data covariance matrix K(X,X), which is used to calculate the predicted mean, and Y is the target value of the observed data X (i.e., the crack history characteristic data);

[0154] Prediction covariance: prediction covariance matrix cov(f * ), which is used to estimate the variance of the prediction results and represents the uncertainty of the prediction results, and is expressed as:

[0155] cov(f * )=K(X * ,X * )-K(X * ,X)K(X,X) -1 K(X,X * );

[0156] Among them, cov(f * ) is the prediction covariance matrix, which is used to estimate the variance (uncertainty) of the prediction results. K(X * ,X * ) is the new input X * and its own covariance, K(X * ,X) is the new input X * and the covariance matrix between the observed data X, K(X,X * ) is the observed data X and the new input X * The covariance matrix between ;

[0157] S534, Uncertainty Measurement and Dynamic Adjustment: Use the variance information in the prediction covariance to evaluate the uncertainty of the prediction results and calculate the prediction variance When the prediction variance When the preset threshold is exceeded, the length scale l is increased to improve the prediction stability and reduce the sensitivity to noise, which is expressed as:

[0158]

[0159] in, is the prediction uncertainty (standard deviation), which represents the variance of the prediction results and is used to judge the reliability of the prediction;

[0160] The preset threshold is set based on the statistical method of historical data. The mean and standard deviation of the prediction variance are calculated using historical crack growth data. The threshold is set to a multiple higher than the average uncertainty level, expressed as:

[0161]

[0162] in, is the forecast uncertainty in historical data The mean of is the standard deviation of uncertainty, k is the adjustment factor, ranging from 1 to 3;

[0163] Through the above content, the multimodal covariance function design and the dynamic adjustment mechanism of uncertainty measurement can effectively integrate crack characteristics and physical state data, accurately predict the crack propagation trend, and set reasonable thresholds to trigger dynamic adjustment of the model when the uncertainty is high, thereby enhancing the adaptability of the model in complex environments, which not only improves the prediction accuracy, but also provides the reliability of risk assessment.

[0164] Evaluation and reporting in S6 include:

[0165] S61, crack distribution map generation: based on the location and distribution characteristics of the cracks, a crack distribution map of the weld joint area is generated to display the spatial distribution of the cracks so as to quickly identify the key monitoring areas;

[0166] S62, maintenance recommendations: Based on the severity of the cracks and the predicted expansion trend, maintenance recommendations are made, including inspection frequency, repair or replacement plans, to prevent potential structural risks;

[0167] S63, report generation and export: crack distribution map generation and maintenance recommendations, generate an assessment report, the report content includes crack distribution, severity level and maintenance recommendations;

[0168] Through the above content, the health status of the welded structure can be analyzed comprehensively and accurately. The generated assessment report not only contains the distribution map and severity classification of cracks, but also provides targeted maintenance suggestions, allowing engineers to quickly identify high-risk areas and take effective preventive measures, thereby improving the efficiency and accuracy of detection and providing a scientific basis for structural safety management.

[0169] The present invention covers any substitution, modification, equivalent method and scheme made on the essence and scope of the present invention. In order to make the public have a thorough understanding of the present invention, specific details are described in detail in the following preferred embodiments of the present invention, but those skilled in the art can fully understand the present invention without the description of these details. In addition, in order to avoid unnecessary confusion about the essence of the present invention, well-known methods, processes, procedures, components and circuits are not described in detail.

[0170] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principle of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. The acoustic emission signal detection method for micro cracks in the welding joint area of ​​aluminum profiles is characterized by: The following steps are involved: S1, welding joint preparation: clean the aluminum profile to be tested, and use ultrasonic cleaning method to remove surface impurities and oxide layer; S2, acoustic emission sensor arrangement: acoustic emission sensors are arranged in the welding joint area and its surroundings to comprehensively monitor and collect signals of the welding joint; S3, signal acquisition and processing: after welding is completed, the acoustic emission sensor is started to collect acoustic emission signals, and the collected acoustic emission signals are denoised; S4, multimodal data fusion: after signal processing, the multimodal data fusion algorithm is used to fuse the acoustic emission signal and the physical state data into comprehensive data; S5, Adaptive Crack Growth Prediction: Based on the integrated data, potential micro crack characteristics are identified and the future development of micro cracks is predicted through the adaptive crack growth prediction model, including: S51, micro crack feature recognition: extract micro crack features from acoustic emission signals based on the fused comprehensive data, including crack location, length, expansion rate, morphology and width; S52, crack location and severity determination: using the identified crack features, determine the location and severity of the crack; S53, Adaptive Crack Extension Prediction: Through the adaptive crack extension prediction model, combined with the current environment and stress state, the identified micro crack characteristics are input to predict the future development of the crack; S6, evaluation and report generation: Based on the identified micro crack characteristics and the predicted results of crack propagation, an evaluation report is generated. The report content includes the distribution and severity of the cracks and corresponding maintenance recommendations.

2. The acoustic emission signal detection method for micro cracks in the aluminum profile welding joint area according to claim 1 is characterized in that: The acoustic emission sensor arrangement in S2 comprises: S21, select sensor type: select acoustic emission sensor according to the material characteristics of the welding joint and monitoring requirements; S22, determine the layout position: evenly install the acoustic emission sensors around the welding joint; S23, sensor installation: Use adhesive or mechanical fixing method to install the acoustic emission sensor at the predetermined location.

3. The acoustic emission signal detection method for micro cracks in the aluminum profile welding joint area according to claim 1 is characterized in that: The signal acquisition and processing in S3 includes: S31, signal acquisition: start the arranged acoustic emission sensors to collect the acoustic emission signals of the welding joint area in real time; S32, denoising: decomposing the acoustic emission signal by wavelet transform to extract different frequency components, and applying soft threshold method to denoise the wavelet coefficients; S33, signal reconstruction: reconstructing the denoised wavelet coefficients into an acoustic emission signal through inverse wavelet transform.

4. The acoustic emission signal detection method for micro cracks in the aluminum profile welding joint area according to claim 1 is characterized in that: The multimodal data fusion in S4 includes: S41, data acquisition and preprocessing: real-time acquisition of physical state data, including temperature, strain and vibration data, and preprocessing of the denoised acoustic emission signal and the acquired physical state data, including normalization and time alignment; S42, multimodal data fusion: combine the preprocessed acoustic emission signal with the physical state data to form comprehensive data.

5. The acoustic emission signal detection method for micro cracks in the aluminum profile welding joint area according to claim 4 is characterized in that: The data collection and preprocessing in S41 include: S411, physical state data collection: using temperature sensors, strain sensors, and vibration sensors to collect temperature, strain, and vibration in the welding process in real time; S412, data preprocessing: preprocessing the denoised acoustic emission signal and the collected physical state data, specifically including: Normalization: standardize different data sources to the same range; Time alignment: The acoustic emission signal and the physical state data are aligned in time by linear interpolation.

6. The acoustic emission signal detection method for micro cracks in the aluminum profile welding joint area according to claim 5 is characterized in that: The multimodal data fusion in S42 adopts a weighted fusion algorithm to combine the pre-processed acoustic emission signal with the physical state data to form a comprehensive data D fused .

7. The acoustic emission signal detection method for micro cracks in the aluminum profile welding joint area according to claim 1 is characterized in that: The micro crack feature identification in S51 includes: S511, Crack location identification: Use the peak detection algorithm to identify abnormal peaks in the acoustic emission signal and determine the location of the potential crack P crack ; S512, Crack length estimation: By calculating the time interval between the crack characteristic peaks and combining the wave velocity c, the crack length L is estimated. crack ; S513, expansion rate calculation: Calculate the crack expansion rate R according to the change of the crack in the time series crack ; S514, Crack morphology analysis: Use wavelet transform to analyze the morphological characteristics of crack signals and obtain the frequency distribution F of the signal. shape (t); S515, Crack Width Estimation: By analyzing the amplitude of the acoustic emission signal and combining the spectrum characteristics, the crack width W is estimated. crack .

8. The acoustic emission signal detection method for micro cracks in the aluminum profile welding joint area according to claim 7 is characterized in that: The determination of the crack location and severity in S52 includes: S521, Crack location location: Based on the result of crack location identification P crack , determine the location of the crack; S522, Crack severity assessment: Use the length, growth rate and amplitude of the acoustic emission signal to assess the severity of the crack and calculate the crack severity score S severity ; S523, Severity classification: The severity of the crack is graded S severity The upper limit of the preset severity threshold Severity lower threshold Compare and classify the severity of cracks. When When It indicates a serious crack.

9. The acoustic emission signal detection method for micro cracks in the aluminum profile welding joint area according to claim 8, characterized in that: The adaptive crack growth prediction model in S53 is constructed based on a Gaussian process regression model, and the adaptive crack growth prediction model includes: S531, define a multi-input feature space: considering that crack extension is affected by multiple physical states, a multi-input feature space is introduced, including crack length L, extension rate R, temperature T, strain S and vibration V, which constitute an input vector X; S532, Adaptive multi-modal covariance function design: Considering the influence of different characteristics on crack propagation, define the combined covariance function k(x,x ′ ); S533, multimodal prediction mean and covariance calculation: based on the adaptive multimodal covariance function, the prediction mean and covariance matrix are calculated, specifically including: Covariance matrix calculation: Let the observed data X = [X1, X2, ..., X n ], the corresponding target value is Y=[Y1,Y2,...,Y n ], calculate the covariance matrix K; Predicted mean: For a new input X * , predicted mean That is, the estimated value of the predicted crack characteristics at a future point in time; Prediction covariance: prediction covariance matrix cov(f * ), which is used to estimate the variance of the prediction results and represents the uncertainty of the prediction results; S534, Uncertainty measurement and dynamic adjustment: Use the variance information in the prediction covariance to evaluate the uncertainty of the prediction results and calculate the prediction variance σ f* , when the prediction variance σ f* When the preset threshold is exceeded, the length scale l is increased.

10. The acoustic emission signal detection method for micro cracks in the aluminum profile welding joint area according to claim 1, characterized in that: The evaluation and report generation in S6 includes: S61, crack distribution map generation: based on the location and distribution characteristics of the cracks, a crack distribution map of the weld joint area is generated to show the spatial distribution of the cracks; S62, maintenance recommendations: Based on the severity of the crack and the predicted expansion trend, maintenance recommendations are made, including inspection frequency, repair or replacement plans; S63, report generation and export: Crack distribution map generation and maintenance recommendations are provided, and an assessment report is generated. The report content includes the distribution of cracks, severity level and maintenance recommendations.

Images