A monitoring method and system for welded structural components based on the vibro-acoustic method

Abstract

The present application provides a method and system for monitoring welded structural parts based on a vibration acoustics method, relating to the technical field of welding monitoring. The method includes: constructing a defect classification model and a defect strategy model, using an ultrasonic phased array to detect a weld to be detected of the welded structural part to obtain a first image, inputting the first image into the defect classification model to obtain a second image, and determining whether there are welding defects in the second image. If not, the welding quality evaluation result of the weld to be detected is set to excellent, and the welding quality evaluation result is output. If so, the second image is input into the defect strategy model to obtain a defect repair strategy for the second image. By monitoring the defects of the welded structural part in real time, classifying the welding defects based on the defect classification model, evaluating the welding defects, and obtaining a targeted repair strategy through the defect strategy model based on the evaluation result, the present application achieves the purpose of timely adjusting the welding process for repair based on the welding defects occurring during the welding process.

Description

Technical Field

[0001] The present application relates to the technical field of welding monitoring, and in particular to a method and system for monitoring welded structural components based on a vibration acoustics method. Background Art

[0002] Welding technology plays a crucial role in modern industrial production and is widely used in many fields such as automotive, aerospace, shipbuilding, construction, and energy. During the welding process, common welding defects include cracks, pores, slag inclusions, lack of fusion, excessive or insufficient weld seams, etc. Traditional welding quality monitoring methods such as ultrasonic, X-ray, and magnetic particle monitoring often focus on the ex-post inspection of the finished welded structural components. Although these methods can detect defects in the welded structural components, due to the fact that they monitor based on the completed welded structural components, there is a certain time lag. If problems are not detected in a timely manner during the welding process, the finished welded structural components with welding defects produced by the welding technology will continue to accumulate, not only increasing the production cost but also affecting the welding quality and welding production efficiency.

[0003] The Chinese invention patent with the application publication number CN118533972A provides an intelligent ultrasonic-based weld detection system. This patent sends ultrasonic signals to the weld of the welded component to be detected through an ultrasonic sensor and records the feedback information of the ultrasonic waves as an ultrasonic acquisition information group; based on the ultrasonic acquisition information group, preprocessing is carried out to obtain a weld feature data set and an ultrasonic feedback signal feature data set; the ultrasonic feedback signal is extracted from the ultrasonic acquisition information group, and the extracted ultrasonic feedback signal is converted into a digital signal to form an ultrasonic signal digital set. The difference information of the ultrasonic signal digital set is synchronously counted to obtain a digital signal feedback coefficient Szxs. Then, the weld feature data set and the ultrasonic feedback signal feature data set are normalized to form a first data set and a second data set; a weld feature extraction model is established for the first data set and the second data set, trained and extracted, and then fitted with the digital signal feedback coefficient Szxs to obtain a weld feature adjustment index Tzzs; by comparing the preset weld evaluation adjustment threshold T with the weld feature adjustment index Tzzs, a weld detection adjustment evaluation strategy scheme is obtained, and specific execution and notification are carried out according to the content of the weld detection adjustment evaluation strategy scheme.

[0004] For the above technical solution, a weld adjustment plan is obtained based on the weld features of ultrasonic signals, but a linkage adjustment mechanism between the welding process and welding defects is not established, and the welding process cannot be adjusted and repaired in a timely manner based on the detected welding defects, thus affecting the stability and long-term reliability of the welding quality of the welded structural components. Summary of the Invention

[0005] In order to adjust the welding process in a timely manner according to the detected welding defects for repair, the present application provides a monitoring method and system for welded structural parts based on the vibration acoustics method.

[0006] In the first aspect, the present application provides a monitoring method for welded structural parts based on the vibration acoustics method, adopting the following technical solution:

[0007] A monitoring method for welded structural parts based on the vibration acoustics method includes the following steps:

[0008] Data acquisition: Collect historical welding process parameters, historical welding defect images and corresponding defect category labels, and defect repair strategies;

[0009] Model construction: Construct a defect classification model and a defect strategy model;

[0010] Model training: Use historical welding process parameters, historical welding defect images and corresponding defect category labels to train the defect classification model to obtain a trained defect classification model, and use historical welding process parameters, historical welding defect images and defect repair strategies to train the defect strategy model to obtain a trained defect strategy model;

[0011] First monitoring: Use ultrasonic phased array to detect the weld to be detected of the welded structural part to obtain a first image;

[0012] First discrimination: Obtain the current welding process parameters, input the current welding process parameters and the first image into the trained defect classification model to obtain a first image marked with defect category labels, denoted as the second image, and discriminate whether there are welding defects in the second image:

[0013] If not, then execute the steps of the first setting;

[0014] If so, then execute the steps of strategy acquisition;

[0015] First setting: Set the welding quality evaluation result of the weld to be detected to excellent and output the welding quality evaluation result;

[0016] Strategy acquisition: Input the current welding process parameters and the second image into the trained defect strategy model to obtain the defect repair strategy of the second image.

[0017] By adopting the above technical solution, the ultrasonic phased array technology is used to monitor the defects of welded structural parts in real time, and the defects are identified and classified based on the current welding process parameters and the detected images, improving the efficiency and accuracy of welding defect detection and evaluation. In addition, based on the evaluation results of welding defects and the current welding process parameters, they are input into the defect repair strategy model to obtain a targeted repair strategy, achieving the technical effect of adjusting the welding process according to the defect type, which helps to quickly identify welding defects occurring during the welding process and timely adjust the welding process according to the situation of welding defects, achieving the purpose of efficiently and accurately repairing welding defects.

[0018] Optionally, after the step of performing the first monitoring and before the step of performing the first discrimination, the following steps are further included:

[0019] First acquisition: Collect the current ambient temperature and ambient humidity; obtain the standard temperature and standard humidity;

[0020] First calculation: Calculate the difference between the ambient temperature and the standard temperature, denoted as the first difference, and calculate the difference between the ambient humidity and the standard humidity, denoted as the second difference;

[0021] Speed adjustment: Obtain the standard ultrasonic propagation speed, calculate the ultrasonic propagation speed adjustment value of the ultrasonic phased array based on the ambient compensation formula, and correct the ultrasonic propagation speed of the ultrasonic phased array based on the calculated ultrasonic propagation speed adjustment value;

[0022] The calculation model of the ambient compensation formula is:

[0023] ;

[0024] Wherein, is the ultrasonic propagation speed adjustment value, is the standard ultrasonic propagation speed, is the influence coefficient of temperature on the ultrasonic propagation speed, is the first difference, is the influence coefficient of humidity on the ultrasonic propagation speed, is the second difference, is the comprehensive influence coefficient of temperature and humidity;

[0025] Second monitoring: Based on the corrected ultrasonic propagation speed, correct the first image to obtain the corrected first image, and update the corrected first image as the first image.

[0026] By adopting the above technical solution, the propagation speed of ultrasonic waves is adjusted based on the environmental compensation mechanism, improving the accuracy of ultrasonic detection results, making the welding defect detection more stable and reliable, not being interfered by environmental changes. At the same time, based on the corrected ultrasonic wave propagation speed and the corrected first image, a more accurate image basis is provided for subsequent defect classification and defect evaluation, optimizing the subsequent defect repair strategy.

[0027] Optionally, after the step of performing the first monitoring and before the step of performing the first discrimination, it further includes:

[0028] First construction: Construct a defect recognition model;

[0029] Sample collection: Collect images containing complete welding defects and their corresponding complete labels, denoted as the first sample, and collect images containing incomplete welding defects and their corresponding incomplete labels, denoted as the second sample; The integrity labels of welding defects include complete labels and incomplete labels;

[0030] First training: Use the first sample and the second sample to train the defect recognition model to obtain the trained defect recognition model;

[0031] Image segmentation: Extract the welding defects in the first image, and segment the first image based on the extracted welding defects to obtain several sub-images containing welding defects;

[0032] Defect monitoring: Input the sub-images containing welding defects into the trained defect recognition model, and output the sub-images labeled with integrity labels;

[0033] Defect screening: Screen out the sub-images labeled with incomplete labels from the sub-images labeled with integrity labels, denoted as incomplete images;

[0034] Defect filling: Adopt the fractal image generation technology to fill the incomplete welding defects in the incomplete images based on the preset fractal dimension to obtain complete welding defects, and denote the filled incomplete images as filled images;

[0035] Parameter adjustment: Extract the features of the welding defects in the filled images, adjust the parameters of the ultrasonic phased array probe based on the extracted features, and perform the step of the first monitoring using the ultrasonic phased array with adjusted parameters.

[0036] By adopting the above technical solution, incomplete defects are identified, and the fractal image generation technology is used to fill the defects in the incomplete image. The ultrasonic probe parameters are adjusted by extracting the welding defect characteristics after filling and the welding defect characteristics before filling. Based on the adjusted ultrasonic probe parameters, the accuracy and sensitivity of ultrasonic waves during the detection process are improved, making the detection of welding defects more accurate, providing a more accurate image basis for subsequent defect classification and defect evaluation, optimizing the subsequent defect repair strategy, and further improving the welding quality of welding defect repair. Moreover, adjusting the ultrasonic probe parameters based on filling the defects in the incomplete image helps to reduce the number of adjustments of the ultrasonic probe parameters, more accurately locate the ultrasonic probe parameters, and improve the detection efficiency of welding defects.

[0037] Optionally, after performing the step of defect repair and before performing the step of parameter adjustment, it further includes:

[0038] First verification: The gray-level co-occurrence matrix of the incomplete welding defects in the incomplete image is constructed by using the gray-level co-occurrence matrix method, denoted as the first matrix. The gray-level co-occurrence matrix of the remaining part containing defects in the filled image is constructed by using the gray-level co-occurrence matrix method, denoted as the second matrix;

[0039] Second verification: Calculate the contrast of the first matrix, denoted as the first contrast, calculate the contrast of the second matrix, denoted as the second contrast, calculate the contrast difference between the first contrast and the second contrast, denoted as the third difference, and determine whether the third difference is less than the preset contrast difference threshold:

[0040] If so, perform the step of parameter adjustment;

[0041] If not, adjust the preset fractal dimension and perform the step of defect filling.

[0042] By adopting the above technical solution, the spatial relationship between the pixel gray values of the welding defect image is analyzed by using the gray-level co-occurrence matrix method, the structural and texture differences between the filled image and the original image are identified, and the filling effect can be quantified by calculating the contrast of the gray-level co-occurrence matrix of the images before and after filling, verifying the filled image, and improving the consistency and accuracy of defect filling. In addition, by dynamically verifying and adjusting the preset fractal dimension, the filled defect image more conforms to the characteristics of the defect before filling, which helps to improve the accuracy and sensitivity of ultrasonic waves during the detection process based on the filled defect image, making the detection of welding defects more accurate, providing a more accurate image basis for subsequent defect classification and defect evaluation, optimizing the subsequent defect repair strategy, and further improving the welding quality of welding defect repair.

[0043] Optionally, the step of parameter adjustment further includes:

[0044] First processing: In the filled image, the incomplete welding defect before filling is denoted as the original part, and the remaining part containing defects in the filled image is denoted as the filled part;

[0045] Coordinate construction: Taking any pixel point in the filled image as the origin, taking the length direction of the filled image as the x-axis, and taking the width direction of the filled image as the y-axis, a coordinate system is constructed;

[0046] Second calculation: Using the edge detection method to obtain the coordinates of the boundary points of the original part, based on the coordinates of the boundary points of the original part, calculating the central coordinates of the original part, and denoting the calculated central coordinates of the original part as the first center;

[0047] Third calculation: Using the edge detection method to obtain the coordinates of the boundary points of the complete welding defect in the filled image, based on the coordinates of the boundary points of the complete welding defect, calculating the central coordinates of the defect, and denoting the calculated central coordinates of the defect as the second center;

[0048] Angle calculation: Denoting the vector from the origin to the first center as the first vector, calculating the angle between the direction of the first vector and the positive direction of the x-axis, denoted as the first angle, denoting the vector from the origin to the second center as the second vector, calculating the angle between the direction of the second vector and the positive direction of the x-axis, denoted as the second angle;

[0049] First adjustment: Obtaining the initial angle of the ultrasonic phased array probe, based on the initial angle, the first angle and the second angle, calculating the adjustment angle through proportional operation, and taking the adjustment angle as the adjusted angle of the ultrasonic phased array probe.

[0050] By adopting the above technical solution, calculating the central coordinates of the welding defect before filling and the central coordinates of the complete welding defect after filling, constructing the corresponding vectors based on the central coordinates of the welding defect before and after filling, calculating the central angle difference of the defect before and after filling through the vectors of the welding defect before and after filling, and calculating the angle of the ultrasonic phased array probe based on the angle difference, so as to improve the accuracy and detection range of ultrasonic waves in the detection process, better adapt to welding defects in different directions or shapes, and enhance the reliability of ultrasonic detection.

[0051] Optionally, after performing the steps of the first adjustment, it further includes:

[0052] Fourth calculation: Based on the second angle, in the filled image, updating the second center to the origin of the coordinate system, taking the direction of the second vector as the positive direction of the new x-axis, and constructing a new y-axis and a new z-axis based on the new x-axis;

[0053] First projection: Projecting the boundary points of the complete welding defect onto the new x-axis, calculating the difference between the coordinate value of the projection point with the maximum coordinate value on the new x-axis and the coordinate value of the projection point with the minimum coordinate value on the new x-axis, and denoting the obtained difference as the first length;

[0054] Second projection: Project the boundary points of the complete welding defect onto the new y-axis, calculate the difference between the coordinate value of the projection point with the maximum coordinate value on the new y-axis and the coordinate value of the projection point with the minimum coordinate value on the new y-axis, and denote the obtained difference as the first width;

[0055] Third projection: Project the boundary points of the complete welding defect onto the new z-axis, calculate the difference between the coordinate value of the projection point with the maximum coordinate value on the new z-axis and the coordinate value of the projection point with the minimum coordinate value on the new z-axis, and denote the obtained difference as the first height;

[0056] Second adjustment: Multiply the first length, the first width and the first height to obtain the first volume, obtain the initial frequency of the ultrasonic phased array probe and the preset standard volume, based on the initial frequency, the preset standard volume and the first volume, calculate the first frequency according to the frequency formula, and use the calculated first frequency as the adjusted frequency of the ultrasonic phased array probe;

[0057] The calculation model of the frequency formula is:

[0058] ;

[0059] Wherein, is the first frequency, is the initial frequency of the ultrasonic phased array probe, is the preset standard volume, is the first volume.

[0060] By adopting the above technical solution, a coordinate system is constructed with the geometric center of the filled welding defect as the origin. Based on the filled welding defect, the maximum values projected by the filled welding defect on different coordinate axes are calculated in the constructed coordinate system. Based on the maximum values projected on different coordinate axes, the volume of the repaired welding defect is calculated. Based on the calculated volume, the frequency of the ultrasonic phased array probe is adjusted, so that the frequency of the ultrasonic probe better adapts to the actual size of the defect, ensuring the accuracy and efficiency of ultrasonic detection. Especially in the case of complex welding structures and diverse defect types, through the above dynamic adjustment mechanism, the adaptability of ultrasonic detection is improved, and the possibility of false alarms and missed alarms of welding defects is reduced.

[0061] Optionally, after the step of performing the first discrimination and before the step of obtaining the strategy, it further includes:

[0062] First extraction: Extract the geometric features of the welding defect in the second image and denote them as the first features;

[0063] Second extraction: Extract the size features of the welding defect in the second image and denote them as the second features;

[0064] Third extraction: Obtain the position data of the boundary points of the key load-bearing area at the weld to be monitored, extract the position data of the boundary points of the corresponding area where the welding defect is located in the second image, calculate the Euclidean distance between the boundary points of the corresponding area where the welding defect is located in the second image and the boundary points of the key load-bearing area, and record the minimum Euclidean distance as the third feature;

[0065] Second processing: Perform encoding processing on the first feature, second feature, and third feature of the welding defect to obtain the encoded features, and perform weighted summation on the encoded features to obtain the first data of the welding defect in the second image;

[0066] Second discrimination: Determine whether the first data of the welding defect is not higher than the preset parameter threshold:

[0067] If so, execute the steps of the first setting;

[0068] If not, execute the steps of strategy acquisition.

[0069] By adopting the above technical solution, by extracting the geometric features, size features of the welding defect, and the Euclidean distance between the defect and the key load-bearing area, a comprehensive quantitative analysis of the welding defect can be carried out. Based on the quantitative analysis, the evaluation steps of the welding defect are optimized, the accuracy of the welding quality evaluation is improved, so that the welds with welding defects but meeting the corresponding quality standards do not need to execute the defect repair steps, achieving the purpose of saving repair costs and time, avoiding unnecessary repair work, and improving the timeliness and efficiency of adjusting the welding process to repair welding defects.

[0070] Optionally, after executing the steps of the second discrimination and before executing the steps of the first setting, it further includes:

[0071] Third discrimination: Judge whether the number of defect category labels of the second image is 1:

[0072] If so, execute the steps of the first setting;

[0073] If not, execute the steps of the third processing;

[0074] Third processing: Segment the second image based on the defect category label to obtain several welding defect images, and integrate the obtained several welding defect images into a third image set;

[0075] Fifth calculation: Use the box-counting method to calculate the fractal dimension of each welding defect image in the third image set to obtain several fractal dimensions, and collect them to obtain the first dimension set;

[0076] First judgment: Judge whether each fractal dimension in the first dimension set is lower than the preset dimension threshold:

[0077] If so, execute the steps of the first setting;

[0078] If not, execute the step of obtaining the strategy.

[0079] By adopting the above technical solution, the evaluation of welding defects is further optimized. Classification processing is carried out based on the number of welding defects to distinguish between single defects and multiple defects, which helps to carry out refined evaluation of welding defects in welds with different complexities. The box counting method is used to calculate the fractal dimension of the welding defect image, providing more refined feature data for each welding defect, which helps to distinguish different types and forms of defects and identify the nature and complexity of each welding defect, improving the accuracy of welding defect evaluation. Welds with welding defects but meeting the corresponding quality standards do not need to execute the defect repair step, achieving the purpose of saving repair costs and time, avoiding unnecessary repair work, and improving the timeliness and efficiency of adjusting the welding process to repair welding defects.

[0080] Optionally, after executing the step of obtaining the strategy, it further includes:

[0081] Third monitoring: Execute the defect repair strategy of the obtained second image, record the weld after executing the defect repair strategy as the first weld, use ultrasonic phased array to detect the first weld of the welded structural member, and obtain the first verification image; Obtain the welding process parameters after executing the defect repair strategy, denoted as the first welding process parameters;

[0082] Second judgment: Input the first verification image and the first welding process parameters into the trained defect classification model to obtain the first verification image marked with the defect category label, denoted as the second verification image, and determine whether there are welding defects in the second verification image:

[0083] If not, execute the step of the first setting;

[0084] If so, execute the step of quality discrimination;

[0085] Quality discrimination: Use the second verification image as the new second image, sequentially execute the steps from the first extraction to the second processing to obtain the first data of the welding defects in the new second image, and determine whether the first data of the welding defects in the new second image is not higher than the preset parameter threshold:

[0086] If so, execute the step of the third judgment;

[0087] If not, execute the step of model update;

[0088] Third judgment: Count the number of defect category labels in the second verification image, denoted as the first quantity, count the number of defect category labels in the second image, denoted as the second quantity, and determine whether the first quantity is less than the second quantity;

[0089] If so, perform the steps of the first setting;

[0090] If not, perform the steps of model update;

[0091] Model update: Use the second image, the corresponding defect repair strategy, and the first welding process parameters as negative sample data for the defect strategy model. Train the model based on the negative sample data to obtain a new trained defect strategy model, and then re - execute the steps of strategy acquisition.

[0092] By adopting the above - mentioned technical solution, re - identify and detect the defects of the repaired weld seam, evaluate the quality of the repaired weld seam, and judge whether the obtained repair strategy can repair the welding defects. If the defect repair does not meet the standard, input the defect repair strategy and welding process parameters as negative sample data into the defect strategy model for model update, so that the defect repair strategy model can be continuously adjusted according to the actual situation to provide a more accurate repair strategy, reduce the risk of repair failure, and improve the accuracy and stability of the welding process for repairing welding defects.

[0093] In a second aspect, the present application provides a welding structure monitoring system based on the vibration - acoustics method, adopting the following technical solution:

[0094] A welding structure monitoring system based on the vibration - acoustics method, the system includes:

[0095] A processor and a memory;

[0096] Program code is stored in the memory;

[0097] When the processor calls the program code in the memory, it executes the steps of the method described in the first aspect above.

[0098] By adopting the above - mentioned technical solution, the defects of the welding structure are monitored in real - time, the defects are identified and classified, and the efficiency and accuracy of welding defect detection and evaluation are improved. In addition, based on the evaluation results of welding defects and the current welding process parameters, input them into the defect repair strategy model to obtain a targeted repair strategy, achieving the technical effect of adjusting the welding process according to the defect type, which helps to quickly identify the welding defects occurring in the welding process and timely adjust the welding process according to the situation of the welding defects, achieving the purpose of efficiently and accurately repairing welding defects.

[0099] In summary, the present application includes at least one of the following beneficial technical effects:

[0100] 1. Real-time monitoring of the defects of welded structural parts, defect identification and classification are carried out, which improves the efficiency and accuracy of welding defect detection and evaluation. In addition, based on the evaluation results of welding defects and the current welding process parameters, they are input into the defect repair strategy model to obtain targeted repair strategies, achieving the technical effect of adjusting the welding process according to the defect type, which helps to quickly identify the welding defects occurring during the welding process and timely adjust the welding process according to the situation of the welding defects, achieving the purpose of efficiently and accurately repairing welding defects.

[0101] 2. Identify incomplete defects, use fractal image generation technology to fill the defects in the incomplete image, and verify the filled image. Based on the verified image, extract the welding defect characteristics after filling and the welding defect characteristics before filling to adjust the angle and frequency of the ultrasonic probe, so as to improve the accuracy and detection range of ultrasonic waves during the detection process, which helps the ultrasonic phased array probe to better adapt to welding defects in different directions or forms, improves the reliability of ultrasonic detection, makes the detection of welding defects more accurate, provides a more accurate image basis for subsequent defect classification and defect evaluation, optimizes the subsequent defect repair strategy, and further improves the welding quality of welding defect repair.

[0102] 3. By extracting the geometric characteristics, size characteristics of welding defects and the Euclidean distance between the defects and the key load-bearing areas, a comprehensive quantitative analysis of welding defects can be carried out. Based on the quantitative analysis, the evaluation steps of welding defects are optimized, the accuracy of welding quality evaluation is improved, so that the welds with welding defects but meeting the corresponding quality standards do not need to perform the steps of defect repair, achieving the purpose of saving repair costs and time, avoiding unnecessary repair work, and improving the timeliness and efficiency of adjusting the welding process to repair welding defects.

[0103] 4. Re-identify and detect the defects of the repaired weld, evaluate the quality of the repaired weld, and judge whether the obtained repair strategy can repair the welding defects. If the defect repair fails to meet the standard, the defect repair strategy and welding process parameters are used as negative sample data to input into the defect strategy model for model update, so that the defect repair strategy model is continuously adjusted according to the actual situation to provide a more accurate repair strategy, reducing the risk of repair failure and improving the accuracy and stability of the welding process to repair welding defects. Brief Description of the Drawings

[0104] Figure 1 is the flowchart of Embodiment 1 of the present application;

[0105] Figure 2 is the flowchart of S41 image correction of Embodiment 2 of the present application;

[0106] Figure 3It is a flowchart of adjusting parameters in S42 of Embodiment 3 of this application;

[0107] Figure 4 It is a flowchart of Embodiment 4 of this application;

[0108] Figure 5 It is a flowchart of Embodiment 5 of this application. Detailed implementation manners

[0109] The following is combined with Figures 1 to 5 to further elaborate on this application in detail.

[0110] Embodiment 1: This embodiment discloses a monitoring method for welded structural parts based on the vibration acoustics method. As Figure 1 shown, the method includes: collecting historical welding process parameters, historical welding defect images and corresponding defect category labels, defect repair strategies, constructing a defect classification model and a defect strategy model, training the defect classification model and the defect strategy model based on the collected data to obtain a trained defect classification model and a trained defect strategy model, using ultrasonic phased array to detect the weld to be detected of the welded structural part to obtain a first image, obtaining the current welding process parameters, inputting the current welding process parameters and the first image into the trained defect classification model to obtain a second image, and determining whether there are welding defects in the second image: if not, setting the welding quality evaluation result of the weld to be detected as excellent and outputting the welding quality evaluation result; if so, inputting the current welding process parameters and the second image into the trained defect strategy model to obtain the defect repair strategy of the second image. This embodiment includes the following steps:

[0111] S1 Data collection: Collect historical welding process parameters, historical welding defect images, and corresponding defect category labels, defect repair strategies.

[0112] The historical welding process parameters include welding current, welding voltage, welding machine speed, heat input, wire diameter, welding position, shielding gas type, shielding gas flow rate, welding electrode, welding material, welding preheating temperature, post-weld heat treatment temperature, welding gap, welding duration, ambient temperature, and ambient humidity.

[0113] The historical welding defect image is a three-dimensional image of the weld welded under the historical welding process parameters collected by ultrasonic phased array technology.

[0114] The welding defects include porosity, slag inclusions, cracks, lack of fusion, and lack of penetration. Porosity includes single porosity, stringy porosity, needle-like porosity, and root porosity. Slag inclusions include internal slag inclusions, root slag inclusions, interlayer slag inclusions, and groove slag inclusions. Cracks include hot cracks, cold cracks, and shrinkage cracks. Lack of fusion includes root lack of fusion and sidewall lack of fusion. Lack of penetration includes partial lack of penetration and complete lack of penetration. Defect category labels include porosity - single porosity, porosity - stringy porosity, porosity - needle-like porosity, porosity - root porosity, slag inclusion - internal slag inclusion, slag inclusion - root slag inclusion, slag inclusion - interlayer slag inclusion, slag inclusion - groove slag inclusion, crack - hot crack, crack - cold crack, crack - shrinkage crack, lack of fusion - root lack of fusion, lack of fusion - sidewall lack of fusion, lack of penetration - partial lack of penetration, and lack of penetration - complete lack of penetration.

[0115] The defect repair strategy is a repair strategy that can solve the above various welding defects.

[0116] In this embodiment, it is necessary to preprocess the collected data. The steps of the data preprocessing include data cleaning and data normalization.

[0117] Data cleaning: Perform operations such as denoising, handling missing values, and removing outliers on the collected data.

[0118] Data normalization: Perform normalization processing on the collected data.

[0119] S2 Model construction: Use a deep convolutional neural network model as the basic model of the defect classification model to construct a defect classification model. Models such as convolutional neural networks, recurrent neural networks, and long short-term memory network models can be used as the basic models of the defect strategy model to construct a defect strategy model.

[0120] In this embodiment, a defect classification model is constructed using the deep convolutional neural network model VGGNet. The defect strategy model can have the function of outputting a defect repair strategy after being trained with historical welding process parameters, historical welding defect images, corresponding defect category labels, and defect repair strategies.

[0121] S3 Model training: Use historical welding process parameters, historical welding defect images, and corresponding defect category labels to train the defect classification model to obtain a trained defect classification model. Use historical welding process parameters, historical welding defect images, corresponding defect category labels, and defect repair strategies to train the defect strategy model to obtain a trained defect strategy model.

[0122] The model training process of the defect classification model includes: labeling defect category labels on corresponding historical welding defect images to obtain first classification samples, labeling historical welding process parameters as parameter labels on the first classification samples to obtain second classification samples, collecting each second classification sample to obtain a classification sample set, and using the classification sample set to train the defect classification model to obtain a trained defect classification model. The trained defect classification model has the function of outputting a welding defect image labeled with a defect category label according to the input welding defect image and the corresponding welding process parameters.

[0123] The model training process of the defect strategy model includes: labeling defect category labels on corresponding historical welding defect images to obtain first strategy samples, labeling historical welding process parameters as parameter labels on the first strategy samples to obtain second strategy samples, collecting the second strategy samples and the corresponding defect repair strategies as a group of samples to obtain a strategy sample set, and using the strategy sample set to train the defect strategy model to obtain a trained defect strategy model. The trained defect strategy model has the function of outputting the defect repair strategy of the welding defect image according to the welding defect image, the corresponding defect category label, and the welding process parameters.

[0124] S4 First monitoring: Use ultrasonic phased array technology to scan the weld to be detected of the welded structural part from different angles to obtain a three-dimensional detection image of the weld to be detected, denoted as the first image.

[0125] S5 First discrimination: Obtain the current welding process parameters, input the current welding process parameters and the first image into the trained defect classification model to obtain a first image labeled with a defect category label, denoted as the second image, and determine whether there is a welding defect in the second image by determining whether there is a defect category label in the second image.

[0126] If there is no defect category label in the second image, that is, there is no welding defect in the second image, then execute S6 First setting.

[0127] If there is a defect category label in the second image, that is, there is a welding defect in the second image, then execute S7 Strategy acquisition.

[0128] S6 First setting: Set the welding quality evaluation result of the weld to be detected to excellent and output the welding quality evaluation result. The welding quality evaluation result being excellent indicates that the weld to be detected is qualified and there is no need to perform welding defect repair operations.

[0129] S7 Strategy acquisition: Input the current welding process parameters and the second image into the trained defect strategy model to obtain the defect repair strategy of the second image, and perform defect repair operations according to the output defect repair strategy to correct the repaired defects.

[0130] In this embodiment, the ultrasonic phased array technology is used to monitor the defects of welded structural parts in real time. Based on the current welding process parameters and the detected images, the defects are identified and classified, which improves the efficiency and accuracy of welding defect detection and evaluation. In addition, based on the evaluation results of welding defects and the current welding process parameters, they are input into the defect repair strategy model to obtain a targeted repair strategy, achieving the technical effect of adjusting the welding process according to the defect type. This helps to quickly identify welding defects occurring during the welding process and timely adjust the welding process according to the situation of welding defects, achieving the purpose of efficiently and accurately repairing welding defects.

[0131] Embodiment 2: The difference from Embodiment 1 is that:

[0132] As Figure 2 shown, after performing the first monitoring in S4 and before performing the first discrimination in S5, it further includes S41 image correction. S41 image correction includes S411 first acquisition, S412 first calculation, S413 speed adjustment, and S414 second monitoring.

[0133] S411 First acquisition: The current ambient temperature and ambient humidity are collected through a temperature sensor and a humidity sensor. The standard atmospheric conditions for ultrasonic waves in the air under standard propagation conditions are obtained. The standard atmospheric conditions include standard temperature and standard humidity. In this embodiment, the standard temperature is 20°C and the standard humidity is 50%.

[0134] S412 First calculation: Calculate the difference between the collected ambient temperature and the standard temperature, denoted as the first difference. Calculate the difference between the collected ambient humidity and the standard humidity, denoted as the second difference.

[0135] S413 Speed adjustment: Obtain the standard ultrasonic wave propagation speed under standard atmospheric conditions. In this embodiment, the standard ultrasonic wave propagation speed is 343.2 m / s. Based on the environmental compensation formula, calculate the ultrasonic wave propagation speed adjustment value of the ultrasonic phased array. Based on the obtained ultrasonic wave propagation speed adjustment value, correct the ultrasonic wave propagation speed of the ultrasonic phased array, that is, add the ultrasonic wave propagation speed adjustment value to the standard ultrasonic wave propagation speed to obtain the actual ultrasonic wave propagation speed under the current ambient temperature and ambient humidity conditions.

[0136] The calculation model of the environmental compensation formula is:

[0137] .

[0138] Among them, is the ultrasonic wave propagation speed adjustment value, is the standard ultrasonic wave propagation speed, is the influence coefficient of temperature on the ultrasonic wave propagation speed, is the first difference, is the influence coefficient of humidity on the ultrasonic wave propagation speed, is the second difference, is the comprehensive influence coefficient of temperature and humidity, that is, the comprehensive influence coefficient of temperature and humidity on the ultrasonic wave propagation speed, . In this embodiment, the influence coefficient of temperature on the ultrasonic wave propagation speed and the influence coefficient of humidity on the ultrasonic wave propagation speed can be obtained through experimental tests and theoretical derivation based on the experimental test data. The comprehensive influence coefficient of temperature and humidity can be obtained through experimental tests and fitted by the multiple linear regression method based on the experimental test data. In this embodiment, the influence coefficient of temperature on the ultrasonic wave propagation speed takes a value of 0.6 m / s∙°C, the influence coefficient of humidity on the ultrasonic wave propagation speed takes a value of 0.01 m / s, and the comprehensive influence coefficient of temperature and humidity takes a value of 0.01 m / s∙°C.

[0139] In this embodiment, the unit of humidity is %, and when calculating, the humidity expressed as a percentage is converted into a decimal form to participate in the calculation of the environmental compensation formula.

[0140] S414 Second monitoring: Based on the corrected ultrasonic wave propagation speed, obtain the ultrasonic wave propagation time before correction. Based on the corrected ultrasonic wave propagation speed and the ultrasonic wave propagation time before correction, calculate the distance from each pixel point in the corrected first image to the probe. Apply the distance from each pixel point in the corrected first image to the first image, update the position information of each point in the first image, take the updated first image as the corrected first image, and update the corrected first image as the first image.

[0141] In this embodiment, based on the environmental compensation mechanism, the changes in temperature and humidity are compensated, the ultrasonic wave propagation speed is corrected, the accuracy, stability, and accuracy of the ultrasonic detection result are improved, the anti-interference ability of the welding defect detection is improved, and at the same time, based on the corrected ultrasonic wave propagation speed and the corrected first image, a more accurate image basis is provided for subsequent defect classification and defect evaluation, and the subsequent defect repair strategy is optimized.

[0142] Embodiment 3: The difference from Embodiment 1 is that:

[0143] As Figure 3As shown, after the first monitoring in S4 and before the first discrimination in S5, it further includes S42 parameter adjustment. S42 parameter adjustment includes S421 first construction, S422 sample collection, S423 first training, S424 image segmentation, S425 defect monitoring, S426 defect screening, S427 defect filling, S428 filling verification, and S429 parameter adjustment.

[0144] S421 first construction: Use the deep convolutional neural network model as the basic model of the defect recognition model to construct the defect recognition model. In this embodiment, the deep convolutional neural network model VGGNet is used to construct the defect recognition model.

[0145] S422 sample collection: Collect images containing complete welding defects and their corresponding complete labels, denoted as the first sample, and collect images containing incomplete welding defects and their corresponding incomplete labels, denoted as the second sample; the integrity labels of welding defects include complete labels and incomplete labels. For example: Collect an image containing a complete bubble, and the complete label of the image of the complete bubble is "complete", that is, the first sample is the complete bubble image - complete. Collect an image with only a bubble, and the incomplete label of the image of the bubble is "incomplete", that is, the second sample is the image of the bubble - incomplete.

[0146] S423 first training: Pool the collected first samples and second samples, and denote the pooled samples as the recognition sample set. Use the recognition sample set to train the defect recognition model to obtain the trained defect recognition model. The trained defect recognition model has the function of outputting a welding defect image marked with an integrity label according to the input welding defect image.

[0147] S424 image segmentation: Extract the welding defects in the first image, and segment the first image based on the extracted welding defects to obtain several sub-images containing welding defects.

[0148] Perform image preprocessing on the first image. The preprocessing includes using three-dimensional Gaussian filtering to remove noise in the first image; using histogram equalization method and Canny edge detection algorithm to enhance the first image.

[0149] Use methods such as threshold segmentation, edge detection, or constructing an image segmentation model to extract each welding defect in the first image, and segment the first image based on each extracted welding defect to obtain several sub-images containing only one welding defect. In this embodiment, the Canny edge detection algorithm is used to extract each welding defect in the first image.

[0150] S425 Defect Monitoring: Input all the obtained sub-images each containing only one welding defect into the trained defect recognition model, and output the sub-images marked with integrity labels accordingly.

[0151] S426 Defect Screening: Screen out the sub-images marked with incomplete labels from the obtained sub-images marked with integrity labels, record them as incomplete images, and gather all the screened incomplete images to obtain an incomplete image set.

[0152] S427 Defect Filling: Adopt the fractal image generation technology, and based on the preset fractal dimension, fill the incomplete welding defects in the incomplete images to obtain complete welding defects, and record the filled incomplete images as filled images.

[0153] Based on the different defect categories of the welding defects in each incomplete image in the incomplete image set, set different fractal dimension ranges for each type of welding defect. For example, the set fractal dimension range for porosity defects is [1.5, 2.0], the set fractal dimension range for slag inclusions defects is [2.0, 2.5], the set fractal dimension range for crack defects is [1.7, 2.2], the set fractal dimension range for lack of fusion defects is [2.0, 2.5], and the set fractal dimension range for lack of penetration defects is [2.0, 2.5]. Based on the set fractal dimensions of different welding defects, compare the fractal dimension of the welding defects in each incomplete image with the preset filling dimension threshold. For the welding defects with a fractal dimension greater than or equal to the preset filling dimension threshold, use methods such as fractal noise interpolation, fractal model interpolation, or multi-scale interpolation to generate a filling area that conforms to the surrounding area of the welding defect and fill the welding defect; for the welding defects with a fractal dimension less than the preset filling dimension threshold, use methods such as linear interpolation or mean interpolation to generate a filling area that conforms to the surrounding area of the welding defect and fill the welding defect. Smooth the filling area to eliminate the phenomenon that the boundary of the filling area is too obvious, obtain complete welding defects, and record the filled incomplete images as filled images.

[0154] In order to more precisely fill the welding defects in the incomplete images, in other embodiments, classification processing of the fractal dimension can be performed based on the structure of the welding defects. For the welding defects with a simple structure, defect filling is performed according to the preset fractal dimension, and for the welding defects with a complex structure, the fractal dimension of the welding defects can be calculated by methods such as the box-counting method for defect filling.

[0155] S428 Filling Verification: Includes S4281 First Verification and S4282 Second Verification.

[0156] S4281 First verification: The gray-level co-occurrence matrix of the incomplete welding defects in the incomplete image is constructed using the gray-level co-occurrence matrix method, denoted as the first matrix. The gray-level co-occurrence matrix of the remaining part containing defects in the filled image is constructed using the gray-level co-occurrence matrix method, denoted as the second matrix.

[0157] Set the gray levels of the incomplete welding defects in the incomplete image to L discrete values, and set 8 directions, with the horizontal direction from left to right as 0 o , rotate clockwise, and the 8 directions are 0 o , 45 o , 90 o , 135 o , 22.5 o , 67.5 o , 112.5 o and 157.5 o . Set the distance between pixel pairs. In this embodiment, the distance between pixel pairs is set to 1 pixel. In the incomplete welding defects, count the gray-level combinations between each pixel and its adjacent pixels in 8 directions. Based on different directions, count the probability of each gray-level combination appearing in the incomplete welding defects, and fill the statistical results into the L×L co-occurrence matrix to obtain the co-occurrence matrix in each direction. Denote the co-occurrence matrix in each obtained direction as the first matrix. According to the above steps, use the gray-level co-occurrence matrix method to construct the gray-level co-occurrence matrix of the remaining part containing defects in the filled image to obtain the second matrix.

[0158] S4282 Second verification: Calculate the contrast of each co-occurrence matrix in the first matrix respectively according to the contrast calculation formula, and calculate using the arithmetic mean method or the weighted average method based on the contrast of each co-occurrence matrix in the first matrix. Denote the calculation result as the first contrast. Calculate the contrast of each co-occurrence matrix in the second matrix respectively according to the contrast calculation formula, and calculate the contrast of each co-occurrence matrix in the second matrix using the arithmetic mean method or the weighted average method. Denote the calculation result as the second contrast. Calculate the contrast difference between the first contrast and the second contrast, denoted as the third difference, and determine whether the third difference is less than the preset contrast difference threshold:

[0159] If so, execute S429 parameter adjustment;

[0160] If not, in the fractal dimension range of the corresponding welding defect, select other preset fractal dimensions, and based on the selected preset fractal dimensions, execute S427 defect filling again.

[0161] In this embodiment, the arithmetic mean method is used to calculate the first contrast and the second contrast.

[0162] The contrast calculation formula is:

[0163] 。

[0164] Among them, represents the gray level of the th row in the co-occurrence matrix, represents the gray level of the th column in the co-occurrence matrix, represents the probability that the gray level combination of gray level and gray level appears.

[0165] S429 parameter adjustment: Extract the features of the welding defects in the filled image, adjust the parameters of the ultrasonic phased array probe based on the extracted features, and perform the first monitoring of S4 using the ultrasonic phased array with the adjusted parameters. The S429 parameter adjustment includes S4291 angle adjustment and S4292 frequency adjustment.

[0166] S4291 angle adjustment: Includes S42911 first processing, S42912 coordinate construction, S42913 second calculation, S42914 third calculation, S42915 angle calculation, and S42916 first adjustment.

[0167] S42911 first processing: In the filled image, mark the incomplete welding defect before filling as the original part, and mark the remaining part containing the defect in the filled image as the filled part, that is, the filled area in S427 defect filling.

[0168] S42912 coordinate construction: Take any pixel point in the filled image as the origin, take the length direction of the filled image as the x-axis, take the width direction of the filled image as the y-axis, and take the depth direction of the filled image as the z-axis to construct a coordinate system.

[0169] S42913 second calculation: Use the edge detection method to obtain the coordinates of the boundary points of the original part. Based on the coordinates of the boundary points of the original part, calculate the central coordinates of the original part, and mark the calculated central coordinates of the original part as the first center.

[0170] Use the edge detection method to obtain the coordinates of the boundary points of the original part, collect the coordinate data of the boundary points of the obtained original part, and mark it as the first data set. Based on each coordinate data in the first data set, calculate the average coordinates of all the boundary points of the original part, and use the calculated average coordinates as the central coordinates of the original part, and mark it as the first center.

[0171] S42914 third calculation: Use the edge detection method to obtain the coordinates of the boundary points of the complete welding defect in the filled image. Based on the coordinates of the boundary points of the complete welding defect, calculate the central coordinates of the complete welding defect, and mark the calculated central coordinates of the complete welding defect as the second center.

[0172] The edge detection method is used to obtain the coordinates of the boundary points of the complete welding defects in the filled image, and the coordinate data of the boundary points of the obtained complete welding defects are collected and recorded as the second data set. Based on each coordinate data in the second data set, the average coordinates of all the boundary points of the complete welding defects are calculated, and the calculated average coordinates are used as the center coordinates of the complete welding defects and recorded as the second center.

[0173] S42915 Angle calculation: Denote the vector from the origin to the first center as the first vector, and use the arctangent function to calculate the angle between the direction of the first vector and the positive x-axis direction, which is denoted as the first angle. Denote the vector from the origin to the second center as the second vector, and use the arctangent function to calculate the angle between the direction of the second vector and the positive x-axis direction, which is denoted as the second angle.

[0174] S42916 First adjustment: Obtain the initial angle of the ultrasonic phased array probe. Based on the initial angle, the first angle, and the second angle, calculate the adjustment angle through proportional operation, and use the adjustment angle as the angle of the ultrasonic phased array probe after adjustment.

[0175] Obtain the beam angle of each ultrasonic phased array probe. Based on the first center, calculate the direction angle of each probe to the first center, calculate the difference between the beam angle of each ultrasonic phased array probe and the direction angle of the probe to the first center, which is denoted as the angle difference of each probe, and use the probe corresponding to the calculated minimum angle difference as the ultrasonic phased array probe to be adjusted. Use the beam angle of the ultrasonic phased array probe to be adjusted as the initial angle. Based on the initial angle, the first angle, and the second angle, calculate the adjustment angle through proportional operation, and use the adjustment angle as the angle of the ultrasonic phased array probe after adjustment.

[0176] S4292 Frequency adjustment: It includes S42921 Fourth calculation, S42922 First projection, S42923 Second projection, S42924 Third projection, and S42925 Second adjustment.

[0177] S42921 Fourth calculation: Based on the second angle, in the filled image, update the second center as the origin of the coordinate system, use the direction of the second vector as the new positive x-axis direction, and construct a new y-axis and a new z-axis based on the new x-axis. Perform normalization processing on the second vector to obtain the unit vector of the new x-axis. Based on the second vector and the vector of the z-axis, calculate the unit vector of the new z-axis through cross product operation. Based on the second vector and the unit vector of the new z-axis, calculate the unit vector of the new y-axis through cross product operation. Construct a transformation matrix based on the unit vector of the new x-axis, the unit vector of the new y-axis, and the unit vector of the new z-axis, and calculate the new coordinates of each boundary point in the updated coordinate system in the coordinate system before update through matrix multiplication operation, and record the updated new coordinates as the coordinates of the corresponding boundary points.

[0178] S42922 First Projection: Project the boundary points of the complete welding defect onto the new x-axis, calculate the difference between the coordinate value of the projection point with the maximum coordinate value on the new x-axis and the coordinate value of the projection point with the minimum coordinate value on the new x-axis, and denote the obtained difference as the first length.

[0179] S42923 Second Projection: Project the boundary points of the complete welding defect onto the new y-axis, calculate the difference between the coordinate value of the projection point with the maximum coordinate value on the new y-axis and the coordinate value of the projection point with the minimum coordinate value on the new y-axis, and denote the obtained difference as the first width.

[0180] S42924 Third Projection: Project the boundary points of the complete welding defect onto the new z-axis, calculate the difference between the coordinate value of the projection point with the maximum coordinate value on the new z-axis and the coordinate value of the projection point with the minimum coordinate value on the new z-axis, and denote the obtained difference as the first height.

[0181] S42925 Second Adjustment: Perform a multiplication operation on the first length, the first width, and the first height to obtain the first volume. Obtain the initial frequency of the ultrasonic phased array probe and the preset standard volume. Based on the initial frequency, the preset standard volume, and the first volume, calculate the first frequency according to the frequency formula, and use the calculated first frequency as the adjusted frequency of the ultrasonic phased array probe.

[0182] The calculation model of the frequency formula is:

[0183] ;

[0184] Where, is the first frequency, is the initial frequency of the ultrasonic phased array probe, is the preset standard volume, that is, the volume of the area that the ultrasonic phased array probe can detect under standard working conditions, is the first volume.

[0185] In this embodiment, the ultrasonic phased array probe in the S42925 Second Adjustment is the same probe as the ultrasonic phased array probe in the S42916 First Adjustment.

[0186] In this embodiment, incomplete defects are identified, the fractal image generation technology is used to fill the defects in the incomplete image, and the filled image is verified. Based on the verified image, the characteristics of the welded defects after filling and before filling are extracted to adjust the angle and frequency of the ultrasonic probe, so as to improve the accuracy and detection range of ultrasonic waves during the detection process, which helps the ultrasonic phased array probe better adapt to welded defects in different directions or forms, improves the reliability of ultrasonic detection, makes the detection of welded defects more accurate, provides a more accurate image basis for subsequent defect classification and defect evaluation, optimizes the subsequent defect repair strategy, and further improves the welding quality of welded defect repair.

[0187] Embodiment 4: The difference from Embodiment 1 is that:

[0188] As Figure 4 shown, after the first discrimination in S5 and before obtaining the S7 strategy, it further includes: S51 first evaluation and S52 second evaluation.

[0189] The S51 first evaluation includes S511 first extraction, S512 second extraction, S513 third extraction, S514 second processing, and S515 second discrimination.

[0190] S511 first extraction: Extract the geometric features of the welded defects in the second image, denoted as the first feature. The geometric features include the geometric shape and position of the welded defects, and the geometric features of the welded defects can be obtained through image processing algorithms (such as edge detection method, contour extraction method, etc.).

[0191] S512 second extraction: Extract the size features of the welded defects in the second image, denoted as the second feature. The size features are mainly the length, width, height, and volume of the welded defects. The length, width, and height of the welded defects can be obtained through image processing algorithms (such as edge detection method, contour extraction method, etc.). Calculate the volume of the welded defects according to the shape of the welded defects and the corresponding length, width, and height in the geometric features of the welded defects obtained. For example, for regular-shaped welded defects (such as pores and slag inclusions, etc.), that is, the shape of the welded defects can be approximated as regular geometric shapes (such as spherical, elliptical, etc.), the volume of the welded defects can be calculated by the geometric formula corresponding to the shape. For irregular-shaped welded defects (such as cracks, etc.), the voxelization method is used to calculate the volume of the irregular-shaped welded defects.

[0192] S513 third extraction: Obtain the position data of the boundary points of the key load-bearing area at the weld to be monitored, extract the position data of the boundary points of the corresponding area where the welded defects are located in the second image, calculate the Euclidean distance between the boundary points of the corresponding area where the welded defects are located in the second image and the boundary points of the key load-bearing area, and denote the minimum Euclidean distance as the third feature.

[0193] Geometric modeling is carried out based on the welded structural member, the material properties of the welded structural member are defined, and the geometric model of the welded structural member is obtained. The material properties include Young's modulus, Poisson's ratio, yield strength, density, and coefficient of thermal expansion. Based on the preset grid volume, the geometric model of the welded structural member is divided into several grids, and the divided grids are denoted as finite elements. Gravity loads and pressure loads are applied to each finite element, the finite element analysis is defined as a static analysis, a finite element equation is constructed, and the finite element equation is solved to obtain the key load-bearing areas of the welded structural member. Among the obtained key load-bearing areas of the welded structural member, the key load-bearing area closest to the weld to be monitored is selected as the key load-bearing area of the weld to be monitored. The position data of all boundary points of the key load-bearing area of the weld to be monitored is obtained, the position data of the boundary points of the corresponding area where the welding defect is located in the second image is extracted, the Euclidean distance method is used to calculate the Euclidean distance between the boundary points of the corresponding area where the welding defect is located in the second image and the boundary points of the key load-bearing area, all the calculated Euclidean distances are collected, and the minimum Euclidean distance is selected and denoted as the third feature.

[0194] S514 Second processing: Normalize the first feature, second feature, and third feature of the welding defect, then perform encoding processing on the normalized features to obtain the encoded features, and perform weighted summation on the encoded features to obtain the first data of the welding defect in the second image.

[0195] In this embodiment, the method of hot encoding is used to perform encoding processing on the normalized features.

[0196] In this embodiment, in the weighted summation of the encoded features, the weight of each feature is set according to expert experience. The weight of the first feature of the welding defect is 0.4, the weight of the second feature of the welding defect is 0.3, and the weight of the third feature of the welding defect is 0.3.

[0197] S515 Second discrimination: Determine whether the first data of the welding defect is not higher than the preset parameter threshold:

[0198] If so, execute S52 Second evaluation;

[0199] If not, execute S7 strategy acquisition.

[0200] S52 Second evaluation: Includes S521 Third discrimination, S522 Quantity judgment, S523 Third processing, S524 Fifth calculation, and S525 First judgment.

[0201] S521 Third discrimination: Determine whether the number of defect category labels of the second image is 1:

[0202] If so, execute S6 First setting;

[0203] If not, perform the S522 quantity judgment.

[0204] S522 quantity judgment: Judge whether the number of defect category labels of the second image is less than the preset label number threshold:

[0205] If so, perform the S523 third processing;

[0206] If not, perform the S7 strategy acquisition.

[0207] S523 third processing: Based on the defect category labels, use the edge detection method to segment the second image to obtain several welding defect images each containing only one welding defect, and integrate the obtained several welding defect images into a third image set.

[0208] S524 fifth calculation: Use the box-counting method to calculate the fractal dimension of each welding defect image in the third image set to obtain several fractal dimensions, and gather them to obtain a first dimension set.

[0209] Set different sizes of the boxes, successively segment the welding defect images in descending order based on the size of the boxes, count the number of boxes that can completely cover the welding defects in the welding defect images at each box scale, calculate the logarithm of the box size based on the box size, calculate the logarithm of the number of boxes corresponding to the box size, use the logarithm of the box size as the x-axis and the logarithm of the number of boxes as the y-axis to construct a double-logarithm graph, and based on the constructed double-logarithm graph, use the linear fitting method to calculate the slope of the double-logarithm graph, and take the calculated slope as the fractal dimension of the welding defect image, and gather the fractal dimensions of all welding defect images to obtain a first dimension set.

[0210] S525 first judgment: Judge whether each fractal dimension in the first dimension set is lower than the preset dimension threshold:

[0211] If so, perform the S6 first setting;

[0212] If not, perform the S7 strategy acquisition.

[0213] In this embodiment, by extracting the geometric features, size features of the welding defects and the Euclidean distance between the defects and the key load-bearing areas, a comprehensive quantitative analysis of the welding defects can be carried out. Based on the quantitative analysis, the evaluation steps of the welding defects are optimized, the accuracy of the welding quality evaluation is improved, so that the welds with welding defects but meeting the corresponding quality standards do not need to perform the defect repair steps, achieving the purpose of saving repair costs and time, avoiding unnecessary repair work, and improving the timeliness and efficiency of adjusting the welding process to repair the welding defects.

[0214] Embodiment 5: The difference from Embodiment 1 is that:

[0215] As Figure 5 shown, after the S7 strategy acquisition is executed, it further includes S8 strategy verification. The S8 strategy verification includes S81 third monitoring, S82 second judgment, S83 quality discrimination, S84 third judgment, and S85 model update.

[0216] S81 third monitoring: Execute the defect repair strategy of the obtained second image. Denote the weld seam after executing the defect repair strategy as the first weld seam. Use ultrasonic phased array to detect the first weld seam of the welded structural part to obtain the first verification image. Obtain the welding process parameters after executing the defect repair strategy and denote them as the first welding process parameters.

[0217] S82 second judgment: Input the first verification image and the first welding process parameters into the trained defect classification model to obtain the first verification image marked with the defect category label, denoted as the second verification image, and determine whether there are welding defects in the second verification image.

[0218] If not, execute S6 first setting.

[0219] If so, execute S83 quality discrimination.

[0220] S83 quality discrimination: Take the second verification image as the new second image, and sequentially execute S511 first extraction to S514 second processing to obtain the first data of the welding defects in the new second image, and determine whether the first data of the welding defects in the new second image is not higher than the preset parameter threshold.

[0221] If so, execute S84 third judgment.

[0222] If not, execute S85 model update.

[0223] S84 third judgment: Count the number of defect category labels in the second verification image, denoted as the first quantity, count the number of defect category labels in the second image, denoted as the second quantity, and determine whether the first quantity is less than the second quantity.

[0224] If so, execute S6 first setting.

[0225] If not, execute S85 model update.

[0226] S85 model update: Take the second image, the corresponding defect repair strategy, and the first welding process parameters as the negative sample data of the defect strategy model, perform model training according to the negative sample data to obtain the new trained defect strategy model, and then re-execute the S7 strategy acquisition.

[0227] In this embodiment, the repaired weld seam is re-detected for defects, and the quality of the repaired weld seam is evaluated to determine whether the obtained repair strategy can repair the welding defects. If the defect repair fails to meet the standard, the defect repair strategy and welding process parameters are input as negative sample data into the defect strategy model for model update, so that the defect repair strategy model can be continuously adjusted according to the actual situation to provide a more accurate repair strategy, reducing the risk of repair failure and improving the accuracy and stability of the welding process for repairing welding defects.

[0228] Embodiment 6: This embodiment discloses a monitoring system for welded structural parts based on the vibration acoustics method. The system includes:

[0229] a processor and a memory,

[0230] wherein program codes are stored in the memory;

[0231] when the processor calls the program codes in the memory, it executes the steps of the method described in the above embodiment.

[0232] In this embodiment, the defects of the welded structural parts are monitored in real time, the defects are identified and classified, improving the efficiency and accuracy of welding defect detection and evaluation. In addition, based on the evaluation results of the welding defects and the current welding process parameters, they are input into the defect repair strategy model to obtain a targeted repair strategy, achieving the technical effect of adjusting the welding process according to the defect type, which helps to quickly identify the welding defects occurring during the welding process and timely adjust the welding process according to the situation of the welding defects, achieving the purpose of efficiently and accurately repairing welding defects.

[0233] The above are all the preferred embodiments of this application. Without limiting the protection scope of this application accordingly, therefore: All equivalent changes made according to the structure, shape, and principle of this application shall be covered within the protection scope of this application.

Claims

1. A monitoring method for welded structural components based on the vibration-acoustic method, characterized in that Including: Data collection: Collect historical welding process parameters, historical welding defect images, corresponding defect category labels, and defect repair strategies. Model construction: Construct a defect classification model and a defect strategy model. Model training: Use historical welding process parameters, historical welding defect images, and corresponding defect category labels to train the defect classification model to obtain a trained defect classification model. Use historical welding process parameters, historical welding defect images, and defect repair strategies to train the defect strategy model to obtain a trained defect strategy model. First monitoring: Use ultrasonic phased array to detect the weld to be inspected of the welded structural member to obtain a first image. First construction: Construct a defect recognition model. Sample collection: Collect images containing complete welding defects and corresponding complete labels, denoted as the first sample. Collect images containing incomplete welding defects and corresponding incomplete labels, denoted as the second sample. The integrity labels of welding defects include complete labels and incomplete labels. First training: Use the first sample and the second sample to train the defect recognition model to obtain a trained defect recognition model. Image segmentation: Extract the welding defects in the first image, and segment the first image based on the extracted welding defects to obtain several sub-images containing welding defects. Defect monitoring: Input the sub-images containing welding defects into the trained defect recognition model to output sub-images labeled with integrity labels. Defect screening: Screen out the sub-images labeled with incomplete labels from the sub-images labeled with integrity labels, denoted as incomplete images. Defect filling: Use the fractal image generation technology to fill the incomplete welding defects in the incomplete images based on the preset fractal dimension to obtain complete welding defects, and denote the filled incomplete images as filled images. Parameter adjustment: Extract the features of the welding defects in the filled images, adjust the parameters of the ultrasonic phased array probe based on the extracted features, and perform the steps of the first monitoring using the ultrasonic phased array with adjusted parameters. First discrimination: Obtain the current welding process parameters, input the current welding process parameters and the first image into the trained defect classification model to obtain a first image labeled with defect category labels, denoted as the second image, and determine whether there are welding defects in the second image: If not, execute the steps of the first setting. If so, execute the steps of strategy acquisition. First setting: Set the welding quality evaluation result of the weld to be inspected as excellent and output the welding quality evaluation result. Strategy acquisition: Input the current welding process parameters and the second image into the trained defect strategy model to obtain the defect repair strategy of the second image.

2. The monitoring method for welded structural members based on the vibro-acoustic method according to claim 1, characterized in that Before executing the steps of the first discrimination after executing the steps of the first monitoring, it also includes: First collection: Collect the current ambient temperature and ambient humidity; obtain the standard temperature and standard humidity. First calculation: Calculate the difference between the ambient temperature and the standard temperature, denoted as the first difference, and calculate the difference between the ambient humidity and the standard humidity, denoted as the second difference. Speed adjustment: Obtain the standard ultrasonic propagation speed, calculate the ultrasonic propagation speed adjustment value of the ultrasonic phased array based on the environmental compensation formula, and correct the ultrasonic propagation speed of the ultrasonic phased array based on the calculated ultrasonic propagation speed adjustment value; The calculation model of the environmental compensation formula is: ； Among them, is the ultrasonic wave propagation speed adjustment value, is the standard ultrasonic wave propagation speed, is the influence coefficient of temperature on the ultrasonic wave propagation speed, is the first difference, is the influence coefficient of humidity on the ultrasonic wave propagation speed, is the second difference and is the comprehensive influence coefficient of temperature and humidity; Second monitoring: Based on the corrected ultrasonic propagation speed, correct the first image to obtain the corrected first image, and update the corrected first image as the first image.

3. The monitoring method for welded structural members based on the vibration acoustic method according to claim 1, characterized in that Before performing the parameter adjustment step after the defect repair step, it also includes: First verification: Use the gray-level co-occurrence matrix method to construct the gray-level co-occurrence matrix of the incomplete welding defects in the incomplete image, denoted as the first matrix, and use the gray-level co-occurrence matrix method to construct the gray-level co-occurrence matrix of the remaining defect-containing part in the filled image, denoted as the second matrix; Second verification: Calculate the contrast of the first matrix, denoted as the first contrast, calculate the contrast of the second matrix, denoted as the second contrast, calculate the contrast difference between the first contrast and the second contrast, denoted as the third difference, and determine whether the third difference is less than the preset contrast difference threshold: If so, perform the parameter adjustment step; If not, adjust the preset fractal dimension and perform the defect filling step.

4. The method for monitoring a welded structural member based on a vibro-acoustic method according to claim 1, wherein The parameter adjustment step also includes: First processing: In the filled image, denote the incomplete welding defect before filling as the original part, and denote the remaining defect-containing part in the filled image as the filled part; Construct coordinates: Take any pixel point in the filled image as the origin, use the length direction of the filled image as the x-axis, and use the width direction of the filled image as the y-axis to construct a coordinate system; Second calculation: Use the edge detection method to obtain the coordinates of the boundary points of the original part, and calculate the center coordinates of the original part based on the coordinates of the boundary points of the original part. Denote the calculated center coordinates of the original part as the first center; Third calculation: Use the edge detection method to obtain the coordinates of the boundary points of the complete welding defect obtained in the filled image, and calculate the center coordinates of the defect based on the coordinates of the boundary points of the complete welding defect. Denote the calculated center coordinates of the defect as the second center; Angle calculation: Denote the vector from the origin to the first center as the first vector, calculate the angle between the direction of the first vector and the positive direction of the x-axis, denoted as the first angle, denote the vector from the origin to the second center as the second vector, and calculate the angle between the direction of the second vector and the positive direction of the x-axis, denoted as the second angle; First adjustment: Obtain the initial angle of the ultrasonic phased array probe, and calculate the adjustment angle through proportional operation based on the initial angle, the first angle, and the second angle. Use the adjustment angle as the adjusted angle of the ultrasonic phased array probe.

5. The method for monitoring a welded structural member based on a vibro-acoustic method according to claim 3, wherein After performing the first adjustment step, it also includes: Fourth calculation: Based on the second angle, in the filled image, update the second center as the origin of the coordinate system, use the direction of the second vector as the new positive direction of the x-axis, and construct a new y-axis and a new z-axis based on the new x-axis; First Projection: Project the boundary points of the complete welding defect onto the new x-axis, calculate the difference between the coordinate value of the projection point with the maximum coordinate value on the new x-axis and the coordinate value of the projection point with the minimum coordinate value on the new x-axis, and denote the obtained difference as the first length; Second Projection: Project the boundary points of the complete welding defect onto the new y-axis, calculate the difference between the coordinate value of the projection point with the maximum coordinate value on the new y-axis and the coordinate value of the projection point with the minimum coordinate value on the new y-axis, and denote the obtained difference as the first width; Third Projection: Project the boundary points of the complete welding defect onto the new z-axis, calculate the difference between the coordinate value of the projection point with the maximum coordinate value on the new z-axis and the coordinate value of the projection point with the minimum coordinate value on the new z-axis, and denote the obtained difference as the first height; Second Adjustment: Perform a multiplication operation on the first length, the first width, and the first height to obtain the first volume. Obtain the initial frequency of the ultrasonic phased array probe and the preset standard volume. Based on the initial frequency, the preset standard volume, and the first volume, calculate the first frequency according to the frequency formula, and use the calculated first frequency as the adjusted frequency of the ultrasonic phased array probe; The calculation model of the frequency formula is: ； Among them, is the first frequency, is the initial frequency of the ultrasonic phased array probe, is the preset standard volume, is the first volume.

6. The welding structure monitoring method based on the vibro-acoustic method according to claim 1, characterized in that, After performing the steps of the first discrimination and before performing the steps of policy acquisition, it further includes: First Extraction: Extract the geometric features of the welding defect in the second image, denoted as the first feature; Second Extraction: Extract the size features of the welding defect in the second image, denoted as the second feature; Third Extraction: Obtain the position data of the boundary points of the key load-bearing area at the weld to be monitored, extract the position data of the boundary points of the corresponding area where the welding defect is located in the second image, calculate the Euclidean distance between the boundary points of the corresponding area where the welding defect is located in the second image and the boundary points of the key load-bearing area, and denote the minimum Euclidean distance as the third feature; Second Processing: Perform encoding processing on the first feature, the second feature, and the third feature of the welding defect to obtain the encoded features, and perform weighted summation on the encoded features to obtain the first data of the welding defect in the second image; Second Discrimination: Determine whether the first data of the welding defect is not higher than the preset parameter threshold: If so, perform the steps of the first setting; If not, perform the steps of policy acquisition.

7. The monitoring method for welded structural members based on the vibration acoustic method according to claim 6, characterized in that, After performing the steps of the second discrimination and before performing the steps of the first setting, it further includes: Third Discrimination: Determine whether the number of defect category labels of the second image is 1: If so, perform the steps of the first setting; If not, perform the steps of the third processing; Third Processing: Segment the second image based on the defect category label to obtain several welding defect images, and integrate the obtained several welding defect images into a third image set; Fifth Calculation: Use the box-counting method to calculate the fractal dimension of each welding defect image in the third image set to obtain several fractal dimensions, and collect them to obtain a first dimension set; First Judgment: Determine whether each fractal dimension in the first dimension set is lower than the preset dimension threshold: If so, perform the steps of the first setting; If not, perform the steps of policy acquisition.

8. The method for monitoring a welded structural member based on a vibro-acoustic method according to claim 6, characterized in that, After performing the steps of policy acquisition, it further includes: Third monitoring: Implement the defect repair strategy for the obtained second image, and record the weld seam after implementing the defect repair strategy as the first weld seam. Use ultrasonic phased array to detect the first weld seam of the welded structural part to obtain the first verification image; Obtain the welding process parameters after implementing the defect repair strategy, denoted as the first welding process parameters; Second judgment: Input the first verification image and the first welding process parameters into the trained defect classification model to obtain the first verification image labeled with defect category labels, denoted as the second verification image, and determine whether there are welding defects in the second verification image: If not, perform the steps of the first setting; If so, perform the steps of quality discrimination; Quality discrimination: Take the second verification image as the new second image, and sequentially perform the steps from the first extraction to the second processing to obtain the first data of the welding defects in the new second image, and determine whether the first data of the welding defects in the new second image is not higher than the preset parameter threshold: If so, perform the steps of the third judgment; If not, perform the steps of model update; Third judgment: Count the number of defect category labels in the second verification image, denoted as the first quantity, count the number of defect category labels in the second image, denoted as the second quantity, and determine whether the first quantity is less than the second quantity; If so, perform the steps of the first setting; If not, perform the steps of model update; Model update: Use the second image, the corresponding defect repair strategy, and the first welding process parameters as the negative sample data of the defect strategy model, and perform model training based on the negative sample data to obtain a new trained defect strategy model, and then re-execute the steps of strategy acquisition.

9. A monitoring system for welded structural components based on the vibration acoustics method, characterized in that, The system includes: A processor and a memory; Program code is stored in the memory; When the processor calls the program code in the memory, it executes the steps of the method according to any one of claims 1-8.

Images