Welding quality analysis method and device, electronic equipment and storage medium

Through the deep learning network model, welding current and voltage data is trained to monitor welding defects in real time, solving the subjectivity and delay problems of welding quality evaluation in the existing technology, and achieving rapid and accurate analysis in the welding process.

CN119035859APending Publication Date: 2024-11-29BEIJING BO TSING TECH CO LTD
View PDF 0 Cites 1 Cited by

Patent Information

Application Number
CN202411276273.9
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-09-12
Publication Date
2024-11-29

AI Technical Summary

Technical Problem

The quality evaluation of existing circuit board welding relies on manual experience, the evaluation results are subjective, lack objective accuracy, and cannot monitor defects in the welding process in real time, resulting in high delay.

Method used

By training welding current and voltage data based on deep learning network model, a welding quality analysis model is generated, and defect types in the welding process are monitored in real time, including unmelted, welded and slag inclusion, etc., and using profile curve comparison and process effective value calculation Determine the defect type.

Benefits of technology

It realizes rapid and accurate analysis of welding quality, can monitor defects in real time during welding, reduce the subjectivity and delay of manual evaluation, and improve welding efficiency and quality.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119035859A_ABST
    Figure CN119035859A_ABST
Patent Text Reader

Abstract

The invention provides a welding quality analysis method and device, electronic equipment and a storage medium, and the method comprises the steps that a first welding defect type of each welding stage is determined based on standard welding current and voltage data and real-time welding current and voltage data of a sample welding object; comparing the welding process effective value data with the welding process effective value reference data to determine a second welding defect type corresponding to each welding stage; marking a first welding defect type and a second welding defect type of the sample welding object to obtain contour sample data and welding sample data, training the deep learning network model by using the sample data, and generating a welding quality analysis model, and the welding quality analysis result of the target welding object in the welding process is determined in real time based on the welding quality analysis model. The defect condition can be monitored in real time in the welding process, so that an operator can adjust a welding strategy in time.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the technical field of welding detection, and in particular to a welding quality analysis method, device, electronic equipment and storage medium. Background Art

[0002] The current soldering quality assessment of circuit boards is generally based on manual experience and has many shortcomings. First, this assessment method often only considers the number of solder joints on the circuit board and determines whether there is solder leakage, etc., and the assessment dimension is single. Secondly, the assessment results are highly subjective because the professional level and experience of the evaluator have an important impact on the assessment results. This makes it difficult for the assessment results to meet objective and accurate standards and lack a true reflection of the soldering defects of the circuit board. And when evaluating the soldering quality, the defects in the soldering quality are often discovered only after the entire soldering process is completed, and the delay is high. Therefore, how to accurately detect soldering defects has become a technical issue that cannot be underestimated. Summary of the invention

[0003] In view of this, the purpose of the present application is to provide a welding quality analysis method, device, electronic device and storage medium, and to obtain a welding quality analysis model by training the model with contour sample data and welding sample data, so that the welding quality analysis model can quickly and accurately determine the welding defect type corresponding to the welding object, and can monitor the defect situation in real time during the welding process, so that the operator can adjust the welding strategy in time.

[0004] The present application provides a welding quality analysis method, which includes:

[0005] Determining a first welding defect type corresponding to each welding stage of the sample welding object based on standard welding current and voltage data and real-time welding current and voltage data of the sample welding object;

[0006] Performing real-time effective value calculation processing on the real-time welding current and voltage data to determine welding process effective value data, and comparing the welding process effective value data with preset welding process effective value reference data to determine the second welding defect type corresponding to each welding stage of the sample welding object;

[0007] Determine contour sample data and welding sample data based on the real-time welding current and voltage data, the standard welding current and voltage data, the first welding defect type, the welding process effective value reference data, the welding process effective value data, and the second welding defect type;

[0008] The deep learning network model is trained based on the contour sample data and the welding sample data to generate a welding quality analysis model, so that the welding quality analysis result of the target welding object during the welding process can be determined in real time based on the welding quality analysis model.

[0009] In a possible implementation manner, determining the first welding defect type corresponding to each welding stage of the sample welding object based on the standard welding current and voltage data and the real-time welding current and voltage data of the sample welding object includes:

[0010] Performing UI fitting processing on the standard welding current and voltage data and the real-time welding current and voltage data to generate a standard contour curve and a real-time contour curve respectively;

[0011] Dividing the standard profile curve and the real-time profile curve into a plurality of welding stages based on a welding sequence;

[0012] The real-time contour curve is compared with the standard contour curve to determine the welding stage where the contour gap appears in the real-time contour curve, and to determine the first welding defect type corresponding to the welding stage where the contour gap appears.

[0013] In a possible implementation manner, the comparing the welding process effective value data with preset welding process effective value reference data to determine the second welding defect type corresponding to each welding stage of the sample welding object includes:

[0014] Detecting whether the current effective data value in the welding process effective value data is less than or equal to a first preset threshold value of the current effective reference data base value in the welding process effective value reference data;

[0015] If yes, it is determined that the second welding defect type existing at the welding stage where the current effective data value is located is unmelted or slag inclusion;

[0016] Detecting whether the current effective data value in the welding process effective value data is less than or equal to a second preset threshold of the current effective reference data base value in the welding process effective value reference data; wherein the second preset threshold is greater than the first preset threshold;

[0017] If so, it is determined that the second welding defect type existing at the welding stage where the current effective data value is located is weld penetration.

[0018] In a possible implementation manner, the determining of the contour sample data and the welding sample data based on the real-time welding current and voltage data, the standard welding current and voltage data, the first welding defect type, the welding process effective value reference data, the welding process effective value data and the second welding defect type includes:

[0019] Based on the real-time contour curve of the real-time welding current and voltage data and the standard contour curve of the standard welding current and voltage data, determining the contour shape change characteristics and the contour closed figure quantity characteristics;

[0020] Labeling the contour shape change characteristics and the contour closed figure quantity characteristics with corresponding first welding defect type labels, and using the contour shape change characteristics and the contour closed figure quantity characteristics after labeling as the contour sample data;

[0021] Based on the welding process effective value reference data and the welding process effective value data, the process data change characteristics are determined, and the process data change characteristics are labeled with corresponding second welding defect type labels, and the labeled process data change characteristics are used as the welding sample data.

[0022] In a possible implementation manner, the training of a deep learning network model based on the contour sample data and the welding sample data to generate a welding quality analysis model includes:

[0023] Inputting the contour sample data and the welding sample data into the deep learning network model, the deep learning network model learns the features corresponding to the contour sample data and the welding sample data, and predicts the predicted welding defect type corresponding to the contour sample data or the welding sample data;

[0024] Determining a loss value of the deep learning network model based on the predicted welding defect type, the contour sample data, or a label corresponding to the welding sample data;

[0025] The deep learning network model is continuously iteratively trained based on the loss value, the contour sample data and the welding sample data to generate the welding quality analysis model.

[0026] In a possible implementation manner, the iterative training of the deep learning network model based on the loss value, the contour sample data, and the welding sample data to generate the welding quality analysis model includes:

[0027] Detecting whether the loss value is less than or equal to a preset loss value;

[0028] If yes, the deep learning network model is determined as the welding quality analysis model;

[0029] If not, the network parameters of the deep learning network model are updated through the back propagation algorithm, and the deep learning network model with updated network parameters is iteratively trained based on the contour sample data and the welding sample data until the loss value is less than or equal to the preset loss value, and the training is stopped to generate the welding quality analysis model.

[0030] In a possible implementation manner, determining in real time the welding quality analysis result of the target welding object during the welding process based on the welding quality analysis model includes:

[0031] The target welding current and voltage data of the target welding object at a preset time interval are input into the welding quality analysis model, and the target welding current and voltage data are respectively subjected to contour comparison processing and effective value data comparison processing to predict the welding quality analysis result of the target welding object at each time interval.

[0032] The present application also provides a welding quality analysis device, the welding quality analysis device comprising:

[0033] A first data preprocessing module, configured to determine a first welding defect type corresponding to each welding stage of the sample welding object based on standard welding current and voltage data and real-time welding current and voltage data of the sample welding object;

[0034] a second data preprocessing module, configured to perform real-time effective value calculation processing on the real-time welding current and voltage data, determine welding process effective value data, and determine a second welding defect type corresponding to each welding stage of the sample welding object based on a comparison between the welding process effective value data and preset welding process effective value reference data;

[0035] a sample data determination module, configured to determine contour sample data and welding sample data based on the real-time welding current and voltage data, the standard welding current and voltage data, the first welding defect type, the welding process effective value reference data, the welding process effective value data, and the second welding defect type;

[0036] A training module is used to train a deep learning network model based on the contour sample data and the welding sample data to generate a welding quality analysis model so that the welding quality analysis result of the target welding object during the welding process can be determined in real time based on the welding quality analysis model.

[0037] An embodiment of the present application also provides an electronic device, comprising: a processor, a memory and a bus, wherein the memory stores machine-readable instructions executable by the processor, and when the electronic device is running, the processor and the memory communicate through the bus, and when the machine-readable instructions are executed by the processor, the steps of the welding quality analysis method as described above are performed.

[0038] An embodiment of the present application further provides a computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the steps of the welding quality analysis method as described above are executed.

[0039] The embodiments of the present application provide a welding quality analysis method, device, electronic device and storage medium, the welding quality analysis method comprising: determining a first welding defect type corresponding to each welding stage of the sample welding object based on standard welding current and voltage data and real-time welding current and voltage data of the sample welding object; performing real-time effective value calculation processing on the real-time welding current and voltage data to determine welding process effective value data, and determining a second welding defect type corresponding to each welding stage of the sample welding object based on comparison of the welding process effective value data and preset welding process effective value reference data; determining contour sample data and welding sample data based on the real-time welding current and voltage data, the standard welding current and voltage data, the first welding defect type, the welding process effective value reference data, the welding process effective value data and the second welding defect type; training a deep learning network model based on the contour sample data and the welding sample data to generate a welding quality analysis model, so that the welding quality analysis result of the target welding object in the welding process can be determined in real time based on the welding quality analysis model. The welding quality analysis model is obtained by training the model with contour sample data and welding sample data, so that the welding quality analysis model can quickly and accurately determine the welding defect type corresponding to the welding object, and can monitor the defect situation in real time during the welding process so that the operator can adjust the welding strategy in time.

[0040] In order to make the above-mentioned objects, features and advantages of the present application more obvious and easy to understand, preferred embodiments are specifically cited below and described in detail with reference to the attached drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for use in the embodiments will be briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present application and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying creative work.

[0042] Figure 1 A flow chart of a welding quality analysis method provided in an embodiment of the present application;

[0043] Figure 2 A schematic diagram of a standard profile curve provided in an embodiment of the present application;

[0044] Figure 3 A schematic diagram of a welding quality analysis method provided in an embodiment of the present application;

[0045] Figure 4 A structural schematic diagram of a welding quality analysis device provided in an embodiment of the present application;

[0046] Figure 5 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0047] To make the purpose, technical scheme and advantages of the embodiments of the present application clearer, the technical scheme in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all of the embodiments. The components of the embodiments of the present application usually described and shown in the drawings here can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present application provided in the drawings is not intended to limit the scope of the application claimed for protection, but merely represents the selected embodiments of the present application. Based on the embodiments of the present application, each other embodiment obtained by those skilled in the art without making creative work belongs to the scope of protection of the present application.

[0048] First, the application scenarios to which the present application is applicable are introduced. The present application can be applied in the field of welding detection technology.

[0049] Research has found that the current soldering quality assessment of circuit boards is generally based on manual experience and has many shortcomings. First, this assessment method often only considers the number of solder joints on the circuit board and determines whether there is solder leakage, etc., and the assessment dimension is single. Secondly, the assessment results are highly subjective because the professional level and experience of the evaluator have an important impact on the assessment results. This makes it difficult for the assessment results to meet objective and accurate standards and lack a true reflection of the soldering defects of the circuit board. And when evaluating the soldering quality, the defects in the soldering quality are often discovered only after the entire soldering process is completed, and the delay is high. Therefore, how to accurately detect soldering defects has become a technical issue that cannot be underestimated.

[0050] Based on this, an embodiment of the present application provides a welding quality analysis method, which obtains a welding quality analysis model by training the model with contour sample data and welding sample data, so that the welding quality analysis model can quickly and accurately determine the welding defect type corresponding to the welding object, and can monitor the defect situation in real time during the welding process, so that the operator can adjust the welding strategy in time.

[0051] See also Figure 1 , Figure 1 This is a flow chart of a welding quality analysis method provided in an embodiment of the present application. Figure 1 As shown in , the welding quality analysis method provided by the embodiment of the present application includes:

[0052] S101: Based on standard welding current and voltage data and real-time welding current and voltage data of a sample welding object, determine a first welding defect type corresponding to each welding stage of the sample welding object.

[0053] In this step, the first welding defect type corresponding to each welding stage of the sample welding object is determined according to the standard welding current and voltage data and the real-time welding current and voltage data of the sample welding object.

[0054] The standard welding current and voltage data are obtained by collecting the standard raw current and voltage data in the time period of 5-10S through the current and voltage Hall sensor (frequency ≥ 5000Hz). The real-time welding current and voltage data are obtained by collecting the raw current and voltage data in real time through the current and voltage Hall sensor (frequency ≥ 5000Hz).

[0055] Here, welding defect types include lack of fusion, weld penetration, slag inclusion and other types.

[0056] In a possible implementation manner, determining the first welding defect type corresponding to each welding stage of the sample welding object based on the standard welding current and voltage data and the real-time welding current and voltage data of the sample welding object includes:

[0057] A: Perform UI fitting processing on the standard welding current and voltage data and the real-time welding current and voltage data to generate a standard contour curve and a real-time contour curve respectively.

[0058] Here, the standard welding current and voltage data and the real-time welding current and voltage data are input to the device end for UI fitting processing to generate a standard contour curve and a real-time contour curve respectively.

[0059] B: Dividing the standard profile curve and the real-time profile curve into a plurality of welding stages based on the welding sequence.

[0060] Here, the standard profile curve and the real-time profile curve are divided into a plurality of welding stages according to the welding sequence.

[0061] C: performing curve comparison processing on the real-time contour curve and the standard contour curve to determine the welding stage where the contour gap appears in the real-time contour curve, and determining the first welding defect type corresponding to the welding stage where the contour gap appears.

[0062] Here, a curve comparison process is performed between the real-time contour curve and the standard contour curve to determine the welding stage where the contour gap appears in the real-time contour curve, and to determine the first welding defect type corresponding to the welding stage where the contour gap appears.

[0063] For further information, see Figure 2 , Figure 2 This is a schematic diagram of the standard contour curve provided in the embodiment of the present application. Figure 2 As shown, it is divided into four sections AD-DC-CB-BA according to the welding sequence, and the standard contour curve is compared with the real-time contour curve. When a contour gap appears in the AD section of the real-time contour curve, the welding defect is manifested as slag inclusion; when a contour gap appears between the AD and CB sections of the real-time contour curve, the welding defect is manifested as weld penetration; when a closed contour gap appears in the AD section of the real-time contour curve, the welding defect is manifested as unmelted.

[0064] S102: Perform real-time effective value calculation processing on the real-time welding current and voltage data to determine welding process effective value data, and compare the welding process effective value data with preset welding process effective value reference data to determine the second welding defect type corresponding to each welding stage of the sample welding object.

[0065] In this step, real-time effective value calculation and processing is performed on the real-time welding current and voltage data to determine the welding process effective value data. The welding process effective value data and the preset welding process effective value reference data are compared to determine the second welding defect type corresponding to each welding stage of the sample welding object.

[0066] Here, the real-time effective value is calculated by the following formula to extract the effective value data of the weld welding process:

[0067]

[0068] Where I is the current effective data value corresponding to a certain period of time, N = Π / Δwt is the number of sampling times within the period Π, i(wt j ) is the original current corresponding to a certain sampling time interval Δwt.

[0069] In a possible implementation manner, the comparing the welding process effective value data with preset welding process effective value reference data to determine the second welding defect type corresponding to each welding stage of the sample welding object includes:

[0070] a: Detect whether the current effective data value in the welding process effective value data is less than or equal to a first preset threshold value of the current effective reference data base value in the welding process effective value reference data.

[0071] Here, it is detected whether the current effective data value in the welding process effective value data is less than or equal to a first preset threshold of the current effective reference data base value in the welding process effective value reference data.

[0072] Among them, the current effective reference data base value is preset.

[0073] b: If yes, it is determined that the second welding defect type existing at the welding stage where the current effective data value is located is unmelted or slag inclusion.

[0074] Here, if it is less than or equal to, it is determined that the second welding defect type existing at the welding stage where the current effective data value is located is unmelted or slag inclusion.

[0075] c: Detect whether the current effective data value in the welding process effective value data is less than or equal to a second preset threshold of the current effective reference data base value in the welding process effective value reference data; wherein the second preset threshold is greater than the first preset threshold.

[0076] Here, it is detected whether the current effective data value in the welding process effective value data is less than or equal to a second preset threshold of the current effective reference data base value in the welding process effective value reference data.

[0077] Among them, the first preset threshold is 10A-20A, and the second preset threshold is above 20A.

[0078] d: If so, it is determined that the second welding defect type existing at the welding stage where the current effective data value is located is weld penetration.

[0079] Here, if yes, it is determined that the second welding defect type existing at the welding stage where the current effective data value is located is weld penetration.

[0080] In a specific embodiment, a base value of the current effective value reference data is pre-set, and the welding defect type is obtained by comparing the current effective data value with the base value of the current effective value reference data. If the current effective value data is lower than the base value by 10-20A, the welding defect manifests as unmelted or slag inclusion. If the current effective value data is lower than the base value by more than 20A, the welding defect manifests as weld penetration.

[0081] S103: Determine contour sample data and welding sample data based on the real-time welding current and voltage data, the standard welding current and voltage data, the first welding defect type, the welding process effective value reference data, the welding process effective value data and the second welding defect type.

[0082] In this step, the contour sample data and the welding sample data are determined according to the real-time welding current and voltage data, the standard welding current and voltage data, the first welding defect type, the welding process effective value reference data, the welding process effective value data and the second welding defect type.

[0083] In a possible implementation manner, the determining of the contour sample data and the welding sample data based on the real-time welding current and voltage data, the standard welding current and voltage data, the first welding defect type, the welding process effective value reference data, the welding process effective value data and the second welding defect type includes:

[0084] (1): Based on the real-time contour curve of the real-time welding current and voltage data and the standard contour curve of the standard welding current and voltage data, the contour shape change characteristics and the contour closed figure quantity characteristics are determined.

[0085] Here, the contour shape change characteristics and the contour closed figure quantity characteristics are determined based on the real-time contour curve of the real-time welding current and voltage data and the standard contour curve of the standard welding current and voltage data.

[0086] (2) Labeling the contour shape change characteristics and the contour closed figure quantity characteristics with corresponding first welding defect type labels, and using the contour shape change characteristics and the contour closed figure quantity characteristics after labeling as the contour sample data.

[0087] Here, the contour shape change characteristics and the contour closed figure quantity characteristics are labeled with corresponding first welding defect type labels, and the labeled contour shape change characteristics and contour closed figure quantity characteristics are used as contour sample data.

[0088] (3): Based on the welding process effective value reference data and the welding process effective value data, the process data change characteristics are determined, and the process data change characteristics are labeled with corresponding second welding defect type labels, and the labeled process data change characteristics are used as the welding sample data.

[0089] Here, based on the welding process effective value reference data and the welding process effective value data, the process data change characteristics are determined, and the process data change characteristics are labeled with the corresponding second welding defect type labels. The labeled process data change characteristics are used as welding sample data, so that the deep learning network model is trained according to the welding sample data and the contour sample data. The generated welding quality analysis model can accurately predict the welding defect type of the welding object from the process data changes and the contour curve changes.

[0090] S104: Training a deep learning network model based on the contour sample data and the welding sample data to generate a welding quality analysis model, so that the welding quality analysis result of the target welding object during the welding process can be determined in real time based on the welding quality analysis model.

[0091] In this step, the deep learning network model is trained according to the contour sample data and the welding sample data to generate a welding quality analysis model, so that the welding quality analysis results of the target welding object during the welding process can be determined in real time according to the welding quality analysis model.

[0092] In a possible implementation manner, the training of a deep learning network model based on the contour sample data and the welding sample data to generate a welding quality analysis model includes:

[0093] I: Input the contour sample data and the welding sample data into the deep learning network model, and the deep learning network model learns the features corresponding to the contour sample data and the welding sample data, and predicts the predicted welding defect type corresponding to the contour sample data or the welding sample data.

[0094] Here, the contour sample data and welding sample data are input into the deep learning network model. The deep learning network model learns the features corresponding to the contour sample data and the welding sample data, and predicts the predicted welding defect type corresponding to the contour sample data or the welding sample data.

[0095] II: Determine the loss value of the deep learning network model based on the predicted welding defect type, the contour sample data or the label corresponding to the welding sample data.

[0096] Here, the cross loss function calculation formula is used to calculate the labels corresponding to the predicted welding defect type, contour sample data or welding sample data to determine the loss value of the deep learning network model.

[0097] III: Continue to iteratively train the deep learning network model based on the loss value, the contour sample data and the welding sample data to generate the welding quality analysis model.

[0098] Here, the deep learning network model is iteratively trained based on the loss value, contour sample data, and welding sample data to generate a welding quality analysis model.

[0099] In a possible implementation manner, the iterative training of the deep learning network model based on the loss value, the contour sample data, and the welding sample data to generate the welding quality analysis model includes:

[0100] i: Check whether the loss value is less than or equal to a preset loss value.

[0101] ii: If so, the deep learning network model is determined as the welding quality analysis model.

[0102] III: If not, the network parameters of the deep learning network model are updated through the back propagation algorithm, and the deep learning network model with updated network parameters is iteratively trained based on the contour sample data and the welding sample data until the loss value is less than or equal to the preset loss value, and the training is stopped to generate the welding quality analysis model.

[0103] Here, if not, the network parameters of the deep learning network model are updated through the back propagation algorithm, and the deep learning network model with updated network parameters is iteratively trained according to the contour sample data and welding sample data until the loss value is less than or equal to the preset loss value, and the training is stopped to generate a welding quality analysis model.

[0104] In a specific embodiment, the real-time data is preprocessed and sliced, and the corresponding effective value average, contour shape change characteristics, and contour closed figure quantity characteristics are calculated for every 50ms of the original current and voltage UI data, and the corresponding defect labels are marked, where the normal weld label is 0, the slag inclusion label is 1, the weld penetration label is 2, and the unmelted label is 3. The ratio of the training set to the test set is 8:2 and is put into a deep learning network model such as LSTM for training. During the training process, the weights are adjusted by the back propagation algorithm to minimize the classification error to obtain a welding quality analysis model.

[0105] In a possible implementation manner, determining in real time the welding quality analysis result of the target welding object during the welding process based on the welding quality analysis model includes:

[0106] The target welding current and voltage data of the target welding object at a preset time interval are input into the welding quality analysis model, and the target welding current and voltage data are respectively subjected to contour comparison processing and effective value data comparison processing to predict the welding quality analysis result of the target welding object at each time interval.

[0107] Here, the target welding current and voltage data of the target welding object at a preset time interval are input into the welding quality analysis model, and the target welding current and voltage data are subjected to contour comparison processing and effective value data comparison processing respectively to obtain the prediction results under the contour comparison processing and the prediction results under the effective value data comparison processing. If the two prediction results are inconsistent, the prediction result under the contour comparison processing is used as the welding quality analysis result of the target welding object.

[0108] Among them, the standard contour curves and current effective reference data base values ​​corresponding to different welding objects are different.

[0109] For further information, see Figure 3 , Figure 3 A schematic diagram of a welding quality analysis method provided in an embodiment of the present application. Figure 3 As shown, S1: continuously obtain welding process effective value data, including current effective data value; S2: detect whether the current effective data value exceeds the current effective reference data base value; S3: if so, determine that the sample welding object has welding defects; if not, determine that the welding of the sample welding object is normal; S4: continuously obtain real-time welding current and voltage data of the sample welding object, and display it in the real-time contour curve of the UI image, where the horizontal axis of the real-time contour curve is current and the vertical axis is voltage; S5: detect whether the real-time contour curve is consistent with the standard contour curve; S6: if they are consistent, the welding is normal, if they are inconsistent, there is a welding defect; S7: input the sample welding object and the corresponding welding information into the deep learning network model for training to generate a welding quality analysis model; S8: determine the welding quality analysis result of the target welding object during the welding process according to the welding quality analysis model.

[0110] A welding quality analysis method provided in an embodiment of the present application comprises: determining a first welding defect type corresponding to each welding stage of the sample welding object based on standard welding current and voltage data and real-time welding current and voltage data of the sample welding object; performing real-time effective value calculation processing on the real-time welding current and voltage data to determine welding process effective value data, and determining a second welding defect type corresponding to each welding stage of the sample welding object based on comparison of the welding process effective value data and preset welding process effective value reference data; determining contour sample data and welding sample data based on the real-time welding current and voltage data, the standard welding current and voltage data, the first welding defect type, the welding process effective value reference data, the welding process effective value data and the second welding defect type; training a deep learning network model based on the contour sample data and the welding sample data to generate a welding quality analysis model, so that the welding quality analysis result of the target welding object in the welding process can be determined in real time based on the welding quality analysis model. The welding quality analysis model is obtained by training the model with contour sample data and welding sample data, so that the welding quality analysis model can quickly and accurately determine the welding defect type corresponding to the welding object, and can monitor the defect situation in real time during the welding process so that the operator can adjust the welding strategy in time.

[0111] See also Figure 4 , Figure 4 This is a schematic diagram of the structure of a welding quality analysis device provided in an embodiment of the present application. Figure 4 As shown in , the welding quality analysis device 400 includes:

[0112] A first data preprocessing module 410, for determining a first welding defect type corresponding to each welding stage of the sample welding object based on standard welding current and voltage data and real-time welding current and voltage data of the sample welding object;

[0113] A second data preprocessing module 420 is used to perform real-time effective value calculation processing on the real-time welding current and voltage data to determine welding process effective value data, and to determine a second welding defect type corresponding to each welding stage of the sample welding object based on a comparison between the welding process effective value data and preset welding process effective value reference data;

[0114] A sample data determination module 430, for determining contour sample data and welding sample data based on the real-time welding current and voltage data, the standard welding current and voltage data, the first welding defect type, the welding process effective value reference data, the welding process effective value data and the second welding defect type;

[0115] The training module 440 is used to train the deep learning network model based on the contour sample data and the welding sample data to generate a welding quality analysis model so as to determine the welding quality analysis results of the target welding object during the welding process in real time based on the welding quality analysis model.

[0116] Further, when the first data preprocessing module is used to determine the first welding defect type corresponding to each welding stage of the sample welding object based on the standard welding current and voltage data and the real-time welding current and voltage data of the sample welding object, the first data preprocessing module is specifically used to:

[0117] Performing UI fitting processing on the standard welding current and voltage data and the real-time welding current and voltage data to generate a standard contour curve and a real-time contour curve respectively;

[0118] Dividing the standard profile curve and the real-time profile curve into a plurality of welding stages based on a welding sequence;

[0119] The real-time contour curve is compared with the standard contour curve to determine the welding stage where the contour gap appears in the real-time contour curve, and to determine the first welding defect type corresponding to the welding stage where the contour gap appears.

[0120] Furthermore, when the second data preprocessing module is used to compare the welding process effective value data with the preset welding process effective value reference data to determine the second welding defect type corresponding to each welding stage of the sample welding object, the second data preprocessing module is specifically used to:

[0121] Detecting whether the current effective data value in the welding process effective value data is less than or equal to a first preset threshold value of the current effective reference data base value in the welding process effective value reference data;

[0122] If yes, it is determined that the second welding defect type existing at the welding stage where the current effective data value is located is unmelted or slag inclusion;

[0123] Detecting whether the current effective data value in the welding process effective value data is less than or equal to a second preset threshold of the current effective reference data base value in the welding process effective value reference data; wherein the second preset threshold is greater than the first preset threshold;

[0124] If so, it is determined that the second welding defect type existing at the welding stage where the current effective data value is located is weld penetration.

[0125] Further, when the sample data determination module is used to determine the contour sample data and the welding sample data based on the real-time welding current and voltage data, the standard welding current and voltage data, the first welding defect type, the welding process effective value reference data, the welding process effective value data and the second welding defect type, the sample data determination module is specifically used to:

[0126] Based on the real-time contour curve of the real-time welding current and voltage data and the standard contour curve of the standard welding current and voltage data, determining the contour shape change characteristics and the contour closed figure quantity characteristics;

[0127] Labeling the contour shape change characteristics and the contour closed figure quantity characteristics with corresponding first welding defect type labels, and using the contour shape change characteristics and the contour closed figure quantity characteristics after labeling as the contour sample data;

[0128] Based on the welding process effective value reference data and the welding process effective value data, the process data change characteristics are determined, and the process data change characteristics are labeled with corresponding second welding defect type labels, and the labeled process data change characteristics are used as the welding sample data.

[0129] Furthermore, when the training module is used to train the deep learning network model based on the contour sample data and the welding sample data to generate a welding quality analysis model, the training module is specifically used to:

[0130] Inputting the contour sample data and the welding sample data into the deep learning network model, the deep learning network model learns the features corresponding to the contour sample data and the welding sample data, and predicts the predicted welding defect type corresponding to the contour sample data or the welding sample data;

[0131] Determining a loss value of the deep learning network model based on the predicted welding defect type, the contour sample data, or a label corresponding to the welding sample data;

[0132] The deep learning network model is continuously iteratively trained based on the loss value, the contour sample data and the welding sample data to generate the welding quality analysis model.

[0133] Furthermore, when the training module is used to continue iteratively training the deep learning network model based on the loss value, the contour sample data, and the welding sample data to generate the welding quality analysis model, the training module is specifically used to:

[0134] Detecting whether the loss value is less than or equal to a preset loss value;

[0135] If yes, the deep learning network model is determined as the welding quality analysis model;

[0136] If not, the network parameters of the deep learning network model are updated through the back propagation algorithm, and the deep learning network model with updated network parameters is iteratively trained based on the contour sample data and the welding sample data until the loss value is less than or equal to the preset loss value, and the training is stopped to generate the welding quality analysis model.

[0137] Furthermore, when the training module is used to determine the welding quality analysis result of the target welding object in the welding process in real time based on the welding quality analysis model, the training module is specifically used to:

[0138] The target welding current and voltage data of the target welding object at a preset time interval are input into the welding quality analysis model, and the target welding current and voltage data are respectively subjected to contour comparison processing and effective value data comparison processing to predict the welding quality analysis result of the target welding object at each time interval.

[0139] A welding quality analysis device provided in an embodiment of the present application comprises: a first data preprocessing module, which is used to determine a first welding defect type corresponding to each welding stage of the sample welding object based on standard welding current and voltage data and real-time welding current and voltage data of the sample welding object; a second data preprocessing module, which is used to perform real-time effective value calculation processing on the real-time welding current and voltage data to determine welding process effective value data, and determine a second welding defect type corresponding to each welding stage of the sample welding object based on comparison of the welding process effective value data and preset welding process effective value reference data; a sample data determination module, which is used to determine contour sample data and welding sample data based on the real-time welding current and voltage data, the standard welding current and voltage data, the first welding defect type, the welding process effective value reference data, the welding process effective value data and the second welding defect type; and a training module, which is used to train a deep learning network model based on the contour sample data and the welding sample data to generate a welding quality analysis model, so that the welding quality analysis result of the target welding object in the welding process can be determined in real time based on the welding quality analysis model. The welding quality analysis model is obtained by training the model with contour sample data and welding sample data, so that the welding quality analysis model can quickly and accurately determine the welding defect type corresponding to the welding object, and can monitor the defect situation in real time during the welding process so that the operator can adjust the welding strategy in time.

[0140] See also Figure 5 , Figure 5 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present application. Figure 5 As shown in , the electronic device 500 includes a processor 510 , a memory 520 and a bus 530 .

[0141] The memory 520 stores machine-readable instructions executable by the processor 510. When the electronic device 500 is running, the processor 510 communicates with the memory 520 via the bus 530. When the machine-readable instructions are executed by the processor 510, the above-mentioned Figure 1 as well as Figure 3 The steps of the welding quality analysis method in the method embodiment shown and the specific implementation methods can be found in the method embodiment, which will not be repeated here.

[0142] The present application also provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the computer program can execute the above-mentioned Figure 1 as well as Figure 3 The steps of the welding quality analysis method in the method embodiment shown and the specific implementation methods can be found in the method embodiment, which will not be repeated here.

[0143] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0144] In the several embodiments provided in the present application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. The device embodiments described above are merely schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some communication interfaces, and the indirect coupling or communication connection of devices or units can be electrical, mechanical or other forms.

[0145] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0146] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0147] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a non-volatile computer-readable storage medium that is executable by a processor. Based on this understanding, the technical solution of the present application can essentially be embodied in the form of a software product, or in other words, the part that contributes to the prior art or the part of the technical solution. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

[0148] Finally, it should be noted that the above-described embodiments are only specific implementation methods of the present application, which are used to illustrate the technical solutions of the present application, rather than to limit them. The protection scope of the present application is not limited thereto. Although the present application is described in detail with reference to the above-mentioned embodiments, ordinary technicians in the field should understand that any technician familiar with the technical field can still modify the technical solutions recorded in the above-mentioned embodiments within the technical scope disclosed in the present application, or can easily think of changes, or make equivalent replacements for some of the technical features therein; and these modifications, changes or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application, and should be included in the protection scope of the present application. Therefore, the protection scope of the present application shall be based on the protection scope of the claims.

Claims

1. A welding quality analysis method, characterized in that: The welding quality analysis method comprises: Determining a first welding defect type corresponding to each welding stage of the sample welding object based on standard welding current and voltage data and real-time welding current and voltage data of the sample welding object; Performing real-time effective value calculation processing on the real-time welding current and voltage data to determine welding process effective value data, and comparing the welding process effective value data with preset welding process effective value reference data to determine the second welding defect type corresponding to each welding stage of the sample welding object; Determine contour sample data and welding sample data based on the real-time welding current and voltage data, the standard welding current and voltage data, the first welding defect type, the welding process effective value reference data, the welding process effective value data, and the second welding defect type; The deep learning network model is trained based on the contour sample data and the welding sample data to generate a welding quality analysis model, so that the welding quality analysis result of the target welding object during the welding process can be determined in real time based on the welding quality analysis model.

2. The welding quality analysis method according to claim 1, characterized in that: The method of determining the first welding defect type corresponding to each welding stage of the sample welding object based on the standard welding current and voltage data and the real-time welding current and voltage data of the sample welding object includes: Performing UI fitting processing on the standard welding current and voltage data and the real-time welding current and voltage data to generate a standard contour curve and a real-time contour curve respectively; Dividing the standard profile curve and the real-time profile curve into a plurality of welding stages based on a welding sequence; The real-time contour curve is compared with the standard contour curve to determine the welding stage where the contour gap appears in the real-time contour curve, and to determine the first welding defect type corresponding to the welding stage where the contour gap appears.

3. The welding quality analysis method according to claim 1, characterized in that: The comparing the welding process effective value data with the preset welding process effective value reference data to determine the second welding defect type corresponding to each welding stage of the sample welding object includes: Detecting whether the current effective data value in the welding process effective value data is less than or equal to a first preset threshold value of the current effective reference data base value in the welding process effective value reference data; If yes, it is determined that the second welding defect type existing at the welding stage where the current effective data value is located is unmelted or slag inclusion; Detecting whether the current effective data value in the welding process effective value data is less than or equal to a second preset threshold of the current effective reference data base value in the welding process effective value reference data; wherein the second preset threshold is greater than the first preset threshold; If so, it is determined that the second welding defect type existing at the welding stage where the current effective data value is located is weld penetration.

4. The welding quality analysis method according to claim 1, characterized in that: The determining of the contour sample data and the welding sample data based on the real-time welding current and voltage data, the standard welding current and voltage data, the first welding defect type, the welding process effective value reference data, the welding process effective value data and the second welding defect type comprises: Based on the real-time contour curve of the real-time welding current and voltage data and the standard contour curve of the standard welding current and voltage data, determining the contour shape change characteristics and the contour closed figure quantity characteristics; Labeling the contour shape change characteristics and the contour closed figure quantity characteristics with corresponding first welding defect type labels, and using the contour shape change characteristics and the contour closed figure quantity characteristics after labeling as the contour sample data; Based on the welding process effective value reference data and the welding process effective value data, the process data change characteristics are determined, and the process data change characteristics are labeled with corresponding second welding defect type labels, and the labeled process data change characteristics are used as the welding sample data.

5. The welding quality analysis method according to claim 1, characterized in that: The deep learning network model is trained based on the contour sample data and the welding sample data to generate a welding quality analysis model, including: Inputting the contour sample data and the welding sample data into the deep learning network model, the deep learning network model learns the features corresponding to the contour sample data and the welding sample data, and predicts the predicted welding defect type corresponding to the contour sample data or the welding sample data; Determining a loss value of the deep learning network model based on the predicted welding defect type, the contour sample data, or a label corresponding to the welding sample data; The deep learning network model is continuously iteratively trained based on the loss value, the contour sample data and the welding sample data to generate the welding quality analysis model.

6. The welding quality analysis method according to claim 5, characterized in that: The step of continuing to iteratively train the deep learning network model based on the loss value, the contour sample data, and the welding sample data to generate the welding quality analysis model includes: Detecting whether the loss value is less than or equal to a preset loss value; If yes, the deep learning network model is determined as the welding quality analysis model; If not, the network parameters of the deep learning network model are updated through the back propagation algorithm, and the deep learning network model with updated network parameters is iteratively trained based on the contour sample data and the welding sample data until the loss value is less than or equal to the preset loss value, and the training is stopped to generate the welding quality analysis model.

7. The welding quality analysis method according to claim 1, characterized in that: The method of determining the welding quality analysis result of the target welding object in the welding process in real time based on the welding quality analysis model includes: The target welding current and voltage data of the target welding object at a preset time interval are input into the welding quality analysis model, and the target welding current and voltage data are respectively subjected to contour comparison processing and effective value data comparison processing to predict the welding quality analysis result of the target welding object at each time interval.

8. A welding quality analysis device, characterized in that: The welding quality analysis device comprises: A first data preprocessing module, configured to determine a first welding defect type corresponding to each welding stage of the sample welding object based on standard welding current and voltage data and real-time welding current and voltage data of the sample welding object; a second data preprocessing module, configured to perform real-time effective value calculation processing on the real-time welding current and voltage data, determine welding process effective value data, and determine a second welding defect type corresponding to each welding stage of the sample welding object based on a comparison between the welding process effective value data and preset welding process effective value reference data; a sample data determination module, configured to determine contour sample data and welding sample data based on the real-time welding current and voltage data, the standard welding current and voltage data, the first welding defect type, the welding process effective value reference data, the welding process effective value data, and the second welding defect type; A training module is used to train a deep learning network model based on the contour sample data and the welding sample data to generate a welding quality analysis model so that the welding quality analysis result of the target welding object during the welding process can be determined in real time based on the welding quality analysis model.

9. An electronic device, characterized in that: include: A processor, a memory and a bus, wherein the memory stores machine-readable instructions executable by the processor, and when the electronic device is running, the processor and the memory communicate through the bus, and the machine-readable instructions are executed by the processor to execute the steps of the welding quality analysis method as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the welding quality analysis method according to any one of claims 1 to 7 are executed.

Citation Information

Cited By

  • Welding process parameter determination method, device, equipment, medium and program product

    CN119703273A