Welding strategy determination method, device, non-volatile storage medium and electronic equipment

Abstract

This application discloses a welding strategy determination method, apparatus, non-volatile storage medium, and electronic device. The method comprises: determining multiple target welding parameters based on first correlation data, wherein the first correlation data represents the degree of correlation between different candidate welding parameters and welding results; determining target parameter values ​​corresponding to each of the multiple target welding parameters based on historical welding results and the historical parameter values ​​corresponding to each of the multiple target welding parameters; and obtaining a target control strategy based on the target parameter values ​​corresponding to each of the multiple target welding parameters. This application addresses the technical problem in related arts of reduced welding yields caused by unsatisfactory control strategy settings.

Description

Technical Field

[0001] The present application relates to the field of battery manufacturing, and more specifically, to a welding strategy determination method, device, non-volatile storage medium, and electronic device. Background Art

[0002] In the production process of lithium batteries, the welding process is an extremely critical link. The welding quality directly determines the performance, safety and reliability of the lithium battery as a whole. A tiny welding defect may lead to a decline in battery performance, an increase in safety hazards, and even affect the service life of the entire product. In the related art, the quality control method in the lithium battery welding process often adopts a post-inspection method, that is, after the battery welding is completed, the good and defective products are identified and only batteries with good quality are retained. Although the related technology can ensure the quality of battery products entering the market, it does not fundamentally solve the problem of improving the welding yield rate, and defective products will still cause unnecessary waste of manpower and resources. Therefore, the related technology has the problem of being unable to analyze relevant data, make future predictions, lacking real-time closed-loop control, and being unable to control the production of defective products.

[0003] To address the above-mentioned problems, no effective solutions have been proposed so far. Summary of the Invention

[0004] The embodiments of the present application provide a welding strategy determination method, device, non-volatile storage medium and electronic device to at least solve the technical problem in the related art that the control strategy setting is not ideal, resulting in a reduced welding yield.

[0005] According to one aspect of an embodiment of the present application, a method for determining a welding strategy is provided, comprising: determining a plurality of target welding parameters according to first associated data, wherein the first associated data represents the degree of association between different candidate welding parameters and welding results; determining target parameter values ​​corresponding to the plurality of target welding parameters based on historical welding results and historical parameter values ​​corresponding to the plurality of target welding parameters; and obtaining a target control strategy based on the target parameter values ​​corresponding to the plurality of target welding parameters. Through the above-mentioned processing, while completing quality inspection, it is possible to provide real-time feedback to the production line based on the inspection results, thereby automatically adjusting and optimizing production parameters, effectively preventing potential quality problems, and reducing the defective rate. This not only enhances the accuracy of product quality control, but also forms a virtuous cycle in improving the yield rate. Through continuous data analysis and feedback, the production line can continuously optimize itself and achieve continuous improvement in product quality.

[0006] Optionally, based on historical welding results and historical parameter values ​​corresponding to multiple target welding parameters, target parameter values ​​corresponding to the multiple target welding parameters are determined, including: determining a fitting relationship between the historical welding results and the historical parameter values ​​corresponding to the multiple target welding parameters; determining prediction results corresponding to multiple candidate parameter combinations based on the fitting relationship, wherein the multiple candidate parameter combinations are obtained by setting the multiple target welding parameters to different candidate parameter values; determining welding strength scores corresponding to the multiple candidate parameter combinations based on the prediction results; determining the parameter combination with the highest welding strength score among the multiple candidate parameter combinations as the target parameter combination; and obtaining target parameter values ​​corresponding to the target welding parameters based on the target parameter combination. Based on historical data and real-time feedback, the standards for welding process parameters are continuously optimized to continuously improve the yield rate of products.

[0007] Optionally, the method further includes: capturing images of historical welding areas to obtain historical images; processing the historical images using corresponding image processing methods based on predetermined defect extraction requirements to obtain two-dimensional welding information; obtaining welding defect characteristics based on the two-dimensional welding information; and labeling the welding defect characteristics with defect types to obtain historical welding results. After training the model using historical welding results, data from the current welding process is input into the model in real time, enabling rapid defect type predictions and process parameter optimization recommendations. Continuously collecting and analyzing historical welding data not only enables immediate optimization of current production but also provides data support for future process improvements.

[0008] Optionally, weld defect signatures can be derived based on 2D welding information. This includes: laser scanning the historical weld area to generate 3D point cloud data; determining 3D surface features of the historical weld area based on the 3D point cloud data; and performing feature stitching based on the 2D welding information and 3D surface features to determine weld defect signatures. By integrating 2D and 3D features, a more comprehensive understanding of the weld area's condition can be achieved, improving the accuracy of weld defect recognition and classification.

[0009] Optionally, determining multiple target welding parameters based on the first correlation data includes: determining second correlation data indicating the degree of correlation between multiple candidate welding parameters; and screening the multiple candidate welding parameters based on the first and second correlation data to determine the multiple target welding parameters. Through this process, the multiple candidate welding parameters are screened to achieve an optimally balanced combination. This data-driven decision-making approach reduces production interruptions and time waste caused by improper parameter adjustments, thereby improving the efficiency of the entire production process.

[0010] Optionally, based on the first and second correlation data, multiple candidate welding parameters are screened to determine multiple target welding parameters, including: determining third correlation data representing the degree of correlation between each of the multiple candidate welding parameters and timing characteristics, wherein the timing characteristics include at least the operating time of the welding equipment and seasonal characteristics; determining fourth correlation data representing the degree of correlation between the multiple candidate welding parameters and welding stability based on the data distribution corresponding to each of the multiple candidate welding parameters; and screening the multiple candidate welding parameters based on the first, second, third, and fourth correlation data to determine multiple target welding parameters. Through the above processing, through the analysis and correlation of multi-source data, key factors in welding parameters can be more accurately identified, thereby improving the accuracy of welding quality control. Taking into account timing characteristics such as equipment operating time and seasonal changes helps to enhance the stability of the welding process and reduce welding quality fluctuations caused by environmental changes.

[0011] Optionally, the method further includes: determining a policy update frequency based on the device type of the welding device; and updating the target control policy to the welding device according to the policy update frequency. The policy update frequency is set taking into account the device response time, data processing speed, and production rhythm to ensure that the update frequency is neither too frequent to avoid unnecessary consumption of computing resources nor too sparse to ensure timely response to production changes.

[0012] According to another aspect of an embodiment of the present application, a welding strategy determination device is provided, including: a parameter screening module for determining multiple target welding parameters according to first associated data, wherein the first associated data represents the degree of association between different candidate welding parameters and welding results; a parameter determination module for determining target parameter values ​​corresponding to multiple target welding parameters based on historical welding results and historical parameter values ​​corresponding to the multiple target welding parameters; and a strategy generation module for obtaining a target control strategy based on the target parameter values ​​corresponding to the multiple target welding parameters.

[0013] According to another aspect of an embodiment of the present application, a non-volatile storage medium is provided. The non-volatile storage medium stores a plurality of instructions, and the instructions are suitable for being loaded by a processor and executing any one of the welding strategy determination methods.

[0014] According to another aspect of an embodiment of the present application, an electronic device is provided, comprising: one or more processors and a memory, the memory being used to store one or more programs, wherein when the one or more programs are executed by one or more processors, the one or more processors implement any one of the welding strategy determination methods.

[0015] In an embodiment of the present application, a closed-loop control approach is employed to determine multiple target welding parameters based on first correlation data, wherein the first correlation data represents the degree of correlation between different candidate welding parameters and welding results; determine target parameter values ​​corresponding to each of the multiple target welding parameters based on historical welding results and the historical parameter values ​​corresponding to each of the multiple target welding parameters; and obtain a target control strategy based on the target parameter values ​​corresponding to each of the multiple target welding parameters. This achieves the goal of optimizing the welding control strategy based on the welding results, improving the accuracy of the welding control strategy, and thereby improving the welding yield rate. This addresses the technical problem in related arts of reduced welding yield due to unsatisfactory control strategy settings. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0017] Figure 1 is a flow chart of an optional welding strategy determination method provided according to an embodiment of the present application;

[0018] Figure 2 is a first schematic diagram of an optional welding strategy determination method provided according to an embodiment of the present application;

[0019] Figure 3 is a second schematic diagram of an optional welding strategy determination method provided according to an embodiment of the present application;

[0020] Figure 4 is a third schematic diagram of an optional welding strategy determination method provided according to an embodiment of the present application;

[0021] Figure 5 This is a schematic diagram of an optional welding strategy determination device provided according to an embodiment of the present application. DETAILED DESCRIPTION

[0022] In order to enable those skilled in the art to better understand the present invention, the following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments in the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts should fall within the scope of protection of this application.

[0023] It should be noted that the terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequential order. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in a sequence other than those illustrated or described herein. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0024] According to an embodiment of the present application, an embodiment of a method for determining a welding strategy is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0025] Figure 1 is a flow chart of an optional method for determining a welding strategy according to an embodiment of the present application, such as Figure 1 As shown, the method includes the following steps:

[0026] Step S102, determining a plurality of target welding parameters according to first correlation data, wherein the first correlation data represents the degree of correlation between different candidate welding parameters and welding results;

[0027] It is understandable that the correlation between all candidate welding parameters in the welding process and the welding results is analyzed. Candidate welding parameters may include welding speed, defocus, weld width, and ram pressure, while welding results involve characteristics such as weld length, penetration depth, and weld width. Through statistical analysis or machine learning algorithms (such as Pearson correlation coefficient and multiple regression analysis), the degree of correlation between candidate welding parameters and welding results can be effectively determined, that is, the first correlation data. By quantifying the correlation between parameters and results, it is ensured that parameter optimization decisions are based on objective data rather than subjective experience, thereby improving the accuracy of control decisions.

[0028] Optionally, the Pearson correlation coefficient is referenced and a functional component in a predetermined programming language is used to calculate the correlation between the feature (candidate welding parameter) and the target variable (welding result) as the first correlation data to preliminarily screen the candidate welding parameters.

[0029] Optionally, multiple candidate welding parameters can be selected from key welding process parameters and key characteristic parameters, including, for example, welding speed, defocus, weld width, ram pressure, weld length, penetration depth, weld width, etc.

[0030] Optionally, the big data system collects image data of the welding area captured by industrial cameras and 3D point cloud data scanned by LiDAR, performing data preprocessing. This includes data cleaning, feature selection, time series data preparation, and categorical variable encoding. This forms training and test sets for subsequent algorithm use. The system also generates work-in-progress welding archives, visually displaying production process traceability data and key process analysis results. Table 1 shows the initial data, and the specific values ​​below are for illustrative purposes only and are not intended to be limiting.

[0031] Table 1

[0032]

[0033] After identifying and correcting errors, removing duplications, filling missing values, converting data types, and formatting to obtain an accurate, consistent, and easy-to-analyze data set, Table 2 shows the corrected data. The missing part of the welding speed is filled with 600, and the welding speed that deviates significantly from 1800 is corrected to 800.

[0034] Table 2

[0035]

[0036]

[0037] The dataset is divided into a training set and a test set (70%:30%). The training set is used to train the model, and the test set is used to evaluate the performance of the model.

[0038] In an optional embodiment, multiple target welding parameters are determined according to the first correlation data, including: determining second correlation data representing the degree of correlation between multiple candidate welding parameters; based on the first correlation data and the second correlation data, screening multiple candidate welding parameters to determine multiple target welding parameters.

[0039] It can be understood that the second correlation data refers to the statistical relationship established between multiple candidate welding parameters, which is used to quantify the mutual influence between these parameters. For example, the correlation between welding speed and welding power, and the correlation between defocus and weld width. By calculating statistical indicators such as the Pearson correlation coefficient and the Spearman rank correlation coefficient, the second correlation data can be obtained to understand the strength and direction of the relationship between parameters. The first correlation data refers to the relationship between welding parameters and the quality indicators of welding results, such as the direct correlation between welding speed and weld quality. Combined with the second correlation data (the relationship between parameters), the comprehensive impact of welding parameters on welding quality can be more comprehensively analyzed. For example, if the weld quality decreases when the welding speed increases, but the negative impact of the welding speed on the weld quality can be partially offset when the welding power is also increased, then when selecting the target welding parameters, it is necessary to consider the balance and optimization between the two. Through the above process, multiple candidate welding parameters are screened to find a combination that can optimize welding quality and achieve the best balance between parameters. Using a data-driven decision-making approach, based on the role of parameters in welding process control, decision-making based on correlation data is achieved, improving the objectivity of welding control.

[0040] Alternatively, an algorithmic model (such as a multi-objective optimization algorithm) can be used to determine an appropriate screening method and then determine the target welding parameters. After data analysis and model screening, the parameter combination that can most effectively improve welding quality is found, providing data support for the subsequent determination of parameter values. By finding the optimal welding parameter combination, welding quality can be significantly improved and the occurrence of welding defects such as cold welds and hot spots can be reduced. Once the target welding parameters are determined, the welding process can be more stable and efficient, reducing production interruptions and time waste caused by improper parameter adjustments, thereby improving the efficiency of the entire production process.

[0041] Optionally, assuming that the correlation between welding speed (marked as V), defocus (marked as F), weld length (marked as L), weld width (marked as W) and penetration (marked as D) is evaluated, the following is the Pearson correlation coefficient between these feature pairs (i.e., paired candidate welding parameters), i.e., the second correlation data. The specific values ​​in Table 3 are only for illustration and are not specifically limited.

[0042] Table 3

[0043] Feature pairs Second related data V and F -0.25 V and L 0.72 V and W 0.51 V and D 0.68 F and L -0.18 F and W -0.05 F and D -0.12 L and W 0.83 L and D 0.75 W and D 0.55

[0044] According to the second correlation data in Table 3, there is a high positive correlation (0.72 and 0.68) between welding speed (V) and weld length (L) and penetration (D), indicating that welding speed may directly affect weld length and penetration. There is a weak negative correlation (-0.25) between welding speed (V) and defocus (F), indicating that welding speed may have an opposite trend with defocus. There is a high positive correlation (0.83) between weld length (L) and weld width (W), which may mean that longer welds tend to be wider. The correlations between other feature pairs are low, indicating that their associations may be weak or there is no obvious linear relationship.

[0045] In an optional embodiment, based on the first correlation data and the second correlation data, multiple candidate welding parameters are screened to determine multiple target welding parameters, including: determining third correlation data representing the degree of correlation between the multiple candidate welding parameters and the timing characteristics, wherein the timing characteristics include at least the operating time of the welding equipment and the seasonal characteristics; based on the data distribution corresponding to the multiple candidate welding parameters, determining fourth correlation data representing the degree of correlation between the multiple candidate welding parameters and welding stability; based on the first correlation data, the second correlation data, the third correlation data, and the fourth correlation data, screening the multiple candidate welding parameters to determine the multiple target welding parameters.

[0046] It can be understood that through data analysis, such as using statistical methods (such as the Pearson correlation coefficient) or machine learning models to analyze the mutual influence between parameters such as welding speed, defocus, weld width, and pressure head pressure, it is helpful to identify which parameters have a more significant impact on welding quality when combined together. Time series features are introduced, such as the operating time of welding equipment and seasonal characteristics. For example, long-term operation of the equipment may cause temperature changes, which in turn affect the welding parameters. Seasonal changes may affect the ambient humidity and also have a potential impact on welding quality. The data distribution of each candidate welding parameter is analyzed to understand its volatility or stability, and then evaluate its impact on welding stability. Based on the data distribution of the parameters, the fourth correlation data is calculated, which reflects the relationship between parameter stability and welding quality. Combining the first, second, third and fourth correlation data, a multi-dimensional analysis is performed to identify the parameter combination that has the greatest impact on welding quality and the highest stability. Based on the comprehensive evaluation results, the parameters that have the greatest impact on improving welding quality are screened out, namely the target welding parameters.

[0047] Optionally, the third associated data can be determined by analyzing the association between the time series characteristics and the candidate welding parameters. The ARIMA (Autoregressive Integrated Moving Average) model, which is a statistical model used for time series analysis and prediction, or the simple moving average (SMA) can be used to complete the time series analysis. The ARIMA model combines the three methods of autoregression, difference and moving average, and can be used to describe the trend, seasonality and randomness of time series data. The regularity, trend and seasonality of the simple moving average over time can be used for time series analysis to analyze and monitor and predict the patterns and trends of key quality parameters so that process parameters can be adjusted in a timely manner to ensure product quality.

[0048] For ease of understanding, the SMA method is used to give a specific example of the weld length (L). Figure 2 This is a first schematic diagram of an optional welding strategy determination method provided in an embodiment of the present application, such as Figure 2 As shown, if the data of weld length begins to show a gradual downward trend over a period of time, even if this decline may not be obvious in a short period of time, SMA can reveal this trend in advance, thereby ensuring the stability of the entire production process and product quality.

[0049] The formula for calculating the Simple Moving Average (SMA) is:

[0050] Among them, MA t Y represents the simple moving average at time t, n represents the window size of the moving average, that is, how many observations at the past time points are considered to calculate the average value. The choice of window size depends on the length of the period that needs to be smoothed. For example, if you want to capture the trend within a week, n ​​may be set to 7. t-i Represents the observation value at time ti, where i is the time identifier before and after time t.

[0051] For each time point t, observations from the past n time points are collected, that is, all observations from t-n+1 to t. Windowing is repeated, and a corresponding simple moving average is calculated for each time point t, forming a smoothed data series. The resulting simple moving average series can be used to identify long-term trends, seasonal patterns, or cyclical changes, and process parameters or other quality control measures can be adjusted accordingly. The simple moving average can effectively filter out random fluctuations and provide clear insight into data trends. For quality control in the lithium battery production process, especially in the welding process, this method can help achieve more accurate and timely process parameter adjustments.

[0052] Optionally, the data distribution can be a normal distribution. For example, for the weld length, the W statistic is 0.987 and the p-value is 0.234. The W value and the p-value are related to the normality test. The W value is a statistic in the Shapiro-Wilk test, a statistical method used to test whether the data conforms to the normal distribution, and is used to measure the degree of fit between the sample data and the theoretical normal distribution. The calculation of the W value is based on the linear relationship between the sample data and the normal distribution, and the range of the W value is between 0 and 1. When the W value is equal to 1, it means that the sample data is completely consistent with the normal distribution. The closer the W value is to 1, the higher the degree of fit between the sample data and the normal distribution, and vice versa. In the Shapiro-Wilk test, the W value is used to determine whether the data conforms to the normal distribution assumption. The p-value is used in hypothesis testing to measure the inconsistency between the result and the null hypothesis (the data conforms to the normal distribution). It is the probability of obtaining the currently observed or more extreme results, assuming that the null hypothesis is true. When the p-value is less than the preset significance level (usually 0.05), we reject the null hypothesis and conclude that the data are unlikely to be normally distributed. If the p-value is greater than the significance level, there is insufficient evidence to reject the null hypothesis, and the data are likely to come from a normal distribution. In the Shapiro-Wilk test, if the p-value is greater than the significance level (for example, 0.05), the hypothesis that the data follow a normal distribution is accepted. Since the p-value is greater than 0.05, we can conclude that the weld length data follow a normal distribution. Based on the weld length (L) data, calculate the mean (μ), standard deviation (σ), and variance (σ²) of this data set.

[0053] The calculation formula for the mean (μ) is: Among them, x1 to x m They represent m welding parameters, such as weld length, penetration, welding speed, etc.

[0054] The calculation formula of standard deviation (σ) (for a given normal distribution, the sample standard deviation formula can be used) is used to estimate the degree of fluctuation of the data set, x j is the data set x1 to x m The jth data point in :

[0055]

[0056] The formula for calculating the variance (σ2) (for a given normal distribution, the variance is directly equal to the parameter σ2), which is also used to measure the fluctuation of the data set:

[0057]

[0058] Using the calculated normal distribution related data and the upper specification limit (USL) and lower specification limit (LSL) of the weld length maintained by the system, the statistical data Cp and Cpk are calculated to evaluate the process capability. Cp and Cpk can be used as two statistics to evaluate process capability. Cp (Process Capability Index) is a capability index that evaluates the relationship between the process center (mean) and the specification limits. It measures the relationship between process fluctuations and specification limits, without considering the process offset (i.e., the relationship between the mean and the specification center). The calculation formula is: Where δ is the correlation coefficient.

[0059] Cpk (Process Capability Index with K factor) is a capability index that takes into account process shifts. It considers both the center and dispersion of the process and is calculated as follows:

[0060] The Cp and Cpk values ​​can be used to assess the process capability of a welding process. If the Cp or Cpk value is greater than 1.33, the process capability is sufficient, and the majority of products produced by the welding process will fall within the specification limits. If the value is between 1.0 and 1.33, the process capability may not be sufficient to ensure that all products fall within the specification limits, but it is acceptable. If the value is less than 1.0, the process capability is severely insufficient, the welding stability is unsatisfactory, and improvement measures are required to ensure product quality.

[0061] Step S104, determining target parameter values ​​corresponding to the plurality of target welding parameters based on historical welding results and the historical parameter values ​​corresponding to the plurality of target welding parameters;

[0062] It can be understood that based on historical welding results and the historical parameter values ​​corresponding to multiple target welding parameters, the target parameter values ​​that best match high-quality welding results in the historical data are found. The target parameter values ​​corresponding to the multiple target welding parameters are then determined to predict possible welding results under different parameter combinations, thereby determining the optimal parameters. By analyzing historical data, the parameter combination most likely to produce high-quality welding results can be identified, thereby achieving parameter optimization.

[0063] In an optional embodiment, based on historical welding results and historical parameter values ​​corresponding to multiple target welding parameters, target parameter values ​​corresponding to multiple target welding parameters are determined, including: determining a fitting relationship between historical welding results and historical parameter values ​​corresponding to multiple target welding parameters; determining prediction results corresponding to multiple candidate parameter combinations according to the fitting relationship, wherein the multiple candidate parameter combinations are obtained by setting multiple target welding parameters to different candidate parameter values; determining welding strength scores corresponding to the multiple candidate parameter combinations based on the prediction results corresponding to the multiple candidate parameter combinations; determining the parameter combination with the highest welding strength score among the multiple candidate parameter combinations as the target parameter combination; and obtaining the target parameter values ​​corresponding to the target welding parameters according to the target parameter combination.

[0064] It can be understood that the fitting relationship is determined by analyzing a large amount of historical welding data as input (i.e., the historical parameter values ​​corresponding to the target welding parameters) and the quality of the welding results as output. The mathematical relationship between these input parameters and the output results, i.e., the fitting relationship, is determined. The above-mentioned fitting relationship can be linear, nonlinear, or in the form of a function. Each candidate parameter combination is input into the established model to predict the welding strength score of the welding process. Using the relationship between welding parameters and welding results, the output score reflects the expected quality level of the welding result under a given parameter combination. From multiple candidate parameter combinations, the combination with the highest predicted welding strength score is selected as the target parameter combination. The target parameter combination represents the optimal parameter combination that can theoretically produce the best welding quality under the current model and historical data. By predicting and selecting the optimal parameter combination, it is ensured that the welding process operates under the optimal parameters, thereby significantly improving the quality of welded products.

[0065] Optionally, the historical welding data may include welding speed, defocusing amount, weld width, pressure head pressure, etc.

[0066] Optionally, the welding results include, for example, weld penetration depth, weld width, whether there are welding defects, etc.

[0067] Optionally, statistical analysis or machine learning techniques, such as linear regression, support vector machine, neural network, etc., are used to determine the fitting relationship.

[0068] Optionally, the values ​​of each parameter in the target parameter combination are determined as target parameter values ​​corresponding to the target welding parameters. The updated parameter values ​​of these key parameters will be used in the actual welding production process to obtain the best welding results and product yield.

[0069] Optionally, based on the determined fitting relationship, multiple candidate parameter combinations are generated. Each candidate parameter combination represents a specific set of parameter values ​​that may be used in the welding process, such as different welding speeds and defocus values. These combinations can be generated through exhaustive enumeration, randomly generated, or set based on historical experience. Each set of parameter values ​​is selected within the given upper and lower parameter limits, in a fixed step size or randomly, to form a set of candidate parameter combinations. Through continuous learning and optimization, production efficiency and product yield can be improved in the long term.

[0070] In an optional embodiment, the method further includes: performing image acquisition on the historical welding area to obtain a historical acquired image; processing the historical acquired image using a corresponding image processing method according to predetermined defect extraction requirements to obtain two-dimensional welding information; obtaining welding defect characteristics based on the two-dimensional welding information; and labeling the welding defect characteristics with defect types to obtain historical welding results.

[0071] It can be understood that images of historical welding areas are collected, and specific image processing techniques such as edge detection, grayscale transformation, and filtering are used to extract two-dimensional welding information from the collected historical welding images. Two-dimensional welding information includes the shape, size, surface quality, etc. of the weld, which is the basis for subsequent feature extraction and defect type judgment. Based on the extracted two-dimensional welding information, the image features are further analyzed to identify and quantify the specific manifestations of welding defects. These features will be used to train machine learning models to identify different types of welding defects. The identified welding defect features are classified and labeled, and the defect type of each historical welding area is clearly marked. These annotation results, together with the defect features and corresponding process parameters, constitute the historical welding result dataset, which is a key data source for subsequent model training and optimization.

[0072] By collecting and annotating a large amount of historical welding data, we provide a rich source of learning material for the intelligent diagnosis algorithm model, helping it better understand the patterns of welding defects under different process parameter settings, thereby improving the model's accuracy and generalization capabilities. As data continues to accumulate, the model's performance will be further improved, achieving long-term stability and continuous improvement in welding quality.

[0073] Alternatively, within a welding production line, images of product areas that have been welded over a period of time are collected. These images are from welded products that have been recorded. The purpose of collecting these images is to build a historical database of various welding defects for subsequent analysis and model training.

[0074] Optionally, two-dimensional welding can characterize features such as the size of the weld spot, the continuity of the weld, the smoothness of the surface, and the presence of cracks or pores. An industrial camera is used to capture images of the weld area, and feature extraction technology is applied to filter out key information from the image, such as basic weld spot location, shape, texture, etc. More detailed features can also be extracted for specific defect types, such as edge irregularities, color changes, and other features. For pores, dark or light spots in the image can be detected.

[0075] Alternatively, a convolutional neural network (CNN) can be used to build a machine learning or deep learning model for defect identification based on weld defect characteristics. During training, a loss function (cross entropy loss) and an optimization algorithm (gradient descent) are used to optimize model parameters. For the trained model, evaluation metrics including accuracy, precision, recall, and F1 score are introduced. Based on the evaluation results, the model is iteratively optimized until satisfactory performance is achieved.

[0076]

[0077] TP (True Positive) is the number of true positive examples predicted by the model as positive. FP (False Positive) is the number of false positive examples predicted by the model as positive. FN (False Negative) is the number of false negative examples predicted by the model as negative. TN (True Negative) is the number of true negative examples predicted by the model as negative.

[0078] Figure 3 This is a second schematic diagram of an optional welding strategy determination method provided in an embodiment of the present application, such as Figure 3 As shown in the figure, as recall increases, accuracy decreases. This relies on a well-trained data model. During the welding process, data is collected and features are extracted in real time. The extracted features are then fed into the corresponding model in the model library for diagnosis. Based on the model's output, welding quality is determined and annotated, forming a data result set. The predicted data and actual results are then cyclically written into the model training set, achieving the machine's autonomous learning capability.

[0079] In an optional embodiment, welding defect characteristics are obtained based on two-dimensional welding information, including: laser scanning the historical welding area to obtain three-dimensional point cloud data; determining the three-dimensional surface characteristics of the historical welding area based on the three-dimensional point cloud data; and performing feature stitching based on the two-dimensional welding information and the three-dimensional surface characteristics to determine the welding defect characteristics.

[0080] It can be understood that a 3D (three-dimensional) laser scanning device can be used to scan the same historical welding area to obtain its three-dimensional point cloud data. Point cloud data can accurately reflect the three-dimensional surface features of the welding area. Based on the three-dimensional point cloud data, the three-dimensional surface features of the historical welding area are determined through the point cloud processing algorithm. The three-dimensional surface features can supplement the depth and spatial information that cannot be reflected in the two-dimensional image information, and provide more comprehensive data support for the evaluation of welding quality. The features extracted based on the two-dimensional welding information are fused with the data obtained based on the three-dimensional surface features, and a comprehensive welding defect feature is formed through feature stitching technology, which can be represented in the form of a feature vector. Based on the analysis of the comprehensive welding defect features, the welding process parameters, such as welding speed, defocus, welding gun pressure, etc., can be adjusted in reverse to achieve the optimal welding effect and reduce the occurrence of welding defects.

[0081] Through real-time data collection and feature analysis, the system automatically adjusts welding equipment parameters and continuously optimizes control models based on weld quality feedback, forming an intelligent closed-loop control process that effectively improves product yield and reduces defective rates. The application of automated feature extraction and defect recognition technology reduces reliance on manual visual inspection, improves production efficiency and stability, and reduces production costs.

[0082] Optionally, the three-dimensional surface features may include: the degree of protrusion of the welding point, surface smoothness, spatial distribution of holes or irregular shapes, etc.

[0083] Optionally, the above-mentioned point cloud processing algorithm can be multiple, such as calculating normal vectors, curvature, surface area, volume, etc. LiDAR is used to scan the welding area to obtain three-dimensional point cloud data, which is then processed to extract geometric features (such as normal vectors, curvature, etc.) and descriptor features (such as normal histograms, surface area, etc.) to obtain information such as the three-dimensional shape, smoothness, and concave-convex changes of the welding surface.

[0084] Optionally, the above-mentioned welding defect features include geometric information, texture information and spatial morphological information of the welding area, which can more accurately identify and classify welding defects such as hot spots, cold welds, leaking welds, weld penetrations, pores, etc.

[0085] Optionally, the above-mentioned feature stitching techniques can be of various types, such as position-based alignment and weighted feature fusion. This is achieved by stitching the two feature vectors together and fusing them using a weighted algorithm in conjunction with a vector threshold maintained by the system. The fused feature vector will contain comprehensive information about the weld area, including both 2D image information (i.e., 2D weld information) and 3D morphological information (3D surface features), facilitating the establishment of a labeled dataset containing samples of various defect types in subsequent data processing.

[0086] In step S106 , a target control strategy is obtained according to target parameter values ​​corresponding to the plurality of target welding parameters.

[0087] As you can see, the target control strategy allows real-time adjustment of key parameters during the welding process to address production changes and ensure consistent welding quality. This achieves a closed-loop control process from parameter analysis to optimization and real-time adjustment, ensuring continuous improvement and optimization of the welding process. This automated control strategy reduces the need for continuous manual monitoring and adjustments, improving production efficiency.

[0088] In an optional embodiment, the method further includes: determining a strategy update frequency based on a device type of the welding device; and updating the target control strategy to the welding device according to the strategy update frequency.

[0089] It's understandable that the IoT platform can collect device data to identify the type of welding equipment, including manufacturer, model, power, and level of automation. Based on the equipment type, a range for policy update frequency can be pre-set or determined through historical data analysis. For example, for equipment with a high degree of automation and strong data processing capabilities, a shorter policy update cycle can be set to achieve faster response and more refined control.

[0090] Optionally, a reasonable strategy update frequency can be determined based on the specific conditions of the welding equipment, such as a fixed time interval (such as hourly, daily, or per batch), or a dynamic frequency based on the equipment status. Based on the predicted optimal process parameter combination, a target control strategy is generated. The target control strategy needs to match the actual status and capabilities of the welding equipment to ensure its feasibility and effectiveness. By continuously collecting 3D inspection data and using machine learning algorithms, the target control strategy is continuously optimized and adjusted to adapt to changes in the production environment and improvements in equipment performance.

[0091] Based on the established strategy update frequency, the optimized target control strategy is updated in real time to the welding equipment, enabling automatic adjustment and optimization of equipment parameters such as welding speed, defocus, and indenter pressure to ensure continuity and quality stability during the welding process. Through real-time data feedback, the equipment can dynamically respond to changes in optimal parameters, further improving the controllability of welding quality and production efficiency.

[0092] In step S102, multiple target welding parameters are determined according to the first correlation data, where the first correlation data indicates the degree of correlation between different candidate welding parameters and welding results. In step S104, target parameter values ​​corresponding to the multiple target welding parameters are determined based on historical welding results and the historical parameter values ​​corresponding to the multiple target welding parameters. In step S106, a target control strategy is obtained based on the target parameter values ​​corresponding to the multiple target welding parameters. This achieves the goal of optimizing the welding control strategy according to the welding results, improves the accuracy of the welding control strategy, and thereby improves the welding yield rate. This solves the technical problem in related arts of reduced welding yield due to unsatisfactory control strategy settings.

[0093] Based on the above-mentioned embodiments and alternative embodiments, this application proposes an alternative implementation method for closed-loop quality control of the lithium battery welding production process. This model designs an intelligent diagnostic algorithm model for the diverse defect classification and model result prediction of current welding production. This model uses feature vectorization of industrial cameras and 3D laser inspection to predict welding result trends and reversely optimize the combination of key welding process parameters to form a closed-loop quality control loop, thereby improving welding production quality.

[0094] Step S1, confirming key welding process parameters, such as welding speed, defocusing amount, and press head pressure.

[0095] Step S2: confirm key welding characteristic parameters, such as weld length, weld width, penetration depth, weld width, etc.

[0096] Step S3: Use the IoT platform to create the device type and corresponding data model, and configure the collection parameters and frequency. This is done using the IoT platform. The collection parameters and frequency are configured based on the device type and data model. These parameters include the type of data to be collected, the data format, and the data unit. To ensure real-time data, the collection frequency is set to a "change-as-you-go" mode.

[0097] Step S4: Automatically collect data using communication methods such as TCP / IP (Transmission Control Protocol / Internet Protocol), MQTT (Message Queuing Telemetry Transport), and OPCUA (Open Platform Communications Unified Architecture). High-performance storage platforms are used for big data storage. Configure TCP / IP, MQTT, OPCUA, and other communication methods to automatically collect data, send device data to a central server or cloud platform, and use high-performance storage platforms such as distributed file systems or databases to store big data.

[0098] Step S5: configuring process standards and setting upper and lower parameter thresholds. The process standards and upper and lower parameter thresholds can be configured using a manufacturing system based on the welding production process file.

[0099] In step S6, data preprocessing is performed on the data collected by the big data system, including data cleaning, feature selection, time series data preparation, classification variable encoding, etc., to form training sets and test sets for use by subsequent algorithms. At the same time, the system generates work-in-progress welding archives to display production process traceability data and key process analysis results in an intuitive way.

[0100] In step S7, based on multi-source feature fusion, a model library containing diagnostic algorithms is established to accurately classify, efficiently label, and intelligently index welding features, thereby realizing online real-time diagnosis and judgment of welding quality.

[0101] Step S8: construct a quality inspection and process parameter feature algorithm model, and predict and reversely analyze the optimal process parameter combination by analyzing the inherent relationship between the features and the process parameters.

[0102] Based on the feature sets of the training set and test set in step S6 and the result data set generated in step S7, a quality inspection and process parameter feature algorithm model is constructed. By analyzing the intrinsic relationship between the features and process parameters, the optimal process parameter combination is predicted and reversely analyzed.

[0103] A linear regression algorithm is used to predict weld strength. A linear regression model is trained based on stored historical data, including key coating quality parameters (inner ring weld power, outer ring weld power, weld height, weld speed, cooling water flow) and one or more target variables (such as weld strength score).

[0104] The linear regression model can be formulated as follows: y = β0 + β1X1 + β2X2 + β3X3 + β4X4 + β5X5. Here, y is the target variable (i.e., weld strength), and X1, X2, X3, X4, and X5 are the characteristic values ​​of different welding parameters, such as welding power (inner loop), welding power (outer loop), welding height, welding speed, and cooling water flow. β0, β1, β2, β3, β4, and β5 are the model parameters (intercept and slope). Figure 4 is a third schematic diagram of an optional welding strategy determination method provided in an embodiment of the present application, such as Figure 4 As shown in the figure, a fitting trend can be fitted from multiple data points, with the horizontal axis representing welding parameters and the vertical axis representing welding strength. The welding strength score is predicted based on the new data points, and the optimal process parameter combination is found in reverse according to the prediction results.

[0105] Train the linear regression model using the training set and the selected result feature set, setting the model learning rate and number of iterations. During training, the model will continuously adjust the intercept and slope terms. The model's predicted results will be compared with the true values ​​and displayed in a scatter plot, drawing the best fit line, and showing the model's prediction trend.

[0106] At the same time, the mean square error (MSE) loss function is introduced to measure the prediction performance of the model. The calculation formula of the mean square error (MSE) is:

[0107] Where q represents the number of samples. i represents the true value, y pred Represents the predicted value. The accuracy of the prediction model is measured by calculating the mean of the squared differences between the predicted value and the true value. A smaller MSE value indicates a better prediction model, meaning the model is more accurate in predicting the true value.

[0108] In step S9, based on the predicted optimal parameters, the key process parameters of the welding equipment are dynamically adjusted, and the control model is continuously optimized through continuous data feedback to achieve closed-loop management and improvement of welding quality.

[0109] The above optional implementation method achieves at least the following effects: receiving various data in the welding production process in real time, extracting local image features and point cloud data features, and using advanced algorithm models for rapid analysis to accurately identify potential quality problem points, such as blast points, cold welds, leaking welds, weld penetrations, pores, etc. The system automatically summarizes production data, and through the intelligent diagnostic algorithm model of the present invention, combined with automatic analysis of normal distribution, it can accurately evaluate the manufacturing capacity of welding production, ensure that product quality conforms to the law of normal distribution, and further improve production stability and product quality. The algorithm model has the ability to continuously learn and can continuously adapt to changes in the welding production process. Through real-time data regression feedback analysis, the algorithm and model are optimized to further improve the accuracy and efficiency of welding quality detection. The intelligent diagnostic algorithm model can realize linkage control with welding equipment, adjust equipment parameters in real time according to the optimal parameter combination of reverse analysis of the model, realize automated and intelligent management of the welding production process, and ensure efficient, stable and high-quality welding production.

[0110] It should be noted that the steps shown in the flowcharts of the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions, and that, although a logical order is shown in the flowcharts, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0111] This embodiment also provides a welding strategy determination device for implementing the aforementioned embodiments and preferred implementations. Details already described will not be repeated. As used below, the terms "module" and "device" may refer to a combination of software and / or hardware that implements a predetermined function. While the devices described in the following embodiments are preferably implemented using software, implementation using hardware, or a combination of software and hardware, is also possible and contemplated.

[0112] According to an embodiment of the present application, there is also provided an embodiment of a device for implementing a welding strategy determination method. Figure 5 is a schematic diagram of a welding strategy determination device according to an embodiment of the present application, such as Figure 5 As shown, the welding strategy determination device includes: a parameter screening module 502, a parameter determination module 504, and a strategy generation module 506. The device is described below.

[0113] a parameter screening module 502 for determining a plurality of target welding parameters according to first correlation data, wherein the first correlation data indicates a degree of correlation between different candidate welding parameters and the welding result;

[0114] a parameter determination module 504 connected to the parameter screening module 502 and configured to determine target parameter values ​​corresponding to the plurality of target welding parameters based on historical welding results and the historical parameter values ​​corresponding to the plurality of target welding parameters;

[0115] The strategy generation module 506 is connected to the parameter determination module 504 and is used to obtain a target control strategy according to target parameter values ​​corresponding to a plurality of target welding parameters.

[0116] In a welding strategy determination device provided in an embodiment of the present application, a parameter screening module 502 is provided for determining multiple target welding parameters according to first associated data, wherein the first associated data represents the degree of association between different candidate welding parameters and welding results; a parameter determination module 504 is connected to the parameter screening module 502 and is used to determine target parameter values ​​corresponding to the multiple target welding parameters based on historical welding results and historical parameter values ​​corresponding to the multiple target welding parameters; and a strategy generation module 506 is connected to the parameter determination module 504 and is used to obtain a target control strategy based on the target parameter values ​​corresponding to the multiple target welding parameters. The device achieves the purpose of optimizing the welding control strategy according to the welding results, improves the accuracy of the welding control strategy, and thereby improves the welding yield rate, thereby solving the technical problem in the related art of poor control strategy setting leading to a reduced welding yield rate.

[0117] It should be noted that the above modules can be implemented by software or hardware. For example, for the latter, it can be implemented in the following ways: the above modules can be located in the same processor; or the above modules can be located in different processors in any combination.

[0118] It should be noted that the parameter screening module 502, parameter determination module 504, and strategy generation module 506 correspond to steps S102 to S106 in the embodiment. The examples and application scenarios implemented by these modules and the corresponding steps are the same, but are not limited to the contents disclosed in the above embodiment. It should be noted that these modules, as part of the device, can be run on a computer terminal.

[0119] It should be noted that the optional or preferred implementation of this embodiment can be found in the relevant description in the embodiment, which will not be repeated here.

[0120] The above-mentioned welding strategy determination device may also include a processor and a memory. The parameter screening module 502, the parameter determination module 504, the strategy generation module 506, etc. are all stored in the memory as program units, and the processor executes the above-mentioned program units stored in the memory to realize the corresponding functions.

[0121] The processor includes a kernel, which retrieves the corresponding program unit from memory. There can be one or more kernels. Memory may include non-permanent memory in a computer-readable medium, random access memory (RAM), and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. Memory includes at least one memory chip.

[0122] An embodiment of the present application provides a non-volatile storage medium having a program stored thereon, which implements a welding strategy determination method when executed by a processor.

[0123] An embodiment of the present application provides an electronic device comprising a processor, a memory, and a program stored in the memory and executable on the processor. When the processor executes the program, the following steps are performed: determining multiple target welding parameters according to first association data, wherein the first association data indicates the degree of association between different candidate welding parameters and welding results; determining target parameter values ​​corresponding to the multiple target welding parameters based on historical welding results and historical parameter values ​​corresponding to the multiple target welding parameters; and obtaining a target control strategy based on the target parameter values ​​corresponding to the multiple target welding parameters. The device herein may be a server, a PC, or the like.

[0124] The present application also provides a computer program product, which, when executed on a data processing device, is suitable for executing a program initialized with the following method steps: determining multiple target welding parameters according to first associated data, wherein the first associated data represents the degree of association between different candidate welding parameters and welding results; determining target parameter values ​​corresponding to the multiple target welding parameters based on historical welding results and historical parameter values ​​corresponding to the multiple target welding parameters; and obtaining a target control strategy based on the target parameter values ​​corresponding to the multiple target welding parameters.

[0125] Those skilled in the art will appreciate that the embodiments of the present application can be provided as methods, systems, or computer program products. Therefore, the present application can adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application can adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) that contain computer-usable program code.

[0126] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the steps in the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0127] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0128] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0129] In a typical configuration, a computing device includes one or more processors (CPUs), input / output interfaces, network interfaces, and memory.

[0130] The memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. The memory is an example of a computer-readable medium.

[0131] Computer-readable media includes permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. The information can be computer-readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disc read-only memory (CD-ROM), digital versatile disc (DVD) or other optical storage, magnetic cassettes, magnetic disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer-readable media does not include transitory computer-readable media (transitory media), such as modulated data signals and carrier waves.

[0132] It should also be noted that the terms "comprises," "includes," or any other variations thereof are intended to encompass non-exclusive inclusion, such that a process, method, commodity, or apparatus that includes a series of elements includes not only those elements but also other elements not explicitly listed, or includes elements inherent to such process, method, commodity, or apparatus. In the absence of further limitations, an element defined by the phrase "comprises a ..." does not exclude the presence of other identical elements in the process, method, commodity, or apparatus that includes the element.

[0133] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Furthermore, the present application may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0134] The above are merely embodiments of the present application and are not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present application should all be included within the scope of the claims of the present application.

Claims

1. A method for determining a welding strategy, characterized in that: include: Determining a plurality of target welding parameters according to first correlation data includes: determining second correlation data indicating a degree of correlation between a plurality of candidate welding parameters; screening the plurality of candidate welding parameters based on the first correlation data and the second correlation data to determine the plurality of target welding parameters, wherein the first correlation data indicates a degree of correlation between different candidate welding parameters and welding results; Determining target parameter values ​​corresponding to the multiple target welding parameters based on historical welding results and historical parameter values ​​corresponding to the multiple target welding parameters, wherein the target parameter values ​​are obtained by determining a parameter combination with the highest welding strength score among multiple candidate parameter combinations as the target parameter combination, and based on the target parameter combination, wherein the multiple candidate parameter combinations are respectively obtained by setting the multiple target welding parameters to different candidate parameter values; A target control strategy is obtained according to the target parameter values ​​respectively corresponding to the multiple target welding parameters.

2. The method according to claim 1, characterized in that The determining, based on historical welding results and historical parameter values ​​corresponding to the plurality of target welding parameters, target parameter values ​​corresponding to the plurality of target welding parameters respectively include: Determining a fitting relationship between the historical welding results and the historical parameter values ​​corresponding to the plurality of target welding parameters; Determining prediction results corresponding to a plurality of candidate parameter combinations according to the fitting relationship; Determining welding strength scores corresponding to the plurality of candidate parameter combinations based on prediction results corresponding to the plurality of candidate parameter combinations; Determine the parameter combination with the highest welding strength score among the multiple candidate parameter combinations as the target parameter combination; According to the target parameter combination, target parameter values ​​corresponding to the target welding parameters are obtained.

3. The method according to claim 1, characterized in that The method further comprises: Perform image acquisition on the historical welding area to obtain a historical acquisition image; According to the predetermined defect extraction requirements, the historically collected images are processed using a corresponding image processing method to obtain two-dimensional welding information; Obtaining welding defect characteristics based on the two-dimensional welding information; The welding defect characteristics are labeled with defect types to obtain the historical welding results.

4. The method according to claim 3, characterized in that The obtaining of welding defect characteristics based on the two-dimensional welding information includes: Performing laser scanning on the historical welding area to obtain three-dimensional point cloud data; Determining three-dimensional surface features of the historical welding area based on the three-dimensional point cloud data; Feature splicing is performed based on the two-dimensional welding information and the three-dimensional surface features to determine the welding defect features.

5. The method according to claim 1, characterized in that The screening of the plurality of candidate welding parameters based on the first associated data and the second associated data to determine the plurality of target welding parameters includes: Determining third correlation data indicating the degree of correlation between the plurality of candidate welding parameters and the time series characteristics, wherein the time series characteristics include at least the operation time of the welding equipment and seasonal characteristics; determining fourth correlation data indicating a correlation degree between the plurality of candidate welding parameters and welding stability based on data distributions corresponding to the plurality of candidate welding parameters; Based on the first correlation data, the second correlation data, the third correlation data, and the fourth correlation data, the plurality of candidate welding parameters are screened to determine the plurality of target welding parameters.

6. The method according to any one of claims 1 to 5, characterized in that The method further comprises: Determine the policy update frequency based on the equipment type of the welding equipment; The target control strategy is updated to the welding equipment according to the strategy update frequency.

7. A welding strategy determination device, characterized in that: include: a parameter screening module, configured to determine a plurality of target welding parameters according to first correlation data, comprising: determining second correlation data indicating a degree of correlation between a plurality of candidate welding parameters; screening the plurality of candidate welding parameters based on the first correlation data and the second correlation data to determine the plurality of target welding parameters, wherein the first correlation data indicates a degree of correlation between different candidate welding parameters and welding results; a parameter determination module, configured to determine target parameter values ​​corresponding to the plurality of target welding parameters based on historical welding results and the historical parameter values ​​corresponding to the plurality of target welding parameters; A strategy generation module is used to obtain a target control strategy based on the target parameter values ​​corresponding to the multiple target welding parameters, wherein the target parameter value is obtained by determining the parameter combination with the highest welding strength score among multiple candidate parameter combinations as the target parameter combination, and the multiple candidate parameter combinations are obtained by setting the multiple target welding parameters to different candidate parameter values.

8. A non-volatile storage medium, characterized in that: The non-volatile storage medium stores a plurality of instructions, and the instructions are suitable for being loaded by a processor and executed by the welding strategy determination method according to any one of claims 1 to 6.

9. An electronic device, characterized in that: include: One or more processors and a memory, the memory being used to store one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors implement the welding strategy determination method according to any one of claims 1 to 6.

Images