Welding production process monitoring method and related equipment

Abstract

The invention provides a welding production process monitoring method and related equipment, and relates to the technical field of energy storage. The method comprises the steps that welding data in the welding production process are obtained, the welding production process comprises a plurality of welding procedures, and the welding data comprise data corresponding to a plurality of welding parameters of each welding procedure; processing the welding data corresponding to the plurality of welding parameters of each welding process through a first anomaly detection model to obtain single-process anomaly data of each welding process; processing the welding data corresponding to the plurality of welding parameters of the plurality of welding procedures through a second anomaly detection model to obtain multi-procedure anomaly data of the plurality of welding procedures; and determining and marking target abnormal data according to the single-process abnormal data and the multi-process abnormal data.

Description

Background Art

[0002] Welding is a key manufacturing process in product production, and welding quality is crucial to the function and reliability of the product.

[0003] In the related art, welding process monitoring usually relies on a fixed threshold method or a rule-based method to identify anomalies. However, the above monitoring methods are not effective when dealing with different materials and welding conditions, and are prone to false positives and false negatives, reducing the accuracy of monitoring. Summary of the invention

[0004] The present disclosure provides a welding production process monitoring method, device, equipment and storage medium, which at least to a certain extent overcome the problem that the monitoring method of the welding production process in the related art has the risk of underreporting or false reporting, and reduces the monitoring accuracy.

[0005] Other features and advantages of the present disclosure will become apparent from the following detailed description, or may be learned in part by the practice of the present disclosure.

[0006] According to one aspect of the present disclosure, a welding production process monitoring method is provided, comprising: acquiring welding data in the welding production process, the welding production process comprising a plurality of welding processes, the welding data comprising data corresponding to a plurality of welding parameters of each welding process; processing the welding data corresponding to the plurality of welding parameters of each welding process by a first abnormality detection model to obtain single-process abnormal data of each welding process; processing the welding data corresponding to the plurality of welding parameters of the plurality of welding processes by a second abnormality detection model to obtain multi-process abnormal data of the plurality of welding processes; and determining and marking target abnormal data based on the single-process abnormal data and the multi-process abnormal data.

[0007] In one embodiment of the present disclosure, the first anomaly detection model is constructed based on an unsupervised learning algorithm, and the unsupervised learning algorithm includes at least one of a clustering algorithm and an anomaly detection algorithm.

[0008] In one embodiment of the present disclosure, the unsupervised learning algorithm is trained based on unlabeled historical welding data with abnormal data of a preset ratio threshold as the target.

[0009] In one embodiment of the present disclosure, the second anomaly detection model is constructed based on a single classification algorithm, and the single classification algorithm includes at least one of an isolation forest, a single-class support vector machine, and a deep random neural network.

[0010] In one embodiment of the present disclosure, determining and marking target abnormal data according to the single-process abnormal data and the multiple-process abnormal data includes: determining the intersection of the single-process abnormal data and the multiple-process abnormal data as the target abnormal data.

[0011] In one embodiment of the present disclosure, after determining and marking the target abnormal data, the method further includes: if the deviation of the target abnormal data from the preset process control value is greater than a preset deviation threshold, triggering a parameter adjustment instruction, the parameter adjustment instruction including the target abnormal data and the welding parameters corresponding to the target abnormal data, so as to adjust the value of the welding parameter corresponding to the target abnormal data to within a preset parameter range according to the target abnormal data and the preset adjustment value.

[0012] In one embodiment of the present disclosure, the method also includes: performing trend analysis on multiple welding data corresponding to target welding parameters of each welding process to obtain a trend analysis result, wherein the target welding parameter is one of the multiple welding parameters; if the trend analysis result is that the multiple welding data corresponding to the target welding parameter show a continuous upward trend or a continuous downward trend, then generating early warning information, which includes the target welding parameter and the trend analysis result.

[0013] In one embodiment of the present disclosure, the method also includes: determining a welding result corresponding to welding data corresponding to multiple welding parameters; updating the training data sets of the first anomaly detection model and the second anomaly detection model with the welding result, triggering an incremental learning task to perform model optimization on the first anomaly detection model and the second anomaly detection model.

[0014] In one embodiment of the present disclosure, after obtaining the welding data in the welding production process, the method further includes: preprocessing the welding data, wherein the preprocessing includes at least one of data cleaning, data filling, data standardization, and feature selection.

[0015] In one embodiment of the present disclosure, the method further includes: storing the welding data, the target abnormality data and the adjusted welding data in a database.

[0016] According to another aspect of the present disclosure, a welding production process monitoring device is provided, including: a data acquisition module, used to acquire welding data in the welding production process, the welding production process includes multiple welding processes, and the welding data includes data corresponding to multiple welding parameters of each welding process; a first detection module, used to process the welding data corresponding to the multiple welding parameters of each welding process through a first abnormality detection model to obtain single-process abnormal data of each welding process; a second detection module, used to process the welding data corresponding to the multiple welding parameters of the multiple welding processes through a second abnormality detection model to obtain multi-process abnormal data of the multiple welding processes; an abnormality determination module, used to determine and mark target abnormal data based on the single-process abnormal data and the multi-process abnormal data.

[0017] According to another aspect of the present disclosure, an electronic device is also provided, which includes: a processor; and a memory for storing executable instructions of the processor; wherein the processor is configured to execute any one of the above-mentioned welding production process monitoring methods by executing the executable instructions.

[0018] According to another aspect of the present disclosure, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the welding production process monitoring method described in any one of the above is implemented.

[0019] According to another aspect of the present disclosure, a computer program product is also provided, including a computer program, and when the computer program is executed by a processor, the welding production process monitoring method described above is implemented.

[0020] In an embodiment of the present disclosure, welding data in a welding production process is obtained, the welding production process includes multiple welding processes, and the welding data includes data corresponding to multiple welding parameters of each welding process; the welding data corresponding to the multiple welding parameters of each welding process is processed by a first abnormality detection model to obtain single-process abnormal data of each welding process; the welding data corresponding to the multiple welding parameters of multiple welding processes is processed by a second abnormality detection model to obtain multi-process abnormal data of multiple welding processes; according to the single-process abnormal data and the multi-process abnormal data, the target abnormal data is determined and marked, and various abnormal situations in the welding process can be accurately identified by the first abnormality detection model and the second abnormality detection model, thereby reducing false alarms and missed alarms. Compared with the monitoring method using fixed thresholds or based on rules in the related art, the monitoring method of the present disclosure can better adapt to different welding environments and welding conditions, and improve the reliability and stability of the welding process.

[0021] It is to be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the present disclosure. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] The accompanying drawings herein are incorporated into the specification and constitute a part of the specification, illustrate embodiments consistent with the present disclosure, and together with the specification are used to explain the principles of the present disclosure. Obviously, the accompanying drawings described below are only some embodiments of the present disclosure, and for ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without creative work.

[0023] Figure 1 A structural schematic diagram of a welding production process monitoring system in an embodiment of the present disclosure is shown.

[0024] Figure 2 A flow chart of a welding production process monitoring method in an embodiment of the present disclosure is shown.

[0025] Figure 3 A flow chart of another welding production process monitoring method in an embodiment of the present disclosure is shown.

[0026] Figure 4 A flow chart of another welding production process monitoring method in an embodiment of the present disclosure is shown.

[0027] Figure 5 A flow chart of another welding production process monitoring method in an embodiment of the present disclosure is shown.

[0028] Figure 6 A flow chart of the model optimization method in an embodiment of the present disclosure is shown.

[0029] Figure 7 A flowchart showing an example of a data preprocessing method in an embodiment of the present disclosure is shown.

[0030] Figure 8 An example flow chart of a welding production process monitoring method in an embodiment of the present disclosure is shown.

[0031] Fig. 9 An example flow chart of an automated welding parameter adjustment method in an embodiment of the present disclosure is shown.

[0032] Fig.10 A schematic diagram showing an example of a model optimization method in an embodiment of the present disclosure.

[0033] Fig.11 A schematic diagram of a welding production process monitoring device in an embodiment of the present disclosure is shown.

[0034] Fig.12 A structural block diagram of an electronic device in an embodiment of the present disclosure is shown. DETAILED DESCRIPTION

[0035] Example embodiments will now be described more fully with reference to the accompanying drawings. However, example embodiments can be implemented in a variety of forms and should not be construed as limited to the examples set forth herein; rather, these embodiments are provided so that the disclosure will be more comprehensive and complete and to fully convey the concepts of the example embodiments to those skilled in the art. The described features, structures, or characteristics may be combined in any suitable manner in one or more embodiments.

[0036] In addition, the accompanying drawings are only schematic illustrations of the present disclosure and are not necessarily drawn to scale. The same reference numerals in the figures represent the same or similar parts, and their repeated description will be omitted. Some of the block diagrams shown in the accompanying drawings are functional entities and do not necessarily correspond to physically or logically independent entities. These functional entities can be implemented in software form, or implemented in one or more hardware modules or integrated circuits, or implemented in different networks and / or processor devices and / or microcontroller devices.

[0037] Welding usually refers to the welding of metals. It is a forming method that uses heating or pressure, or both, to create an atomic bond between two separate objects and connect them into one.

[0038] According to the degree of heating and process characteristics during the welding process, welding can be divided into fusion welding, pressure welding, brazing, etc.

[0039] Fusion welding is to heat the welding part of the workpiece locally to a molten state to form a molten pool, usually adding filler metal, which forms a weld after cooling and crystallization, and the welded workpieces are combined into an inseparable whole. Fusion welding includes gas welding, arc welding, electroslag welding, plasma arc welding, electron beam welding, laser welding, etc.

[0040] Pressure welding is a welding method that requires pressure regardless of whether heating is used during the welding process. Pressure welding includes resistance welding, friction welding, cold pressure welding, diffusion welding, explosion welding, etc.

[0041] Brazing uses a brazing material with a melting point lower than the metal being welded. After melting, it fills the joint gap and diffuses with the metal being welded to achieve connection. During the brazing process, the workpieces being welded do not melt and generally do not have plastic deformation.

[0042] Welding can be used in machine manufacturing, shipbuilding, construction engineering, power equipment production, energy storage, aviation and aerospace industries, and other application scenarios.

[0043] The following is an explanation using the preparation of a lithium battery as an example.

[0044] like Figure 1As shown, the lithium battery production process mainly includes the following steps: positive electrode ingredients, positive electrode coating, positive electrode sheeting, positive electrode sheet preparation, negative electrode ingredients, negative electrode coating, negative electrode sheeting, negative electrode sheet preparation, winding, shelling, spot welding, baking, liquid injection, cap welding, cleaning, and storage.

[0045] Positive electrode ingredients: The positive electrode materials of lithium batteries include active substances, conductive agents, binders, etc. The positive electrode materials are mixed according to a certain adding order, adding time, and adding ratio to obtain positive electrode slurry.

[0046] Negative electrode ingredients: The negative electrode materials of the lithium electrode include active substances, conductive agents, binders, dispersants, etc. The negative electrode materials are prepared in a similar way to the positive electrode slurry to obtain the negative electrode slurry.

[0047] Positive electrode coating: The positive electrode slurry is attached to the AB surface of the aluminum current collector by extrusion coating or spraying.

[0048] Negative electrode coating: The negative electrode slurry is attached to the AB surface of the copper current collector by extrusion coating or spraying.

[0049] Positive electrode sheeting: After the positive electrode is coated and dried, the rollers are used to compact the electrode sheets and divide the positive electrode into strips. Hot pressing or cold pressing can be used.

[0050] Negative electrode production: A similar process to the positive electrode production is used for compaction and negative electrode striping.

[0051] Preparation of positive electrode sheet: Dry the positive electrode sheet, weld the aluminum tabs and perform the tab encapsulation process to obtain the positive electrode sheet.

[0052] Preparation of negative electrode sheet: Dry the negative electrode sheet, weld the nickel tabs and coat the tabs with glue to obtain the negative electrode sheet.

[0053] Winding: The separator, positive electrode sheet and negative electrode sheet are prepared into a winding core by a winding machine, and the separator is arranged between the positive electrode sheet and the negative electrode sheet.

[0054] Insert into the shell: After inserting the lower pad into the bottom of the winding core, bend the negative electrode ear so that the ear surface is facing the winding pin hole of the winding core, and insert the winding core vertically into the shell.

[0055] Spot welding: Insert the welding needle into the middle hole of the coil core, roll the groove to fix the coil core in the shell, and perform spot bottom welding.

[0056] Baking: After tumbling, the battery cells are baked to reduce the moisture content in the battery cells.

[0057] Injection: Inject the electrolyte between the positive and negative electrodes.

[0058] Welding cap: Place the cap inlet into the lower mold of the super welding machine, align the positive ear of the battery cell with the ear of the cap, and spot weld the positive ear to the cap.

[0059] Cleaning: Clean the surface of the lithium battery.

[0060] Storage: Dry lithium batteries and store them.

[0061] In the welding process of lithium batteries, different welding parameter values ​​can be used for different types of lithium batteries, and the appropriate welding current, time, pressure, temperature, etc. can be determined according to the battery type and material.

[0062] from Figure 1 It can be found that when a lithium battery is produced by welding, the welding production process may include multiple welding processes. After a single welding process, there may be no abnormality in the monitoring of welding parameters, but the influence of subsequent welding processes will lead to a decrease in the aforementioned welding quality. For example, after the bottom end of the welding pin is welded to the shell, the subsequent welding of the cap may cause the bottom end of the welding pin to become loose, affecting the performance of the battery cell.

[0063] In the related art, the welding process can be monitored by a monitoring system to identify abnormalities, for example, by a fixed threshold method or a rule-based method.

[0064] For the fixed threshold method to monitor whether the welding parameters are abnormal, such as monitoring a single welding parameter such as welding current, welding voltage or welding temperature, when the value of a welding parameter exceeds the threshold, the monitoring system will issue an alarm. This method is simple and easy to implement, but for complex welding environments, such as the impact of the subsequent welding process on the previous welding process, the threshold setting is difficult to cover various complex situations, which can easily lead to false alarms and missed alarms.

[0065] For the empirical rule method, the welding quality is usually judged by the operator's experience and rules. This method is highly dependent on the operator's skills and experience, is difficult to standardize and automate, and reduces welding efficiency.

[0066] In addition, historical data can be used for statistical analysis to identify abnormal situations. This method requires a large amount of data support and is not suitable for scenarios with high real-time requirements.

[0067] It can be seen that the above method has the following problems:

[0068] 1. Limited accuracy: Fixed thresholds and empirical rules perform poorly when dealing with different materials and welding conditions and lack adaptive capabilities.

[0069] 2. Response lag: After detecting an abnormality, the monitoring system has a slow response speed to make adjustments, making it difficult to adjust welding parameters in real time, which may lead to product quality problems.

[0070] 3. Insufficient data utilization: The monitoring system cannot fully utilize historical welding data and process data for in-depth analysis and trend prediction.

[0071] In order to solve at least some of the above-mentioned technical problems, the welding production process monitoring method provided by the present invention obtains welding data in the welding production process, the welding production process includes multiple welding processes, and the welding data includes data corresponding to multiple welding parameters of each welding process; the welding data corresponding to the multiple welding parameters of each welding process is processed by a first abnormality detection model to obtain single-process abnormal data of each welding process; the welding data corresponding to the multiple welding parameters of multiple welding processes is processed by a second abnormality detection model to obtain multi-process abnormal data of multiple welding processes; according to the single-process abnormal data and the multi-process abnormal data, the target abnormal data is determined and marked, and various abnormal situations in the welding process can be accurately identified by the first abnormality detection model and the second abnormality detection model, thereby reducing false alarms and missed alarms. Compared with the monitoring method using fixed thresholds or based on rules in the related art, the monitoring method of the present invention can better adapt to different welding environments and welding conditions, and improve the reliability and stability of the welding process.

[0072] The welding production process monitoring method provided by the present disclosure can be applied to the case where lithium batteries are produced through multiple welding processes, and can also be applied to the case where other products are produced through multiple welding processes.

[0073] Under the above system architecture, a welding production process monitoring method is provided in the embodiment of the present disclosure, and the method can be executed by any electronic device with computing and processing capabilities, such as a control device, which can be a programmable logic controller (PLC) or a single-chip microcomputer, as long as it can realize the corresponding functions.

[0074] Figure 2 A flow chart of a welding production process monitoring method according to an embodiment of the present disclosure is shown as follows: Figure 2 As shown, the welding production process monitoring method provided in the embodiment of the present disclosure includes the following steps:

[0075] S202, acquiring welding data in a welding production process, where the welding production process includes a plurality of welding processes, and the welding data includes data corresponding to a plurality of welding parameters of each welding process.

[0076] During the welding production process, sensors can be used to collect welding data corresponding to various welding parameters in the welding process in real time.

[0077] Welding parameters are parameters that affect the quality of the product. Welding parameters may include but are not limited to welding current, welding voltage, welding temperature, welding pressure, welding speed, gas flow, etc. Corresponding sensors may be used to collect welding data.

[0078] The sensor collects welding data and can transmit it to the control device in real time. The control device can perform the preprocessing mentioned below, and can also perform model training or reasoning of the first anomaly detection model and the second anomaly detection model mentioned below.

[0079] The sensor collects welding data and can also transmit the welding data to a database so that the control device can pull the welding data from the database.

[0080] In one embodiment, after acquiring welding data in the welding production process in S202, the method further includes: preprocessing the welding data, wherein the preprocessing includes at least one of data cleaning, data filling, data standardization, and feature selection.

[0081] Data cleaning is used to process, clean and convert welding data to ensure the accuracy and reliability of the data.

[0082] Data filling is used to fill in missing values ​​in welding data by using mean filling, interpolation and other filling methods to ensure the integrity of the data.

[0083] Data standardization can make welding data with different welding parameters comparable, which is convenient for subsequent data analysis and processing.

[0084] Feature selection is used to select key welding parameters that have a greater impact on welding quality. For example, welding current, welding voltage, welding temperature, etc. can be selected as input data for subsequent analysis and adjustment.

[0085] It should be noted that the above preprocessing methods can be selected according to actual conditions, and the present disclosure does not specifically limit the order of preprocessing.

[0086] S204. Processing welding data corresponding to a plurality of welding parameters of each welding process through a first abnormality detection model to obtain single process abnormality data of each welding process.

[0087] In one embodiment, the first anomaly detection model is constructed based on an unsupervised learning algorithm, and the unsupervised learning algorithm includes at least one of a clustering algorithm and an anomaly detection algorithm.

[0088] Unsupervised learning algorithms are a type of machine learning technique that refers to algorithms that learn on unlabeled training data sets, trying to find the structure or patterns in the welding data in the training data sets.

[0089] The clustering algorithm can classify the historical welding data in the training data set into normal data and abnormal data according to the similarity between them.

[0090] The clustering algorithm may include K-means Clustering Algorithm (K-means Clustering Algorithm, K-means clustering), density-based clustering algorithm (Density-Based Spatial Clustering of Applications with Noise, DBSCAN), hierarchical clustering, etc.

[0091] K-means clustering is an iterative clustering analysis method. The welding data in the training data set can be divided into K groups, K objects are randomly selected as the initial cluster centers, and then the distance between each object and each seed cluster center is calculated, and each object is assigned to the nearest cluster center. The cluster center and the object assigned to the cluster center are a cluster. Each time an object is assigned, the cluster center of the cluster will be recalculated according to the existing objects in the cluster until no object is reassigned to a different cluster, or the cluster center no longer changes, and the error sum of squares is locally minimized.

[0092] DBSCAN can find clusters of arbitrary shapes in a noisy spatial database by finding dense areas of data points and dividing them into clusters. It defines a cluster as the largest set of density-connected points to form clusters. It can be applied to both convex and non-convex sample sets, and can identify clusters of arbitrary shapes and process noisy data.

[0093] Hierarchical clustering can create a hierarchical nested clustering tree by calculating the similarity of data points of different categories. In the clustering tree, the original data points of different categories are the bottom layer of the tree, and the top layer is the root node of a cluster. The clustering tree can be created by merging from the bottom up or splitting from the top down.

[0094] Anomaly detection algorithms can identify outliers or outliers in data. Anomaly detection algorithms include Isolation Forest and Local Outlier Factor (LOF).

[0095] Isolation Forest is an unsupervised learning algorithm based on a tree structure. It isolates anomalies in data by building random decision trees. It is suitable for anomaly recognition and detection of high-dimensional data. Isolation Forest randomly selects features and partition thresholds, so that most normal data points are quickly separated. Abnormal data is isolated only after multiple partitioning steps, which can quickly identify abnormal data.

[0096] LOF identifies outliers by calculating the local density deviation of data points. The LOF algorithm compares the local density of each data point with the local density of its neighbors. If the local density of a data point is larger than that of most data points, it is considered as an outlier. It is suitable for multidimensional data sets and can effectively identify low-density outliers.

[0097] In one embodiment, the unsupervised learning algorithm is trained based on unlabeled historical welding data with abnormal data of a preset ratio threshold as the target. The preset ratio threshold can be determined according to actual training requirements, for example, 2%, that is, 2% of abnormal data is selected from the training data set as the target to train the first abnormality detection model.

[0098] It should be noted that the first anomaly detection model can be constructed by using one unsupervised learning algorithm or by using multiple unsupervised learning algorithms, and the present disclosure does not make any specific limitation on this.

[0099] In one embodiment, the welding production process includes multiple welding steps, and the values ​​of the welding parameters used in each welding step can be the same or different.

[0100] In one embodiment, the first anomaly detection model can be trained separately using the historical welding data corresponding to the welding parameters of each welding process, so as to identify anomalies in the welding data of the welding process collected in real time through the first anomaly detection model to obtain single-process abnormal data.

[0101] The single process abnormal data may include information such as the type of welding parameters corresponding to the abnormal data, abnormal time, abnormal location, and abnormal data value. The single process abnormal data may be represented in a set form. For example, the welding parameters corresponding to the single process abnormal data include welding current I and welding voltage U. The single process abnormal parameters are represented as {Yv: (I1, U1, T1), (Ii, Ui, Ti), ... (Im, Um, Tm)}, where Ii, Ui, Ti are the current value, voltage value, and abnormal time corresponding to the i-th abnormal data, respectively, i = {1, 2, ... m}, m is the number of abnormal data of the welding process Yv, and Yv is the process identifier of the abnormal data, which is determined according to the actual situation.

[0102] S206. Process the welding data corresponding to the multiple welding parameters of the multiple welding processes through the second abnormality detection model to obtain multi-process abnormality data of the multiple welding processes.

[0103] In one embodiment, the second anomaly detection model is constructed based on a single classification algorithm, and the single classification algorithm includes at least one of an isolation forest, a one-class support vector machine (One-Class Support Vector Machine, One-Class SVM), and a deep random neural network.

[0104] One-Class SVM can be trained by applying only normal data, identifying abnormal data that is different from normal data, and can be used for anomaly identification and outlier detection. It can be used to circle positive samples by finding a hyperplane, which is suitable for scenarios with massive data.

[0105] Deep random neural networks include One-Class SVM, Support Vector Data Description (SVDD), etc., which are suitable for complex data sets and anomaly recognition.

[0106] SVDD is a single value classification algorithm that can distinguish target samples from non-target samples.

[0107] In one embodiment, the welding production process includes multiple welding steps, and the values ​​of the welding parameters used in each welding step can be the same or different.

[0108] When the values ​​of welding parameters used in multiple welding processes are different, data standardization and data normalization can be performed on the historical welding data corresponding to the multiple welding parameters of the multiple welding processes, and the processed historical welding data can be used to train the second anomaly detection model, so that the welding data corresponding to the multiple welding parameters of the multiple welding processes can be processed by the trained second anomaly detection model to identify multi-process abnormal data of the multiple welding processes.

[0109] The multi-process abnormal data may include information such as the welding parameter type corresponding to the abnormal data, the abnormal time, the abnormal location, the value of the abnormal data, etc. Multi-process abnormal data can be represented in a set form. For example, the welding parameters corresponding to the multi-process abnormal data include welding current I and welding voltage U. The multi-process abnormal parameters are expressed as {Yv1: (I1, U1, T1), (Ih, Uh, Th), … (Iq, Uq, Tq); Yv2: (I1, U1, T1), (Ij, Uj, Tj), … (Ip, Up, Tp)}}, wherein Ih, Uh, Th are the current value, voltage value and abnormal time corresponding to the h-th abnormal data of the welding process Yv1, h = {1, 2, … q}, q is the number of abnormal data of the welding process Yv1, Ij, Uj, Tj are the current value, voltage value and abnormal time corresponding to the j-th abnormal data of the welding process Yv2, j = {1, 2, … p}, p is the number of abnormal data of the welding process Yv2, which is determined according to actual conditions.

[0110] S208. Determine and mark target abnormal data according to the single-process abnormal data and the multi-process abnormal data.

[0111] In one embodiment, welding data that appears simultaneously in single-process abnormal data and multi-process abnormal data can be determined as target abnormal data, and the target abnormal data is also the final abnormal data determined by the present disclosure.

[0112] In another embodiment, the welding data in the single-process abnormal data and the multi-process abnormal data may also be clustered based on the abnormal time and abnormal position, and the clustered data is determined as the target abnormal data.

[0113] Normal welding data is marked as the first data state; when the welding data is determined to be abnormal data, it can be marked as the second data state. The first data state and the second data state can be represented by binary numbers, for example, 0 represents the first data state and 1 represents the second data state.

[0114] It should be noted that the present disclosure can perform abnormality identification for a situation where the welding production process includes multiple welding processes. For a situation where only one welding process is included, the first abnormality detection model or the second abnormality detection model can be used for abnormality identification.

[0115] In an embodiment of the present disclosure, welding data in a welding production process is obtained, the welding production process includes multiple welding processes, and the welding data includes data corresponding to multiple welding parameters of each welding process; the welding data corresponding to the multiple welding parameters of each welding process is processed by a first abnormality detection model to obtain single-process abnormal data of each welding process; the welding data corresponding to the multiple welding parameters of multiple welding processes is processed by a second abnormality detection model to obtain multi-process abnormal data of multiple welding processes; according to the single-process abnormal data and the multi-process abnormal data, the target abnormal data is determined and marked, and various abnormal situations in the welding process can be accurately identified by the first abnormality detection model and the second abnormality detection model, thereby reducing false alarms and missed alarms. Compared with the monitoring method using fixed thresholds or based on rules in the related art, the monitoring method of the present disclosure can better adapt to different welding environments and welding conditions, and improve the reliability and stability of the welding process.

[0116] Figure 3 Another welding production process monitoring method flow chart of the present disclosure embodiment is shown. Figure 2 Based on the embodiment, S208 is further refined into S2082 to limit the situation of determining abnormal data through single-process abnormal data and multi-process abnormal data. Figure 3 As shown, the above S208 determines and marks the target abnormal data according to the single-process abnormal data and the multi-process abnormal data, including:

[0117] S2082. Determine the intersection of the single-process abnormal data and the multi-process abnormal data as the target abnormal data.

[0118] In one embodiment, the set of single-process abnormal data and the set of multi-process abnormal data of each welding process obtained in the above embodiment are intersected to determine target abnormal data.

[0119] For example, single-process abnormal data include: {Yv1: (I1, U1, T1), (I3, U3, T3), (I7, U7, T7), (I8, U8, T8)} and {Yv2: (I2, U2, T2), (I3, U3, T3), (I5, U5, T5), (I9, U9, T9)}; multi-process abnormal data include: {Yv1: (I1, U1, T1), (I5, U5, T5), (I7, U7, T7), (I8, U8, T8)} 8,U8,T8); Yv2: (I1,U1,T1), (I2,U2,T2), (I3,U3,T3), (I5,U5,T5)}, the intersection of single-process abnormal data and multi-process abnormal data is the target abnormal data: {Yv1: (I1,U1,T1), (I7,U7,T7), (I8,U8,T8); Yv2: (I2,U2,T2), (I3,U3,T3), (I5,U5,T5)}.

[0120] In one embodiment, the obtained sets of single-process abnormal data of each welding process may be combined, and then intersected with the sets of multi-process abnormal data to determine the target abnormal data.

[0121] In the embodiments of the present disclosure, it is possible to solve the problem that in the related art, equipment parameter monitoring uses SPC to control the upper and lower limit specifications of a single parameter without considering the comprehensive impact of multiple welding parameters on the welding quality of the product. By integrating unsupervised learning algorithms and single-classification learning algorithms, the impact of multiple welding parameters on product quality is comprehensively considered, and the quality inspection coordination and cross-influence of each welding parameter are taken into account, thereby effectively improving the monitoring efficiency and effectiveness; by performing intersection processing on single-process abnormal data and multi-process abnormal data, anomalies can be effectively identified.

[0122] Figure 4 FIG. 2 is a flow chart of another method for monitoring a welding production process according to an embodiment of the present disclosure. Figure 4 As shown, in one embodiment, after determining and marking the target abnormal data in S208, the method of the embodiment of the present disclosure further includes:

[0123] S210. If the deviation between the target abnormal data and the preset process control value is greater than the preset deviation threshold, a parameter adjustment instruction is triggered, and the parameter adjustment instruction includes the target abnormal data and the welding parameters corresponding to the target abnormal data, so as to adjust the value of the welding parameter corresponding to the target abnormal data to within the preset parameter range according to the target abnormal data and the preset adjustment value.

[0124] In one embodiment, the preset process control value can be pre-configured in the control device, and the preset process control value may include a first process control value and a second process control value, the first process control value is greater than the second process control value, the first process value is less than the upper limit value of the welding parameter in the SPC, and the second process control value is greater than the lower limit value of the welding parameter in the SPC.

[0125] The preset deviation threshold is smaller than the difference between the above upper limit value and the first process control value, or smaller than the difference between the second process control value and the above lower limit value. The preset deviation threshold can also be determined based on experience to prevent the target abnormal data from exceeding the control range of SPC.

[0126] The deviation between the target abnormal data and the preset process control value may be the difference between the target abnormal data and the first process control value, or the difference between the second process control value and the target abnormal data, and the deviation is a positive value.

[0127] It should be noted that only the first process control value or the second process control value may be set, or both the first process control value and the second process control value may be set at the same time, and the present disclosure does not make any specific limitation on this.

[0128] When the deviation between the abnormal data and the preset process control value is greater than the preset deviation threshold, it indicates that the quality of the product produced using the target abnormal data may not meet the control requirements, triggering the host computer to send a parameter adjustment instruction, and sending the parameter adjustment instruction to the PLC program, and the PLC program executes the parameter adjustment task corresponding to the parameter adjustment instruction.

[0129] The parameter adjustment instruction is used to transmit relevant information of the abnormal data to the PLC program, including but not limited to the above-mentioned target abnormal data and the welding parameters corresponding to the target abnormal data.

[0130] The PLC program automatically adjusts the values ​​of the welding parameters according to the preset adjustment values ​​and abnormal data, so that the values ​​of the welding data are adjusted to within the preset parameter range. The preset adjustment value can be the increment of each adjustment of the abnormal data. For example, when the difference between the first process control value and the welding current is greater than the preset deviation threshold, the welding current can be gradually adjusted by reducing 0.2A each time based on the abnormal data of the welding current until the current value corresponding to the welding current is adjusted to within the preset current range.

[0131] The preset parameter range may be a numerical range corresponding to the first process control value and the second process control value.

[0132] In one embodiment, after adjusting the welding parameters, the welding data corresponding to the welding parameters are continuously monitored to verify the adjustment effect and ensure that the values ​​of the welding parameters are stable within the preset parameter range.

[0133] In the embodiments of the present disclosure, the values ​​of welding parameters can be automatically adjusted according to real-time abnormal data, thereby reducing manual intervention and improving production efficiency and product quality.

[0134] Figure 5 A flow chart of another welding production process monitoring method in an embodiment of the present disclosure is shown. In one embodiment, the welding production process monitoring method provided by the present disclosure further includes:

[0135] S502, performing trend analysis on a plurality of welding data corresponding to a target welding parameter of each welding process to obtain a trend analysis result, wherein the target welding parameter is one of the plurality of welding parameters;

[0136] S504. If the trend analysis result shows that the plurality of welding data corresponding to the target welding parameters present a continuous upward trend or a continuous downward trend, a warning message is generated, and the warning message includes the target welding parameters and the trend analysis result.

[0137] The target welding parameter is any one of the welding parameters such as welding current, welding voltage, welding temperature, welding pressure, welding speed, gas flow rate, etc. in the above embodiments.

[0138] The multiple welding data corresponding to the target welding parameters of a welding process refer to the welding data of each product collected when a batch of products is welded using the same welding process.

[0139] The trend analysis result includes that the multiple welding data corresponding to the target welding parameters present a continuous upward trend, a continuous downward trend, or a random fluctuation. Among them, the continuous upward trend means that the next welding data is greater than the previous welding data, or the difference between the next welding data and the previous welding data is greater than the first preset parameter threshold; the continuous downward trend means that the next welding data is less than the previous welding data, or the difference between the previous welding data and the next welding data is greater than the second preset parameter threshold.

[0140] The first preset parameter threshold and the second preset parameter threshold can be determined according to actual needs, and the present disclosure does not make any specific limitations on this.

[0141] The number of welding data used for trend analysis can be determined according to actual needs, for example, the number of welding data can be 5, 10, etc.

[0142] The warning information is used to remind the user to pay attention to the target welding parameters, so that the user can focus on the changes in the target welding parameters.

[0143] In the disclosed embodiments, based on the trend analysis results of multiple welding data corresponding to the target welding parameters, the user can be promptly reminded to pay attention to the changes in the parameters, thereby avoiding the occurrence of a large number of unqualified products and affecting product quality.

[0144] Figure 6 The flow chart of the model optimization method in the embodiment of the present disclosure is shown. Figure 6 As shown, in one embodiment, the welding generation process monitoring method provided by the present disclosure further includes:

[0145] S602, determining a welding result corresponding to welding data corresponding to a plurality of welding parameters;

[0146] S604. Update the training data sets of the first anomaly detection model and the second anomaly detection model with the welding data corresponding to the welding results, triggering an incremental learning task to optimize the first anomaly detection model and the second anomaly detection model.

[0147] During the welding production process, the production products can be quality inspected to obtain welding results, which include qualified product data (spot inspection OK), critical product data (OK / NG critical) and unqualified product data (full inspection NG). Among them, qualified product data is determined by spot inspection, and the spot inspection ratio can be customized, for example 1%; critical product data is the unqualified / qualified product critical.

[0148] Updating refers to adding welding data corresponding to the welding results to the training data set. The training data set can be updated regularly with welding data corresponding to qualified product data, or the training data set can be updated when the welding data corresponding to qualified product data reaches a preset number.

[0149] After the training data set is updated, the incremental learning task is triggered to optimize the first anomaly detection model and the second anomaly detection model.

[0150] In the disclosed embodiment, the welding data corresponding to the qualified product data is added to the training data set for model training, so that the model is optimized regularly, so that the model can recognize the data fluctuations and process fluctuations in the real-time production process, improve the model prediction accuracy, and make it more applicable.

[0151] In one embodiment, the method further includes storing the welding data, the target abnormality data, and the adjusted welding data in a database.

[0152] During the entire welding production process, all monitoring data such as welding parameters, welding data, abnormal detection results and adjustment records can be stored in the database to provide data support for subsequent quality analysis and optimization.

[0153] The system can feed back real-time monitoring data and adjustment results to operators so that they can understand the operating status of the production line.

[0154] In addition, statistical analysis can be performed regularly on the stored monitoring data to identify potential problems and improvement points, generate detailed monitoring reports, record abnormal situations and adjustment measures, and provide decision support for users.

[0155] In order to deepen the understanding of the welding production process monitoring method provided by the present disclosure, the following Figure 7-10 Provide detailed explanation.

[0156] Figure 7 A flow chart showing an example of a data preprocessing method in an embodiment of the present disclosure is shown. Figure 7 As shown, the data preprocessing method disclosed in the present invention comprises the following steps:

[0157] S701, data acquisition; for example, a high-precision sensor can be used to collect the values ​​of welding parameters in the welding process, such as welding current, welding voltage, welding temperature, welding speed, gas flow, etc., in real time, and transmit them to a control device or a database in real time through the sensor.

[0158] S702, data cleaning: Clean the collected data to remove noise or outliers in the data to ensure the accuracy and reliability of the data. The outliers are mutation points in the data, etc.

[0159] S703, data filling: Missing values ​​in the data are filled by filling methods such as mean filling and interpolation to ensure data integrity.

[0160] S704, data standardization: standardize the data to make data of different parameters comparable, so as to facilitate subsequent analysis and processing.

[0161] S705, feature selection: select key parameters that have a greater impact on welding quality, such as welding current, welding voltage, welding temperature, etc., as input data for subsequent analysis and adjustment.

[0162] It should be noted that at least one of the above steps can be executed according to the actual situation of the welding data. When multiple steps are executed, data preprocessing can be performed in the above order.

[0163] Figure 8 FIG. 2 is a flow chart showing an example of a method for monitoring a welding production process according to an embodiment of the present disclosure. Figure 8 As shown, the welding production process monitoring method disclosed in the present invention comprises the following steps:

[0164] S801, data input. The welding data input into the model is data corresponding to multiple welding parameters of multiple welding processes. The pre-processed welding data is input into Figure 8 In the set model, training data sets and cognitive data are provided for model training.

[0165] S802. For single-process multiple parameters, DBSCAN is used to process the welding data corresponding to multiple welding parameters (single-process multiple parameters) of each welding process to obtain single-process abnormal data, also known as out-of-cluster data.

[0166] S803. For multiple processes and multiple parameters, Isolation Forest is used to process welding data corresponding to multiple welding parameters of multiple welding processes (multiple processes and multiple parameters) to obtain multi-process abnormal data, also known as outliers.

[0167] S804. Take the intersection of the single-process abnormal data and the multi-process abnormal data.

[0168] S805: Determine the data after the intersection as target abnormal data.

[0169] S806: Mark target abnormal data.

[0170] It should be noted that DBSCAN and Isolation Forest use historical data to train unsupervised learning algorithms to obtain the corresponding first anomaly detection model and second anomaly detection model.

[0171] The real-time collected welding data is input into the above-mentioned DBSCAN and Isolation Forest to identify abnormal data.

[0172] Fig. 9 FIG. 2 shows an example flow chart of the automatic welding parameter adjustment method in the embodiment of the present disclosure. Fig. 9 As shown, the automatic welding parameter adjustment method disclosed in the present invention comprises the following steps:

[0173] S910, data monitoring: real-time monitoring of various welding parameters during the welding process and obtaining current welding data.

[0174] S920, trend analysis: trend analysis is performed on real-time welding data, and abnormal data showing an upward (or downward) trend at multiple consecutive points or a sudden change point is detected to identify welding parameters that deviate from the preset process control values.

[0175] S930, adjustment trigger; when the deviation between the current welding data and the preset process control value exceeds the preset deviation threshold, the host computer is triggered to send instructions to the PLC program, automatically triggering the parameter adjustment mechanism.

[0176] S940, issuing a PLC parameter adjustment instruction; the parameter adjustment instruction includes the abnormal data and the welding parameters corresponding to the abnormal data.

[0177] S950, automatic adjustment: according to the preset adjustment value, the welding parameter value is automatically adjusted to restore it to the preset parameter range.

[0178] S960, adjustment verification: continue to monitor parameter changes after adjustment, verify the adjustment effect, and ensure that the welding parameters are stable within the preset parameter range.

[0179] Fig.10 A schematic diagram showing an example of a model optimization method in an embodiment of the present disclosure is shown. Fig.10 As shown, after the first anomaly detection model and the second anomaly detection model identify the anomaly of the welding data, the product is quality tested to obtain three types of welding results, including qualified product data, critical product data, and unqualified product data. The welding data corresponding to the above welding results are used to regularly update the training data set to trigger incremental learning tasks, keep the model advanced, and make the model aware of data fluctuations and process fluctuations in the real-time production process, ensuring real-time updates of the model and making it more practical.

[0180] Based on the same inventive concept, the present disclosure also provides a welding production process monitoring device in the following embodiments. Since the principle of solving the problem in the device embodiment is similar to that in the above method embodiment, the implementation of the device embodiment can refer to the implementation of the above method embodiment, and the repeated parts will not be repeated.

[0181] Fig.11 A schematic diagram of a welding production process monitoring device in an embodiment of the present disclosure is shown. Fig.11 As shown, the device includes a data acquisition module 1110 , a first detection module 1120 , a second detection module 1130 and an abnormality determination module 1140 .

[0182] The data acquisition module 1110 is used to acquire welding data in a welding production process, where the welding production process includes multiple welding processes, and the welding data includes data corresponding to multiple welding parameters of each welding process;

[0183] A first detection module 1120 is used to process welding data corresponding to multiple welding parameters of each welding process through a first abnormality detection model to obtain single process abnormality data of each welding process;

[0184] The second detection module 1130 is used to process welding data corresponding to multiple welding parameters of multiple welding processes through a second abnormality detection model to obtain multi-process abnormality data of the multiple welding processes;

[0185] The abnormality determination module 1140 is used to determine and mark target abnormal data according to the single-process abnormal data and the multi-process abnormal data.

[0186] It should be noted that the data acquisition module 1110, the first detection module 1120, the second detection module 1130 and the abnormality determination module 1140 correspond to S202 to S208 in the method embodiment, and the examples and application scenarios implemented by the above modules and the corresponding steps are the same, but are not limited to the contents disclosed in the above method embodiment. It should be noted that the above modules as part of the device can be executed in an electronic device such as a set of computer executable instructions.

[0187] In one embodiment, the first anomaly detection model is constructed based on an unsupervised learning algorithm, and the unsupervised learning algorithm includes at least one of a clustering algorithm and an anomaly detection algorithm.

[0188] It should be noted that the unsupervised learning algorithm is trained based on unlabeled historical welding data and abnormal data with a preset ratio threshold.

[0189] In one embodiment, the second anomaly detection model is constructed based on a single classification algorithm, and the single classification algorithm includes at least one of an isolation forest, a single-class support vector machine, and a deep random neural network.

[0190] In one embodiment, the abnormality determination module 1140 is used to determine the intersection of the single-process abnormality data and the multi-process abnormality data as the target abnormality data.

[0191] In one embodiment, the device also includes a parameter adjustment module not shown in the accompanying drawings. The parameter adjustment module is used to trigger a parameter adjustment instruction after determining and marking the target abnormal data. If the deviation between the target abnormal data and the preset process control value is greater than a preset deviation threshold, the parameter adjustment instruction includes the target abnormal data and the welding parameters corresponding to the target abnormal data, so as to adjust the value of the welding parameter corresponding to the target abnormal data to within the preset parameter range according to the target abnormal data and the preset adjustment value.

[0192] In one embodiment, the device also includes a trend analysis module and an early warning module not shown in the accompanying drawings, wherein the trend analysis module is used to perform trend analysis on multiple welding data corresponding to the target welding parameters of each welding process to obtain a trend analysis result, and the target welding parameter is one of the multiple welding parameters; the early warning module is used to generate early warning information if the trend analysis result is that the multiple welding data corresponding to the target welding parameter show a continuous upward trend or a continuous downward trend, and the early warning information includes the target welding parameter and the trend analysis result.

[0193] In one embodiment, the device also includes a model optimization module not shown in the accompanying drawings, and the model optimization module is used to determine the welding results corresponding to the welding data corresponding to multiple welding parameters; update the training data sets of the first anomaly detection model and the second anomaly detection model with the welding data corresponding to the welding results, triggering an incremental learning task to perform model optimization on the first anomaly detection model and the second anomaly detection model.

[0194] In one embodiment, the device also includes a preprocessing module not shown in the accompanying drawings, which is used together with the processing module to preprocess the welding data after acquiring the welding data in the welding production process. The preprocessing includes at least one of data cleaning, data filling, data standardization, and feature selection.

[0195] In one embodiment, the device further includes a storage module not shown in the accompanying drawings, which is used to store welding data, target abnormality data and adjusted welding data into a database.

[0196] In an embodiment of the present disclosure, welding data in a welding production process is obtained, the welding production process includes multiple welding processes, and the welding data includes data corresponding to multiple welding parameters of each welding process; the welding data corresponding to the multiple welding parameters of each welding process is processed by a first abnormality detection model to obtain single-process abnormal data of each welding process; the welding data corresponding to the multiple welding parameters of multiple welding processes is processed by a second abnormality detection model to obtain multi-process abnormal data of multiple welding processes; according to the single-process abnormal data and the multi-process abnormal data, the target abnormal data is determined and marked, and various abnormal situations in the welding process can be accurately identified by the first abnormality detection model and the second abnormality detection model, thereby reducing false alarms and missed alarms. Compared with the monitoring method using fixed thresholds or based on rules in the related art, the monitoring method of the present disclosure can better adapt to different welding environments and welding conditions, and improve the reliability and stability of the welding process.

[0197] Those skilled in the art will appreciate that various aspects of the present disclosure may be implemented as systems, methods or program products. Therefore, various aspects of the present disclosure may be specifically implemented in the following forms, namely: complete hardware implementation, complete software implementation (including firmware, microcode, etc.), or a combination of hardware and software, which may be collectively referred to herein as "circuits", "modules" or "systems".

[0198] Refer to the following Fig.12 1200 according to this embodiment of the present disclosure is described. Fig.12 The electronic device 1200 shown is merely an example and should not bring any limitation to the functions and scope of use of the embodiments of the present disclosure.

[0199] like Fig.12 As shown, the electronic device 1200 is in the form of a general computing device. The components of the electronic device 1200 may include but are not limited to: at least one processing unit 1210, at least one storage unit 1220, and a bus 1230 connecting different system components (including the storage unit 1220 and the processing unit 1210).

[0200] The storage unit stores a program code, and the program code can be executed by the processing unit 1210, so that the processing unit 1210 executes the steps described in the "Exemplary Method" section of the specification according to various exemplary embodiments of the present disclosure. For example, the processing unit 1210 can execute the following steps of the above method embodiment: obtaining welding data in a welding production process, the welding production process includes multiple welding processes, and the welding data includes data corresponding to multiple welding parameters of each welding process; processing the welding data corresponding to multiple welding parameters of each welding process through a first abnormality detection model to obtain single-process abnormal data of each welding process; processing the welding data corresponding to multiple welding parameters of multiple welding processes through a second abnormality detection model to obtain multi-process abnormal data of multiple welding processes; determining and marking target abnormal data based on the single-process abnormal data and the multi-process abnormal data.

[0201] The storage unit 1220 may include a readable medium in the form of a volatile storage unit, such as a random access storage unit (RAM) 12201 and / or a cache storage unit 12202 , and may further include a read-only storage unit (ROM) 12203 .

[0202] The storage unit 1220 may also include a program / utility 12204 having a set (at least one) of program modules 12205, such program modules 12205 including but not limited to: an operating system, one or more application programs, other program modules, and program data, each of which or some combination may include an implementation of a network environment.

[0203] Bus 1230 may represent one or more of several types of bus structures, including a memory unit bus or memory unit controller, a peripheral bus, an accelerated graphics port, a processing unit, or a local bus using any of a variety of bus architectures.

[0204] The electronic device 1200 may also communicate with one or more external devices 1240 (e.g., keyboards, pointing devices, Bluetooth devices, etc.), may also communicate with one or more devices that enable a user to interact with the electronic device 1200, and / or communicate with any device that enables the electronic device 1200 to communicate with one or more other computing devices (e.g., routers, modems, etc.). Such communication may be performed via an input / output (I / O) interface 1250. Furthermore, the electronic device 1200 may also communicate with one or more networks (e.g., local area networks (LANs), wide area networks (WANs), and / or public networks, such as the Internet) via a network adapter 1260. As shown, the network adapter 1260 communicates with other modules of the electronic device 1200 via a bus 1230. It should be understood that, although not shown in the figure, other hardware and / or software modules may be used in conjunction with the electronic device 1200, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage systems.

[0205] Through the description of the above implementation, it is easy for those skilled in the art to understand that the example implementation described here can be implemented by software, or by software combined with necessary hardware. Therefore, the technical solution according to the implementation of the present disclosure can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (which can be a CD-ROM, a USB flash drive, a mobile hard disk, etc.) or on a network, including several instructions to enable a computing device (which can be a personal computer, a server, a terminal device, or a network device, etc.) to execute the method according to the implementation of the present disclosure.

[0206] In particular, according to an embodiment of the present disclosure, the process described in the above reference flow chart may be implemented as a computer program product, which includes: a computer program, which implements the above-mentioned battery cell capacity calibration method when executed by a processor.

[0207] In an exemplary embodiment of the present disclosure, a computer-readable storage medium is also provided, which may be a readable signal medium or a readable storage medium. In some possible implementations, various aspects of the present disclosure may also be implemented in the form of a program product, which includes a program code, and when the program product is run on a terminal device, the program code is used to enable the terminal device to execute the steps according to various exemplary implementations of the present disclosure described in the above “Exemplary Method” section of this specification.

[0208] More specific examples of computer-readable storage media in the present disclosure may include, but are not limited to, an electrical connection having one or more conductors, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0209] In the present disclosure, a computer readable storage medium may include a data signal propagated in baseband or as part of a carrier wave, wherein a readable program code is carried. Such propagated data signals may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. A readable signal medium may also be any readable medium other than a readable storage medium, which may send, propagate, or transmit a program for use by or in conjunction with an instruction execution system, apparatus, or device.

[0210] Alternatively, the program code contained on the computer-readable storage medium may be transmitted using any appropriate medium, including but not limited to wireless, wired, optical cable, RF, etc., or any suitable combination of the foregoing.

[0211] In a specific implementation, the program code for performing the operations of the present disclosure may be written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Java, C++, etc., and conventional procedural programming languages ​​such as "C" or similar programming languages. The program code may be executed entirely on the user computing device, partially on the user device, as a separate software package, partially on the user computing device and partially on a remote computing device, or entirely on a remote computing device or server. In the case of a remote computing device, the remote computing device may be connected to the user computing device through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computing device (e.g., using an Internet service provider to connect through the Internet).

[0212] It should be noted that, although several modules or units of the device for action execution are mentioned in the above detailed description, this division is not mandatory. In fact, according to the embodiments of the present disclosure, the features and functions of two or more modules or units described above can be embodied in one module or unit. On the contrary, the features and functions of one module or unit described above can be further divided into multiple modules or units to be embodied.

[0213] In addition, although the steps of the method in the present disclosure are described in a specific order in the drawings, this does not require or imply that the steps must be performed in this specific order, or that all the steps shown must be performed to achieve the desired results. Additionally or alternatively, some steps may be omitted, multiple steps may be combined into one step, and / or one step may be decomposed into multiple steps, etc.

[0214] Through the description of the above implementation, it is easy for those skilled in the art to understand that the example implementation described here can be implemented by software, or by software combined with necessary hardware. Therefore, the technical solution according to the implementation of the present disclosure can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (which can be a CD-ROM, a USB flash drive, a mobile hard disk, etc.) or on a network, including several instructions to enable a computing device (which can be a personal computer, a server, a mobile terminal, or a network device, etc.) to execute the method according to the implementation of the present disclosure.

[0215] Those skilled in the art will readily appreciate other embodiments of the present disclosure after considering the specification and practicing the invention disclosed herein. The present disclosure is intended to cover any variations, uses or adaptations of the present disclosure, which follow the general principles of the present disclosure and include common knowledge or customary techniques in the art that are not disclosed in the present disclosure. The description and examples are intended to be exemplary only, and the true scope and spirit of the present disclosure are indicated by the appended claims.

Claims

1. A welding production process monitoring method, characterized in that: include: Acquire welding data in a welding production process, wherein the welding production process includes a plurality of welding processes, and the welding data includes data corresponding to a plurality of welding parameters of each welding process; Processing welding data corresponding to a plurality of welding parameters of each welding process by a first anomaly detection model to obtain single process anomaly data of each welding process; Processing the welding data corresponding to the multiple welding parameters of the multiple welding processes by a second abnormality detection model to obtain multi-process abnormality data of the multiple welding processes; According to the single-process abnormal data and the multi-process abnormal data, target abnormal data are determined and marked.

2. The welding production process monitoring method according to claim 1, characterized in that: The first anomaly detection model is constructed based on an unsupervised learning algorithm, and the unsupervised learning algorithm includes at least one of a clustering algorithm and an anomaly detection algorithm.

3. The welding production process monitoring method according to claim 2, characterized in that: The unsupervised learning algorithm is trained based on unlabeled historical welding data and taking abnormal data with a preset ratio threshold as the target.

4. The welding production process monitoring method according to claim 1, characterized in that: The second anomaly detection model is constructed based on a single classification algorithm, and the single classification algorithm includes at least one of an isolation forest, a single-class support vector machine, and a deep random neural network.

5. The welding production process monitoring method according to claim 1, characterized in that: The determining and marking target abnormal data according to the single-process abnormal data and the multi-process abnormal data includes: The intersection of the single-process abnormal data and the multiple-process abnormal data is determined as the target abnormal data.

6. The welding production process monitoring method according to claim 1, characterized in that: After determining and marking the target abnormal data, the method further includes: If the deviation between the target abnormal data and the preset process control value is greater than the preset deviation threshold, a parameter adjustment instruction is triggered, and the parameter adjustment instruction includes the target abnormal data and the welding parameters corresponding to the target abnormal data, so as to adjust the value of the welding parameter corresponding to the target abnormal data to within the preset parameter range according to the target abnormal data and the preset adjustment value.

7. The welding production process monitoring method according to claim 1, characterized in that: The method further comprises: Performing trend analysis on a plurality of welding data corresponding to a target welding parameter of each welding process to obtain a trend analysis result, wherein the target welding parameter is one of the plurality of welding parameters; If the trend analysis result shows that the plurality of welding data corresponding to the target welding parameter present a continuous upward trend or a continuous downward trend, an early warning message is generated, and the early warning message includes the target welding parameter and the trend analysis result.

8. The welding production process monitoring method according to claim 1, characterized in that: The method further comprises: determining a welding result corresponding to welding data corresponding to a plurality of welding parameters; The welding data corresponding to the welding result is used to update the training data sets of the first anomaly detection model and the second anomaly detection model, triggering an incremental learning task to optimize the first anomaly detection model and the second anomaly detection model.

9. The welding production process monitoring method according to claim 1, characterized in that: After obtaining the welding data in the welding production process, the method further includes: The welding data is preprocessed, and the preprocessing includes at least one of data cleaning, data filling, data standardization, and feature selection.

10. The welding production process monitoring method according to claim 6, characterized in that: The method further comprises: The welding data, the target abnormality data and the adjusted welding data are stored in a database.

11. A welding production process monitoring device, characterized in that: include: A data acquisition module, used to acquire welding data in a welding production process, wherein the welding production process includes a plurality of welding processes, and the welding data includes data corresponding to a plurality of welding parameters of each welding process; A first detection module, configured to process welding data corresponding to a plurality of welding parameters of each welding process through a first abnormality detection model to obtain single process abnormality data of each welding process; A second detection module, configured to process welding data corresponding to a plurality of welding parameters of the plurality of welding processes by using a second abnormality detection model to obtain multi-process abnormality data of the plurality of welding processes; The abnormality determination module is used to determine and mark target abnormal data according to the single-process abnormal data and the multi-process abnormal data.

12. An electronic device, characterized in that: include: processor; as well as A memory, configured to store executable instructions of the processor; Wherein, the processor is configured to execute the welding production process monitoring method according to any one of claims 1 to 10 by executing the executable instructions.

13. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the welding production process monitoring method according to any one of claims 1 to 10 is implemented.

14. A computer program product, characterized in that The computer program product includes a computer program or computer instructions, and the computer program or the computer instructions are loaded and executed by a processor so that a computer implements the welding production process monitoring method as described in any one of claims 1-10.

Images