Laser welding quality detection method and system

Abstract

The invention provides a laser welding quality detection method and system, and relates to the technical field of quality detection.The method comprises the steps that equipment operation parameters are obtained, a welding quality evaluation model is established on the basis of a historical operation parameter set in combination with self-adaptive topology learning, real-time image data are obtained through image collection equipment and uploaded to a welding defect recognition layer, and the welding defect recognition layer recognizes welding defects. And embedding a welding defect identification layer into the welding quality evaluation model, and determining a welding quality detection result in combination with the equipment operation parameters, so that the technical problem that hidden defects such as cooling shrinkage cracks and residual stress accumulation are difficult to effectively detect due to off-line analysis in laser welding quality detection is solved, and the detection efficiency is improved. The technical effects of dynamically monitoring the whole welding process by collecting the equipment operation parameters of preheating, welding and cooling stages, improving the detection accuracy of the hidden defects and providing powerful support for optimization of the laser welding process are achieved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of quality detection, and particularly relates to a laser welding quality detection method and system. BACKGROUND

[0002] Laser welding has the advantages of small weld, small deformation, high quality, fast welding speed, etc., and is applied in the fields of aviation, aerospace, electronics, automobiles, etc. The welding quality is a key factor for evaluating the performance and reliability of a welded joint, and the quality of the welded joint is directly related to the safety and service life of a product.

[0003] At present, a conventional laser welding monitoring system is usually offline analysis, cannot access dynamic image data of a molten pool state, a heat affected zone, etc. in real time, has insufficient recognition accuracy for typical defects such as incomplete fusion and pores, and lags behind the welding process. In addition, since welding parameters are set according to experience or fixed process specifications, the adaptability to structural differences of a combined welding product and morphological changes of a small welding point is poor, and misjudgment and missed judgment often occur in complex component welding, which seriously affects the production yield and efficiency.

[0004] In summary, the laser welding quality detection in the prior art is offline analysis, and it is difficult to effectively detect hidden defects such as cooling shrinkage cracks and residual stress accumulation. SUMMARY

[0005] The purpose of the present application is to provide a laser welding quality detection method and system, so as to solve the technical problem that the laser welding quality detection in the prior art is offline analysis, and it is difficult to effectively detect hidden defects such as cooling shrinkage cracks and residual stress accumulation.

[0006] In view of the above problems, the present application provides a laser welding quality detection method and system.

[0007] In the first aspect, the present application provides a laser welding quality detection method, which is implemented by a laser welding quality detection system, wherein the method includes: connecting a laser welding machine to obtain equipment operating parameters, wherein the equipment operating parameters include laser power, welding time node, and shielding gas flow rate; based on the historical operating parameter set corresponding to the preheating stage, the historical operating parameter set corresponding to the welding stage, and the historical operating parameter set corresponding to the cooling stage, combined with adaptive topology learning, a welding quality evaluation model is established, and the welding quality evaluation model is used to dynamically evaluate the welding quality of each welding point; based on the welding defect records in the laser welding database, a welding defect identification layer under the welding defect type limitation is determined, wherein the welding defect types include unfusion, pores, slag inclusions, and cracks, and an image acquisition device is used to obtain real-time image data during the welding process, and the real-time image data is uploaded to the welding defect identification layer, wherein the real-time image data includes the molten pool state and the heat-affected zone; the welding defect identification layer is embedded in the welding quality evaluation model, and the welding quality detection results are determined in combination with the equipment operating parameters.

[0008] In the second aspect, the present application also provides a laser welding quality detection system for executing the laser welding quality detection method as described in the first aspect, wherein the system includes: an operating parameter acquisition module, the operating parameter acquisition module is used to connect to the laser welding machine and obtain the equipment operating parameters, the equipment operating parameters include laser power, welding time node, and shielding gas flow rate; an evaluation model establishment module, the evaluation model establishment module is used to establish a welding quality evaluation model based on the historical operating parameter set corresponding to the preheating stage, the historical operating parameter set corresponding to the welding stage, and the historical operating parameter set corresponding to the cooling stage, combined with adaptive topology learning, the welding quality evaluation model is used for dynamic Evaluate the welding quality of each welding point; an image acquisition module, which is used to determine the welding defect identification layer under the welding defect type limitation based on the welding defect records in the laser welding database, wherein the welding defect types include unfusion, pores, slag inclusions, and cracks. Use image acquisition equipment to obtain real-time image data during the welding process, and upload the real-time image data to the welding defect identification layer, wherein the real-time image data includes the molten pool state and the heat-affected zone; a quality inspection result acquisition module, which is used to embed the welding defect identification layer into the welding quality evaluation model and determine the welding quality inspection results in combination with the equipment operating parameters.

[0009] One or more technical solutions provided in this application have at least the following technical effects or advantages: By setting the welding operation nodes corresponding to each welding point based on the welding seam based on the laser welding process, connecting the laser welding machine to obtain the equipment operation parameters, the equipment operation parameters include laser power, welding time node, and protective gas flow; based on the historical operation parameter set corresponding to the preheating stage, the historical operation parameter set corresponding to the welding stage, and the historical operation parameter set corresponding to the cooling stage, combined with adaptive topology learning, a welding quality evaluation model is established, which is used for dynamically evaluating the welding quality of each welding point; based on the welding defect records in the laser welding database, a welding defect identification layer under the limitation of the welding defect type is determined, the welding defect type includes incomplete fusion, porosity, slag inclusion, and crack, and an image acquisition device is used to obtain real-time image data in the welding process, and the real-time image data is uploaded to the welding defect identification layer, wherein the real-time image data includes the molten pool state and the heat affected zone; the welding defect identification layer is embedded into the welding quality evaluation model, and the equipment operation parameters are combined to determine the welding quality detection result, which achieves the technical effect of dynamically monitoring the whole welding process by collecting the equipment operation parameters in the preheating, welding, and cooling stages, improving the detection accuracy of hidden defects, and providing strong support for laser welding process optimization.

[0010] The above description is only a summary of the technical solutions of the present application. In order to more clearly understand the technical means of the present application, the specific embodiments of the present application can be implemented according to the content of the specification, and in order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the following specific embodiments of the present application are described. It should be understood that the content described in this part is not intended to identify the key or important features of the embodiments of the present application, nor is it intended to limit the scope of the present application. Other features of the present application will become apparent from the following description. BRIEF DESCRIPTION OF DRAWINGS

[0011] In order to more clearly illustrate the technical solutions in the present application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiments or the prior art description. Obviously, the drawings in the following description are only exemplary, and other drawings can be obtained by the provided drawings without creative labor for those skilled in the art.

[0012] Figure 1 The flowchart of the laser welding quality detection method of the present application.

[0013] Figure 2 The structure diagram of the laser welding quality detection system of the present application.

[0014] Explanation of reference signs: operation parameter acquisition module 11, evaluation model establishment module 12, image acquisition module 13, quality detection result acquisition module 14. DETAILED DESCRIPTION

[0015] The application solves the technical problem that laser welding quality detection is mostly offline analysis, and that hidden defects such as cooling shrinkage cracks and residual stress accumulation are difficult to effectively detect, and achieves the technical effect of improving the detection accuracy of hidden defects by collecting equipment operating parameters in the preheating, welding and cooling stages, dynamically monitoring the whole welding process, and providing strong support for laser welding process optimization.

[0016] The technical solutions in the application will be described clearly and completely below with reference to the drawings. Obviously, the described embodiments are only some of the embodiments of the application, rather than all the embodiments of the application. It should be understood that the application is not limited by the example embodiments described herein. Based on the embodiments of the application, all other embodiments obtained by a person of ordinary skill in the art without creative work fall within the scope of protection of the application. In addition, it should be noted that, for convenience of description, only parts related to the application are shown in the drawings, rather than all parts.

[0017] Embodiment one Please refer to the drawings Figure 1 The application provides a laser welding quality detection method, which is applied to a laser welding quality detection system and specifically includes the following steps. S1: Connect a laser welding machine, and acquire equipment operating parameters, wherein the equipment operating parameters include laser power, welding time nodes and protective gas flow.

[0018] Specifically, a laser welding machine is connected to an upper computer, such as a computer, by using a suitable data line or communication interface. Data transmission is performed by using a USB, RS232 or Ethernet interface. The control software is opened, and a connection with the laser welding machine is established. The device parameter option is searched to enter the device parameter setting interface. The equipment operating parameters include laser power, welding time nodes and protective gas flow. The laser power displays the output power of the current laser. This parameter determines the energy density and welding speed during welding. The welding time nodes display the data of each key time point in the welding process, such as the welding start time and end time.

[0019] The protective gas flow displays the flow value of the protective gas, which plays a role in preventing oxidation and protecting the quality of the weld in the welding process.

[0020] S2: Based on a historical operating parameter set corresponding to a preheating stage, a historical operating parameter set corresponding to a welding stage and a historical operating parameter set corresponding to a cooling stage, a welding quality evaluation model is established by combining adaptive topology learning, and the welding quality evaluation model is used to dynamically evaluate the welding quality of each welding point.

[0021] Specifically, outliers, duplicates, and missing values ​​are removed to ensure data accuracy and completeness. Features highly correlated with weld quality, such as laser power, welding speed, and shielding gas flow rate, are selected. By analyzing the relationship between parameter changes and weld quality in historical data, the topological structure of the data is learned. As new data is added, the model automatically adjusts its topology to adapt to the new data distribution and feature relationships. Features that represent weld quality are extracted using statistical methods, signal processing techniques, or machine learning algorithms. For example, features can be extracted by calculating the fluctuation range of laser power or the distribution of welding time. Machine learning algorithms, such as support vector machines (SVMs), random forests, or neural networks, are selected to train the model. Preprocessed historical data is used as the training set to develop a model capable of predicting weld quality. The preprocessed dataset is divided into a training set and a validation set. The model is trained using the training set, and its parameters and structure are adjusted to optimize its performance. The model is evaluated using an independent test set to verify its performance in real-world applications. Metrics such as accuracy and recall are calculated to assess the model's predictive capabilities. During the actual welding process, operating parameters at each welding point are collected in real time. By inputting real-time data into the trained model, the weld quality prediction value of each weld point is obtained. Based on the prediction results, welding parameters can be adjusted in a timely manner or other measures can be taken to improve the weld quality.

[0022] S3: Based on the welding defect records in the laser welding database, determine the welding defect identification layer under the welding defect type limitation, the welding defect type includes lack of fusion, porosity, slag inclusion, and crack, use image acquisition equipment to obtain real-time image data during the welding process, and upload the real-time image data to the welding defect identification layer, wherein the real-time image data includes the molten pool state and heat-affected zone.

[0023] Specifically, the types of welding defects to be detected are identified, which typically include lack of fusion, porosity, slag inclusion, and cracks. These defects have a serious impact on the quality and performance of the welded joint, so they need to be identified and addressed in a timely and accurate manner. Historical image data containing various welding defects is extracted from a laser welding database. These data will be used to train and verify the welding defect recognition model. According to the characteristics of the welding defects, appropriate image recognition and machine learning algorithms are selected to build the recognition model. For example, a convolutional neural network in deep learning can be used to train the model, enabling it to automatically identify welding defects from image data. The trained model is verified using an independent validation set to evaluate its accuracy and reliability in identifying welding defects. The model is adjusted and optimized as necessary based on the verification results. High-resolution, high-speed image acquisition devices are used to capture real-time image data of key areas such as the molten pool and heat-affected zone during the welding process. Real-time image data is uploaded to the welding defect recognition layer for processing. This is achieved through wired or wireless network connections, ensuring real-time and accurate data.

[0024] S4: Embed the welding defect recognition layer into the welding quality evaluation model, combine the device operating parameters, and determine the welding quality detection results.

[0025] Specifically, the welding defect recognition layer is embedded into the welding quality evaluation model. When evaluating the welding quality, the recognition results of the welding defects will also be considered. Ensure that the welding defect recognition layer can receive real-time image data from the image acquisition device and output the recognition results to other parts of the welding quality evaluation model. Ensure that the data from the welding operation node is time-synchronized with the collection of real-time image data, so that each welding stage operation can be accurately matched with possible defects. Combine the welding operation node data, such as welding speed, welding current, voltage, etc., with the welding defect recognition results to analyze the potential relationship between operation and defects. Collect the operating parameters of the welding equipment during the welding process, such as laser power, protective gas flow, welding head movement trajectory, etc. Compare and analyze these device operating parameters with the welding defect recognition results to find out the device operation mode or parameter settings that may cause defects. Consider the welding defect recognition results, welding operation node data, and device operating parameters to comprehensively evaluate the welding quality. For example, if a serious welding defect is identified and the device operating parameters show that the laser power is unstable, it can be judged that the welding quality is poor.

[0026] Further, the present application also includes: The welding stage is determined by dividing the running state of the equipment running parameter, including a preheating stage, a welding stage and a cooling stage; a laser welding database is connected, and data extraction is performed based on the welding stage to obtain a historical running parameter set corresponding to the preheating stage, a historical running parameter set corresponding to the welding stage and a historical running parameter set corresponding to the cooling stage.

[0027] Specifically, the welding stage is determined by dividing the running state of the equipment running parameter, including a preheating stage, a welding stage and a cooling stage, and a laser welding database is connected, and data extraction is performed based on the welding stage to obtain a historical running parameter set corresponding to each stage. The division of the running state of the equipment running parameter refers to determining the stage of welding according to the real-time change of parameters such as laser power, welding time node and protective gas flow according to a preset division rule; connecting the laser welding database refers to establishing a communication link with the database storing a large amount of historical welding data so as to extract the required data therefrom; and the data extraction refers to screening the historical running parameter set corresponding to the determined welding stage from the database for subsequent model establishment and the like.

[0028] The welding stage is divided by monitoring the running state of the equipment running parameter, which can accurately divide the entire welding process into three stages of preheating, welding and cooling, providing clear time node division basis for subsequent targeted quality detection and model establishment, and accurate division of the welding stage can improve the controllability of the welding process; the laser welding database is connected and the historical running parameter set corresponding to each stage is extracted, and the historical running parameter set contains rich past welding experience data, which is combined with the current welding process to provide massive data support for establishing a welding quality evaluation model with learning and prediction capabilities.

[0029] Further, the historical running parameter set corresponding to the welding stage is obtained, and the method described in the application further comprises: Based on the laser welding process, welding operation nodes corresponding to each welding point based on a weld are set; a parameter time sequence matrix is constructed according to the welding operation nodes corresponding to each welding point based on a weld and the equipment running parameter sequence corresponding to each welding point; based on the parameter time sequence matrix, welding quality influence factors are configured in combination with a welding process standard, and the welding quality influence factors are strongly associated with the input layer weight and node activation function of the welding quality evaluation model.

[0030] Specifically, a series of specific welding operation position points are determined on the weld along the weld direction and welding process requirements, and corresponding welding operation instructions are assigned to each point, such as setting points of welding speed, laser power and other parameters. Further, the device operation parameter sequence refers to a sequence formed by arranging a series of parameter values of the device operation in time sequence at each welding operation node, such as a sequence of laser power changing over time at a certain welding point. The parameter time sequence matrix is formed by integrating the device operation parameter sequences of each welding point into a matrix form, where the rows represent different welding points and the columns represent parameter values at each time point, thereby forming a multi-dimensional data structure, which helps to present the parameter changes in the welding process in the form of a mathematical model, facilitating subsequent analysis and processing. The welding quality influence factor refers to a characteristic quantity of a parameter or parameter combination that has a significant impact on the welding quality. The strong correlation between these factors and the input layer weight and node activation function of the welding quality evaluation model means that these influence factors will directly affect the model's evaluation of the welding quality, i.e., the model will judge the welding quality according to the weight and activation function of these factors.

[0031] First, the welding operation nodes are set based on the laser welding process, ensuring fine management of the welding process and enabling precise welding operations at each key position of the weld. Then, the parameter time sequence matrix is constructed based on the welding operation nodes and the device operation parameter sequences, effectively integrating the parameter change information in the welding process and converting the welding process into a quantifiable and analyzable data structure. Preferably, by constructing the parameter time sequence matrix, the device operation parameter sequences of multiple welding points are integrated, and each device operation parameter sequence contains parameter values corresponding to timestamps, providing a detailed data basis for subsequent quality analysis.

[0032] Finally, based on the parameter time sequence matrix and combined with the welding process standard, the welding quality influence factor is configured, the data in the parameter time sequence matrix is mined to find features that have a significant impact on the welding quality, and these features are associated with the input layer weight and node activation function of the evaluation model, enabling the model to more accurately evaluate the welding quality. Preferably, by configuring the welding quality influence factor, the accuracy of welding quality evaluation is improved. Through the above steps, key data support and model optimization basis are provided for establishing an accurate welding quality evaluation model, ensuring the accuracy and reliability of welding quality detection.

[0033] Further, the present application also includes: Based on the laser welding requirements, the cooperative operation strategy of each welding point and the welding robot is set. The laser welding machine and the robot control system are connected, and the welding parameter-robot trajectory information is obtained in combination with the cooperative operation strategy. The welding quality detection result is dynamically evaluated by the welding quality evaluation model, and the welding cooperative operation is performed according to the welding parameter-robot trajectory information.

[0034] Specifically, the position of each welding point is determined according to the structure of the workpiece and the welding requirements. For each welding point, the movement trajectory of the robot is planned to ensure that the laser welding head can accurately and efficiently reach and complete the welding task. According to the material and thickness of the workpiece, appropriate laser power, welding speed and welding distance and other parameters are set. Ensure that the welding parameters match the robot trajectory to achieve high-quality welding. Through appropriate interfaces and protocols, the laser welding machine is connected with the robot control system. Ensure stable and real-time data transmission between the two. Synchronize welding parameters and robot trajectory information to the control system. Ensure that the laser welding machine and the robot can operate cooperatively according to the set parameters and trajectories. Set specific welding parameters in the control system, such as laser power, welding speed, etc. Input the trajectory information of the robot, including the coordinates and movement paths of each welding point. Verify the accuracy of the welding parameters and robot trajectory through simulation or test run. Adjust the parameters and trajectory until the desired effect is achieved. During the welding process, the welding quality is dynamically evaluated through the welding quality evaluation model. Real-time monitoring of various parameters during the welding process, such as weld shape, molten pool state, etc. According to the feedback of the quality evaluation model, adjust the welding parameters or robot trajectory in real time. Ensure that the welding quality always remains within the preset standard range.

[0035] Further, the present application also includes: By real-time monitoring equipment, the temperature field and stress field distribution of the current welding area are monitored in real time; based on the temperature field and stress field distribution of the current welding area, the microstructure and mechanical properties of the welded joint are predicted; according to the prediction result, the welding adjustment parameter is obtained, and the welding adjustment parameter meets the laser welding process parameter corresponding to the laser welding requirement.

[0036] Specifically, high-precision temperature sensors and stress sensors are installed in the welding area, which can monitor the temperature and stress changes in real time during the welding process. The sensors send the real-time collected temperature and stress data to the data processing system. Using the temperature field and stress field data obtained by real-time monitoring, combined with the physical properties of the welding material and the thermodynamic model of the welding process, the microstructure changes of the welded joint are simulated and analyzed. Based on the processed data, a prediction model of the microstructure and mechanical properties of the welded joint is established using numerical simulation techniques or machine learning algorithms. This model can predict the microstructure and possible mechanical properties of the welded joint based on the current temperature field and stress field distribution. According to the prediction results of the microstructure and mechanical properties output by the prediction model, it is analyzed whether the quality of the joint under the current welding parameters meets the expectations. If not, adjustments need to be made. Based on the prediction results and analysis, the welding parameters that need to be adjusted are determined through algorithms. These parameters include laser power, welding speed, welding sequence, etc. The goal of adjustment is to make the microstructure and mechanical properties of the welded joint meet the requirements of laser welding.

[0037] Further, the present application also includes: determining a target welding product, the target welding product including a combined welding product, a micro welding product; based on the target welding product, optimizing the layout of each welding point corresponding to the laser welding process, and updating the welding operation node.

[0038] Specifically, after determining the target welding product, the layout optimization of the laser welding process and the update of the welding operation node can be carried out according to the characteristics of the product. Identify the connection method between each component and the welding requirements. Determine the components that need to be welded and the strength and sealing requirements of the welding. According to the product structure, reasonably plan the layout of the welding points to reduce welding deformation and stress concentration. Optimize the welding sequence to ensure uniform heat distribution during the welding process and avoid the generation of thermal cracks and residual stress. According to the results of layout optimization, adjust the position and sequence of the welding operation node. Ensure that each welding operation node has clear operation guidance and quality control standards. For micro welding products, higher welding precision is required, so accurate positioning of the welding points must be ensured. Use high-precision positioning and clamping equipment to ensure the stability and accuracy of the welding process. According to the characteristics of the micro product, optimize the layout of the welding points to reduce the heat affected zone and welding deformation. Use micro welding technology or laser micro welding to improve welding quality and precision. According to the characteristics of the micro product and the results of layout optimization, adjust the parameter settings of the welding operation node, such as laser power, welding speed, etc. Strengthen the quality control and detection of the welding operation node to ensure that the welding quality of each node meets the requirements of the micro product.

[0039] Further, the present application also includes: Obtain the welding assembly structure of the combined welding product, which includes a plurality of welding components; based on the welding assembly structure, set the structure-differentiated laser welding process parameters; add the structure-differentiated laser welding process parameters as constraint information to the welding defect identification layer for real-time classification and identification of welding defects, and obtain a type of identification results.

[0040] Specifically, the detailed welding assembly structure of the combined welding product is obtained, which includes the material, thickness, shape, and connection method of each welding component. According to the structural characteristics of the assembly, such as thickness and material type, the welding components are classified. For different categories of welding components, appropriate laser welding process parameters are determined according to their structural characteristics. These parameters include laser power, welding speed, spot diameter, and welding angle. Through simulation, the initially determined process parameters are optimized to ensure welding quality and efficiency. The structure-differentiated laser welding process parameters are added as constraint information to the welding defect identification layer. When identifying welding defects, the process parameter differences of different welding assemblies are considered. In the welding process, the updated welding defect identification layer is used to identify welding defects in real time. After real-time classification and identification, a type of identification results is output, i.e., the specific welding defect types identified according to the welding assembly structure and differentiated process parameters.

[0041] Further, the present application also includes: Based on the micro welding product, the welding point morphology changes are synchronously output; based on the welding point morphology changes, the distributed welding point process parameters are set; the distributed welding point process parameters are added as constraint information to the welding defect identification layer for real-time classification and identification of welding defects, and a type of identification results is obtained.

[0042] Specifically, the formation of the welding points, the dynamics of the molten pool, the fusion of the welding joint, etc. of the micro welding products are monitored in real time using high-resolution camera systems or sensors. The monitored welding point morphology change data is recorded in real time. According to the real-time monitoring and recording of the welding point morphology change data, the formation process and characteristics of the welding points are analyzed. According to different welding point morphologies and change characteristics, corresponding distributed welding point process parameters are set. These parameters include fine-tuning of laser power, adjustment of welding speed, spacing and arrangement of welding points, etc. The distributed welding point process parameters are added to the welding defect identification layer as new constraint information. In this way, the morphology change of the welding points and the corresponding process parameters are considered simultaneously when identifying welding defects. Ensure that the welding defect identification layer can receive and process these new process parameters in order to more accurately perform real-time classification and identification of welding defects. During the welding process, the updated welding defect identification layer is used in combination with the distributed welding point process parameters to perform real-time classification and identification of welding defects. Output the two-class recognition result, that is, the specific welding defect type identified according to the welding point morphology change of the micro welding product and the distributed process parameters.

[0043] Further, the present application also includes: According to the structure difference laser welding process parameters, a critical recognition margin is set; based on each welding point position, an initial population is obtained, and the search space is set according to the critical recognition margin; based on the initial population, the layout of each welding point position is optimized in the search space, and the welding operation node is updated after the layout optimization is completed.

[0044] Specifically, the laser welding process parameters are differentiated according to the structure. Based on these parameters, a critical recognition margin is set. This margin is a threshold value used to determine the acceptable adjustment range of the weld point layout during optimization. It should be reasonably set according to the specific welding requirements and product quality standards to ensure a balance between welding quality and recognition accuracy. Based on the current weld points, an initial weld point layout scheme is obtained, which constitutes the initial population. Each weld point in this initial population is considered as an individual. Next, the search space is set according to the critical recognition margin. This search space defines the range and manner in which the weld points can be adjusted during optimization. For example, the maximum movement distance of each weld point in a specific direction can be set, or the minimum and maximum spacing between weld points can be set, etc. After setting the initial population and search space, optimization algorithms such as genetic algorithm, particle swarm optimization, etc. are used to search for the optimal weld point layout within the search space. During this process, the algorithm continuously adjusts the positions of the weld points to seek the optimal solution that meets all welding requirements and process parameter constraints. After the layout optimization is completed, an optimized weld point layout scheme is obtained. According to this scheme, the welding operation nodes are updated, including the specific position of each weld point, the welding sequence, the laser welding process parameters used, etc. These updates should be accurately reflected in the welding operation plan to ensure that the actual welding process can be carried out according to the optimized scheme.

[0045] In summary, the laser welding quality detection method provided by the present application has the following technical effects: Connect the laser welding machine, obtain the equipment running parameters, the equipment running parameters include laser power, welding time node, protective gas flow; based on the historical running parameter set corresponding to the preheating stage, the historical running parameter set corresponding to the welding stage, the historical running parameter set corresponding to the cooling stage, combined with adaptive topology learning, establish a welding quality evaluation model, the welding quality evaluation model is used for dynamically evaluating the welding quality of each weld point; based on the welding defect records in the laser welding database, determine the welding defect recognition layer under the limitation of the welding defect type, the welding defect type includes incomplete fusion, porosity, slag inclusion, crack, use an image acquisition device to obtain real-time image data in the welding process, upload the real-time image data to the welding defect recognition layer, wherein the real-time image data includes the molten pool state, the heat affected zone; embed the welding defect recognition layer into the welding quality evaluation model, determine the welding quality detection result combined with the equipment running parameters, achieve the technical effect of dynamically monitoring the whole process of welding by collecting the equipment running parameters of the preheating, welding and cooling stages, improving the detection accuracy of hidden defects, and providing strong support for laser welding process optimization.

[0046] Embodiment Two Based on the laser welding quality detection method in the foregoing embodiments, based on the same inventive concept, the application also provides a laser welding quality detection system, please refer to the accompanying Figure 2 , the system comprises: An operating parameter acquisition module 11 is configured to be connected to a laser welding machine and acquire device operating parameters, including laser power, welding time nodes, and protective gas flow.

[0047] An evaluation model establishment module 12 is configured to establish a welding quality evaluation model based on a set of historical operating parameters corresponding to the preheating stage, a set of historical operating parameters corresponding to the welding stage, and a set of historical operating parameters corresponding to the cooling stage, combined with adaptive topology learning, the welding quality evaluation model is used to dynamically evaluate the welding quality of each welding point.

[0048] An image acquisition module 13 is configured to determine a welding defect identification layer under the limitation of welding defect types based on welding defect records in a laser welding database, the welding defect types include incomplete fusion, porosity, slag inclusion, and cracks, use an image acquisition device to acquire real-time image data in the welding process, and upload the real-time image data to the welding defect identification layer, wherein the real-time image data includes molten pool state and heat affected zone.

[0049] A quality detection result acquisition module 14 is configured to embed the welding defect identification layer into the welding quality evaluation model, determine the welding quality detection result combined with the device operating parameters.

[0050] Further, the system comprises: By dividing the operating state of the device operating parameters, determine the welding stage, the welding stage includes the preheating stage, the welding stage, and the cooling stage; connect the laser welding database, and extract data based on the welding stage, acquire a set of historical operating parameters corresponding to the preheating stage, a set of historical operating parameters corresponding to the welding stage, and a set of historical operating parameters corresponding to the cooling stage.

[0051] Further, the system further comprises: Based on the laser welding process, set the welding operation nodes corresponding to each welding point based on the weld; according to the welding operation nodes corresponding to each welding point based on the weld, and the device operating parameter sequence corresponding to each welding point, construct a parameter time sequence matrix; based on the parameter time sequence matrix, combined with the welding process standard, configure the welding quality influence factor, the welding quality influence factor is strongly associated with the input layer weight and node activation function of the welding quality evaluation model.

[0052] Further, the system further comprises: Based on the laser welding requirement, a cooperative operation strategy of each welding point and a welding mechanical arm is set; a laser welding machine and a mechanical arm control system are connected, welding parameter-mechanical arm trajectory information is obtained in combination with the cooperative operation strategy; and welding quality detection results obtained by the welding quality evaluation model are dynamically evaluated, and welding cooperative operation is performed according to the welding parameter-mechanical arm trajectory information.

[0053] Further, the system further comprises: A real-time monitoring device is used to monitor the temperature field and stress field distribution of the current welding area in real time; the microstructure and mechanical properties of the welded joint are predicted based on the temperature field and stress field distribution of the current welding area; and welding adjustment parameters are obtained according to the prediction results, wherein the welding adjustment parameters meet the laser welding process parameters corresponding to the laser welding requirement.

[0054] Further, the system further comprises: A target welding product is determined, wherein the target welding product includes a combined welding product and a micro welding product; each welding point corresponding to the laser welding process is layout optimized based on the target welding product, and the welding operation node is updated.

[0055] Further, the system further comprises: A welding assembly structure of a combined welding product is obtained, wherein the welding assembly structure includes a plurality of welding components; structure-differentiated laser welding process parameters are set based on the welding assembly structure; and the structure-differentiated laser welding process parameters are added to the welding defect identification layer as constraint information to perform real-time classification and identification of welding defects, and a first type of identification result is obtained.

[0056] Further, the system further comprises: Based on a micro welding product, a welding point shape change is synchronously output; based on the welding point shape change, distributed welding point process parameters are set; and the distributed welding point process parameters are added to the welding defect identification layer as constraint information to perform real-time classification and identification of welding defects, and a second type of identification result is obtained.

[0057] Further, the system further comprises: A critical identification margin is set by comparing the structure-differentiated laser welding process parameters; an initial population is obtained based on each welding point, and a search space is set by comparing the critical identification margin; based on the initial population, each welding point is layout optimized in the search space, and the welding operation node is updated after the layout optimization is completed.

[0058] In the specification, each embodiment is described in a progressive manner, and each embodiment focuses on the difference from other embodiments.Figure 1 The laser welding quality inspection method and specific examples in Example 1 are also applicable to the laser welding quality inspection system of this embodiment. The detailed description of the laser welding quality inspection method above will clearly indicate the laser welding quality inspection system of this embodiment to those skilled in the art. For the sake of brevity, a detailed description will not be given here. The system disclosed in this embodiment will be described briefly, as it corresponds to the method disclosed in this embodiment. For relevant details, refer to the method description.

[0059] The above description of the disclosed embodiments is intended to enable one skilled in the art to implement or use the present application. Various modifications to these embodiments will be readily apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application is not limited to the embodiments shown herein, but is intended to conform to the widest scope consistent with the principles and novel features disclosed herein.

[0060] Obviously, those skilled in the art may make various changes and modifications to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the present application and its equivalents, the present application is intended to include these modifications and variations.

Claims

1. Laser welding quality detection method, characterized in that, The method comprises: Connecting to a laser welding machine to obtain equipment operating parameters, including laser power, welding time node, and shielding gas flow rate; Based on the historical operating parameter sets corresponding to the preheating stage, the welding stage, and the cooling stage, combined with adaptive topology learning, a welding quality evaluation model is established. The welding quality evaluation model is used to dynamically evaluate the welding quality of each welding point. Based on the welding defect records in the laser welding database, a welding defect identification layer is determined under the welding defect type definition, wherein the welding defect types include lack of fusion, porosity, slag inclusion, and cracks. Real-time image data of the welding process is acquired using an image acquisition device, and the real-time image data is uploaded to the welding defect identification layer, wherein the real-time image data includes the molten pool state and the heat-affected zone; The welding defect identification layer is embedded in the welding quality evaluation model, and the welding quality detection result is determined in combination with the equipment operating parameters.

2. The laser welding quality detection method according to claim 1, wherein: The method comprises: Determine the welding stage by dividing the equipment operation parameters and operation states, wherein the welding stage includes a preheating stage, a welding stage, and a cooling stage; Connect to the laser welding database, and extract data based on the welding stage to obtain a set of historical operating parameters corresponding to the preheating stage, a set of historical operating parameters corresponding to the welding stage, and a set of historical operating parameters corresponding to the cooling stage.

3. The laser welding quality detection method according to claim 2, wherein: Obtaining a set of historical operating parameters corresponding to the welding stage, the method further includes: Based on the laser welding process, set the welding operation nodes corresponding to each welding point based on the weld; According to the welding operation nodes corresponding to each welding point based on the weld, and the equipment operation parameter sequence corresponding to each welding point, a parameter timing matrix is ​​constructed; Based on the parameter time series matrix and in combination with the welding process standard, a welding quality influencing factor is configured, and the welding quality influencing factor is strongly associated with the input layer weight and node activation function of the welding quality evaluation model.

4. The laser welding quality detection method according to claim 2, wherein: Determining a welding quality test result, the method further comprises: Based on the laser welding requirements, set the collaborative operation strategy between each welding point and the welding robot arm; Connecting the laser welding machine to the robotic arm control system, combining the collaborative operation strategy to obtain welding parameter-robotic arm trajectory information; The welding quality detection results obtained by dynamic evaluation of the welding quality evaluation model are used to perform welding collaborative operations based on the welding parameter-robot arm trajectory information.

5. The laser welding quality detection method according to claim 4, characterized in that: Using an image acquisition device to obtain real-time image data during the welding process, the method further includes: Through real-time monitoring equipment, the temperature field and stress field distribution of the current welding area are monitored in real time; Predicting the microstructure and mechanical properties of the weld joint based on the temperature field and stress field distribution of the current welding area; According to the prediction result, welding adjustment parameters are obtained, where the welding adjustment parameters meet the laser welding process parameters corresponding to the laser welding requirements.

6. The laser welding quality detection method according to claim 3, wherein: Based on the laser welding process, the method further includes: Determining target welding products, wherein the target welding products include combination welding products and micro welding products; Based on the target welding product, the layout of each welding point corresponding to the laser welding process is optimized, and the welding operation node is updated.

7. The laser welding quality detection method according to claim 6, wherein: The target welding product includes a combined welding product, and the method includes: obtaining a welded assembly structure of a combined welded product, the welded assembly structure comprising a plurality of welded components; Based on the structure of the welding component, setting structurally differentiated laser welding process parameters; The structurally differentiated laser welding process parameters are used as constraint information and added to the welding defect identification layer to perform real-time classification and identification of welding defects to obtain a class of identification results.

8. The laser welding quality detection method according to claim 6, wherein: The target welding product includes a micro welding product, and the method further includes: Based on micro welding products, synchronously output the changes in welding point shape; Setting distributed welding point process parameters based on the welding point morphology changes; The distributed welding point process parameters are used as constraint information and added to the welding defect identification layer to perform real-time classification and identification of welding defects to obtain a second-category identification result.

9. The laser welding quality detection method according to claim 7, wherein: Optimizing the layout of each welding point corresponding to the laser welding process and updating the welding operation node, the method includes: According to the structure-differentiated laser welding process parameters, a critical identification margin is set; Based on each welding point, an initial population is obtained, and a search space is set according to the critical recognition margin; Based on the initial population, the layout of each welding point is optimized in the search space, and after the layout optimization is completed, the welding operation node is updated.

10. Laser welding quality detection system, characterized in that, For implementing the steps of the method according to any one of claims 1 to 9, the system comprises: An operating parameter acquisition module, which is used to connect to the laser welding machine and obtain equipment operating parameters, including laser power, welding time node, and shielding gas flow rate; An evaluation model establishment module, which is used to establish a welding quality evaluation model based on a set of historical operating parameters corresponding to the preheating stage, a set of historical operating parameters corresponding to the welding stage, and a set of historical operating parameters corresponding to the cooling stage, in combination with adaptive topology learning. The welding quality evaluation model is used to dynamically evaluate the welding quality of each welding point; An image acquisition module, the image acquisition module being used to determine a welding defect identification layer defined by welding defect types based on welding defect records in a laser welding database, wherein the welding defect types include lack of fusion, porosity, slag inclusion, and cracks; and to use an image acquisition device to acquire real-time image data during the welding process, and upload the real-time image data to the welding defect identification layer, wherein the real-time image data includes the state of the molten pool and the heat-affected zone; The quality inspection result acquisition module is used to embed the welding defect identification layer into the welding quality evaluation model and determine the welding quality inspection result in combination with the equipment operating parameters.

Images