Welding methods, systems, and storage media for crystallizer copper tubes

Abstract

The application discloses a crystallizer copper pipe welding method and system and a storage medium, and relates to the technical field of welding. The method comprises the following steps: performing correlation analysis on a historical welding data set of a crystallizer copper pipe, establishing a copper pipe characteristic parameter-welding parameter-quality mapping relationship library; performing matching analysis and deviation optimization on a target crystallizer copper pipe, determining initial welding parameters; deploying a monitoring and sensing network; based on the initial welding parameters, performing welding control on the target crystallizer copper pipe, simultaneously collecting multi-modal welding data streams through the monitoring and sensing network, performing defect identification on the multi-modal welding data streams, and obtaining copper pipe welding defect parameters; performing compensation analysis on the initial welding parameters, determining target welding parameters, and performing welding compensation control through the target welding parameters. The technical problem of low welding quality of the crystallizer copper pipe in the prior art is solved, and the technical effect of improving the welding quality by identifying welding defects in real time and dynamically adjusting welding parameters is achieved.

Description

Technical Field

[0001] This invention relates to the field of welding technology, and more specifically to welding methods, systems, and storage media for copper tubes in crystallizers. Background Technology

[0002] In the welding and manufacturing process of copper tubes for crystallizers, welding quality directly affects the service life and overall performance of the crystallizer. However, existing technologies typically rely on manual experience to set welding parameters, lacking systematic data support. This not only makes it difficult to accurately match the characteristic parameters of different copper tubes but also easily leads to weld defects such as cracks, porosity, and lack of fusion due to unreasonable parameter settings. Furthermore, the detection of welding defects in traditional welding processes often lags behind the welding operation itself, making real-time monitoring and dynamic correction impossible. This results in unstable welding quality, a high rework rate, and seriously affects production efficiency and product consistency. Summary of the Invention

[0003] This application provides a welding method, system, and storage medium for crystallizer copper tubes, which solves the technical problem of low welding quality of crystallizer copper tubes in the prior art.

[0004] The first aspect of this application provides a method for welding copper tubes in a crystallizer, the method comprising:

[0005] Historical welding datasets of copper tubes in a crystallizer are collected, and correlation analysis is performed on these datasets to establish a mapping relationship library between copper tube characteristic parameters, welding parameters, and quality. Based on this library, matching analysis and deviation optimization are performed on the target crystallizer copper tube to determine initial welding parameters. A monitoring and sensing network is deployed in the welding area of ​​the target crystallizer copper tube, integrating an infrared thermal imager, a laser displacement sensor, and an acoustic emission detector. Welding control is performed on the target crystallizer copper tube based on the initial welding parameters. Simultaneously, multimodal welding data streams are collected through the monitoring and sensing network, and defect identification is performed on these data streams to obtain copper tube welding defect parameters. Compensation analysis is then performed on the initial welding parameters based on these defect parameters to determine target welding parameters, and welding compensation control is implemented using these target welding parameters.

[0006] A second aspect of this application provides a welding system for copper tubes in a crystallizer, the system comprising:

[0007] The system comprises the following modules: **Correlation Analysis Module:** Collects historical welding datasets of the crystallizer copper tubes, performs correlation analysis on these datasets, and establishes a mapping relationship library between copper tube characteristic parameters, welding parameters, and quality. **Initial Parameter Determination Module:** Based on the copper tube characteristic parameter-welding parameter-quality mapping relationship library, performs matching analysis and deviation optimization on the target crystallizer copper tube to determine the initial welding parameters. **Monitoring Network Deployment Module:** Deploys a monitoring and sensing network in the welding area of ​​the target crystallizer copper tube. This network integrates an infrared thermal imager, a laser displacement sensor, and an acoustic emission detector. **Defect Identification Module:** Performs welding control on the target crystallizer copper tube based on the initial welding parameters. Simultaneously, it collects multimodal welding data streams through the monitoring and sensing network, identifies defects in the multimodal welding data streams, and obtains copper tube welding defect parameters. **Compensation Control Module:** Performs compensation analysis on the initial welding parameters based on the copper tube welding defect parameters, determines the target welding parameters, and performs welding compensation control based on these target welding parameters.

[0008] A third aspect of this application provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the welding method for the crystallizer copper tube provided in this application.

[0009] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0010] First, historical welding datasets of crystallizer copper tubes are collected, and correlation analysis is performed on these datasets to establish a mapping relationship library between copper tube characteristic parameters, welding parameters, and quality. Next, based on this mapping relationship library, matching analysis and deviation optimization are performed on the target crystallizer copper tube to determine the initial welding parameters. Simultaneously, a monitoring and sensing network is deployed in the welding area of ​​the target crystallizer copper tube, integrating an infrared thermal imager, a laser displacement sensor, and an acoustic emission detector. Then, welding control is executed on the target crystallizer copper tube based on the initial welding parameters. Simultaneously, multimodal welding data streams are collected through the monitoring and sensing network, and defect identification is performed on these data streams to obtain copper tube welding defect parameters. Finally, compensation analysis is performed on the initial welding parameters based on the copper tube welding defect parameters to determine the target welding parameters, and welding compensation control is implemented using these target welding parameters. This solves the technical problem of low welding quality of crystallizer copper tubes in existing technologies, achieving the technical effect of improving welding quality through real-time identification of welding defects and dynamic adjustment of welding parameters. Attached Figure Description

[0011] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0012] Figure 1 A schematic diagram of the welding method for the copper tube of the crystallizer provided in the embodiments of this application;

[0013] Figure 2 This is a schematic diagram of the welding system structure for the copper tube of the crystallizer provided in an embodiment of this application.

[0014] Figure labeling: Correlation analysis module 11, initial parameter determination module 12, monitoring network deployment module 13, defect identification module 14, compensation control module 15. Detailed Implementation

[0015] This application solves the technical problem of low welding quality of crystallizer copper tubes in the prior art by providing a welding method, system and storage medium for crystallizer copper tubes.

[0016] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0017] It should be noted that the terms "comprising" and "having" are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or server that includes a series of steps or units is not necessarily limited to those steps or units that are explicitly listed, but may include other steps or modules that are not explicitly listed or that are inherent to these processes, methods, products, or devices.

[0018] Example 1, as Figure 1 As shown, this application provides a method for welding copper tubes for crystallizers, wherein the method includes:

[0019] Collect historical welding datasets of copper tubes in the crystallizer, perform correlation analysis on the historical welding datasets of copper tubes in the crystallizer, and establish a mapping relationship library of copper tube characteristic parameters, welding parameters, and quality.

[0020] In this embodiment of the application, welding data generated during the previous production process of the crystallizer copper tube is collected to form a historical welding dataset of the crystallizer copper tube. The historical welding dataset of the crystallizer copper tube includes the physical and chemical property parameters of the copper tube (such as the material composition, grain size, geometric dimensions, surface condition, thermal conductivity, hardness, etc.), the corresponding welding process parameters (such as welding current, voltage, welding speed, shielding gas flow rate, wire feed speed, heat input, etc.), and welding quality inspection results (such as weld strength, density, appearance, defect distribution, pass rate, etc.).

[0021] Multidimensional statistical analysis and machine learning modeling methods were used to calculate and filter the correlation between copper tube characteristic parameters, welding parameters, and welding quality results, eliminating invalid or weakly correlated parameters to obtain key copper tube characteristic data, key welding parameters, and welding quality data. Based on these key data, a multidimensional mapping model was further constructed, using methods such as regression analysis, neural networks, or Bayesian inference, to establish a database of copper tube characteristic parameters, welding parameters, and quality mapping relationships. This database allows for the association, storage, and retrieval of copper tubes with different characteristics with corresponding welding process parameters and welding quality results.

[0022] Furthermore, a database mapping relationship between copper tube characteristic parameters, welding parameters, and quality is established, including:

[0023] The historical welding dataset of the crystallizer copper tubes is subjected to anomaly cleaning and missing value imputation to obtain a usable historical welding dataset of crystallizer copper tubes. Based on the usable historical welding dataset of crystallizer copper tubes, copper tube characteristic dimension data, welding parameter dimension data, and welding quality dimension data are extracted. Correlation analysis and filtering are performed on the copper tube characteristic dimension data, welding parameter dimension data, and welding quality dimension data to obtain copper tube characteristic data, associated welding parameters, and welding quality data. Based on the copper tube characteristic data, associated welding parameters, and welding quality data, mapping modeling is performed to establish a mapping relationship library of copper tube characteristic parameters, welding parameters, and quality.

[0024] In establishing the mapping relationship database of copper tube characteristic parameters, welding parameters, and quality, the first step is to preprocess the collected historical welding dataset of crystallizer copper tubes. Outliers are identified and removed, and missing data items are filled using mean interpolation, interpolation fitting, or completion methods based on similar samples to obtain a usable historical welding dataset of crystallizer copper tubes. Based on this, the usable historical welding dataset of crystallizer copper tubes is dimensionally split, extracting copper tube characteristic data (including geometric dimensions, material composition, thermal conductivity, hardness, surface roughness, etc.), welding parameter data (including welding current, voltage, welding speed, wire feed speed, shielding gas flow rate, heat input, etc.), and welding quality data (including weld density, forming consistency, strength, defect rate, rework rate, etc.). Subsequently, correlation analysis methods, such as Pearson correlation coefficient analysis, principal component analysis, and mutual information entropy calculation, are used to determine and filter the feature correlations of the three types of data, removing redundant or weakly correlated terms to obtain effective copper tube characteristic data, key associated welding parameters, and welding quality data. Finally, based on copper tube characteristic data, key associated welding parameters, and welding quality data, a mapping function relationship is constructed and stored using multivariate regression modeling, neural network modeling, or Bayesian network modeling methods, forming a copper tube characteristic parameter-welding parameter-quality mapping relationship library.

[0025] For the copper tube characteristic dimension data, welding parameter dimension data, and welding quality dimension data extracted from the historical welding dataset of crystallizer copper tubes, further feature correlation determination and screening were performed. Specifically, Pearson correlation coefficient analysis was used to calculate the linear correlation between various characteristic parameters and welding quality indicators, and redundant features with correlation coefficients below a preset threshold were eliminated. Subsequently, principal component analysis (PCA) was used to reduce the dimensionality of multidimensional features, extracting principal component features that contribute significantly to welding quality and avoiding multicollinearity among features. At the same time, mutual information entropy calculation was introduced to mine nonlinear relationships and potential implicit correlations, ensuring that features with weak linear correlations but high information contributions are retained.

[0026] Optionally, a neural network modeling method is employed, using copper pipe characteristic data and welding parameter data as the input layer and welding quality indicators as the output layer. Through feature extraction and nonlinear mapping in the hidden layer, the nonlinear relationships between complex multidimensional parameters are learned. Alternatively, a Bayesian network modeling method is used to establish causal relationship links between parameters based on the conditional probability distributions of copper pipe characteristics, welding parameters, and welding quality.

[0027] Based on the copper tube characteristic parameters-welding parameters-quality mapping relationship library, the target crystallizer copper tube is matched and the deviation is adjusted to determine the initial welding parameters.

[0028] Based on a database of copper tube characteristic parameters, welding parameters, and quality mapping relationships, the characteristic parameters of the target crystallizer copper tube are input and analyzed. Specifically, the geometric dimensions, material composition, thermal conductivity, hardness, and surface condition of the target copper tube are used as input features. Similarity matching calculations are performed between these features and historical copper tube characteristic parameters stored in the database. Similarity calculations can employ Euclidean distance, cosine similarity, or cluster center-based matching algorithms to select the historical characteristic samples closest to the target copper tube and their corresponding welding parameters. Based on the matching results, the welding parameters are optimized by comparing the differences between the target and matched copper tube characteristic parameters to obtain a deviation factor. This factor is then used to correct welding parameters such as current, voltage, welding speed, heat input, and wire feed speed. Through matching analysis and deviation optimization, initial welding parameters that meet the actual characteristic requirements of the target copper tube are finally obtained.

[0029] Furthermore, determining the initial welding parameters includes:

[0030] Obtain the copper tube welding quality standard and the characteristic parameters of the target crystallizer copper tube; optimize the copper tube characteristic parameter-welding parameter-quality mapping relationship library according to the copper tube welding quality standard to obtain a set of optional copper tube characteristic parameters-associated welding parameters; calculate the similarity between the characteristic parameters of the target crystallizer copper tube and the set of optional copper tube characteristic parameters-associated welding parameters to obtain the matching copper tube characteristic parameters-associated welding parameters with the highest similarity; perform deviation optimization analysis on the matching copper tube characteristic parameters-associated welding parameters to determine the initial welding parameters.

[0031] Specifically, a preset copper tube welding quality standard is obtained, which includes evaluation indicators such as weld strength, density, uniformity of formation, and defect control rate. Simultaneously, characteristic parameters of the target crystallizer copper tube are acquired, such as geometric dimensions, material composition, thermal conductivity, hardness, and surface condition. Based on the copper tube welding quality standard, the established copper tube characteristic parameter-welding parameter-quality mapping relationship library is screened and optimized, eliminating sample data that does not meet the welding quality standard, resulting in a set of optional copper tube characteristic parameters-associated welding parameters. On this basis, the characteristic parameters of the target copper tube are compared with the optional copper tube characteristic parameter-associated welding parameter set for similarity calculation. Similarity calculation methods can employ Euclidean distance, Mahalanobis distance, cosine similarity, or cluster center-based matching algorithms to obtain the matching copper tube characteristic parameter-associated welding parameter with the highest similarity. Finally, a deviation optimization analysis is performed on the matching copper tube characteristic parameters and associated welding parameters. This involves converting the difference between the actual characteristics of the target copper tube and the characteristics of the matching sample into a deviation factor, and correcting process parameters such as welding current, voltage, speed, heat input, and wire feed speed based on a preset parameter adjustment rule library to obtain initial welding parameters that meet the characteristics requirements of the target copper tube.

[0032] Furthermore, a deviation optimization analysis is performed on the matching copper tube characteristic parameters and associated welding parameters to determine the initial welding parameters, including:

[0033] The deviation parameters of the target crystallizer copper tube are calculated with those of the matching copper tube and associated welding parameters to determine the copper tube characteristic deviation parameters. A set of copper tube characteristic deviation types is obtained, and a welding parameter adjustment strategy is associated with the set of copper tube characteristic deviation types. The adjustment range of the welding parameter adjustment strategy is analyzed based on the set of copper tube characteristic deviation types and the set of optional copper tube characteristic parameters and associated welding parameters to construct a welding parameter optimization strategy library. Based on the welding parameter optimization strategy library, the deviation of the copper tube characteristic deviation parameters is analyzed for deviation tuning to determine the initial welding parameters.

[0034] Specifically, the characteristic parameters of the target crystallizer copper tube are compared item by item with the characteristic parameters of the matching copper tube and the associated welding parameters. Copper tube characteristic deviation parameters are obtained through methods such as difference calculation, proportional difference calculation, or normalized difference degree calculation. These deviation parameters include geometric dimension deviations, material composition deviations, thermal property deviations, and surface condition deviations. A preset set of copper tube characteristic deviation types is obtained. Different types of deviations correspond to different welding parameter adjustment strategies. For example, dimensional deviations can be associated with welding current and voltage adjustment strategies, material deviations can be associated with heat input and welding speed adjustment strategies, thermal conductivity deviations can be associated with shielding gas flow rate and wire feed speed adjustment strategies, and surface condition deviations can be associated with arc stability and forming control strategies.

[0035] Based on the copper pipe characteristic deviation type set, historical data from the optional copper pipe characteristic parameter-related welding parameter set are statistically analyzed and their magnitudes calculated. The parameter correction magnitudes corresponding to different deviation degrees are analyzed to construct a welding parameter optimization strategy library. This library is then used to fine-tune the deviation parameters of the target copper pipe. Based on the magnitude and type of the deviation, corresponding welding parameters are selected and corrected, outputting initial welding parameters that meet the actual requirements of the target copper pipe.

[0036] For example, when the target copper tube's wall thickness is detected to be 5% greater than the matching sample, the system automatically increases the welding current by about 3%, the welding voltage by about 2%, and simultaneously reduces the welding speed by 2% to ensure that the welding heat input meets the fusion requirements of the thickened copper tube. When the target copper tube's material is detected to have a high impurity content, the system automatically increases the shielding gas flow rate by about 5% and reduces the wire feed speed by about 3% to mitigate the generation of porosity defects. When the target copper tube's thermal conductivity is detected to be 10% higher than the matching sample, the system automatically increases the welding current by about 4%, increases the heat input, and appropriately reduces the welding speed by 1% to 2% to prevent the weld from developing incomplete fusion defects due to excessive heat dissipation. When the target copper tube's surface roughness is detected to be higher than the standard threshold, the system automatically reduces the arc voltage by about 2% and fine-tunes the wire feed speed to make the arc more stable, thereby ensuring the consistency of the weld formation.

[0037] A monitoring and sensing network is deployed in the welding area of ​​the copper tube of the target crystallizer. The monitoring and sensing network integrates an infrared thermal imager, a laser displacement sensor, and an acoustic emission detector.

[0038] A monitoring and sensing network is deployed in the welding area of ​​the copper tube in the target crystallizer to enable real-time, multi-dimensional monitoring and data acquisition of the welding process. This network integrates multiple sensor units, including an infrared thermal imager, a laser displacement sensor, and an acoustic emission detector. The infrared thermal imager dynamically acquires and analyzes the temperature field in the welding area, providing real-time data on the temperature distribution and gradient changes in the weld and heat-affected zone, thus determining the appropriateness of the welding heat input. The laser displacement sensor detects the weld formation and geometric deformation of the copper tube during welding, obtaining forming characteristic parameters such as weld reinforcement height, weld width, and weld offset through non-contact, high-precision measurement. The acoustic emission detector monitors transient elastic wave signals generated during welding due to defects such as porosity, cracks, or lack of fusion, and achieves early identification and location of potential defects through feature extraction and pattern recognition of the acoustic emission signals.

[0039] Welding control is performed on the target crystallizer copper tube based on the initial welding parameters. At the same time, multimodal welding data streams are collected through the monitoring and sensing network, and defects are identified in the multimodal welding data streams to obtain copper tube welding defect parameters.

[0040] Welding control is performed on the target crystallizer copper tube based on the initial welding parameters. That is, the process parameters such as current, voltage, welding speed, wire feed speed and shielding gas flow rate after deviation optimization are input into the welding control device to drive the welding power supply, wire feed mechanism and shielding gas supply system to operate and carry out welding operations on the welding area of ​​the target copper tube.

[0041] During the welding process, a monitoring and sensing network acquires multimodal welding data streams in real time. The infrared thermal imager outputs the temperature field distribution and heat input time-series curves of the welding area; the laser displacement sensor outputs weld geometry data, including weld width, reinforcement height, weld offset, and welding deformation; and the acoustic emission detector outputs acoustic emission waveform data containing defect signal features such as porosity, cracks, and lack of fusion. After acquisition, the multimodal welding data streams are synchronously transmitted to the defect identification module for feature analysis and pattern recognition. Signal processing, feature extraction, and multimodal fusion algorithms are used to cross-validate different data sources, forming a multidimensional feature set of welding defects. A classification model or deep learning network is then used to determine the defect type and quantify the defect severity, ultimately outputting copper tube welding defect parameters, including defect type, defect location, defect occurrence probability, and severity indicators.

[0042] Furthermore, the parameters for copper tube welding defects are obtained, including:

[0043] A multimodal welding defect dataset of the target crystallizer copper tube is collected. Defect recognition training is performed based on the multimodal welding defect dataset to build a multi-channel for crystallizer copper tube defect recognition. The multimodal welding data stream is matched and mapped to the multi-channel for crystallizer copper tube defect recognition for synchronous defect recognition, and a multimodal feature set of welding defects is output. The multimodal feature set of welding defects is weighted, fused, and quantified to obtain the copper tube welding defect parameters.

[0044] Specifically, a multimodal welding defect dataset is collected during the welding process of the target crystallizer copper tube. This dataset includes time-series welding temperature field data acquired by an infrared thermal imager, weld geometry feature data acquired by a laser displacement sensor, and acoustic signal feature data acquired by an acoustic emission detector. Based on this multimodal welding defect dataset, a defect recognition model is constructed using known defect samples for training. Corresponding recognition branch networks are established for each input channel of different sensor modes, ultimately integrated into a multi-channel system for crystallizer copper tube defect recognition. The real-time acquired multimodal welding data stream is input into the defect recognition multi-channel system for matching, mapping, and parallel processing. Each recognition channel independently extracts temperature anomaly features, forming deviation features, and acoustic emission defect features, generating a multimodal feature set of welding defects at the output. Subsequently, the multimodal feature set of welding defects is weighted and fused according to the accuracy weight factor of each recognition channel, and numerically processed in conjunction with preset quantitative indicators for defect types (such as crack length, porosity, and unfused area) to obtain copper tube welding defect parameters. These parameters include the defect type, location, probability of occurrence, and severity level.

[0045] Furthermore, a multi-channel system for identifying defects in the copper tubes of the crystallizer is established, including:

[0046] Based on the monitoring and sensing network, a multi-channel defect identification architecture is constructed, wherein each identification channel in the multi-channel defect identification architecture corresponds one-to-one with the sensor type in the monitoring and sensing network; modal classification and defect identification training are performed on the multimodal welding defect dataset according to the monitoring and sensing network to obtain a multimodal defect identification branch network set; the multimodal defect identification branch network set is matched and embedded into the multi-channel defect identification architecture to build the crystallizer copper tube defect identification multi-channel.

[0047] In constructing a multi-channel defect identification system for crystallizer copper tubes, the first step is to build a multi-channel defect identification architecture based on the composition of the monitoring and sensing network. This architecture includes several parallel identification channels, each corresponding one-to-one with a sensor type in the monitoring and sensing network. For example, the infrared thermal imaging channel is used to process abnormal temperature field features, the laser displacement channel is used to process weld geometry features, and the acoustic emission channel is used to process acoustic signal features of defects. Subsequently, based on the multimodal welding defect dataset collected by the monitoring and sensing network, the data of different modes are classified and labeled as temperature mode data, geometric mode data, and acoustic mode data, respectively. Known defect samples are then used to separately extract features and train defect identification for each mode of data, resulting in a corresponding multimodal defect identification branch network set. Finally, the multimodal defect identification branch network set is embedded into the multi-channel defect identification architecture according to the corresponding relationships, forming a complete multi-channel crystallizer copper tube defect identification system. This allows data collected by different sensors to be parsed in parallel in independent channels, generating synchronous identification results of multi-source features at the output end.

[0048] Furthermore, the multimodal feature set of welding defects is weighted, fused, and quantified to obtain the welding defect parameters of the copper tube, including:

[0049] Based on the output accuracy of each channel in the multi-channel identification of copper tube defects in the crystallizer, a multimodal feature weighting factor is determined; the welding defect multimodal feature set is weighted and fused based on the multimodal feature weighting factor to obtain a welding defect fusion feature set; the welding defect fusion feature set is then subjected to a defect degree quantification analysis according to the welding defect type to obtain the copper tube welding defect parameters.

[0050] Specifically, based on the training accuracy and real-time output accuracy of each identification channel in the multi-channel identification of copper tube defects in the crystallizer, corresponding multimodal feature weighting factors are determined, such as the temperature feature identification accuracy of the infrared thermography channel, the geometric shape feature identification accuracy of the laser displacement channel, and the acoustic feature identification accuracy of the acoustic emission channel. Based on the multimodal feature weighting factors, the welding defect features output from different channels are weighted and fused, making the feature contribution of high-precision channels greater and the feature contribution of low-precision channels relatively weaker, thereby obtaining a welding defect fusion feature set with higher robustness and reliability. The welding defect fusion feature set is then quantitatively analyzed according to defect type. For example, crack length and propagation rate are calculated for crack defects, pore diameter, number, and volume distribution are calculated for porosity defects, and incomplete fusion area and insufficient fusion depth rate are calculated for non-fusion defects. Through the above quantitative process, the copper tube welding defect parameters are obtained.

[0051] Based on the copper tube welding defect parameters, the initial welding parameters are compensated and analyzed to determine the target welding parameters, and welding compensation control is performed using the target welding parameters.

[0052] The initial welding parameters are compensated based on the welding defect parameters of the copper tube. This involves matching the welding defect parameters with a pre-defined welding defect-parameter compensation database to determine the process parameter adjustment strategy corresponding to the defect characteristics. Specifically, when the welding defect parameters indicate abnormal temperature field or insufficient heat input, the compensation analysis unit automatically increases the welding current and voltage, and reduces the welding speed if necessary, to enhance heat input. When the defect parameters indicate weld formation deviation, the system corrects the weld trajectory by adjusting the welding torch position compensation and the wire feed speed. When the defect parameters indicate porosity or crack risk, the probability of defect occurrence is reduced by increasing the shielding gas flow rate and adjusting the welding arc stability control parameters. The above compensation analysis process can employ a real-time parameter correction algorithm based on a PID controller or fuzzy controller to convert the defect parameters into welding parameter compensation amounts, which are then superimposed and corrected to the initial welding parameters to obtain the target welding parameters. Welding compensation control is then performed on the target crystallizer copper tube based on the target welding parameters to achieve dynamic adjustment and closed-loop optimization of the welding process, thereby continuously suppressing defect propagation and improving welding quality and stability during the welding process.

[0053] Furthermore, determining the target welding parameters includes:

[0054] A welding defect-associated welding parameter library is constructed. Based on the welding defect-associated welding parameter library, a compensation control fitting is performed to obtain a PID controller. The PID controller is used to perform compensation analysis on the welding defect parameters of the copper tube to determine the welding parameter compensation amount. Based on the welding parameter compensation amount, the initial welding parameters are compensated and corrected to determine the target welding parameters.

[0055] The welding defect-associated welding parameter library stores the relationship between different types of welding defects and their corresponding process parameter adjustments. For example, the relationship between crack defects and the adjustment of current, voltage, and welding speed; the relationship between porosity defects and the adjustment of shielding gas flow rate and wire feed speed; and the relationship between incomplete fusion defects and the adjustment of heat input and welding path offset.

[0056] Compensation control fitting is performed based on a welding defect-related welding parameter library to train and obtain a PID controller. This PID controller can output corresponding parameter correction commands in real time according to the changes and dynamic trends of the defect parameters. Specifically, after acquiring the copper tube welding defect parameters, the PID controller uses the defect deviation as input and calculates the corresponding welding parameter compensation amount through proportional (P), integral (I), and derivative (D) components, such as current correction, voltage correction, welding speed correction, or gas flow correction. The welding parameter compensation amount is then superimposed on the initial welding parameters to form the dynamically corrected target welding parameters.

[0057] In summary, the embodiments of this application have at least the following technical effects:

[0058] First, historical welding datasets of crystallizer copper tubes are collected, and correlation analysis is performed on these datasets to establish a mapping relationship library between copper tube characteristic parameters, welding parameters, and quality. Next, based on this mapping relationship library, matching analysis and deviation optimization are performed on the target crystallizer copper tube to determine the initial welding parameters. Simultaneously, a monitoring and sensing network is deployed in the welding area of ​​the target crystallizer copper tube, integrating an infrared thermal imager, a laser displacement sensor, and an acoustic emission detector. Then, welding control is executed on the target crystallizer copper tube based on the initial welding parameters. Simultaneously, multimodal welding data streams are collected through the monitoring and sensing network, and defect identification is performed on these data streams to obtain copper tube welding defect parameters. Finally, compensation analysis is performed on the initial welding parameters based on the copper tube welding defect parameters to determine the target welding parameters, and welding compensation control is implemented using these target welding parameters. This solves the technical problem of low welding quality of crystallizer copper tubes in existing technologies, achieving the technical effect of improving welding quality through real-time identification of welding defects and dynamic adjustment of welding parameters.

[0059] Example 2, based on the same inventive concept as the welding method for the crystallizer copper tube in the aforementioned examples, such as... Figure 2 As shown, this application provides a welding system for crystallizer copper tubes, wherein the system includes:

[0060] Association Analysis Module 11: Collects historical welding datasets of crystallizer copper tubes, performs association analysis on the historical welding datasets of crystallizer copper tubes, and establishes a mapping relationship library of copper tube characteristic parameters, welding parameters, and quality. Initial Parameter Determination Module 12: Based on the mapping relationship library of copper tube characteristic parameters, welding parameters, and quality, performs matching analysis and deviation optimization on the target crystallizer copper tube to determine the initial welding parameters. Monitoring Network Deployment Module 13: Deploys a monitoring and sensing network in the welding area of ​​the target crystallizer copper tube, the monitoring and sensing network integrating an infrared thermal imager, a laser displacement sensor, and an acoustic emission detector. Defect Identification Module 14: Performs welding control on the target crystallizer copper tube based on the initial welding parameters, and simultaneously collects multimodal welding data streams through the monitoring and sensing network, performs defect identification on the multimodal welding data streams, and obtains copper tube welding defect parameters. Compensation Control Module 15: Performs compensation analysis on the initial welding parameters based on the copper tube welding defect parameters, determines the target welding parameters, and performs welding compensation control based on the target welding parameters.

[0061] Furthermore, the correlation analysis module 11 is used to perform the following methods:

[0062] The historical welding dataset of the crystallizer copper tubes is subjected to anomaly cleaning and missing value imputation to obtain a usable historical welding dataset of crystallizer copper tubes. Based on the usable historical welding dataset of crystallizer copper tubes, copper tube characteristic dimension data, welding parameter dimension data, and welding quality dimension data are extracted. Correlation analysis and filtering are performed on the copper tube characteristic dimension data, welding parameter dimension data, and welding quality dimension data to obtain copper tube characteristic data, associated welding parameters, and welding quality data. Based on the copper tube characteristic data, associated welding parameters, and welding quality data, mapping modeling is performed to establish a mapping relationship library of copper tube characteristic parameters, welding parameters, and quality.

[0063] Furthermore, the initial parameter determination module 12 is used to perform the following method:

[0064] Obtain the copper tube welding quality standard and the characteristic parameters of the target crystallizer copper tube; optimize the copper tube characteristic parameter-welding parameter-quality mapping relationship library according to the copper tube welding quality standard to obtain a set of optional copper tube characteristic parameters-associated welding parameters; calculate the similarity between the characteristic parameters of the target crystallizer copper tube and the set of optional copper tube characteristic parameters-associated welding parameters to obtain the matching copper tube characteristic parameters-associated welding parameters with the highest similarity; perform deviation optimization analysis on the matching copper tube characteristic parameters-associated welding parameters to determine the initial welding parameters.

[0065] Furthermore, the initial parameter determination module 12 is used to perform the following method:

[0066] The deviation parameters of the target crystallizer copper tube are calculated with those of the matching copper tube and associated welding parameters to determine the copper tube characteristic deviation parameters. A set of copper tube characteristic deviation types is obtained, and a welding parameter adjustment strategy is associated with the set of copper tube characteristic deviation types. The adjustment range of the welding parameter adjustment strategy is analyzed based on the set of copper tube characteristic deviation types and the set of optional copper tube characteristic parameters and associated welding parameters to construct a welding parameter optimization strategy library. Based on the welding parameter optimization strategy library, the deviation of the copper tube characteristic deviation parameters is analyzed for deviation tuning to determine the initial welding parameters.

[0067] Furthermore, the defect identification module 14 is used to perform the following method:

[0068] A multimodal welding defect dataset of the target crystallizer copper tube is collected. Defect recognition training is performed based on the multimodal welding defect dataset to build a multi-channel for crystallizer copper tube defect recognition. The multimodal welding data stream is matched and mapped to the multi-channel for crystallizer copper tube defect recognition for synchronous defect recognition, and a multimodal feature set of welding defects is output. The multimodal feature set of welding defects is weighted, fused, and quantified to obtain the copper tube welding defect parameters.

[0069] Furthermore, the defect identification module 14 is used to perform the following method:

[0070] Based on the monitoring and sensing network, a multi-channel defect identification architecture is constructed, wherein each identification channel in the multi-channel defect identification architecture corresponds one-to-one with the sensor type in the monitoring and sensing network; modal classification and defect identification training are performed on the multimodal welding defect dataset according to the monitoring and sensing network to obtain a multimodal defect identification branch network set; the multimodal defect identification branch network set is matched and embedded into the multi-channel defect identification architecture to build the crystallizer copper tube defect identification multi-channel.

[0071] Furthermore, the defect identification module 14 is used to perform the following method:

[0072] Based on the output accuracy of each channel in the multi-channel identification of copper tube defects in the crystallizer, a multimodal feature weighting factor is determined; the welding defect multimodal feature set is weighted and fused based on the multimodal feature weighting factor to obtain a welding defect fusion feature set; the welding defect fusion feature set is then subjected to a defect degree quantification analysis according to the welding defect type to obtain the copper tube welding defect parameters.

[0073] Furthermore, the compensation control module 15 is used to perform the following method:

[0074] A welding defect-associated welding parameter library is constructed. Based on the welding defect-associated welding parameter library, a compensation control fitting is performed to obtain a PID controller. The PID controller is used to perform compensation analysis on the welding defect parameters of the copper tube to determine the welding parameter compensation amount. Based on the welding parameter compensation amount, the initial welding parameters are compensated and corrected to determine the target welding parameters.

[0075] Example 3: Based on the same inventive concept as the welding method for the copper tubes of the crystallizer in the preceding examples, this example provides a computer-readable storage medium that can be used to store software programs, computer-executable programs, and modules, such as the program instructions / modules corresponding to the welding method for the copper tubes of the crystallizer in this application. The processor executes various functional applications and data processing of the computer device by running the software programs, instructions, and modules stored in the memory, thereby realizing the aforementioned welding method for the copper tubes of the crystallizer.

[0076] It should be noted that the order of the embodiments described above is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. Furthermore, the above description focuses on specific embodiments of this specification. Additionally, the processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired results. In some implementations, multitasking and parallel processing are possible or may be advantageous.

[0077] The above description is only a preferred embodiment of this application and is not intended to limit this application. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.

[0078] This specification and accompanying drawings are merely illustrative examples of this application and are intended to cover any and all modifications, variations, combinations, or equivalents within the scope of this application. Clearly, those skilled in the art can make various alterations and modifications to this application without departing from its scope. Therefore, if such modifications and modifications fall within the scope of this application and its equivalents, this application intends to include such modifications and modifications.

Claims

1. A method of welding a crystallizer copper tube, characterized by, The method comprises: Collecting a historical welding data set of the crystallizer copper tube, performing correlation analysis on the historical welding data set of the crystallizer copper tube, and establishing a copper tube characteristic parameter-welding parameter-quality mapping relationship library; Based on the copper tube characteristic parameter-welding parameter-quality mapping relationship library, the target crystallizer copper tube is matched and analyzed for deviation, and the initial welding parameter is determined; Deploy a monitoring and sensing network in the welding area of the target crystallizer copper tube, the monitoring and sensing network integrates an infrared thermal imager, a laser displacement sensor and an acoustic emission detector, Based on the initial welding parameter, the target crystallizer copper tube is executed for welding control, and at the same time, the monitoring and sensing network collects a multi-modal welding data stream, identifies defects based on the multi-modal welding data stream, and obtains a copper tube welding defect parameter; Based on the copper tube welding defect parameter, the initial welding parameter is compensated and analyzed to determine the target welding parameter, and welding compensation control is performed through the target welding parameter; Wherein, the initial welding parameter comprises: Obtaining the copper tube welding quality standard and the characteristic parameter of the target crystallizer copper tube; According to the copper tube welding quality standard, the copper tube characteristic parameter-welding parameter-quality mapping relationship library is optimized to obtain a set of optional copper tube characteristic parameters-correlated welding parameters; Based on the characteristic parameters of the target crystallizer copper tube and the set of optional copper tube characteristic parameters-correlated welding parameters, similarity calculation is performed to obtain the most similar matching copper tube characteristic parameter-correlated welding parameter; The matching copper tube characteristic parameter-correlated welding parameter is analyzed for deviation adjustment to determine the initial welding parameter; Wherein, the initial welding parameter comprises: The characteristic parameters of the target crystallizer copper tube are calculated for deviation with the matching copper tube characteristic parameter-correlated welding parameter to determine the copper tube characteristic deviation parameter; Obtaining a set of copper tube characteristic deviation types, and according to the set of copper tube characteristic deviation types, a welding parameter adjustment strategy is associated; According to the set of copper tube characteristic deviation types, the welding parameter adjustment strategy is adjusted in amplitude based on the set of optional copper tube characteristic parameters-correlated welding parameters to construct a welding parameter optimization strategy library; Based on the welding parameter optimization strategy library, the copper tube characteristic deviation parameter is analyzed for deviation adjustment to determine the initial welding parameter; Wherein, the copper tube welding defect parameter comprises: Collecting a multi-modal welding defect data set of the target crystallizer copper tube, performing defect recognition training based on the multi-modal welding defect data set, and building a crystallizer copper tube defect recognition multi-channel; The multi-modal welding data stream is matched and mapped to the crystallizer copper tube defect recognition multi-channel for synchronous defect recognition, and a welding defect multi-modal feature set is output; The welding defect multi-modal feature set is weighted and fused and quantitatively analyzed to obtain a copper tube welding defect parameter; Wherein, the target welding parameter comprises: A welding defect-correlated welding parameter library is constructed, a PID controller is obtained based on compensation control fitting of the welding defect-correlated welding parameter library; The PID controller is used to compensate and analyze the copper pipe welding defect parameters, to determine the welding parameter compensation amount, and to compensate and correct the initial welding parameters based on the welding parameter compensation amount, to determine the target welding parameters.

2. The method of welding a crystallizer copper tube as claimed in claim 1, wherein, The mapping relationship library of the copper pipe characteristic parameters-welding parameters-quality is established, including: The historical welding data set of the crystallizer copper pipe is subjected to abnormal cleaning processing and missing value interpolation processing to obtain the available historical welding data set of the crystallizer copper pipe; According to the available historical welding data set of the crystallizer copper pipe, copper pipe characteristic dimension data, welding parameter dimension data and welding quality dimension data are extracted and obtained; The copper pipe characteristic dimension data, welding parameter dimension data and welding quality dimension data are subjected to correlation analysis and screening to obtain copper pipe characteristic data, associated welding parameters and welding quality data; The mapping relationship library of the copper pipe characteristic parameters-welding parameters-quality is established based on the copper pipe characteristic data, associated welding parameters and welding quality data.

3. The method of claim 1, wherein the welding is performed by using a laser beam. The crystallizer copper pipe defect recognition multi-channel is built, including: According to the monitoring and sensing network, a defect recognition multi-channel architecture is constructed, each recognition channel in the defect recognition multi-channel architecture corresponds to one sensor type in the monitoring and sensing network; According to the monitoring and sensing network, the multi-modal welding defect data set is subjected to modal classification identification and defect recognition training respectively to obtain a multi-modal defect recognition branch network set; The multi-modal defect recognition branch network set is matched and embedded into the defect recognition multi-channel architecture to build the crystallizer copper pipe defect recognition multi-channel.

4. The method of welding a crystallizer copper tube of claim 1, wherein, The welding defect multi-modal feature set is subjected to weighted fusion and quantitative analysis to obtain copper pipe welding defect parameters, including: According to the output accuracy of each channel in the crystallizer copper pipe defect recognition multi-channel, a multi-modal feature weight factor is determined; Based on the multi-modal feature weight factor, the welding defect multi-modal feature set is subjected to weighted fusion to obtain a welding defect fusion feature set; According to the welding defect type, the welding defect fusion feature set is subjected to defect degree quantitative analysis in turn to obtain the copper pipe welding defect parameters.

5. A welding system for crystallizer copper tubes, characterized by, The system for implementing the welding method of the crystallizer copper pipe according to any one of claims 1-4, the system comprising: An association analysis module: collecting a historical welding data set of a crystallizer copper pipe, and performing association analysis on the historical welding data set of the crystallizer copper pipe to establish a mapping relationship library of copper pipe characteristic parameters-welding parameters-quality; An initial parameter determination module: based on the mapping relationship library of copper pipe characteristic parameters-welding parameters-quality, matching analysis and deviation optimization are performed on a target crystallizer copper pipe to determine initial welding parameters; A monitoring network deployment module: deploying a monitoring and sensing network in a welding area of the target crystallizer copper pipe, the monitoring and sensing network integrating an infrared thermal imager, a laser displacement sensor and an acoustic emission detector; A defect recognition module: based on the initial welding parameters, welding control is performed on the target crystallizer copper pipe, and at the same time, multi-modal welding data streams are collected through the monitoring and sensing network, and defect recognition is performed on the multi-modal welding data streams to obtain copper pipe welding defect parameters; The compensation control module compensates and analyzes the initial welding parameters based on the copper tube welding defect parameters, determines target welding parameters, and performs welding compensation control through the target welding parameters.

6. A computer-readable storage medium having stored thereon a computer program, characterized in that, The program is executed by the processor to implement the welding method of the crystallizer copper tube as claimed in any one of claims 1-4.

Images