Steel member high-precision welding data analysis method and system

Abstract

The invention discloses a steel member high-precision welding data analysis method and system, and relates to the technical field of workpiece welding. The method comprises the steps that welding process parameters and welding state parameters are obtained; according to the welding state parameters and the welding target state, state parameter deviation is calculated; according to the state parameter deviation, a parameter adjusting instruction is generated, and the parameter adjusting instruction is used for correcting welding process parameters; associating the parameter adjustment instruction with batch information of the steel members; after association, analyzing the parameter adjustment instructions of the steel members in the same batch, and identifying the integral process deviation; and according to the integral process deviation, welding process parameter adjusting suggestions are generated. According to the method, the state parameter deviation and the parameter adjustment instruction can be combined to identify the integral process deviation, so that welding data analysis is realized, and the accuracy and the welding quality are improved.

Description

Technical Field

[0001] This invention relates to the field of workpiece welding technology, and in particular to a method and system for high-precision welding data analysis of steel components. Background Technology

[0002] In modern industrial production, automated welding production lines play a crucial role. These lines typically process various specialized high-strength steel plates, each with unique material properties. Existing systems adjust welding parameters by monitoring the welding process in real time. However, subtle differences may exist between different batches of raw materials, causing the welding system to continuously perform fixed parameter adjustments. While these adjustments superficially maintain the appearance of the weld, they can lead to inconsistencies in the internal structure of the metal, affecting its long-term strength and reliability. Existing systems struggle to perform different parameter adjustments based on varying differences, resulting in low data analysis accuracy and impacting weld quality.

[0003] In summary, the technical problems existing in the relevant technologies need to be improved. Summary of the Invention

[0004] The main objective of this invention is to propose a high-precision welding data analysis method and system for steel components, which can identify overall process deviations by combining state parameter deviations and parameter adjustment commands, thereby achieving welding data analysis and improving accuracy and welding quality.

[0005] On one hand, embodiments of the present invention provide a method for high-precision welding data analysis of steel components, including the following steps: The welding process parameters and welding status parameters are obtained. The welding process parameters include current, voltage, wire feed speed and actuator moving speed. The welding status parameters include molten pool shape, weld contour, arc sound and temperature distribution. Calculate the state parameter deviation based on the welding state parameters and the welding target state; Based on the deviation of the state parameters, a parameter adjustment instruction is generated, which is used to correct the welding process parameters; The parameter adjustment command is associated with the batch information of the steel components; After correlation, the parameter adjustment instructions of steel components in the same batch are analyzed to identify overall process deviations; Based on the overall process deviation, suggestions for adjusting welding process parameters are generated.

[0006] On the other hand, embodiments of the present invention provide a high-precision welding data analysis system for steel components, comprising: The parameter acquisition module is used to acquire welding process parameters and welding status parameters. The welding process parameters include current, voltage, wire feed speed and actuator moving speed. The welding status parameters include molten pool shape, weld contour, arc sound and temperature distribution. The parameter deviation calculation module is used to calculate the state parameter deviation based on the welding state parameters and the welding target state. The adjustment instruction generation module is used to generate parameter adjustment instructions based on the state parameter deviation, and the parameter adjustment instructions are used to correct the welding process parameters; The batch association module is used to associate the parameter adjustment command with the batch information of the steel components; The process deviation identification module is used to analyze the parameter adjustment instructions of steel components within the same batch after correlation, and to identify overall process deviations. The adjustment suggestion generation module is used to generate welding process parameter adjustment suggestions based on the overall process deviation.

[0007] The embodiments of this application include at least the following beneficial effects: First, the welding process parameters and welding state parameters are obtained. Based on the welding state parameters and the welding target state, the state parameter deviation is calculated. Then, based on the state parameter deviation, parameter adjustment instructions are generated, and the parameter adjustment instructions are associated with the batch information of the steel components. Next, the parameter adjustment instructions of the steel components in the same batch are analyzed to identify overall process deviations. Finally, based on the overall process deviations, welding process parameter adjustment suggestions are generated. Thus, it is possible to identify overall process deviations by combining state parameter deviations and parameter adjustment instructions, thereby realizing welding data analysis and improving accuracy and welding quality.

[0008] Other features and advantages of the invention will be set forth in the following description, and will be apparent in part from the description, or may be learned by practicing the invention. The objects and other advantages of the invention may be realized and obtained by means of the structures particularly pointed out in the description and the drawings. Attached Figure Description

[0009] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below.

[0010] Figure 1 This is a flowchart of a high-precision welding data analysis method for steel components according to an embodiment of the present invention; Figure 2 This is a schematic diagram of a high-precision welding data analysis system for steel components according to an embodiment of the present invention. Detailed Implementation

[0011] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments.

[0012] In related technologies, automated welding production lines play a crucial role in modern industrial production. These lines typically process various special high-strength steel plates, each with unique material properties. Existing systems adjust welding parameters by monitoring the welding process in real time. However, subtle differences may exist between different batches of raw materials, causing the welding system to continuously perform fixed parameter adjustments. While these adjustments superficially maintain the appearance of the weld, they can cause inconsistencies in the internal structure of the metal, thus affecting its long-term strength and reliability. Existing systems struggle to perform different parameter adjustments based on varying differences, resulting in low data analysis accuracy and impacting weld quality.

[0013] For example, in a manufacturing workshop for large bridges or marine engineering equipment, an automated steel component welding production line is in operation. This production line is not for the repetitive production of a single specification product, but rather a flexible line that needs to weld special high-strength steel plates of different grades and thicknesses according to order requirements. These steel plates will be used to construct key nodes in the load-bearing structure. On the production line, a multi-axis welding robot is the core equipment. It integrates vision sensors, laser contour scanners, arc sound collectors, and infrared thermal imagers to continuously acquire information about the arc state, molten pool morphology, weld geometry, and temperature field distribution during the welding process. This information, along with execution-level information such as welding current, arc voltage, wire feed speed, and welding speed fed back by the robot controller, is transmitted in real time to a central data processing unit. The goal of this unit is to analyze this massive amount of information and fine-tune the welding robot's movements and parameters in real time to ensure that each weld achieves the designed penetration depth and width, and is free from defects such as porosity, slag inclusions, and incomplete penetration, ultimately guaranteeing the strength and service life of the entire steel component.

[0014] However, the production line encountered a tricky situation when processing a batch of high-strength weathering steel plates. The supplier provided complete material certificates, and the chemical composition and mechanical properties were within the standard tolerance range. The production line followed the standard procedure, calling up the standard welding process parameters for this grade and thickness. However, during the actual welding process, the data analysis system frequently issued adjustment commands. Infrared thermal imager showed that, under the same initial parameters, the peak temperature field of the molten pool formed by this batch of steel plates was lower, and the high-temperature zone was smaller than before. To achieve the set penetration depth, the system had to continuously increase the welding current or slow down the welding speed. Because the heat input fluctuated repeatedly during welding to compensate, the uniformity of the internal metal structure of the weld and its affected area deteriorated, resulting in coarse grains in some areas. This directly led to a decrease in the impact toughness of the welded joint compared to the standard sample, resulting in low weld quality.

[0015] On flexible production lines, when welding special high-strength steel components of different grades and thicknesses with subtle batch-to-batch variations, existing systems can only passively and immediately correct parameter deviations during individual welding processes. Due to continuous parameter fluctuations, such corrections can lead to uneven internal structures in the welded joints, affecting their deeper strength and toughness. Therefore, it is necessary to effectively address accidental interference during individual welding processes and to analyze the welding process information of multiple consecutive workpieces within the same batch to identify overall process deviations caused by differences in raw material characteristics. Based on this identification, the initial welding process parameters for the entire batch should be proactively adjusted, thereby shifting from passive response to proactive adaptation and ultimately ensuring a high degree of consistency in both the appearance and internal properties of the weld.

[0016] The embodiments of this application will be explained in detail below with reference to the accompanying drawings: Figure 1 This is an optional flowchart of a high-precision welding data analysis method for steel components provided in an embodiment of this application. Figure 1 The method may include, but is not limited to, steps S101 to S106.

[0017] Step S101: Obtain welding process parameters and welding status parameters. Welding process parameters include current, voltage, wire feed speed and actuator moving speed. Welding status parameters include molten pool shape, weld contour, arc sound and temperature distribution. Step S102: Calculate the deviation of the state parameters based on the welding state parameters and the target welding state; Step S103: Generate parameter adjustment instructions based on the deviation of state parameters. The parameter adjustment instructions are used to correct the welding process parameters. Step S104: Associate the parameter adjustment command with the batch information of the steel component; Step S105: After association, analyze the parameter adjustment instructions of steel components in the same batch to identify overall process deviations. Step S106: Based on the overall process deviation, generate welding process parameter adjustment suggestions.

[0018] Steps S101 to S106 as shown in the embodiments of this application can identify overall process deviations by combining state parameter deviations and parameter adjustment instructions, so as to realize welding data analysis and improve accuracy and welding quality.

[0019] In some embodiments, steps S101-S106 may first involve acquiring welding process parameters and welding status parameters. Welding process parameters, including current, voltage, wire feed speed, and actuator movement speed, can be acquired by installing appropriate sensors on the welding equipment. Current and voltage can be acquired using clamp-on ammeters and voltage sensors; the wire feed speed can be calculated by measuring the rotational speed of the wire feed mechanism using an encoder or speed sensor; and the actuator movement speed can be acquired using a laser rangefinder or a vision tracking system. Welding status parameters, including molten pool shape, weld contour, arc sound, and temperature distribution, can be acquired using non-contact sensors. The molten pool shape and weld contour can be monitored using a high-speed camera combined with image processing technology; the arc sound can be acquired using a high-sensitivity microphone; and the temperature distribution can be obtained through non-contact measurement using an infrared thermal imager.

[0020] Then, based on the welding status parameters and the target welding status, the deviation of the status parameters is calculated. The welding status parameters acquired in real time, such as the weld pool shape, weld contour, arc sound, and temperature distribution, can be compared with the preset target welding status. This comparison can be performed using numerical differences, percentage deviations, or feature matching based on machine learning models. For example, if the target weld pool width is 10mm, but the actual measurement is 9.5mm, the deviation is -0.5mm.

[0021] Based on the deviation of the state parameters, parameter adjustment instructions are generated to correct the welding process parameters. These instructions can be generated using a preset control algorithm (such as a PID controller, fuzzy logic controller, or neural network controller) based on the calculated deviation of the state parameters. For example, if the molten pool temperature is too low, the system may generate instructions to increase the current or decrease the wire feed speed.

[0022] The parameter adjustment instructions are associated with the batch information of the steel components. In the production management system, a unique batch information can be assigned to each steel component, and after each welding process, the parameter adjustment instructions generated during that welding process are bound and stored with the corresponding steel component batch information. This can be achieved through a database management system, recording the parameter adjustment instructions as an attribute of the batch information.

[0023] After correlation, the parameter adjustment instructions for steel components within the same batch are analyzed to identify overall process deviations. Parameter adjustment instructions for all steel components in the same batch can be queried in the database, and statistical analysis can be performed on these instructions, such as calculating the average, variance, or trend line, to identify whether there is a persistent, systematic adjustment tendency, thereby determining whether there is an overall process deviation.

[0024] Finally, based on overall process deviations, the system generates recommendations for adjusting welding process parameters. If the analysis results show that the welding current of the same batch of steel components remains consistently high, the system will recommend a general reduction in the initial welding current setting for that batch of steel components. These recommendations can be output as text reports, charts, or by directly modifying process parameter templates.

[0025] This embodiment achieves refined control of welding quality through deep fusion and intelligent analysis of multi-source heterogeneous data during the welding process. First, the system acquires welding process parameters and welding status parameters in real time, which comprehensively reflect the dynamic characteristics and results of the welding process. Then, by comparing the real-time welding status parameters with preset welding target states, the system accurately calculates the deviation of the status parameters, quantifying the degree of deviation in the welding process. Based on these deviations, the system can generate immediate parameter adjustment instructions to fine-tune the welding process parameters to ensure the quality of each weld. Furthermore, these parameter adjustment instructions are associated with batch information of the steel components. Through this association, the system can accumulate and analyze the trend of parameter adjustment instructions for all steel components within the same batch. If a batch of steel components is found to continuously exhibit parameter adjustments in the same direction during the welding process, such as continuously increasing the current or decreasing the wire feed speed, this indicates that there may be an overall characteristic deviation in the raw materials of that batch, causing the initial process parameters to no longer be fully applicable. By identifying this overall process deviation, the system can generate welding process parameter adjustment suggestions for that batch of raw materials. These recommendations involve macroscopic and systematic corrections to initial process parameters, such as overall adjustments to the baseline values ​​of welding current, voltage, or wire feed speed, thereby fundamentally eliminating continuous fluctuations caused by batch differences in raw materials and ensuring the uniformity of the internal structure and the consistency of mechanical properties of the welded joint.

[0026] Through the above technical solution, this embodiment introduces batch information association and a holistic process deviation identification mechanism, enabling the system to identify slight drifts in the inherent properties of raw materials. This embodiment can make higher-level decisions, namely, proactively and holistically modifying the initial process parameters to form a temporary dedicated process for a specific batch of steel plates, thereby fundamentally eliminating continuous fluctuations in heat input and ensuring the stability of the internal structure and consistency of mechanical properties of the welded joint. This shift from passive correction to proactive optimization significantly improves the stability and reliability of welding quality, providing a more solid technical guarantee for the manufacture of high-precision steel components.

[0027] In some embodiments, step S105 involves analyzing the parameter adjustment instructions for steel components within the same batch to identify overall process deviations, which may include, but is not limited to, the following steps: Step S201: Calculate the cumulative integral term caused by the characteristics of the raw material batch according to the parameter adjustment instruction; Step S202: Calculate the cumulative average change and first average change trend of the integral term for all steel components in the same batch; Step S203: Perform an over-limit judgment on the average change and obtain the over-limit judgment result; Step S204: If the result of the over-limit judgment is that it exceeds the upper limit, then the overall process deviation is generated based on the average change and the first average change trend.

[0028] In some embodiments, due to the numerous factors affecting welding quality, batch characteristics of raw materials often lead to systematic and persistent process deviations. If only a general analysis is performed on parameter adjustment instructions, it may be difficult to accurately distinguish between the cumulative effects caused by batch characteristics of raw materials and deviations caused by other factors such as equipment wear and environmental changes, thereby affecting the accuracy of overall process deviation identification and the specificity of subsequent process adjustment recommendations.

[0029] Therefore, the cumulative integral term caused by the batch characteristics of raw materials can be calculated first based on the parameter adjustment instructions. During the welding process, parameter adjustment instructions are typically generated by a proportional-integral-derivative (PID) controller or other similar control algorithms to correct welding process parameters in real time. The cumulative integral term reflects the cumulative adjustment amount made by the system to eliminate persistent deviations. This embodiment analyzes the integral term in the parameter adjustment instructions and isolates the cumulative portion not caused by batch characteristics to calculate the cumulative integral term caused by the batch characteristics of raw materials, aiming to quantify the persistent impact of a specific batch of raw materials on welding process parameters.

[0030] Then, the cumulative average change and the first average trend of change of the integral term for all steel components within the same batch are calculated. The average change reflects the degree of overall and systematic deviation caused by the characteristics of the raw materials during the welding process of the steel components in that batch. The first average trend of change describes the pattern of this deviation as it changes with the sequence of workpieces or time, such as whether it continues to increase, decreases, or remains stable.

[0031] Next, an over-limit judgment is performed on the average change to obtain the over-limit judgment result. This over-limit judgment can be performed on the average change to determine whether it exceeds the preset allowable range. The over-limit judgment result indicates whether the deviation caused by the characteristics of the current batch of raw materials has reached a level requiring attention or intervention.

[0032] If the result of the exceedance judgment is that the limit is exceeded, an overall process deviation is generated based on the average change and the first average change trend. If the result of the exceedance judgment is that the limit is exceeded, it indicates that the characteristics of this batch of raw materials have a significant and unacceptable systematic impact on the welding process. In this case, an overall process deviation is generated based on the average change and the first average change trend. This overall process deviation not only indicates the existence of the deviation but also includes its degree and development trend, providing a crucial basis for subsequently generating targeted welding process parameter adjustment suggestions.

[0033] This embodiment addresses the issues of insufficient accuracy and lack of specificity in identifying overall process deviations by introducing an analysis of the cumulative integral term in parameter adjustment instructions and correlating it with the characteristics of raw material batches. Specifically, the cumulative integral term effectively captures persistent and systematic deviations caused by the characteristics of raw material batches, rather than instantaneous or random fluctuations. By calculating the average change and first average trend of the cumulative integral term for all steel components within the same batch, the overall impact of that batch of raw materials on the welding process can be quantified and tracked. When the average change exceeds a preset upper limit, combined with its trend, the overall process deviation caused by the characteristics of that batch of raw materials can be accurately identified, thus avoiding misjudging deviations caused by non-batch characteristics as batch characteristic problems or ignoring potential risks caused by batch characteristics. This mechanism makes the identified process deviations more root-cause, providing a clear direction for subsequent process adjustments.

[0034] To illustrate this technical solution more clearly, a specific example is used below. Suppose a steel component manufacturing company receives a batch of new steel plate raw materials and uses them to produce a batch of steel components labeled PA. During the welding process, the welding system continuously acquires welding process parameters and welding status parameters, and generates parameter adjustment instructions based on deviations in the status parameters. These parameter adjustment instructions are recorded in real time and associated with the batch information of PA. When analyzing the welding data of the steel components within batch PA, the system extracts integral term data from the parameter adjustment instructions of each steel component. Through further processing, such as removing the cumulative portion caused by non-batch characteristics like welding power fluctuations and shielding gas flow changes, the system calculates the cumulative integral term for each steel component caused by the characteristics of the batch PA raw materials. Subsequently, the system calculates the average change in the cumulative integral term of all steel components within batch PA. For example, it finds that the average value shows a continuous upward trend, indicating that the batch of raw materials may cause systematic deviations in weld penetration or weld formation. Simultaneously, the system calculates the first average trend, showing, for example, that the cumulative integral term value exhibits a linear increase within the batch.

[0035] Next, the system compares the average change with a preset upper limit. If the average change exceeds the upper limit, for example, by exceeding a 0.5% threshold set based on historical data and experience, the system determines that an over-limit situation exists. Based on this over-limit determination, the system combines the average change and the first average change trend to generate an overall process deviation report for the batch of PA. This report may indicate that the batch of PA raw materials has a high heat input sensitivity, requiring a continuous increase in current or a decrease in wire feed speed during welding to maintain a stable molten pool, and that this requirement tends to intensify in the later stages of the batch. Based on this overall process deviation, the system can further generate adjustment suggestions, such as adjusting the initial setting of the welding current for the batch of PA raw materials, or adjusting the compensation curve of the wire feed speed to adapt to its specific material characteristics, thereby ensuring welding quality.

[0036] Through the above technical solution, this embodiment significantly improves the accuracy and specificity of process deviation identification by focusing on the accumulation of integral terms and combining the judgment of exceeding limits of average change and trend. This avoids potential welding quality problems caused by batch characteristics of raw materials and provides more accurate and effective suggestions for subsequent welding process parameter adjustments, thereby improving the welding quality and production efficiency of steel components.

[0037] In some embodiments, step S201, calculating the cumulative integral term caused by the characteristics of the raw material batch according to the parameter adjustment instruction, may include, but is not limited to, the following steps: Extract the first numerical change of the integral term of the proportional-integral-derivative controller from the parameter adjustment command; The second numerical change of non-batch characteristic auxiliary parameters is obtained. These non-batch characteristic auxiliary parameters include the output voltage fluctuation of the welding power source, the instantaneous value of the shielding gas flow rate, the ambient temperature, the ambient humidity, and the sensor zero-point drift data. Identify the correlation between the first and second numerical changes; Based on the correlation, the cumulative part caused by non-batch characteristics is extracted from the first numerical change to obtain the cumulative integral term caused by the batch characteristics of raw materials.

[0038] In some embodiments, in addition to raw material batch characteristics, non-batch characteristic auxiliary parameters such as welding power supply output voltage fluctuations, instantaneous shielding gas flow rates, ambient temperature, ambient humidity, and sensor zero-point drift data may also affect the integral term in the parameter adjustment command. Failure to separate the cumulative effects of these non-batch characteristics may lead to inaccurate assessments of raw material batch characteristics, thereby affecting the accuracy of overall process deviation identification and potentially generating inaccurate welding process parameter adjustment recommendations.

[0039] Therefore, the first numerical change of the integral term of the proportional-integral-derivative (PID) controller can be extracted from the parameter adjustment instructions. This first numerical change refers to the cumulative adjustment of process parameters by the PID controller during welding to eliminate steady-state error. This first numerical change reflects the continuous correction made by the system to maintain the target state, which may include the cumulative effect caused by both batch and non-batch characteristics of the raw materials.

[0040] Then, the second numerical change of the non-batch characteristic auxiliary parameters is obtained. This second numerical change refers to auxiliary factors that are unrelated to the batch of steel component raw materials but affect the stability of the welding process and parameter adjustment commands. Non-batch characteristic auxiliary parameters include the output voltage fluctuation of the welding power source, the instantaneous value of the shielding gas flow rate, ambient temperature, ambient humidity, and sensor zero-point drift data. Changes in these parameters directly or indirectly affect the stability of the welding process, which is then reflected in the integral term of the PID controller.

[0041] Next, identify the correlation between the changes in the first and second numerical values. Statistical methods (such as correlation coefficient analysis and regression analysis) or machine learning algorithms can be used to establish a mathematical model or association rule between the changes in the first and second numerical values. The aim is to quantify the degree and pattern of influence of non-batch characteristic auxiliary parameters on the integral term of the PID controller.

[0042] Finally, based on the correlation, the cumulative portion caused by non-batch characteristics is extracted from the change in the first value, yielding the cumulative integral term caused by the batch characteristics of the raw materials. Using the identified correlation model, the contribution of a given change in the auxiliary parameter of non-batch characteristics to the change in the first value can be predicted, and this contribution can be subtracted from the change in the first value to obtain the cumulative integral term caused solely by the batch characteristics of the raw materials.

[0043] This embodiment addresses the issue of potential interference from various non-batch factors in the accumulation of integral terms by meticulously analyzing the integral term in parameter adjustment instructions and introducing non-batch characteristic auxiliary parameters for correction. Specifically, it first extracts the first numerical change of the PID controller's integral term, which contains all accumulated correction information. Then, it acquires and analyzes the second numerical change of non-batch characteristic auxiliary parameters that are unrelated to the raw material batch but affect the welding process. By identifying the correlation between the first and second numerical changes, this embodiment can quantify and distinguish the portion of the integral term accumulation caused by non-batch characteristics. Finally, this portion of accumulation caused by non-batch characteristics is separated from the total first numerical change, ensuring that the obtained integral term accumulation more accurately and purely reflects the impact of raw material batch characteristics on the welding process, providing a more reliable data foundation for subsequent identification of overall process deviations.

[0044] To illustrate this technical solution more clearly, a specific example is used below. Suppose that during the welding process of a batch of steel components, the parameter adjustment command monitors a continuous, accumulating trend in the first value change of the PID controller's integral term. This embodiment obtains non-batch characteristic auxiliary parameters during the welding process, such as observing continuous low-amplitude fluctuations in the welding power supply's output voltage over a certain period, and the ambient temperature being slightly higher than normal. Analysis reveals a significant positive correlation between these voltage fluctuations and the increase in ambient temperature and the accumulation of the PID integral term. Specifically, when voltage fluctuations and ambient temperature increase, the PID controller tends to increase the integral term to compensate for these external disturbances.

[0045] This embodiment utilizes statistical methods (such as correlation coefficient analysis and regression analysis) to calculate the cumulative integral term caused by these non-batch characteristics (voltage fluctuations and ambient temperature). For example, through a regression analysis model, it is determined that for every 0.1V increase in voltage fluctuation, the cumulative integral term increases by X units; for every 1°C increase in ambient temperature, the cumulative integral term increases by Y units. After separating these cumulative components caused by non-batch characteristics from the total first numerical change, it is found that the remaining cumulative integral term caused by raw material batch characteristics is actually much smaller than the initially observed total cumulative term, and may even be within the normal fluctuation range. Therefore, this embodiment avoids misjudging process deviations caused by external environment or equipment fluctuations as raw material batch problems, thereby enabling a more accurate determination of whether there are indeed problems with the raw material characteristics of the batch of steel components, and generating more precise welding process parameter adjustment suggestions accordingly. For example, if the cumulative component caused by raw material characteristics is found to be very small after separation, it may be recommended to check the power supply stability or environmental control system, rather than directly adjusting the welding parameters for the raw material.

[0046] Through the above technical solution, this embodiment can effectively eliminate the interference of non-batch characteristic auxiliary parameters on the cumulative calculation of integral terms, significantly improving the calculation accuracy of the cumulative integral terms caused by raw material batch characteristics. Therefore, when analyzing parameter adjustment instructions for steel components within the same batch, it can more accurately identify the overall process deviation truly caused by raw material batch characteristics, avoiding misjudging non-batch factors as raw material problems. This accurate deviation identification helps generate more targeted and effective welding process parameter adjustment suggestions, thereby optimizing the welding process, improving the welding quality and production efficiency of steel components, and reducing resource waste and production costs caused by misjudgment.

[0047] In some embodiments, step S203, which involves judging whether the average change exceeds the limit and obtaining the result of the judgment, may include, but is not limited to, the following steps: Obtain the material grade and thickness of the steel components; Retrieve the heat input sensitivity index corresponding to the material grade and thickness from the material property database; Determine the upper limit of variation based on the heat input sensitivity index; The average change and the upper limit of change are compared to obtain the result of the over-limit judgment.

[0048] In some embodiments, the material grade and thickness of the steel component can be obtained first. This information is typically available through a production management system or a bill of materials. Then, the heat input sensitivity index corresponding to the material grade and thickness is retrieved from a material property database. The heat input sensitivity index refers to the degree or trend of change in the internal microstructure, mechanical properties, or macroscopic deformation of a specific material under different welding heat input conditions. For example, some high-strength steels may be more sensitive to changes in heat input, while ordinary carbon steel may be relatively insensitive. This index can be a numerical value, a range, or a functional relationship.

[0049] Next, based on the heat input sensitivity index, an upper limit for variation is determined. This upper limit is not a fixed value but is adjusted according to the material's sensitivity to heat input. For example, for highly sensitive materials, the upper limit may be set narrower to identify potential process deviations earlier. Finally, the average variation is compared with the upper limit to obtain the result of exceeding the limit judgment. This results in a more accurate judgment of exceeding the limit.

[0050] This embodiment introduces the material grade and thickness information of steel components, and combines it with the heat input sensitivity index in the material property database, enabling the upper limit of variation used for judging the average change to be adaptively adjusted according to the specific material properties. Since different materials have different response mechanisms to welding heat input, a fixed upper limit judgment method may not be able to comprehensively cover all situations. This embodiment obtains the material grade and thickness, and queries the corresponding heat input sensitivity index to quantify the material's behavioral characteristics during the welding process. Based on this, the upper limit of variation is dynamically determined according to the sensitivity index, ensuring that the threshold for judging the exceedance matches the actual physical properties of the material. This mechanism allows the judgment result to more accurately reflect whether the cumulative integral term caused by the batch characteristics of raw materials truly exceeds the reasonable range that the material can withstand, thereby avoiding misjudgments or omissions due to material differences.

[0051] To illustrate this technical solution more clearly, a specific example is used below. Assume there are two different batches of steel components: batch PB is made of Q235 steel with a thickness of 10mm; batch PC is made of Q345 steel with a thickness of 20mm. During welding data analysis, the system first obtains the material grade Q235 and thickness 10mm for batch PB, and then queries the material property database to find the heat input sensitivity index V corresponding to the 10mm thickness of Q235 steel. Based on index V, the system determines the upper limit of the average variation of batch PB as L1. Simultaneously, for batch PC, the system obtains its material grade Q345 and thickness 20mm, and queries the heat input sensitivity index K corresponding to the 20mm thickness of Q345 steel. Since Q345 steel is generally more sensitive to heat input, index K may be higher than index V. Therefore, the system determines the upper limit of the average variation of batch PC as L2 based on index K, where L2 may be less than L1. Subsequently, the cumulative average variation of the integral terms for each batch PB and batch PC is compared with L1 and L2, respectively. For example, if the average variation of batch PB is 0.8L1 and the average variation of batch PC is 0.8L2, although both are proportional to their respective upper limits, L1 and L2 are dynamically determined based on their material properties. Therefore, it is possible to more accurately determine which batch has a more significant process deviation, or whether both are within acceptable limits. This adaptive judgment mechanism based on material properties ensures higher accuracy and applicability in identifying process deviations in steel components made of different materials.

[0052] Through the above technical solution, this embodiment can significantly improve the accuracy and reliability of identifying overall process deviations when analyzing parameter adjustment instructions for steel components within the same batch. Since the upper limit of variation is dynamically determined based on the material grade, thickness, and corresponding heat input sensitivity index of the steel component, the result of exceeding the limit judgment can more accurately reflect the true process deviations faced by a specific material batch. This avoids false alarms or missed alarms caused by using general or inaccurate thresholds, making the subsequent welding process parameter adjustment suggestions more targeted and effective, thereby effectively improving the welding quality and production efficiency of steel components.

[0053] In some embodiments, in step S202, after calculating the cumulative average change and first average trend of the integral terms of all steel components in the same batch, the method may also include, but is not limited to, the following steps: Step S301: Determine the workpiece quantity window based on the number of steel components; Step S302: Based on the workpiece quantity window, perform trend analysis on the cumulative integral term to obtain the second average change trend; Step S303: Validate the confidence level of the second average trend and obtain the confidence level validation result; Step S304: If the confidence verification result meets the confidence requirement, then update the first average change trend according to the second average change trend.

[0054] In some embodiments, relying solely on a single average trend may not adequately capture the dynamics and locality of process variations within a batch, especially when the batch size is large or initial data is volatile, which may result in insufficiently accurate or timely identification of overall process deviations.

[0055] To this end, a workpiece quantity window can be determined first based on the number of steel components. A data range for trend analysis can be dynamically set based on the number of steel components that have been welded in the current batch. This workpiece quantity window can be a fixed-size sliding window, for example, always analyzing data from the most recent N steel components; or it can be a window determined proportionally to the total number of batches, for example, analyzing the last R% of steel components completed in the current batch. The aim is to focus on welding data in the recent period or within a specific range to capture more timely or localized process change trends.

[0056] Then, based on the workpiece quantity window, trend analysis is performed on the cumulative integral term to obtain the second average trend. Within the workpiece quantity window, statistical or machine learning methods can be used to analyze the cumulative integral term data to identify its patterns of change over time or workpiece sequence. For example, linear regression, exponential smoothing, and moving average methods can be used to calculate the average trend within this window. This second average trend is a trend representation that is more sensitive to local or recent data, complementing the first average trend calculated based on the entire batch of data.

[0057] The second average trend is then validated with confidence, yielding the validation results. This allows for the assessment of the reliability and stability of the second average trend. Validation is achieved by comparing the results with a pre-defined confidence level, which can be determined based on historical data, expert experience, or welding quality sensitivity indicators in a material property database. The validation process aims to ensure that the identified local trends are not caused by random noise or outliers, but rather represent statistically significant true trends.

[0058] If the confidence level verification result meets the confidence level requirement, the first average trend is updated based on the second average trend. This means that when a local trend is confirmed to be reliable, it is incorporated into the judgment of the overall trend to correct or refine the previously calculated first average trend. The update method can be to directly replace the first average trend with the second average trend, or to combine the two through weighted averaging or other methods, so that the first average trend can reflect the actual process status of the current batch more timely and accurately.

[0059] This embodiment effectively solves the problem of inaccurate identification that may result from relying solely on a single overall trend by introducing a workpiece quantity window, a second average trend, and a confidence verification mechanism. Specifically, the workpiece quantity window allows trend analysis to focus on more representative or timely data segments within a batch, thereby enabling more sensitive detection of subtle changes or local drifts in process parameters. Consequently, the calculated second average trend provides a more sensitive perspective on recent process conditions. Furthermore, confidence verification of the second average trend ensures the reliability of the identified local trends, avoiding misjudgments caused by data noise or random fluctuations. When the second average trend is verified as reliable, it is used to update the first average trend, ensuring that the overall process deviation identification considers not only the global characteristics of the batch but also incorporates the latest, verified local change information, making the identification of overall process deviations more dynamic and accurate.

[0060] To illustrate this technical solution more clearly, a specific example is used below. Assume a batch contains 100 steel components. After analyzing the cumulative integral terms of the first 20 steel components and calculating the first average trend, the system dynamically determines the number of components within a window. For example, a sliding window of size 10 can be set. When the welding data for the 21st steel component is acquired, the system calculates a second average trend based on the cumulative integral terms of the 12th to 21st steel components. Subsequently, the confidence level of this second average trend is verified, for example, by comparing it with a preset confidence level requirement (such as the significance level of the trend slope). If the verification result shows that the second average trend meets the confidence level requirement, the system updates the previously calculated first average trend based on the second average trend. For example, a weighted average method can be used, giving the second average trend a higher weight, allowing it to have a greater impact on the first average trend. As subsequent steel components are welded, the process continues, the workpiece quantity window slides continuously, the second average trend is continuously calculated and verified, and used to dynamically update the first average trend, thereby ensuring that the identification of overall process deviations is always based on the latest and most reliable data.

[0061] Through the above technical solution, this embodiment can significantly improve the accuracy and robustness of identifying overall process deviations in high-precision welding data analysis of steel components. By dynamically adjusting the workpiece quantity window and calculating the second average change trend, the system can respond more promptly to local changes in process parameters within a batch, avoiding lag or bias in trend judgment caused by initial data fluctuations or excessively large batch sizes. Furthermore, the introduction of a confidence verification mechanism effectively filters out unreliable trend information, ensuring that the data used to update the first average change trend has high credibility. This makes the final welding process parameter adjustment recommendations more instructive and effective, further optimizing welding quality and production efficiency.

[0062] In some embodiments, step S303, which involves performing a confidence verification on the second average change trend to obtain a confidence verification result, may include, but is not limited to, the following steps: Step S401: Obtain the material grade and thickness of the steel component; Step S402: Obtain the welding quality sensitivity index corresponding to the material grade and thickness from the material property database; Step S403: Determine the confidence level requirement based on the welding quality sensitivity index; Step S404: Compare the second average trend with the confidence level requirement to obtain the confidence level verification result.

[0063] In some embodiments, if the confidence level requirement fails to adequately consider the material properties of the steel component itself, the accuracy of the confidence level verification results may be insufficient, thereby affecting the reliability of subsequent process parameter adjustment recommendations. For example, steel components of different material grades and thicknesses exhibit significantly different sensitivities to parameters such as heat input and cooling rate during the welding process. If a uniform confidence level requirement is adopted, it may be impossible to effectively identify potential process deviations in specific material batches.

[0064] To do this, the material grade and thickness of the steel components can be obtained first. When analyzing welding data, the specific material properties of the current batch of steel components must first be determined. The material grade could be, for example, Q235, Q345, S355, etc., and the thickness refers to the actual dimensions of the steel component. This information forms the basis for subsequent assessments of weld quality sensitivity. Then, the weld quality sensitivity indices corresponding to the material grade and thickness can be obtained from a material property database. A pre-established database can be queried based on the known material grade and thickness of the steel components. This database stores a large amount of welding performance data for different materials at different thicknesses, including but not limited to quantitative indicators such as crack sensitivity, deformation sensitivity, penetration consistency, and weld formation. These indicators reflect the degree to which a specific material responds to changes in process parameters during welding and their impact on the final weld quality.

[0065] Next, based on the welding quality sensitivity index, the confidence level requirement is determined. Different confidence level thresholds can be set according to the material's sensitivity to parameters such as heat input and cooling rate. For materials with high welding quality sensitivity, the confidence level requirement will be more stringent; that is, the second average trend needs to have a smaller fluctuation range to be considered to meet the confidence level requirement. Conversely, for materials with low welding quality sensitivity, the confidence level requirement can be appropriately relaxed. The purpose is to ensure that the confidence level requirement accurately reflects the actual welding characteristics of the current steel component material.

[0066] Finally, the second average trend is compared with the confidence level requirement to obtain the confidence level verification result. The second average trend obtained through trend analysis can be compared with the confidence level requirement determined based on material properties. For example, it can be determined whether the fluctuation range of the second average trend or its deviation from zero is within the allowable range of the confidence level requirement. If the trend meets the confidence level requirement, the verification result is that the confidence level requirement is met; otherwise, the verification result is that the confidence level requirement is not met.

[0067] This embodiment addresses the problem that confidence requirements might be too general and fail to accurately reflect the welding characteristics of specific materials by introducing the material grade and thickness of the steel components and combining them with welding quality sensitivity indicators from a material property database to determine the confidence level requirements. Since different materials respond differently to changes in process parameters during welding, their welding quality stability also varies. By acquiring the material grade and thickness, the system can extract welding quality sensitivity indicators from the database that highly match the current material properties. These indicators quantify the material's tolerance to various process deviations during welding and their impact on the final quality. Based on this, the confidence level requirements are dynamically determined according to these sensitivity indicators, making the confidence level verification process more targeted and accurate. This material property-based confidence level requirement can more realistically reflect the welding quality risk of the current batch of steel components, thus making the verification results of the second average trend more reliable and providing stronger data support for whether to update the first average trend subsequently.

[0068] To illustrate this technical solution more clearly, a specific example is used below. Suppose there is a batch of steel components made of Q345 material with a thickness of 20mm. During confidence level verification, the system first obtains this information. Then, the system queries the material property database and finds that, for Q345 material at a thickness of 20mm, its weld quality sensitivity index shows that it is highly sensitive to fluctuations in heat input and changes in cooling rate, making it prone to cold cracking. Based on these sensitivity indicators, the system determines a relatively strict confidence level requirement. For example, it requires that the fluctuation range of the second average trend must be less than a certain preset small threshold, and its deviation from zero cannot exceed another extremely small threshold.

[0069] At this point, if the calculated second average trend shows large fluctuations or significant deviations, even if it might be acceptable under a general confidence level standard, it will be judged as not meeting the confidence requirements under this strict, material property-based confidence level requirement. Therefore, the system will not update the first average trend, but may trigger further analysis or alerts, indicating a potential problem in the welding process of the current batch of steel components that is inconsistent with the properties of Q345 material. Conversely, if the second average trend still shows good stability under strict confidence requirements, it can be considered reliable, and the first average trend can be updated accordingly. This approach ensures the accuracy of the confidence verification results, avoids misjudgments caused by differences in material properties, and thus improves the level of refinement in welding data analysis.

[0070] Through the above technical solution, this embodiment achieves more accurate and reliable confidence verification of the second average variation trend. Since the confidence level requirement is dynamically determined based on the material grade and thickness of the steel component and its corresponding welding quality sensitivity index, it effectively avoids misjudgments or omissions caused by using a general confidence level standard. This significantly improves the accuracy of identifying overall process deviations, ensures the effectiveness of subsequent welding process parameter adjustment suggestions, and thus further improves the welding quality and production efficiency of the steel components.

[0071] In some embodiments, step S404, comparing the second average trend of change with the confidence level requirement to obtain the confidence level verification result, may include, but is not limited to, the following steps: Step S501: Obtain the material sensitivity index corresponding to the material grade and thickness from the material property database. The material sensitivity index includes the impact toughness sensitivity index and the fatigue life sensitivity index. Step S502: Calculate the first comprehensive risk assessment value of the current material batch based on the material sensitivity index and the preset weight allocation rules; Step S503: Update the confidence level requirements based on the comprehensive risk assessment value of the first material; Step S504: Compare the second average trend of change with the updated confidence level requirement to obtain the confidence level verification result.

[0072] In some embodiments, relying solely on weld quality sensitivity indicators to determine confidence requirements may not fully reflect the intrinsic properties of the material and its risks under specific welding conditions. For example, some materials may exhibit high sensitivity to impact toughness or fatigue life, and if these deeper material properties are not adequately considered, the determined confidence requirements may be inaccurate or fail to fully cover potential quality risks. Failure to address these issues may compromise the accuracy and reliability of confidence verification results, thereby affecting the identification of overall process deviations and the generation of welding process parameter adjustment recommendations.

[0073] Therefore, material sensitivity indices corresponding to material grades and thicknesses can be obtained from material property databases. Material sensitivity indices are quantitative parameters reflecting changes in material performance under specific stress or environmental conditions. These indices include impact toughness sensitivity indices and fatigue life sensitivity indices. Impact toughness sensitivity indices measure a material's ability to resist impact loads and its sensitivity to factors such as welding heat input, while fatigue life sensitivity indices assess a material's durability under cyclic loading and its sensitivity to welding defects or stress concentration. These indices can be obtained from material property databases, which typically contain detailed mechanical property data and failure mode data for different material grades and thicknesses.

[0074] Then, based on the material sensitivity indicators and preset weighting rules, the first comprehensive material risk assessment value for the current material batch is calculated. The preset weighting rules refer to the relative importance coefficients assigned to different material sensitivity indicators (such as impact toughness sensitivity and fatigue life sensitivity) when calculating the first comprehensive material risk assessment value. For example, in applications with high structural safety requirements, the fatigue life sensitivity indicator may be given a higher weight. These weights can be set according to specific application needs, industry standards, or expert experience. The first comprehensive material risk assessment value is an indicator that quantifies the comprehensive risks that the current material batch may face during the welding process. This assessment value is calculated by combining multiple material sensitivity indicators and considering their relative importance (i.e., the preset weighting rules), aiming to provide a comprehensive risk view.

[0075] The confidence level requirement is then updated based on the first material comprehensive risk assessment value. The confidence level requirement determined by the welding quality sensitivity index can be adjusted based on the calculated first material comprehensive risk assessment value. If the first material comprehensive risk assessment value is high, indicating a high inherent risk in the current material batch, the confidence level requirement may be increased accordingly to ensure stricter verification standards; conversely, if the risk assessment value is low, the confidence level requirement may be appropriately relaxed. Finally, the second average trend and the updated confidence level requirement are compared to obtain the confidence level verification result.

[0076] This embodiment introduces material sensitivity indices, such as impact toughness sensitivity and fatigue life sensitivity, to gain a deeper understanding of the intrinsic properties of materials and their potential impact on weld quality. These indices are closely related to the mechanical properties and long-term reliability of materials, compensating for the shortcomings of relying solely on weld quality sensitivity indices. By combining these material sensitivity indices with preset weighting rules, a first comprehensive material risk assessment value for the current material batch can be calculated. This assessment value comprehensively considers the material's sensitivity under different failure modes, thus providing a more comprehensive and accurate material risk profile. Based on this first comprehensive material risk assessment value, the confidence level requirement is dynamically updated, enabling it to more accurately reflect the actual risk level of the current material batch. For example, when a material batch exhibits high impact toughness sensitivity or fatigue life sensitivity, the confidence level requirement will be increased accordingly, prompting the system to conduct more rigorous verification of the second average trend. Therefore, this embodiment ensures that the confidence verification process not only focuses on the apparent quality of the welding process but also considers the inherent risks of the material itself, making the verification results more representative and reliable.

[0077] To illustrate this technical solution more clearly, a specific example is used below. Assume a batch of steel components uses Q345B material with a thickness of 20mm. First, retrieve the material sensitivity index corresponding to this material grade and thickness from the material property database. For example, the impact toughness sensitivity index might be 0.7 (indicating a significant impact on impact toughness), and the fatigue life sensitivity index might be 0.8 (indicating a significant impact on fatigue life). Simultaneously, a preset weighting rule sets the weight of the impact toughness sensitivity index to 0.4 and the fatigue life sensitivity index to 0.6. Based on this, the first comprehensive material risk assessment value for the current batch can be calculated. For example, the first comprehensive material risk assessment value = (impact toughness sensitivity index × weight) + (fatigue life sensitivity index × weight) = (0.7 × 0.4) + (0.8 × 0.6) = 0.28 + 0.48 = 0.76. Assume the initial confidence level requirement determined based on the welding quality sensitivity index is 0.85. Since the calculated first material comprehensive risk assessment value of 0.76 indicates that the batch of materials has a certain inherent risk, the system updates the confidence requirement based on this assessment value. For example, an update rule can be set: if the first material comprehensive risk assessment value is higher than a certain threshold (e.g., 0.7), the confidence requirement is increased by 0.05. Therefore, the updated confidence requirement becomes 0.85 + 0.05 = 0.90. Subsequently, the calculated second average trend (e.g., 0.88) is compared with the updated confidence requirement of 0.90. Since 0.88 is less than 0.90, the confidence verification result is that the confidence requirement is not met. This indicates that although the second average trend may be close to the initial confidence requirement, considering the inherent sensitivity of the material, the trend is still insufficient to be considered completely reliable, thus prompting the system to take further measures or issue a warning. In this way, this embodiment can more comprehensively and accurately assess the stability of the welding process and avoid potential quality problems caused by ignoring the inherent risks of the materials.

[0078] Through the above technical solution, this embodiment introduces material sensitivity indices such as impact toughness sensitivity and fatigue life sensitivity, and calculates the first material comprehensive risk assessment value by combining them with preset weight allocation rules. This allows for dynamic and precise updates to the confidence level requirements. This enables the confidence level verification process to more comprehensively reflect the actual risk level of the current material batch, thereby significantly improving the accuracy and reliability of the confidence level verification results. Therefore, this embodiment can more effectively identify overall process deviations caused by the characteristics of raw material batches, providing a solid foundation for generating more accurate and effective welding process parameter adjustment suggestions, ultimately contributing to improving the welding quality and reliability of steel components.

[0079] In some embodiments, updating the confidence level requirement based on the first material comprehensive risk assessment value may include, but is not limited to, the following steps: Step S601: Obtain the second material comprehensive risk assessment value corresponding to each preceding material batch in the preceding material batch sequence; Step S602: Smooth the comprehensive risk assessment value of the first material and multiple comprehensive risk assessment values ​​of the second material to obtain a smoothed risk assessment value; Step S603: Update the confidence level requirement based on the smoothed risk assessment value.

[0080] In some embodiments, a second comprehensive risk assessment value for each preceding material batch in the preceding material batch sequence can be obtained first. Comprehensive risk assessment data related to a series of material batches preceding the current material batch can be retrieved and extracted from historical data. These preceding material batches can be arranged in chronological order or filtered based on other relevance (e.g., supplier, production line, etc.). The second comprehensive risk assessment value has the same calculation logic and assessment dimensions as the first comprehensive risk assessment value, but it targets historical batch data.

[0081] Then, the comprehensive risk assessment value of the first material and multiple comprehensive risk assessment values ​​of the second material are smoothed to obtain a smoothed risk assessment value. Statistical or signal processing methods can be used to perform weighted averaging, moving averages, exponential smoothing, etc., on the assessment values ​​of the current batch and historical batches to eliminate random fluctuations and noise in the data and highlight potential trends or patterns. The goal is to obtain a more stable assessment value that better represents the overall risk level of the material batch. For example, the exponentially weighted moving average (EWMA) method can be used, assigning higher weights to recent batches and lower weights to earlier batches, thereby reflecting the latest situation while maintaining a certain degree of historical stability.

[0082] Then, the confidence level requirement is updated based on the smoothed risk assessment value. The threshold or standard required for confidence verification of the second average trend can be dynamically adjusted based on the smoothed risk assessment value. For example, when the smoothed risk assessment value is high, it indicates a higher overall risk for the material batch; in this case, the confidence level requirement can be appropriately increased to ensure that the validity of the second average trend can only be confirmed under more stringent conditions. Conversely, when the smoothed risk assessment value is low, the confidence level requirement can be appropriately relaxed. The aim is to make the confidence level requirement more accurately and stably reflect the actual risk level of the material batch, thereby improving the reliability of welding process adjustments.

[0083] This embodiment introduces a second comprehensive risk assessment value for each preceding material batch in the preceding material batch sequence, and smooths the first comprehensive risk assessment value for the current batch against these historical assessment values, thereby obtaining a more stable and representative smoothed risk assessment value. Because this smoothed risk assessment value comprehensively considers historical data, it effectively suppresses potential random fluctuations or measurement errors in single-batch assessment values, enabling the confidence requirement updated based on this smoothed risk assessment value to more accurately reflect the long-term characteristics and true risk level of the material batch. This avoids frequent fluctuations in confidence requirements caused by anomalies in single-batch data, providing a more solid and reliable foundation for subsequent confidence verification of the second average trend.

[0084] To illustrate this technical solution more clearly, a specific example is used below. Assume the steel component batch currently being processed is batch PN, and its calculated first comprehensive material risk assessment value is 0.75. In the material property database, the system stores the sequence of preceding material batches, such as batch PN-1, batch PN-2, and batch PN-3, whose corresponding second comprehensive material risk assessment values ​​are 0.70, 0.68, and 0.72, respectively. To update the confidence level requirement, the second comprehensive material risk assessment values ​​of these preceding batches are first obtained.

[0085] Next, the comprehensive risk assessment value of the first material in the current batch PN (0.75) and the comprehensive risk assessment values ​​of the second material in the previous batches PN-1, PN-2, and PN-3 (0.70, 0.68, and 0.72) are smoothed. For example, a simple moving average method or a more complex exponentially weighted moving average method can be used. If a simple weighted average is used, for example, with a weight of 0.4 for the current batch and an average weight of 0.6 for the previous batches, the smoothed risk assessment value can be calculated as (0.75 * 0.4) + ((0.70 + 0.68 + 0.72) / 3 * 0.6) = 0.3 + (0.70 * 0.6) = 0.3 + 0.42 = 0.72.

[0086] Finally, the confidence requirement is updated based on this smoothed risk assessment value of 0.72. For example, if the preset rule is that the confidence requirement increases by 5% when the smoothed risk assessment value is higher than 0.70, then the new confidence requirement will be 5% higher than the original value. In this way, the update of the confidence requirement not only considers the current batch but also incorporates the experience of historical batches, making the update results more stable and representative, thereby improving the accuracy and reliability of welding process parameter adjustments.

[0087] Through the above technical solution, this embodiment smooths the comprehensive risk assessment values ​​of the current batch and previous batches, resulting in smoothed risk assessment values ​​that more comprehensively and stably reflect the overall risk status of the material batch. This method makes the update of confidence requirements more robust and reliable, avoiding excessive influence of short-term fluctuations or abnormal data on confidence requirements, thereby improving the accuracy and reliability of the second average trend confidence verification. Ultimately, this helps generate more stable and effective welding process parameter adjustment suggestions, further improving the quality and efficiency of high-precision welding of steel components.

[0088] In some embodiments, step S602 involves smoothing the first material comprehensive risk assessment value and multiple second material comprehensive risk assessment values ​​to obtain a smoothed risk assessment value. This process may include, but is not limited to, the following steps: Obtain the first welding environment parameter in the current material batch and the second welding environment parameter corresponding to each preceding material batch in the preceding material batch sequence; Select one batch from the preceding material batch sequence as the target material batch; Calculate parameter similarity based on the first welding environment parameters and the second welding environment parameters corresponding to the target material batch; Based on parameter similarity, the target material batches are weighted to obtain batch weights; Based on the batch weight corresponding to each preceding material batch in the preceding material batch sequence, the comprehensive risk assessment value of the first material and the comprehensive risk assessment values ​​of multiple second materials are smoothed to obtain a smoothed risk assessment value.

[0089] In some embodiments, different batches of material may be processed under different welding environmental parameters, such as ambient temperature, humidity, and welding equipment status. If these risk assessment values ​​are simply smoothed without fully considering the differences in these environmental parameters, the smoothing results may fail to accurately reflect the true risk level of each batch of material, thereby affecting the accuracy and reliability of the confidence level update.

[0090] To achieve this, we can first obtain the first welding environment parameters for the current material batch and the second welding environment parameters for each preceding material batch in the sequence. The first welding environment parameters refer to the environmental conditions and equipment parameters of the current material batch during the welding process, such as ambient temperature, ambient humidity, the actual output power of the welding power source, the shielding gas flow rate, and the welding torch posture. The second welding environment parameters refer to the environmental conditions and equipment parameters corresponding to each preceding material batch in its respective welding process. These parameters are used to quantify the external influencing factors during the welding of different batches.

[0091] Then, a batch is selected from the preceding material batch sequence as the target material batch. When calculating similarity, the correlation between each preceding batch and the current batch can be evaluated one by one. Parameter similarity is calculated based on the first welding environment parameter and the second welding environment parameter corresponding to the target material batch. Parameter similarity can be obtained by comparing the first welding environment parameter of the current material batch with the second welding environment parameter corresponding to the target material batch, quantifying their similarity in welding environment. For example, methods such as Euclidean distance, cosine similarity, or Pearson correlation coefficient can be used for calculation, with the aim of identifying the preceding batch whose welding environment is most similar to the current batch.

[0092] Next, based on parameter similarity, batches of target materials are weighted to obtain batch weights. This means that batches with higher similarity have a greater influence on the smoothing process. For example, a weighting function can be designed so that higher similarity values ​​result in higher weights, and vice versa. The purpose is to ensure that the smoothing results more accurately reflect the risk assessment information of batches with similar welding environments to the current batch.

[0093] Finally, based on the batch weight corresponding to each preceding material batch in the preceding material batch sequence, the comprehensive risk assessment value of the first material and the comprehensive risk assessment values ​​of multiple second materials are smoothed to obtain a more accurate and reliable smoothed risk assessment value.

[0094] This embodiment introduces welding environment parameters and assigns batch weights based on the similarity of their calculated parameters, thereby performing weighted smoothing on the first material comprehensive risk assessment value and multiple second material comprehensive risk assessment values. This embodiment considers the actual environmental differences between different batches during the welding process, enabling the smoothing process to intelligently assign greater weight to preceding batches with welding environments more similar to the current batch. This weight allocation mechanism based on environmental similarity effectively avoids assessment bias caused by environmental differences, ensuring that the smoothed risk assessment value more realistically and accurately reflects the potential risks of the current material batch, thus providing a more solid data foundation for subsequent updates to confidence level requirements.

[0095] To illustrate this technical solution more clearly, a specific example is used below. Assume the current material batch is batch PM, and its first welding environment parameters include an ambient temperature of 25℃, an ambient humidity of 60%, and a welding power supply output voltage of 24V. The preceding material batches include batches PM-1, PM-2, and PM-3, and their respective second welding environment parameters are: batch PM-1 (ambient temperature 26℃, ambient humidity 58%, welding power supply output voltage 23.8V), batch PM-2 (ambient temperature 20℃, ambient humidity 70%, welding power supply output voltage 25V), and batch PM-3 (ambient temperature 25.5℃, ambient humidity 61%, welding power supply output voltage 24.1V). First, the similarity of the welding environment parameters between batch PM and batch PM-1 can be calculated. For example, using normalized Euclidean distance calculation, the parameter similarity between batch PM-1 and batch PM is found to be 0.95. Similarly, the parameter similarity between batch PM-2 and batch PM is calculated to be 0.70, and the parameter similarity between batch PM-3 and batch PM is 0.98. Next, batch weights are assigned to each preceding batch based on these parameter similarities. For example, a simple linear or exponential function can be used to map the similarity to the weights, such that the higher the similarity, the greater the weight. Assume the weight Q1 of batch PM-1 is 0.4, the weight Q2 of batch PM-2 is 0.1, the weight Q3 of batch PM-3 is 0.5, and the weight Q0 of the current batch PM is 1. Finally, assume the first material comprehensive risk assessment value of the current batch PM is 0.6, and the second material comprehensive risk assessment values ​​of the preceding batches PM-1, PM-2, and PM-3 are 0.5, 0.7, and 0.55, respectively. The smoothed risk assessment value can then be calculated as follows: Smoothed Risk Assessment Value = (0.6 * Q0 + 0.5 * Q1 + 0.7 * Q2 + 0.55 * Q3) / (Q0 + Q1 + Q2 + Q3) = (0.6 * 1 + 0.5 * 0.4 + 0.7 * 0.1 + 0.55 * 0.5) / (1 + 0.4 + 0.1 + 0.5) = 0.5725. Compared to the simple average calculation, the smoothed risk assessment value obtained through this weighted smoothing process is more biased towards the risk assessment value of batches similar to the current batch of PM welding environment (such as batches PM-1 and PM-3), thus providing a more valuable risk assessment result.

[0096] Through the above technical solution, this embodiment can fully consider the differences in welding environment between different batches when smoothing the comprehensive risk assessment value of the first material and multiple comprehensive risk assessment values ​​of the second material. This weighted smoothing process based on the similarity of environmental parameters makes the smoothed risk assessment value more accurate and representative, effectively avoiding assessment distortion caused by insufficient consideration of environmental factors. As a result, the updated confidence level requirement will be more in line with the actual material risk situation, significantly improving the accuracy and reliability of the high-precision welding data analysis method for steel components, thereby optimizing the adjustment suggestions for the overall welding process and reducing potential welding quality risks.

[0097] The beneficial effects of implementing the embodiments of the present invention include: First, the welding process parameters and welding state parameters are obtained. Based on the welding state parameters and the welding target state, the state parameter deviation is calculated. Then, based on the state parameter deviation, a parameter adjustment instruction is generated, and the parameter adjustment instruction is associated with the batch information of the steel components. The parameter adjustment instructions of the steel components in the same batch are then analyzed to identify overall process deviations. Finally, based on the overall process deviations, welding process parameter adjustment suggestions are generated. This enables the identification of overall process deviations by combining state parameter deviations and parameter adjustment instructions, thereby achieving welding data analysis and improving accuracy and welding quality.

[0098] like Figure 2 As shown, this embodiment of the invention also provides a high-precision welding data analysis system for steel components, comprising: The parameter acquisition module 701 is used to acquire welding process parameters and welding status parameters. The welding process parameters include current, voltage, wire feed speed and actuator moving speed, and the welding status parameters include molten pool shape, weld contour, arc sound and temperature distribution. The parameter deviation calculation module 702 is used to calculate the state parameter deviation based on the welding state parameters and the welding target state. The adjustment instruction generation module 703 is used to generate parameter adjustment instructions based on the deviation of the state parameters. The parameter adjustment instructions are used to correct the welding process parameters. Batch association module 704 is used to associate parameter adjustment commands with batch information of steel components; The process deviation identification module 705 is used to analyze the parameter adjustment instructions of steel components in the same batch after correlation to identify overall process deviations. The adjustment suggestion generation module 706 is used to generate welding process parameter adjustment suggestions based on overall process deviations.

[0099] The content of the above method embodiments is applicable to this system embodiment. The specific functions implemented in this system embodiment are the same as those in the above method embodiments, and the beneficial effects achieved are also the same as those achieved in the above method embodiments.

[0100] The embodiments described in this application are for the purpose of more clearly illustrating the technical solutions of the embodiments of this application, and do not constitute a limitation on the technical solutions provided by the embodiments of this application. As those skilled in the art will know, with the evolution of technology and the emergence of new application scenarios, the technical solutions provided by the embodiments of this application are also applicable to similar technical problems.

Claims

1. A method for high-precision welding data analysis of steel components, characterized in that, Includes the following steps: The welding process parameters and welding status parameters are obtained. The welding process parameters include current, voltage, wire feed speed and actuator moving speed. The welding status parameters include molten pool shape, weld contour, arc sound and temperature distribution. Calculate the state parameter deviation based on the welding state parameters and the welding target state; Based on the deviation of the state parameters, a parameter adjustment instruction is generated, which is used to correct the welding process parameters; The parameter adjustment command is associated with the batch information of the steel components; After correlation, the parameter adjustment instructions of steel components in the same batch are analyzed to identify overall process deviations; Based on the overall process deviation, suggestions for adjusting welding process parameters are generated.

2. The method according to claim 1, characterized in that, The analysis of parameter adjustment instructions for steel components within the same batch to identify overall process deviations includes: Based on the parameter adjustment instructions, calculate the cumulative integral term caused by the characteristics of the raw material batch; Calculate the cumulative average change and first average trend of the integral term of all steel components in the same batch; The average change is subjected to an exceedance judgment to obtain the exceedance judgment result; If the result of the over-limit judgment is that it exceeds the upper limit, then the overall process deviation is generated based on the average change and the first average change trend.

3. The method according to claim 2, characterized in that, The step of calculating the cumulative integral term caused by the characteristics of the raw material batch according to the parameter adjustment instruction includes: Extract the first numerical change of the integral term of the proportional-integral-derivative controller from the parameter adjustment command; The second numerical change of non-batch characteristic auxiliary parameters is obtained, including the output voltage fluctuation of the welding power source, the instantaneous value of the shielding gas flow rate, the ambient temperature, the ambient humidity, and the sensor zero-point drift data; Identify the correlation between the first numerical change and the second numerical change; Based on the correlation, the cumulative portion caused by non-batch characteristics is extracted from the first numerical change to obtain the cumulative integral term caused by the raw material batch characteristics.

4. The method according to claim 2, characterized in that, The step of determining whether the average change exceeds the limit and obtaining the result of the limit exceedance judgment includes: Obtain the material grade and thickness of the steel components; Obtain the heat input sensitivity index corresponding to the material grade and thickness from the material property database; The upper limit of variation is determined based on the aforementioned heat input sensitivity index; The average change and the upper limit of change are compared to obtain the result of the over-limit judgment.

5. The method according to claim 2, characterized in that, After calculating the cumulative average change and first average trend of the integral term for all steel components within the same batch, the method further includes: Determine the workpiece quantity window based on the number of steel components; Based on the workpiece quantity window, a trend analysis is performed on the cumulative integral term to obtain the second average change trend; The confidence level of the second average change trend is verified to obtain the confidence level verification result; If the confidence verification result meets the confidence requirement, then the first average change trend is updated according to the second average change trend.

6. The method according to claim 5, characterized in that, The confidence verification of the second average trend to obtain the confidence verification result includes: Obtain the material grade and thickness of the steel components; Obtain the welding quality sensitivity index corresponding to the material grade and thickness from the material property database; Based on the aforementioned welding quality sensitivity index, the confidence level requirement is determined; The second average trend of change is compared with the confidence level requirement to obtain the confidence level verification result.

7. The method according to claim 6, characterized in that, The step of comparing the second average trend with the confidence level requirement to obtain the confidence level verification result includes: Obtain the material sensitivity index corresponding to the material grade and thickness from the material property database. The material sensitivity index includes the impact toughness sensitivity index and the fatigue life sensitivity index. Based on the material sensitivity index and the preset weight allocation rules, calculate the first comprehensive material risk assessment value for the current material batch; The confidence level requirement is updated based on the comprehensive risk assessment value of the first material. The confidence verification result is obtained by comparing the second average trend of change with the updated confidence level requirement.

8. The method according to claim 7, characterized in that, The step of updating the confidence level requirement based on the first material comprehensive risk assessment value includes: Obtain the comprehensive risk assessment value of the second material for each preceding material batch in the preceding material batch sequence; The comprehensive risk assessment value of the first material and multiple comprehensive risk assessment values ​​of the second material are smoothed to obtain a smoothed risk assessment value. The confidence level requirement is updated based on the smoothed risk assessment value.

9. The method according to claim 8, characterized in that, The smoothing process of the first material comprehensive risk assessment value and multiple second material comprehensive risk assessment values ​​to obtain a smoothed risk assessment value includes: Obtain the first welding environment parameter in the current material batch and the second welding environment parameter corresponding to each preceding material batch in the preceding material batch sequence; Select one batch from the preceding material batch sequence as the target material batch; Calculate parameter similarity based on the first welding environment parameter and the second welding environment parameter corresponding to the target material batch; Based on the parameter similarity, the target material batches are weighted to obtain batch weights; Based on the batch weight corresponding to each preceding material batch in the preceding material batch sequence, the comprehensive risk assessment value of the first material and multiple comprehensive risk assessment values ​​of the second material are smoothed to obtain the smoothed risk assessment value.

10. A high-precision welding data analysis system for steel components, characterized in that, include: The parameter acquisition module is used to acquire welding process parameters and welding status parameters. The welding process parameters include current, voltage, wire feed speed and actuator moving speed. The welding status parameters include molten pool shape, weld contour, arc sound and temperature distribution. The parameter deviation calculation module is used to calculate the state parameter deviation based on the welding state parameters and the welding target state. The adjustment instruction generation module is used to generate parameter adjustment instructions based on the state parameter deviation, and the parameter adjustment instructions are used to correct the welding process parameters; The batch association module is used to associate the parameter adjustment command with the batch information of the steel components; The process deviation identification module is used to analyze the parameter adjustment instructions of steel components within the same batch after correlation, and to identify overall process deviations. The adjustment suggestion generation module is used to generate welding process parameter adjustment suggestions based on the overall process deviation.

Images