Differential welding method and system for aluminum alloy door frame

Abstract

According to the differentiated welding method and system for the aluminum alloy door frame, stress distribution data are obtained through finite element analysis, a quantitative model is established, welding parameters are determined according to local strength requirements, in the welding process, the welding depth and density are monitored in real time, laser parameters are adjusted through a feedback control algorithm, stable output is ensured, and then the welding quality is improved. An optical scanning technology is adopted to detect welding quality, a stress distribution model is combined to optimize local strength, meanwhile, a resource allocation algorithm is utilized to analyze the influence of welding parameters on production efficiency, appearance quality and efficiency are balanced through iterative calculation, a final differential welding scheme is obtained, and finally, a digital twinning technology is applied to verify the applicability of the scheme. And parameters are dynamically corrected by adopting a self-adaptive adjustment algorithm, so that process optimization of diversified scenes is realized. The structural strength and the production efficiency of the aluminum alloy door frame are effectively improved, and meanwhile the appearance quality is guaranteed.

Description

Technical Field

[0001] The present invention relates to the field of intelligent processing technology, and in particular to a differentiated welding method and system for aluminum alloy door frames. Background Art

[0002] Laser welding technology for aluminum alloy door frames, an indispensable and critical area in modern manufacturing, is directly related to the structural stability, service life, and improved production efficiency of building components. With the construction industry's growing demand for lightweight, high-strength, and aesthetically pleasing materials, laser welding, due to its high precision and flexibility, has become a core process in aluminum alloy door frame manufacturing. However, research and application in this field still face numerous technical bottlenecks that urgently need to be overcome, and its importance is self-evident, especially in meeting diverse usage scenarios and complex stress environments.

[0003] Currently, traditional welding methods such as argon arc welding (TIG) and resistance welding (RW) are widely used in the manufacture of aluminum alloy door frames, but these technologies generally have limitations. TIG welding has a large heat-affected zone, which can easily lead to material deformation and strength loss. Resistance welding, on the other hand, struggles to achieve precise connections in complex structures and is less adaptable to decorative applications requiring consistent appearance. These shortcomings make it difficult to achieve both reasonable strength distribution and refined design in practical door frames, limiting their potential for application in the high-end market.

[0004] While the introduction of laser welding technology has brought significant improvements, it still faces core challenges. Specifically, controlling weld depth and width, matching weld density and position, and optimizing stress distribution are three key technical factors. Because welding parameters are difficult to flexibly adjust based on the functional requirements of different door frame components, existing technologies often result in localized excess or insufficient strength. For example, in high-stress areas such as hinge joints, insufficient weld depth and density can easily cause deformation or even fracture. Meanwhile, in low-stress areas, excessive welding increases costs and compromises aesthetics. These unresolved technical factors directly contribute to the unique dilemma of uneven strength distribution and wasted resources.

[0005] Therefore, achieving differentiated strength distribution in different locations of aluminum alloy door frames through laser welding technology has become a critical challenge that needs to be overcome. Research needs to focus on adjusting weld depth, width, and density based on stress analysis to ensure a balance between deformation resistance, appearance quality, and production efficiency. Solving this problem will provide a practical technical path for the manufacture of aluminum alloy door frames. Summary of the Invention

[0006] In order to solve the technical problems raised in the above background technology, the first aspect of the present invention provides a differentiated welding method for aluminum alloy door frames, the method comprising: S1, using finite element analysis to obtain stress distribution data for aluminum alloy door frames under different usage scenarios, determine the initial thresholds of welding depth and welding density for high-stress areas and low-stress areas, and obtain a quantitative model of stress distribution; S2, based on the quantitative model of stress distribution, extracts the specific requirements of local strength from the high stress area, uses the laser welding parameter adjustment algorithm to calculate the matching values ​​of welding depth and width, and determines the welding parameter configuration table for each part; S3, after obtaining the welding parameter configuration table, the actual values ​​of welding depth and density are detected during the welding process through the real-time monitoring system. If the actual values ​​deviate from the set values ​​in the configuration table, the laser power and speed are adjusted through the feedback control algorithm to obtain a stable welding output result; S4, for stable welding output results, uses optical scanning technology to detect the deviation of weld width and appearance quality, extracts surface consistency indicators from the scan data, and determines whether it meets the preset appearance quality threshold; S5, by comparing the surface consistency index with the stress distribution model, obtain the distribution characteristics of local strength and deformation resistance. If the deformation resistance of the high stress area is lower than the functional requirement, the welding density is increased through the local repair welding algorithm to obtain the optimized strength distribution data; S6, based on the optimized strength distribution data, a resource allocation algorithm is used to analyze the impact of welding depth and density on production efficiency, from which the parameter redundancy values ​​of the redundant welding area are extracted to determine the quantitative degree of resource waste; S7, after obtaining the quantified degree of resource waste, adjusts the welding depth and density in the low-stress area through a parameter optimization algorithm, performs iterative calculations on the balance point between appearance quality and production efficiency, and obtains the final differentiated welding solution; S8: For the final differentiated welding solution, digital twin technology is used to simulate the door frame's deformation resistance and structural stability under complex stress environments. Verification data on performance distribution is extracted from the simulation results to determine the solution's applicability. S9, through the verification data of performance distribution, obtains the deviation characteristics that may occur in the actual welding process, and uses the adaptive adjustment algorithm to dynamically correct the welding parameter configuration table to obtain a process parameter set suitable for diverse scenarios.

[0007] Optionally, step S1, obtaining stress distribution data of the aluminum alloy door frame in different usage scenarios through finite element analysis, determining welding depth and welding density initial thresholds for high stress areas and low stress areas, respectively, and obtaining a quantitative model of stress distribution, includes: Step S11, obtaining stress distribution data of the aluminum alloy door frame from different usage scenarios through finite element analysis to obtain preliminary stress distribution results; Step S12, dividing the high stress area and the low stress area according to the stress distribution result, and obtaining the stress distribution characteristics after the area division; Step S13, extracting stress distribution characteristics for the high stress area and the low stress area respectively, and determining the initial threshold range of welding depth and welding density; Step S14, using a support vector machine algorithm to classify the initial threshold ranges of welding depth and welding density to obtain adjusted welding depth and welding density thresholds; Step S15, using the adjusted welding depth and welding density threshold as input parameters, performing parameter mapping on the aluminum alloy door frame through finite element analysis to obtain stress distribution data after parameter optimization; Step S16, constructing a quantitative model of stress distribution based on the stress distribution data after parameter optimization; Step S17: Calculate the uniformity index of stress distribution through the quantitative model to determine the structural stability of the aluminum alloy door frame in different usage scenarios.

[0008] Optionally, step S17, calculating the uniformity index of stress distribution by a quantitative model to determine the structural stability of the aluminum alloy door frame in different usage scenarios, includes: Step S171, obtaining original stress distribution data of the aluminum alloy door frame in different usage scenarios, wherein the original stress distribution data is obtained by performing mechanical simulation on the aluminum alloy door frame using a finite element analysis tool; Step S172, extracting distribution feature vectors from the original stress distribution data using principal component analysis; Step S173: input the distribution characteristic vector into a time series analysis tool, calculate the standard deviation of the stress distribution in each scenario, and use the standard deviation of the stress distribution as an indicator of stress distribution uniformity; Step S174, obtaining a 95th percentile threshold of historical data, and comparing the stress distribution uniformity index with the 95th percentile threshold; Step S175: If the stress distribution uniformity index exceeds the 95th percentile threshold, a gradient descent algorithm is called to adjust the weight coefficient of the distribution feature vector to obtain a corrected distribution feature vector; Step S176, re-inputting the corrected distribution characteristic vector into the finite element analysis tool to generate new stress distribution data; Step S177, calculating an updated stress distribution uniformity index based on the new stress distribution data; Step S178, using an allowable stress comparison method to determine structural stability, obtaining a maximum stress value of the new stress distribution data, and comparing the maximum stress value with a yield strength threshold; Step S179: If the maximum stress value exceeds the yield strength threshold, it is determined that the structure is unstable; Step S1710: for unstable structures, extract factors that are strongly correlated with the stress peak from the finite element analysis parameters using the Pearson correlation coefficient; Step S1711: Based on the strongly correlated factors, a stress weighting algorithm is used to locally enhance the distribution characteristics of the key areas; Step S1712: outputting the optimized stress distribution uniformity index and stress distribution stability determination result.

[0009] Optionally, the yield strength threshold is 70% of the material yield strength.

[0010] Optionally, step S5, by comparing the surface consistency index with the stress distribution model, obtains the distribution characteristics of local strength and deformation resistance. If the deformation resistance of the high stress area is lower than the functional requirement, the welding density is increased by a local repair welding algorithm to obtain optimized strength distribution data, including: Step S51, obtaining workpiece surface roughness data, inputting a pre-established elastic-plastic analysis model, and obtaining stress distribution data; Step S52, comparing the stress distribution data with a stress distribution threshold, and marking areas exceeding the stress distribution threshold as high stress areas; Step S53, for the high stress area, using a welding simulation module to calculate the anti-deformation capability under the current welding density; Step S54, if the anti-deformation capability is lower than a preset value, calculating the welding density increment according to a preset material strength formula to generate an adjusted welding density distribution; Step S55, inputting the adjusted welding density distribution into the elastic-plastic analysis model to recalculate the stress distribution; Step S56, extracting the stress value in the recalculated stress distribution and the adjusted welding density as features, and performing classification using a preset classification algorithm, wherein the classification label is whether the deformation resistance meets the standard; Step S57: if the classification result is not up to standard, the welding density of the high stress area is updated with a preset step size, and the welding simulation is repeated until the number of iterations reaches a preset value or the deformation resistance of the high stress area is all up to standard; Step S58: draw a curve showing how the intensity of each region changes with the number of iterations, and determine whether the final intensity of each region reaches the regional intensity threshold.

[0011] Optionally, the preset step size is 10% of the welding density increment.

[0012] Optionally, step S6, based on the optimized strength distribution data, uses a resource allocation algorithm to analyze the impact of welding depth and density on production efficiency, extracts parameter redundancy values ​​of redundant welding areas, and determines the quantitative degree of resource waste, including: Step S61, loading a preset analysis module with the intensity distribution data, and using principal component analysis to decompose and obtain optimized data and distribution data; Step S62, using the simplex method to calculate the influence coefficients of welding depth and welding density based on the optimization data and distribution data, and obtaining the change trend of production efficiency; Step S63: extracting parameter redundancy values ​​based on the changing trend of production efficiency and determining the boundary range of the redundant area; Step S64, calculating the initial quantitative value of resource waste based on the boundary range of the redundant area, and obtaining the adjustment requirements for resource allocation; Step S65: If the initial quantified value of resource waste exceeds the waste quantification threshold, the least square method is used to optimize resource allocation to obtain an adjusted allocation plan; Step S66: updating the distribution data of welding depth and welding density according to the adjusted allocation scheme to determine the quantitative degree of resource waste; Step S67 : Determine the final resource allocation result by comparing the quantified degree of resource waste with the initial quantified value.

[0013] Optionally, in step S7, after obtaining the quantitative degree of resource waste, the welding depth and density in the low-stress area are adjusted using a parameter optimization algorithm, and an iterative calculation is performed on the balance point between appearance quality and production efficiency to obtain a final differentiated welding solution, including: Step S71: obtaining quantitative data on resource waste from production data, calculating the degree of waste using descriptive statistics, and obtaining an initial assessment result; Step S72, based on the initial evaluation results, using a gradient descent algorithm to adjust the welding depth and welding density in the low stress area to obtain a preliminary adjustment plan; Step S73: Calculate the weld surface roughness and weld width as the appearance quality index values ​​for the preliminary adjustment plan, determine whether they meet the preset appearance index threshold, and obtain quality assessment data; Step S74: Calculate the welding length per unit time as the production efficiency index value based on the quality assessment data. If it is lower than the production efficiency threshold, adjust the welding density to obtain the efficiency optimization result. Step S75: Obtain the efficiency optimization result, combine it with the appearance quality index value, use the Newton iteration method to determine the parameter combination of the balance point, and obtain a set of candidate solutions; Step S76: Using the K-means clustering algorithm to analyze the differential features from the candidate solution set to obtain the final differentiated welding solution; In step S77 , the resource waste quantitative data before and after optimization are compared using the final differentiated welding solution to verify the downward trend and obtain the optimized data distribution.

[0014] Optionally, the step S76, using a K-means clustering algorithm to analyze the differential features from the candidate solution set to obtain a final differentiated welding solution, includes: Step S761, obtaining a differentiated feature data set including current, voltage, and speed; Step S762, using a Z-score standardization method to standardize the differential feature dataset to obtain a standardized feature dataset; Step S763, clustering the standardized feature data set using a K-means clustering algorithm to obtain a preliminary grouping result, wherein the preliminary grouping result includes at least two clusters; Step S764, obtaining the mean feature vector of each cluster in the preliminary grouping result; Step S765, calculating the Euclidean distance between the mean eigenvectors. If the Euclidean distance is greater than the Euclidean distance threshold, adjusting the cluster center and re-clustering until the Euclidean distance is less than or equal to the Euclidean distance threshold, thereby obtaining an optimized grouping result. Step S766, obtaining a welding scheme parameter set corresponding to each cluster in the optimization grouping result; Step S767, using a Pearson correlation coefficient matrix to analyze the correlation between the welding scheme parameter set and the welding quality, and obtaining a welding scheme parameter subset having a correlation coefficient greater than a correlation threshold; Step S768, generating a differentiated welding plan according to the welding plan parameter subset; Step S769, obtaining the current, voltage, and speed parameter distribution data of each of the differentiated welding schemes; Step S7610: Use a box plot to analyze the discreteness of the parameter distribution data to obtain the solution distribution characteristics.

[0015] A second aspect of the present invention provides a differentiated welding system for aluminum alloy door frames, wherein the differentiated welding of the aluminum alloy door frames is performed using the above-mentioned method, and the system comprises: The stress analysis module is used to obtain stress distribution data of aluminum alloy door frames under different usage scenarios through finite element analysis. The module determines the initial thresholds of welding depth and welding density for high-stress areas and low-stress areas, respectively, to obtain a quantitative model of stress distribution. The parameter calculation module is used to extract the specific requirements of local strength from high-stress areas based on the quantitative model of stress distribution, calculate the matching values ​​of welding depth and width using the laser welding parameter adjustment algorithm, and determine the welding parameter configuration table for each part; The real-time control module is used to obtain the welding parameter configuration table and detect the actual values ​​of welding depth and density during the welding process through the real-time monitoring system. If the actual values ​​deviate from the set values ​​in the configuration table, the laser power and speed are adjusted through the feedback control algorithm to obtain a stable welding output result; The quality inspection module is used to detect the deviation of weld width and appearance quality based on the stable welding output results using optical scanning technology, extract surface consistency indicators from the scan data, and determine whether the preset appearance quality threshold is met; The strength optimization module is used to obtain the distribution characteristics of local strength and deformation resistance by comparing the surface consistency index with the stress distribution model. If the deformation resistance of the high-stress area is lower than the functional requirement, the welding density is increased through the local repair welding algorithm to obtain the optimized strength distribution data; The resource assessment module is used to analyze the impact of welding depth and density on production efficiency based on the optimized strength distribution data using a resource allocation algorithm, extract parameter redundancy values ​​of excess welding areas, and determine the quantitative degree of resource waste; The parameter balancing module is used to obtain the quantified degree of resource waste, then adjust the welding depth and density in low-stress areas through a parameter optimization algorithm. It iteratively calculates the balance point between appearance quality and production efficiency to obtain the final differentiated welding solution. The performance verification module uses digital twin technology to simulate the door frame's deformation resistance and structural stability under complex stress environments for the final differentiated welding solution. It then extracts performance distribution verification data from the simulation results to determine the solution's applicability. The dynamic correction module is used to obtain the deviation characteristics that may occur in the actual welding process through the verification data of the performance distribution, and dynamically correct the welding parameter configuration table using the adaptive adjustment algorithm to obtain the process parameter set suitable for various scenarios. The technical solution provided by the embodiment of the present invention has the following beneficial effects: The present invention provides a differentiated welding method and system for aluminum alloy door frames. Stress distribution data is obtained through finite element analysis, a quantitative model is established, and welding parameters are determined according to local strength requirements. During the welding process, the welding depth and density are monitored in real time, and the laser parameters are adjusted through a feedback control algorithm to ensure stable output. Subsequently, optical scanning technology is used to detect the welding quality, and the local strength is optimized in combination with the stress distribution model. At the same time, a resource allocation algorithm is used to analyze the impact of welding parameters on production efficiency. The appearance quality and efficiency are balanced through iterative calculation to obtain the final differentiated welding scheme. Finally, digital twin technology is used to verify the applicability of the scheme, and an adaptive adjustment algorithm is used to dynamically correct the parameters to achieve process optimization in diversified scenarios.

[0016] The present invention effectively improves the structural strength and production efficiency of the aluminum alloy door frame while ensuring the appearance quality. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 The present invention is a flowchart of a differentiated welding method for aluminum alloy door frames.

[0018] Figure 2 It is a structural schematic diagram of a differentiated welding system for aluminum alloy door frames of the present invention. DETAILED DESCRIPTION

[0019] The following will describe the technical solutions in the embodiments of the present invention in detail with reference to the accompanying drawings. The described embodiments are only a part of the embodiments of the present invention.

[0020] like Figure 1 As shown, the first aspect of the present invention provides a differentiated welding method for an aluminum alloy door frame, the method comprising: S1, through finite element analysis to obtain the stress distribution data of aluminum alloy door frames in different usage scenarios, determine the initial thresholds of welding depth and welding density for high stress areas and low stress areas respectively, and obtain a quantitative model of stress distribution.

[0021] Optionally, this step also includes: Step S11, obtaining stress distribution data of the aluminum alloy door frame from different usage scenarios through finite element analysis to obtain preliminary stress distribution results; step S12, dividing high stress areas and low stress areas according to the stress distribution results to obtain stress distribution characteristics after regional division; step S13, extracting stress distribution characteristics for high stress areas and low stress areas respectively, and determining the initial threshold ranges of welding depth and welding density; step S14, using the support vector machine algorithm to classify the initial threshold ranges of welding depth and welding density to obtain adjusted welding depth and welding density thresholds; step S15, using the adjusted welding depth and welding density thresholds as input parameters, performing parameter mapping on the aluminum alloy door frame through finite element analysis to obtain stress distribution data after parameter optimization; step S16, constructing a quantitative model of stress distribution based on the stress distribution data after parameter optimization; step S17, calculating the stress distribution uniformity index through the quantitative model to judge the structural stability of the aluminum alloy door frame under different usage scenarios.

[0022] Specifically, obtaining stress distribution data of aluminum alloy door frames from different usage scenarios through finite element analysis is the first step in optimizing the design.

[0023] For example, assuming an aluminum alloy door frame is subjected to the dynamic loads of opening and closing the door in a home, finite element analysis software can be used to simulate the stress distribution of the door frame under forces of 50N and 100N, resulting in stress values ​​of 120MPa and 80MPa at the top and bottom nodes of the door frame, respectively. This method can intuitively reveal the door frame's weaknesses in actual use and provide data support for subsequent analysis. Using these stress distribution results to delineate high-stress and low-stress areas is key to further refining the analysis.

[0024] Specifically, based on the above data, the top region with stress greater than 100 MPa can be classified as a high-stress zone, while the bottom region with stress less than 90 MPa can be classified as a low-stress zone. This division helps designers focus on the reinforcement requirements of high-stress areas while avoiding overdesign in low-stress areas, saving material costs. Extracting stress distribution characteristics for high-stress and low-stress areas, and determining the initial threshold ranges for weld depth and weld density, are the core of the technical implementation.

[0025] In one possible implementation, for high-stress areas, the maximum stress concentration point of 120 MPa can be extracted, with a preliminary weld depth of 5 mm and a weld density of one weld every 10 cm. For low-stress areas, due to lower stress, the weld depth is set at 3 mm and the weld density is set at one weld every 20 cm. This differentiated design effectively balances strength and cost. Using a support vector machine algorithm to classify the initial threshold ranges for weld depth and weld density and adjust welding parameters is a manifestation of intelligent optimization.

[0026] Optimally, sample data from high-stress areas (e.g., stress values, weld depth of 5mm, and weld density of 10cm) and low-stress areas (e.g., stress values, weld depth of 3mm, and weld density of 20cm) can be fed into a support vector machine. The algorithm then adjusts the classification boundaries to optimize the weld depth to 5.5mm and density to 8cm in the high-stress area, and to 3.2mm and 18cm in the low-stress area. This adjustment improves parameter accuracy and reduces trial-and-error costs. Finite element analysis can be used to perform parameter mapping using the adjusted weld depth and density as input parameters to verify the optimization results.

[0027] For example, re-entering the model with parameters for a 5.5mm weld depth and 8cm weld density revealed that the maximum stress in high-stress areas dropped to 110 MPa, resulting in a more uniform distribution. This method provides intuitive verification of the effectiveness of parameter adjustments and lays the foundation for subsequent modeling. Building a quantitative stress distribution model based on the stress distribution data after parameter optimization further deepens the analysis.

[0028] In one embodiment, a model can be constructed by statistically analyzing the mean square error (assuming 15 MPa) of stress values ​​to quantify the degree of dispersion in the stress distribution. This model provides a mathematical basis for stability assessment, enhancing the scientific nature of the design. The ultimate goal is to calculate the uniformity index of the stress distribution using the quantitative model and determine structural stability.

[0029] The stress distribution uniformity index can be defined as the ratio of the mean square error of the stress values ​​to the mean stress. A calculated value of 0.12 indicates a relatively uniform stress distribution and high door frame stability during opening and closing. This assessment method provides reliable assurance for the long-term use of aluminum alloy door frames, while also optimizing production processes and extending product life.

[0030] Optionally, the step S17 of calculating the uniformity index of stress distribution by using a quantitative model to determine the structural stability of the aluminum alloy door frame in different usage scenarios further includes: Step S171, obtaining the original stress distribution data of the aluminum alloy door frame in different usage scenarios, wherein the original stress distribution data is obtained by mechanical simulation of the aluminum alloy door frame using a finite element analysis tool; Step S172, extracting distribution feature vectors from the original stress distribution data using principal component analysis; Step S173, inputting the distribution feature vectors into a time series analysis tool, calculating the standard deviation of the stress distribution under each scenario, and using the standard deviation of the stress distribution as a stress distribution uniformity index; Step S174, obtaining the 95th percentile threshold of historical data, and comparing the stress distribution uniformity index with the 95th percentile threshold; Step S175, if the stress distribution uniformity index exceeds the 95th percentile threshold, calling the gradient descent algorithm to adjust the weight coefficient of the distribution feature vector to obtain a corrected distribution feature vector; Step S176, The corrected distribution characteristic vector is re-input into the finite element analysis tool to generate new stress distribution data; in step S177, based on the new stress distribution data, an updated stress distribution uniformity index is calculated; in step S178, the structural stability is judged by using the allowable stress comparison method, the maximum stress value of the new stress distribution data is obtained, and the maximum stress value is compared with the yield strength threshold; in step S179, if the maximum stress value exceeds the yield strength threshold, the structure is judged to be unstable; in step S1710, for unstable structures, the Pearson correlation coefficient is used to extract factors that are strongly correlated with the stress peak from the finite element analysis parameters; in step S1711, based on the strongly correlated factors, the distribution characteristics of the key area are locally enhanced by using the stress weighting algorithm; in step S1712, the optimized stress distribution uniformity index and stability judgment result are output.

[0031] Specifically, using finite element analysis tools to conduct mechanical simulation of different usage scenarios of aluminum alloy door frames and obtaining original stress distribution data is the basis of the entire process.

[0032] For example, in a home setting, a doorframe may be subject to dynamic loads from frequent opening and closing. Using finite element analysis software, we can simulate the stress distribution of a doorframe under a 50N thrust, obtaining data showing a top node stress of 130MPa and a bottom node stress of 70MPa. This method clearly reflects the stress variations in the doorframe during actual use, providing a reliable starting point for subsequent analysis. Using principal component analysis to extract distribution eigenvectors from gridded stress data is a key step in data dimensionality reduction.

[0033] Specifically, assuming a grid contains 1,000 nodes, the stress values ​​at each node form a high-dimensional dataset. Principal component analysis can convert this data into a few primary eigenvectors. For example, the first three vectors may capture 90% of the stress distribution information. This extraction simplifies subsequent calculations while preserving core features. Inputting the distribution eigenvectors into a time series analysis tool and calculating the standard deviation of the stress distribution under various scenarios as an indicator of stress distribution uniformity is an important step in quantifying the distribution pattern.

[0034] In one possible implementation, for the aforementioned 50N load scenario, time series analysis shows a stress standard deviation of 20 MPa. If the 95th percentile threshold of the historical data is 25 MPa, the current indicator indicates a relatively uniform distribution. This metric can help designers quickly determine the degree of dispersion in the stress distribution. If the indicator value exceeds the 95th percentile threshold of the historical data, a gradient descent algorithm is used to adjust the weight coefficients of the eigenvectors.

[0035] Preferably, assuming the stress standard deviation reaches 30 MPa, exceeding the 95th percentile threshold of 25 MPa, gradient descent can iteratively adjust the weights of the eigenvectors, for example, reducing the weight of the top region from 0.6 to 0.5, thereby achieving a more balanced stress distribution. This adjustment effectively optimizes data features. Re-entering the corrected eigenvectors into the finite element analysis tool to generate new stress distribution data and calculate the updated stress distribution uniformity index is key to verifying the effectiveness of the adjustment.

[0036] For example, after adjustments and re-simulation, the top stress might drop to 115 MPa, with the stress standard deviation reduced to 18 MPa. This repeated verification ensures the correctness of the optimization direction. Using the allowable stress comparison method to determine structural stability is a core method for assessing door frame reliability.

[0037] Preferably, the yield strength threshold is set to 70% of the material yield strength.

[0038] It's understandable that if the aluminum alloy has a yield strength of 200 MPa, the yield strength threshold is set at 70% of the material's yield strength, or 140 MPa. If the maximum stress in the simulation is 130 MPa, it does not exceed the yield strength threshold and is considered stable. This method is intuitive and practical, providing a safe basis for design. For unstable results, using the Pearson correlation coefficient to extract factors strongly correlated with peak stress from the finite element analysis parameters is an effective way to pinpoint problems.

[0039] In one example, assume the maximum stress reaches 150 MPa, exceeding the yield strength threshold of 140 MPa. Analysis shows that the peak stress is highly correlated with mesh density and load angle, with correlation coefficients of 0.85 and 0.9, respectively. This analysis can help designers focus on key influencing factors. Using a stress-weighted algorithm to locally enhance the distribution characteristics of critical areas is the final step in design optimization.

[0040] It's important to note that for the high-stress area at the top, the characteristics of this region can be enhanced by increasing the weight coefficient to 0.7. After simulation, the peak stress can be reduced to 125 MPa, and the stress distribution uniformity index can be improved to 15 MPa. This localized strengthening can significantly improve the overall performance of the door frame and ensure long-term use.

[0041] S2, based on the quantitative model of stress distribution, extracts the specific requirements of local strength from the high stress area, uses the laser welding parameter adjustment algorithm to calculate the matching values ​​of welding depth and width, and determines the welding parameter configuration table for each part.

[0042] Optionally, this step also includes: Step S21, obtain the three-dimensional model data of the component, use the finite element analysis tool to calculate the stress distribution of the component, and obtain stress cloud map data; step S22, obtain the mechanical property data of the material, and determine the area in the stress cloud map where the stress value exceeds the preset percentage of the material yield strength as the high stress area; step S23, perform edge detection on the high stress area to obtain the boundary coordinate data of the high stress area; step S24, use the gradient descent algorithm to calculate the initial matching value of the welding depth and width according to the preset laser parameters; step S25, when the welding depth deviation exceeds the depth deviation threshold or the welding width deviation exceeds the width deviation threshold, trigger the control algorithm to correct the welding parameters; step S26, generate a parameter allocation table including welding speed and defocus amount according to the tensile strength value in the preset standard; step S27, convert the allocation table into the power-speed configuration parameter per unit length of the weld by table lookup; step S28, use non-destructive testing equipment to detect the quality of the weld, and determine that the verification is passed when the defect rate is lower than the preset percentage and the stress-strain value meets the allowable stress condition.

[0043] Specifically, the ANSYS finite element analysis tool is used to calculate the stress distribution of the component, and the output stress cloud map data is the starting point of the design.

[0044] For example, for an aluminum alloy door frame, ANSYS can be used to simulate the forces acting on it during opening and closing. Setting boundary conditions to a fixed top and a 50N pull at the bottom, a stress contour map is generated, visually displaying the stress variations in various parts of the door frame. Colors in the contour map represent stress levels, facilitating subsequent analysis. Obtaining yield strength values ​​from a material mechanical properties database and marking high-stress areas is a method for identifying critical locations.

[0045] It can be understood that, assuming that the yield strength of aluminum alloy is 150 MPa, the area where the stress exceeds 120 MPa (i.e., 80% of the material yield strength) is considered to be a high stress area.

[0046] For example, the top of a door frame shows a high stress level of 130 MPa due to concentrated stress, while the bottom shows only 70 MPa. This method allows for quick identification of areas requiring reinforcement. Using OpenCV to perform Canny edge detection on high-stress areas and extract boundary coordinate data is a step in the refinement process.

[0047] Specifically, after inputting the stress cloud image, OpenCV performs grayscale conversion and edge detection to output the boundary coordinates of the top high-stress area, such as (10, 20) to (15, 25). This provides a precise geometric basis for subsequent welding parameter design. Based on the laser welding wavelength of 1064nm and a power density of 1e6W per square centimeter, a gradient descent algorithm is used to calculate the weld depth and width, which is the core of parameter optimization.

[0048] In one possible implementation, the weld depth is initially set at 5mm and the weld width is 2mm for the high-stress top zone. Gradient descent is then used to iteratively adjust the welding parameters to suit the stress distribution. This approach ensures that the weld quality matches the component requirements. When the weld depth deviation exceeds a depth deviation threshold of ±0.1mm or the width deviation exceeds a width deviation threshold of ±0.2mm, the PID control algorithm is triggered for correction, ensuring accuracy.

[0049] For example, if the actual weld depth reaches 5.15mm, exceeding the set weld depth, the PID controller adjusts the laser power and speed through feedback to return the weld depth to 5mm. This dynamic correction improves process stability. Generating a parameter allocation table based on the tensile strength values ​​in the GB / T228 standard embodies process standardization.

[0050] Optimally, for a door frame with a top tensile strength requirement of 200 MPa, a welding speed of 10 mm / s and a defocus of 2 mm can be set. Lower requirements are imposed on the bottom, so the welding speed can be increased to 15 mm / s. This allocation table provides practical guidance for production. A lookup table method is a convenient way to convert the allocation table into a power-speed combination per millimeter of weld seam.

[0051] In one embodiment, the top weld is configured with 800W of power and a speed of 10mm / s per millimeter, while the bottom is configured with 600W and 15mm / s. This conversion simplifies the parameter application process. Using ultrasonic flaw detectors to inspect welds and verifying them with stress-strain equations is key to quality control.

[0052] For example, if the top weld defect rate is only 2% and the calculated stress value is lower than the allowable stress of 190 MPa, it is considered qualified. This method ensures the reliability of the weld and the long-term performance of the component.

[0053] S3, after obtaining the welding parameter configuration table, the actual values ​​of welding depth and density are detected during the welding process through the real-time monitoring system. If the actual values ​​deviate from the set values ​​in the configuration table, the laser power and speed are adjusted through the feedback control algorithm to obtain a stable welding output result.

[0054] Optionally, this step also includes: Step S31, obtain the set values ​​in the welding parameter configuration table, including welding depth, welding density, laser power and welding speed. Step S32, collect the actual depth and density data during the welding process through the real-time monitoring system to form a continuous data stream. Step S33, compare the actual collected data with the set values ​​in the parameter configuration table, calculate the deviation value and determine whether it exceeds the parameter deviation threshold. Step S34, if the deviation value exceeds the parameter deviation threshold, start the feedback control algorithm and automatically adjust the laser power output according to the degree of deviation. Step S35, calculate and adjust the optimal welding speed according to the deviation value and the adjusted laser power to maintain stable welding quality. Step S36, continuously monitor the effect of the adjusted welding parameters, collect new depth and density data, and perform closed-loop optimization. Step S37, update the parameter configuration table based on the accumulated welding data and adjustment records, optimize the initial setting value, and improve system adaptability.

[0055] Specifically, during the laser welding process, real-time monitoring and parameter adjustment are crucial to ensuring welding quality.

[0056] For example, consider an automobile body welding production line. The welding parameter configuration table sets the weld depth to 3mm, the density to 7.8g / cm³, the laser power to 2000W, and the welding speed to 50mm / s.

[0057] Specifically, the real-time monitoring system uses high-speed cameras and optical sensors to collect weld depth and density data every 0.1 seconds.

[0058] It should be noted that this high-frequency acquisition can form a continuous data stream, which helps to detect welding anomalies in a timely manner.

[0059] In one possible implementation, the system compares the actual depth collected with a set depth value of 3 mm.

[0060] For example, if the actual depth is detected to be 2.8 mm and the deviation is -0.2 mm, which exceeds the preset parameter deviation threshold of ±0.1 mm, feedback control is triggered.

[0061] Preferably, the feedback control algorithm adopts fuzzy PID control to dynamically adjust the laser power according to the deviation.

[0062] It is understandable that for a deviation of -0.2mm, the system may increase the power to 2100W to increase the penetration depth. At the same time, the welding speed also needs to be adjusted accordingly.

[0063] In one embodiment, the system reduces the welding speed to 48 mm / s based on the new power value and material properties to ensure weld quality. This coordinated adjustment of speed and power effectively balances welding efficiency and quality. During the continuous monitoring process, the system continuously collects new data and performs closed-loop optimization.

[0064] For example, the adjusted parameters may stabilize the weld depth between 2.95mm and 3.05mm, and maintain the weld density within the range of 7.75-7.85g / cm³, achieving the desired effect. Finally, the system updates the parameter configuration table based on the accumulated adjustment records.

[0065] In one embodiment, if it is found that a certain material combination always requires a slightly higher initial power, the system may adjust the default power of the combination from 2000W to 2050W, thereby improving the adaptability and efficiency of the system.

[0066] S4, for stable welding output results, uses optical scanning technology to detect the deviation of welding width and appearance quality, extracts surface consistency indicators from the scanning data, and determines whether it meets the preset appearance quality threshold.

[0067] Optionally, this step also includes: Step S41, obtaining the deviation data matrix of the welding output; Step S42, calculating the standard deviation of the deviation data matrix to obtain a preliminary surface consistency index; Step S43, inputting the preliminary consistency index into a preset radial basis kernel function support vector machine; Step S44, determining whether the preliminary consistency index is within the appearance quality threshold range; Step S45, if so, extracting samples with a classification label of 1 from the output of the support vector machine; Step S46, using a box plot to analyze the deviation data distribution corresponding to the samples with a classification label of 1; Step S47, calculating the interval from the 25th percentile to the 75th percentile of the box plot to obtain an adjustment value; Step S48, adding the adjustment value to the preliminary consistency index to obtain an updated consistency index; Step S49, if the updated consistency index is greater than a preset upper limit threshold, performing a preset neighborhood mean filtering on the deviation data matrix; Step S410, recalculating the standard deviation based on the filtered deviation data to obtain a final consistency index; Step S411, if the final consistency index is less than or equal to a preset qualified threshold, determining that the welding appearance quality meets the preset standard.

[0068] Specifically, when optical scanning technology is used to obtain the deviation data matrix of welding output, it can be achieved through a high-precision laser scanner.

[0069] For example, in an automotive body welding production line, optical scanners scan the weld surface 100 times per second, generating 3D deviation data. This method captures minute deformations and ensures comprehensive data.

[0070] It should be noted that the deviation data matrix contains multidimensional information such as weld height and width, providing a basis for subsequent analysis. When calculating the standard deviation of the deviation data matrix as a preliminary indicator of surface consistency, the standard deviation reflects the degree of surface fluctuation of the weld.

[0071] For example, if a standard deviation of 0.2 mm is obtained after scanning a certain section of the weld, it indicates that the surface consistency is good.

[0072] In one possible implementation, the system can further improve the reliability of the indicator by taking an average value through multiple scans.

[0073] Understandably, this preliminary indicator provides a quantitative basis for subsequent classification. When the preliminary consistency indicator is input into a preset radial basis kernel support vector machine to determine the appearance quality threshold range, the support vector machine, trained based on historical data, can distinguish whether the consistency is between 0.1 and 0.3 mm.

[0074] Specifically, if a standard deviation of 0.2 mm is input, the model outputs a classification label of 1, which is in line with expectations.

[0075] In one embodiment, if the standard deviation is 0.35 mm, the label may be 0, indicating that adjustment is required. This method improves judgment efficiency and avoids manual intervention. When samples with a classification label of 1 are extracted from the support vector machine output and their deviation data distribution is analyzed using a boxplot, the boxplot intuitively demonstrates the central tendency of the data.

[0076] For example, the median deviation data of a batch of weld samples is 0.15 mm, and the upper and lower interquartile ranges are small, indicating a uniform distribution.

[0077] Preferably, outliers can be identified through the box plot, such as the point where the deviation suddenly increases to 0.4 mm, so that process parameters can be optimized later. When the distance between the 25th and 75th percentiles of the box plot is calculated as the adjustment value, the distance reflects the degree of data dispersion.

[0078] In one embodiment, if the spacing is 0.1 mm, the adjustment value is 0.1 mm, which is added to the initial consistency index of 0.2 mm to obtain an updated consistency index of 0.3 mm. This adjustment method makes the consistency index more consistent with actual needs and improves assessment accuracy. If the updated consistency index is greater than 0.35 mm, when the original deviation data matrix is ​​subjected to a 3×3 neighborhood mean filter, the filter can smooth out local noise.

[0079] For example, in a certain weld area, the deviation suddenly increased due to spatter. After filtering, the data became smoother and the standard deviation decreased from 0.4 mm to 0.25 mm.

[0080] Understandably, this processing preserves key features while reducing interference, helping to improve appearance quality. When the standard deviation is recalculated from the filtered data as the final consistency metric, if the result is 0.28 mm, which is less than or equal to 0.3 mm, the weld appearance quality is considered to meet the standard.

[0081] In a possible implementation, the system records the indicator changes after each filtering to form a trend analysis.

[0082] For example, the weld seam thickness of multiple weld sections remained stable between 0.25 and 0.3 mm, demonstrating process stability. This approach not only improved judgment accuracy but also provided data support for subsequent parameter optimization.

[0083] S5, by comparing the surface consistency index with the stress distribution model, obtain the distribution characteristics of local strength and deformation resistance. If the deformation resistance of the high-stress area is lower than the functional requirement, the welding density is increased through the local repair welding algorithm to obtain the optimized strength distribution data.

[0084] Optionally, this step also includes: Step S51, obtain the surface roughness data of the workpiece, input the pre-established elastic-plastic analysis model, and obtain stress distribution data; Step S52, compare the stress distribution data with the stress distribution threshold, and mark the area exceeding the stress distribution threshold as a high stress area; Step S53, for the high stress area, use the welding simulation module to calculate the deformation resistance under the current welding density; Step S54, if the deformation resistance is lower than the preset value, calculate the welding density increment according to the preset material strength formula to generate an adjusted welding density distribution; Step S55, input the adjusted welding density distribution into the elastic-plastic analysis model. The stress distribution is recalculated using the analysis model; in step S56, the stress value in the recalculated stress distribution and the adjusted welding density are extracted as features, and a preset classification algorithm is used for classification, where the classification label is whether the anti-deformation ability meets the standard; in step S57, if the classification result is not up to standard, the welding density of the high stress area is updated with a preset step size, and the welding simulation is repeated until the number of iterations reaches a preset value or the anti-deformation ability of the high stress area all meets the standard; in step S58, a curve of the change of the strength of each region with the number of iterations is plotted to determine whether the final strength of each region reaches the regional strength threshold.

[0085] Specifically, when collecting workpiece surface roughness data through industrial cameras, high-resolution CCD cameras can be used to capture the texture features of the workpiece surface.

[0086] For example, in automotive parts processing, a camera captures the weld surface at 50 frames per second, generating a grayscale image that reflects the roughness distribution. This method can intuitively record microscopic surface changes and provide reliable input for subsequent analysis.

[0087] It should be noted that the accuracy of roughness data is closely related to the focal length of the camera lens and the angle of the light source. For example, adjusting the lighting to a 45-degree angle can reduce reflection interference. When the data is input into the ABAQUS elastic-plastic analysis model, the model simulates the stress state of the workpiece based on the finite element method.

[0088] In one example, assuming the workpiece is a steel plate, the model, after inputting roughness data, calculates stress concentration areas, such as the peak stress of 60 MPa near the edge of the weld. This simulation clearly reflects the mechanical response within the material.

[0089] Preferably, the calculation accuracy can be improved by adjusting the grid density, for example, refining the grid to 0.5 mm level. When the stress distribution data is compared with the 50 MPa stress distribution threshold, high stress areas can be marked by color coding.

[0090] For example, if a workpiece's weld seam is displayed red, it means that the stress exceeds the standard.

[0091] It should be noted that this marking method facilitates quick location of problem areas.

[0092] In one possible implementation, if the percentage of areas exceeding the standard exceeds 10%, further analysis of the cause, such as uneven material thickness, is required. When calculating deformation resistance using the ANSYS welding simulation module, the simulation is based on the principle of thermal-structural coupling.

[0093] Specifically, the current welding density is input as 5 kilograms per cubic meter, and the calculation results show that the deformation resistance is 25 kilonewtons per millimeter, which is lower than the standard of 30 kilonewtons per millimeter.

[0094] It can be understood that this simulation can reflect the influence of welding parameters on structural strength.

[0095] For example, increasing the weld density to 6 kg / m³ increases the deformation resistance to 32 kN / mm. When calculating the weld density increment based on the material strength formula, the optimal value can be determined through iterative adjustment.

[0096] In one embodiment, the initial weld density increment was set at 0.5 kg / m³. After adjustment, the stress distribution became more uniform, for example, the maximum stress was reduced from 60 MPa to 48 MPa. This approach helps optimize process parameters. When the adjusted weld density was input into ABAQUS and recalculated, the stress distribution became more stable.

[0097] For example, the stress value in a certain area dropped from 55 MPa to 45 MPa, indicating that the adjustment was effective.

[0098] Preferably, multiple rounds of validation are performed to ensure stable results. When extracting stress values ​​and new weld density as features and inputting them into the random forest algorithm, the classifier is trained based on historical data. For example, samples with a deformation resistance below 30 kilonewtons per millimeter are labeled as 0. The classification results intuitively reflect regional performance. After the classifier outputs substandard regions, the weld density is updated in increments of 10% of the weld density increment, for example, by 0.05 kilograms per cubic meter.

[0099] For example, after three iterations, the deformation resistance of a certain area increased from 28 kN / mm to 31 kN / mm. This iterative approach improves process adaptability. Finally, using MATLAB fitting tools to plot the strength variation curve, the curve shows that the strength gradually flattens with the number of iterations.

[0100] For example, the intensity of a certain region stabilized at 33 kN / mm after the fourth iteration, exceeding the preset regional intensity threshold. This visualization method facilitates the evaluation of optimization results.

[0101] Understandably, the curve can also reveal potential process bottlenecks, such as slow strength improvement in a certain section, which may be related to material properties.

[0102] S6, based on the optimized strength distribution data, uses the resource allocation algorithm to analyze the impact of welding depth and density on production efficiency, extracts the parameter redundancy value of the redundant welding area, and determines the quantitative degree of resource waste.

[0103] Optionally, this step also includes: In step S61, a preset parsing module is loaded with the intensity distribution data, and principal component analysis is used to decompose the data to obtain optimized data and distribution data. In step S62, based on the optimized data and distribution data, the simplex method is used to calculate the influence coefficients of weld depth and weld density to obtain the changing trend of production efficiency. In step S63, based on the changing trend of production efficiency, parameter redundancy values ​​are extracted and the boundary range of the redundant area is determined. In step S64, the initial quantitative value of resource waste is calculated based on the boundary range of the redundant area to obtain the adjustment requirements for resource allocation. In step S65, if the initial quantitative value of resource waste exceeds the preset waste quantification threshold, the least squares method is used to optimize resource allocation to obtain an adjusted allocation plan. In step S66, based on the adjusted allocation plan, the distribution data of weld depth and weld density are updated to determine the quantitative degree of resource waste. In step S67, the final resource allocation result is determined by comparing the quantitative degree of resource waste with the initial quantitative value.

[0104] Specifically, the preset analytical module is loaded through the intensity distribution data, and the principal component analysis is used to decompose the optimized data and the distribution data.

[0105] It is understandable that principal component analysis can reduce the dimensionality of complex multidimensional data and extract key features.

[0106] In one embodiment, assuming that the workpiece is a steel weld component, the strength distribution data includes stress values ​​and deformation amounts.

[0107] For example, principal component analysis decomposes this data into key influencing factors, such as areas of stress concentration and uniform distribution, generating optimized characteristic data. This approach facilitates subsequent analysis and avoids interference from redundant information. Based on the optimized and distributed data, the simplex method is used to calculate the influencing factors of weld depth and weld density. Once the production efficiency trend is determined, the simplex method uses multivariate optimization to find the optimal solution.

[0108] Specifically, in the welding seam processing of automotive parts, the welding depth is set to 3 mm and the density is 5 kg per cubic meter. After iterative calculation using the simplex method, it is shown that when the depth is increased to 4 mm, the production efficiency is improved by about 15%.

[0109] Preferably, this method can quickly reflect the impact of parameter adjustment on production efficiency.

[0110] It should be noted that the influence coefficient can also reveal that welding depth is more sensitive to production efficiency. According to the changing trend of production efficiency, when extracting parameter redundancy values ​​and determining the boundary range of the redundant area, redundancy can be located through statistical analysis.

[0111] For example, the efficiency of a certain weld area is improved by only 2%, which is far lower than the average efficiency of the weld area of ​​10%. Its boundary range can be defined as the area with stress below 20 MPa.

[0112] In one possible implementation, these boundaries can be visualized using a heat map to help visually identify areas where parameter adjustments are less meaningful. This boundary demarcation helps focus on key areas.

[0113] By calculating the initial quantitative value of resource waste based on the boundary range of the redundant area and obtaining the adjustment requirements for resource allocation, for example, assuming that the redundant area accounts for 15% of the total area, the quantitative value of resource waste is 8 units, which exceeds the preset waste quantitative threshold of 5 units.

[0114] It should be noted that the quantified value of resource waste is calculated based on the cost of welding materials and labor hours, reflecting the actual degree of waste. If it exceeds the waste quantification threshold, the least squares method is used to optimize resource allocation to obtain the adjusted allocation plan.

[0115] Specifically, the least squares method adjusts parameters through fitting to reduce the waste quantification value to 3 units. This optimization can balance resource utilization efficiency.

[0116] According to the adjusted allocation plan, the distribution data of welding depth and welding density are updated to judge the quantitative degree of resource waste. In one embodiment, the welding depth is adjusted to 3.5 mm, the welding density is reduced to 4.8 kg per cubic meter, and the waste quantification value is reduced from 8 to 4.

[0117] Preferably, such an update can make resource distribution more reasonable.

[0118] For example, by comparing the initial quantitative value of resource waste of 10 units and the adjusted value of 4 units, the resource waste is significantly reduced.

[0119] It is understandable that this comparison intuitively reflects the effectiveness of the adjustment.

[0120] By comparing the quantified degree of resource waste with the initial quantified value, the final resource allocation result is determined. For example, the final solution stabilizes the welding depth at 3.6 mm, the welding density at 4.9 kg per cubic meter, and the waste quantification value is controlled within 3 units.

[0121] In one possible implementation, multiple rounds of validation show consistent results, proving the stability of the solution. This approach ensures accurate resource allocation and avoids over-investment.

[0122] S7, after obtaining the quantitative degree of resource waste, adjusts the welding depth and density of the low-stress area through the parameter optimization algorithm, and iteratively calculates the balance point between appearance quality and production efficiency to obtain the final differentiated welding solution.

[0123] Optionally, this step also includes: In step S71, quantitative data on resource waste is obtained from production data. The degree of waste is calculated using descriptive statistics to obtain an initial assessment result. In step S72, based on the initial assessment result, the weld depth and weld density in low-stress areas are adjusted using a gradient descent algorithm to obtain a preliminary adjustment plan. In step S73, for the preliminary adjustment plan, the weld surface roughness and weld width are calculated as indicators of appearance quality. A determination is made as to whether they meet a preset appearance threshold to obtain quality assessment data. In step S74, the weld length per unit time is calculated based on the quality assessment data as an indicator of production efficiency. If it is below the production efficiency threshold, the weld density is adjusted to obtain an efficiency optimization result. In step S75, the efficiency optimization result is obtained and, combined with the appearance quality indicator, the Newton iteration method is used to determine the parameter combination of the equilibrium point to obtain a set of candidate solutions. In step S76, the K-means clustering algorithm is used to analyze the differentiated features of the candidate solutions to obtain a final differentiated welding solution. In step S77, the quantitative data on resource waste before and after optimization are compared using the final differentiated welding solution to verify the downward trend and obtain the optimized data distribution.

[0124] Specifically, when obtaining quantitative data on resource waste from production data, descriptive statistical methods can effectively calculate the degree of waste.

[0125] For example, on a welding production line, data on wire usage, power consumption, and working hours can be collected over a week, and statistics such as mean and standard deviation can be calculated to generate an initial assessment result. Based on this initial assessment result, a gradient descent algorithm can be used to adjust welding parameters in low-stress areas.

[0126] In one possible implementation, welding depth and welding density are used as variables, resource waste is used as the objective function, and the optimal solution is found through iterative calculation.

[0127] For example, the initial welding depth of a low-stress area is 4 mm and the welding density is 6 kg per cubic meter. After algorithm adjustment, the welding depth is reduced to 3.5 mm and the welding density is reduced to 5.5 kg per cubic meter, forming a preliminary adjustment plan.

[0128] It should be noted that weld surface roughness and weld width are important appearance quality indicators.

[0129] Specifically, a laser scanner is used to measure the weld surface and calculate the average roughness value. Image processing technology is also used to measure the weld width. If the roughness is less than 5 microns and the width is between 4 and 6 mm, it is considered to meet the preset appearance index threshold. Based on this quality assessment data, the weld length per unit time is calculated as a production efficiency indicator.

[0130] Preferably, if the production efficiency is lower than a preset production efficiency threshold of 20 cm per minute, the welding density needs to be adjusted.

[0131] For example, increasing the density from 5.5 kg per cubic meter to 5.8 kg per cubic meter increases production efficiency to 22 cm per minute.

[0132] In one embodiment, Newton's method is used to determine the parameter combination of the equilibrium point.

[0133] As you can see, this method quickly converges to the optimal solution. Through multiple iterations, candidate solutions were obtained, including a weld depth of 3.6 mm and a weld density of 5.7 kg per cubic meter. Finally, the K-means clustering algorithm was used to analyze the differentiated characteristics of the candidate solutions.

[0134] For example, the solutions were categorized into three groups: high efficiency and low quality, medium efficiency and medium quality, and low efficiency and high quality. The most suitable differentiated welding solution was then selected from these groups. By comparing the data before and after optimization, the downward trend in resource waste was verified, resulting in the optimized data distribution.

[0135] Optionally, the step S76, analyzing the differential features from the candidate solution set using a K-means clustering algorithm to obtain a final differentiated welding solution, further includes: Step S761, obtaining a differential feature data set including current, voltage, and speed; Step S762, using the Z-score standardization method to standardize the differential feature data set to obtain a standardized feature data set; Step S763, using the K-means clustering algorithm to cluster the standardized feature data set to obtain a preliminary grouping result, wherein the preliminary grouping result includes at least two clusters; Step S764, obtaining the mean feature vector of each cluster in the preliminary grouping result; Step S765, calculating the Euclidean distance between the mean feature vectors, and if the Euclidean distance is greater than the Euclidean distance threshold, adjusting the cluster center and re-clustering until the Euclidean distance is greater than the Euclidean distance threshold. The distance is less than or equal to the Euclidean distance threshold, and an optimized grouping result is obtained; in step S766, a welding scheme parameter set corresponding to each cluster in the optimized grouping result is obtained; in step S767, a Pearson correlation coefficient matrix is ​​used to analyze the correlation between the welding scheme parameter set and the welding quality, and a welding scheme parameter subset with a correlation coefficient greater than a correlation threshold is obtained; in step S768, a differentiated welding scheme is generated according to the welding scheme parameter subset; in step S769, current, voltage, and speed parameter distribution data of each scheme in the differentiated welding scheme are obtained; in step S7610, a box plot is used to analyze the discreteness of the parameter distribution data, and a scheme distribution feature is obtained.

[0136] Specifically, when obtaining candidate solutions from a solution set, differentiated feature data can be collected through data extraction.

[0137] For example, on a welding production line, for different candidate solutions, the specific values ​​of current, voltage, and welding speed are recorded. Assuming that the current of a certain solution is 200 amperes, the voltage is 25 volts, and the welding speed is 25 centimeters per minute, an initial differentiated feature data set is formed.

[0138] In one possible implementation, if multiple sets of data are collected—for example, a current of 210 amps, a voltage of 24 volts, and a speed of 23 cm—they can be organized into a data table containing multidimensional features, laying the foundation for subsequent analysis. When using the Z-score normalization method to process datasets with differentiated features, the goal is to eliminate dimensional differences and make different parameters comparable.

[0139] Specifically, for the features such as current, voltage, and speed, their respective means and standard deviations are first calculated, and then the raw data are converted into standardized values.

[0140] For example, if the current mean is 205 amps and the standard deviation is 5, a standardized value of 200 amps can be obtained through simple calculation. This method effectively highlights the relative distribution characteristics of the data, facilitating subsequent clustering. When using the K-means clustering algorithm to group the standardized feature data, the number of clusters can be set to 3, each representing a different combination of welding characteristics.

[0141] For example, after running the algorithm, a set of cluster mean feature vectors may be obtained, such as cluster 1 with high current and high speed, cluster 2 with high voltage and moderate speed, and cluster 3 with low current and slow speed.

[0142] In one embodiment, the mean of cluster 1 may be 215 amps, 23 volts, and 26 cm / s. The corresponding welding plan parameter set can be directly extracted. This grouping can quickly reflect the differentiated features between plans.

[0143] When calculating the Euclidean distance between each parameter set, preferably, if the distance between cluster 1 and cluster 2 exceeds a preset Euclidean distance threshold, such as 5 units, it means that the grouping is not compact enough and the cluster center needs to be adjusted and regrouped.

[0144] It is understandable that this dynamic adjustment can improve the rationality of grouping.

[0145] For example, after adjustment, the mean of cluster 1 might become 210 amps, 24 volts, and 25 cm / s. This optimized grouping result is closer to actual needs. After extracting the eigenvectors of the final solution from the optimized grouping results, the Pearson correlation coefficient matrix was used to analyze its correlation with welding quality.

[0146] Specifically, assuming that welding quality is measured by the weld strength index, the calculation shows that the correlation coefficient between current and strength is 0.75, which exceeds the correlation threshold of 0.7, so the high current solution is retained first.

[0147] For example, the matching parameter combination might be a current of 212 amps, a voltage of 23 volts, and a speed of 24 cm. This analysis can intuitively reveal which features have the greatest impact on quality. When extracting parameter distribution data from differentiated welding solutions, using a boxplot to analyze the degree of dispersion is an intuitive method.

[0148] For example, for current data, a box plot may show a median of 208 amps and an interquartile range of 5 amps, indicating that the current distribution is relatively concentrated.

[0149] In one possible implementation, if the speed box plot shows too many outliers, such as a solution with a speed of 30 cm, the solution can be eliminated to ensure stability. This method helps to screen out more reliable parameter combinations and improve the practicality of the solution.

[0150] S8, for the final differentiated welding solution, uses digital twin technology to simulate the door frame's anti-deformation ability and structural stability in a complex stress environment. Verification data on performance distribution is extracted from the simulation results to determine the applicability of the solution.

[0151] Optionally, this step also includes: Step S81: Use ANSYS to build a door frame model, simulate the behavior under a complex stress environment, and obtain preliminary data on deformation resistance and structural stability. Step S82: Use MATLAB to extract performance distribution characteristics from the simulation results and determine the quantitative representation of the preliminary data. Step S83: If the performance distribution characteristics exceed the preset performance distribution threshold, adjust the welding scheme parameters and generate an updated door frame model. Step S84: Use ABAQUS to simulate the response under the stress environment for the updated door frame model to obtain new performance distribution data. Step S85: Extract verification data from the new performance distribution data to determine whether the adjusted welding scheme meets the structural stability requirements. Step S86: Compare the verification data with the scheme applicability standards to determine the final differentiated welding scheme. Step S87: If the verification data of the final scheme does not meet the preset standards, repeat the simulation and adjustment process to obtain the optimized scheme applicability conclusion.

[0152] Specifically, when building a door frame model using ANSYS, finite element analysis can be used to simulate its behavior in complex stress environments.

[0153] For example, assuming a door frame is subjected to vertical loads and lateral wind pressure, the software can set material properties, such as the steel's Young's modulus and Poisson's ratio, before meshing and applying boundary conditions. Preliminary simulation data might reveal a 2 mm deformation at the top of the door frame and a maximum stress of 150 MPa in the stress concentration area at the bottom. This data provides the foundation for subsequent analysis. When extracting performance distribution characteristics from simulation results using MATLAB, scripts can be used to process the stress and deformation data output by ANSYS.

[0154] Specifically, thermal maps of stress distribution and deformation curves can be generated.

[0155] For example, the heat map shows low stress in the middle of the door frame, with an average value of 80 MPa, while the curve graph indicates a nonlinear distribution of deformation along the height. This quantitative representation provides an intuitive understanding of the door frame's performance characteristics. If the performance distribution exceeds a pre-defined threshold, such as a maximum deformation exceeding 1.5 mm or a stress exceeding 120 MPa, the welding plan parameters need to be adjusted.

[0156] One possible implementation approach is to increase the number of welds or adjust their locations, such as adding a transverse weld in the central, low-stress area. This adjustment generates an updated door frame model for further verification. When using ABAQUS to simulate the response of the updated door frame model under stress conditions, the loading conditions can be re-set and the analysis can be run.

[0157] For example, the new model shows that the top deformation is reduced to 1.2 mm and the maximum stress is reduced to 110 MPa.

[0158] It should be noted that ABAQUS's high-precision meshing can more accurately capture stress changes in welds, thereby providing reliable performance distribution data. When extracting verification data from the new performance distribution data, you can focus on key indicators such as deformation and stress values.

[0159] Preferably, if the verification data shows that the deformation of all areas is less than 1.5 mm and the stress is less than 120 MPa, the adjusted welding scheme preliminarily meets the structural stability requirements.

[0160] For example, the data may show that the stress in the middle has dropped to 75 MPa, proving that the adjustment is effective. When comparing the validation data with the solution suitability criteria, multi-dimensional criteria can be set, such as stability, fatigue resistance, and cost control.

[0161] In one embodiment, if the standard requires deformation below 1.3 mm and stress below 115 MPa, the current solution may need to be fine-tuned.

[0162] This comparison ensures that the final differentiated welding solution balances performance and practicality. If the verification data of the final solution does not meet the requirements, for example, the deformation is still 1.4 mm, the simulation and adjustment process will be repeated.

[0163] For example, the weld layout can be further optimized, such as increasing the weld width from 5 mm to 6 mm, and re-simulating to confirm that the deformation is reduced to 1.2 mm. This iterative process gradually approaches the optimal solution, ensuring the applicability of the solution.

[0164] For example, starting from the core solution, you can first adjust the weld position and then expand to optimize the weld width and number.

[0165] In one embodiment, the initial solution is a single longitudinal weld, and the expanded solution includes transverse auxiliary welds, ultimately forming a differentiated solution with multiple welds working together. This progressive logic ensures solution integrity and improves reliability through multi-faceted verification.

[0166] S9, through the verification data of performance distribution, obtains the deviation characteristics that may occur in the actual welding process, and uses the adaptive adjustment algorithm to dynamically correct the welding parameter configuration table to obtain a process parameter set suitable for diverse scenarios.

[0167] Optionally, this step also includes: Step S91, obtain the real-time sensor data stream of welding current, voltage and speed, and transmit the data stream to the Kafka message queue in JSON format; Step S92, perform sliding window mean filtering on the data stream and remove abnormal data points exceeding 3 times the standard deviation to obtain preprocessed data; Step S93, use OpenCV to perform edge detection on the filtered weld image, extract the weld width and depth pixel values, and convert them into actual millimeter units through a preset calibration plate to obtain the weld size; Step S94, input the preprocessed data and the weld size into a pre-trained random forest regression model model, and obtain the predicted value of weld strength; in step S95, if the predicted value of weld strength is lower than the weld strength threshold of 250MPa, trigger the PID controller to calculate the correction amount of welding current and speed according to the preset correction amount formula, wherein the current correction amount is equal to the integral of the weld strength error multiplied by the coefficient 0.2; in step S96, update the welding parameter configuration table preset in the MySQL database according to the correction amount to obtain the corrected welding parameters; in step S97, query the appropriate plate parameter combination from the welding parameter configuration table, obtain the corresponding welding current, voltage and speed parameters, and output them as the corrected welding parameters.

[0168] Specifically, when obtaining real-time data streams of welding current, voltage, and speed through sensors, high-frequency sensors can be placed on the welding equipment to ensure real-time data collection.

[0169] For example, a sensor collects data 100 times per second, generating JSON-formatted data containing a timestamp, current values ​​such as 180 amps, voltage values ​​such as 22 volts, and speed values ​​such as 45 centimeters per minute. This data is transmitted through a Kafka message queue, and the partitioning design of the queue ensures high throughput and low latency of the data stream.

[0170] Specifically, independent partitions can be assigned to different welding stations to avoid data congestion.

[0171] In one possible implementation, when using Python's Pandas library to perform sliding window mean filtering on the data stream, the window size can be set to 10 data points, and the local mean can be calculated to smooth the noise.

[0172] For example, a segment of current data fluctuates between 175 and 185 amps, then stabilizes at around 180 amps after mean filtering. When removing data points exceeding three standard deviations, outliers reaching 200 amps, far exceeding the normal fluctuation range, are automatically removed. This method effectively eliminates abnormal interference and ensures the reliability of subsequent analysis. When edge detection is performed on the filtered weld image using OpenCV, the Canny algorithm can be used to identify the weld boundary.

[0173] For example, the image shows that the weld width pixel value is 50 pixels, which corresponds to the actual width of 5 mm after conversion by the calibration plate, and the weld depth pixel value of 20 corresponds to 2 mm.

[0174] It's important to note that the accuracy of the calibration plate directly impacts the conversion results. A high-resolution camera is preferred for improved edge detection accuracy. This quantification method provides an intuitive basis for subsequent feature extraction. When inputting feature data into the Scikit-learn random forest regression model, the model, trained on historical data, comprehensively considers the relationships between multiple variables, such as current, voltage, and weld width.

[0175] For example, a set of data is input: current 180 amperes, weld width 5 mm, and the model predicts that the weld strength is 260 MPa.

[0176] Preferably, if the predicted value is lower than the preset weld strength threshold of 250 MPa, such as 240 MPa, it indicates that the current parameters need to be adjusted. This prediction mechanism provides data support for dynamic optimization.

[0177] In one embodiment, if the predicted intensity is insufficient, when the PID controller is triggered to calculate the correction amount, the current may be estimated to be increased by 2 amperes based on an intensity error of, for example, -10 MPa and a coefficient of 0.2.

[0178] For example, if the current is adjusted from 180 amps to 182 amps, the adjusted parameters are updated to the configuration table in the MySQL database, overwriting the original process parameter set. This real-time feedback enables rapid response to process requirements.

[0179] Specifically, when selecting a parameter combination from the configuration table through SQL query, the configuration of 180 amps, 22 volts, and 45 cm / min can be retrieved based on the plate thickness of 5 mm and the welding wire diameter of 1.2 mm.

[0180] Understandably, this query logic facilitates rapid matching of working conditions, avoiding delays caused by manual intervention. While the core solution initially relies on a single sensor to collect data, expansion options can incorporate multiple sensors working together, such as adding a temperature sensor to monitor the heat-affected zone of a weld, generating a more comprehensive data stream.

[0181] In one embodiment, after incorporating temperature data, the model prediction accuracy is improved, and the strength prediction deviation is reduced to within 5 MPa. This progressive optimization ensures the comprehensiveness and applicability of process parameters.

[0182] like Figure 2 As shown, the second aspect of the present invention provides a differentiated welding system for aluminum alloy door frames, which uses the above-mentioned method to perform differentiated welding on aluminum alloy door frames. The system includes: The stress analysis module is used to obtain stress distribution data of aluminum alloy door frames under different usage scenarios through finite element analysis. The module determines the initial thresholds of welding depth and welding density for high-stress areas and low-stress areas, respectively, to obtain a quantitative model of stress distribution. The parameter calculation module is used to extract the specific requirements of local strength from high-stress areas based on the quantitative model of stress distribution, calculate the matching values ​​of welding depth and width using the laser welding parameter adjustment algorithm, and determine the welding parameter configuration table for each part; The real-time control module is used to obtain the welding parameter configuration table and detect the actual values ​​of welding depth and density during the welding process through the real-time monitoring system. If the actual values ​​deviate from the set values ​​in the configuration table, the laser power and speed are adjusted through the feedback control algorithm to obtain a stable welding output result; The quality inspection module is used to detect the deviation of weld width and appearance quality based on the stable welding output results using optical scanning technology, extract surface consistency indicators from the scan data, and determine whether the preset appearance quality threshold is met; The strength optimization module is used to obtain the distribution characteristics of local strength and deformation resistance by comparing the surface consistency index with the stress distribution model. If the deformation resistance of the high-stress area is lower than the functional requirement, the welding density is increased through the local repair welding algorithm to obtain the optimized strength distribution data; The resource assessment module is used to analyze the impact of welding depth and density on production efficiency based on the optimized strength distribution data using a resource allocation algorithm, extract parameter redundancy values ​​of excess welding areas, and determine the quantitative degree of resource waste; The parameter balancing module is used to obtain the quantified degree of resource waste, then adjust the welding depth and density in low-stress areas through a parameter optimization algorithm. It iteratively calculates the balance point between appearance quality and production efficiency to obtain the final differentiated welding solution. The performance verification module uses digital twin technology to simulate the door frame's deformation resistance and structural stability under complex stress environments for the final differentiated welding solution. It then extracts performance distribution verification data from the simulation results to determine the solution's applicability. The dynamic correction module is used to obtain the deviation characteristics that may occur in the actual welding process through the verification data of the performance distribution, and dynamically correct the welding parameter configuration table using an adaptive adjustment algorithm to obtain a set of process parameters suitable for diverse scenarios.

[0183] The above embodiment is only one of the preferred implementation methods of the present invention and should not be used to limit the scope of protection of the present invention. Any changes or modifications that have no substantive meaning made to the main design concept and spirit of the present invention, as long as the technical problems they solve are still consistent with the present invention, should be included in the scope of protection of the present invention.

Claims

1. A differentiated welding method for aluminum alloy door frames, characterized in that: The method comprises: S1, using finite element analysis to obtain stress distribution data for aluminum alloy door frames under different usage scenarios, determine the initial thresholds of welding depth and welding density for high-stress areas and low-stress areas, and obtain a quantitative model of stress distribution; S2, based on the quantitative model of stress distribution, extracts the specific requirements of local strength from the high stress area, uses the laser welding parameter adjustment algorithm to calculate the matching values ​​of welding depth and width, and determines the welding parameter configuration table for each part; S3, after obtaining the welding parameter configuration table, the actual values ​​of welding depth and density are detected during the welding process through the real-time monitoring system. If the actual values ​​deviate from the set values ​​in the configuration table, the laser power and speed are adjusted through the feedback control algorithm to obtain a stable welding output result; S4, for stable welding output results, uses optical scanning technology to detect the deviation of weld width and appearance quality, extracts surface consistency indicators from the scan data, and determines whether it meets the preset appearance quality threshold; S5, by comparing the surface consistency index with the stress distribution model, obtain the distribution characteristics of local strength and deformation resistance. If the deformation resistance of the high stress area is lower than the functional requirement, the welding density is increased through the local repair welding algorithm to obtain the optimized strength distribution data; S6, based on the optimized strength distribution data, a resource allocation algorithm is used to analyze the impact of welding depth and density on production efficiency, from which the parameter redundancy values ​​of the redundant welding area are extracted to determine the quantitative degree of resource waste; S7, after obtaining the quantified degree of resource waste, adjusts the welding depth and density in the low-stress area through a parameter optimization algorithm, performs iterative calculations on the balance point between appearance quality and production efficiency, and obtains the final differentiated welding solution; S8: For the final differentiated welding solution, digital twin technology is used to simulate the door frame's deformation resistance and structural stability under complex stress environments. Verification data on performance distribution is extracted from the simulation results to determine the solution's applicability. S9, through the verification data of performance distribution, obtains the deviation characteristics that may occur in the actual welding process, and uses the adaptive adjustment algorithm to dynamically correct the welding parameter configuration table to obtain a process parameter set suitable for diverse scenarios.

2. The method according to claim 1, characterized in that Step S1, obtaining stress distribution data of the aluminum alloy door frame in different usage scenarios through finite element analysis, determining the initial thresholds of welding depth and welding density for high stress areas and low stress areas respectively, and obtaining a quantitative model of stress distribution, includes: Step S11, obtaining stress distribution data of the aluminum alloy door frame from different usage scenarios through finite element analysis to obtain preliminary stress distribution results; Step S12, dividing the high stress area and the low stress area according to the stress distribution result, and obtaining the stress distribution characteristics after the area division; Step S13, extracting stress distribution characteristics for the high stress area and the low stress area respectively, and determining the initial threshold range of welding depth and welding density; Step S14, using a support vector machine algorithm to classify the initial threshold ranges of welding depth and welding density to obtain adjusted welding depth and welding density thresholds; Step S15, using the adjusted welding depth and welding density threshold as input parameters, performing parameter mapping on the aluminum alloy door frame through finite element analysis to obtain stress distribution data after parameter optimization; Step S16, constructing a quantitative model of stress distribution based on the stress distribution data after parameter optimization; Step S17: Calculate the uniformity index of stress distribution through the quantitative model to determine the structural stability of the aluminum alloy door frame in different usage scenarios.

3. The method according to claim 2, characterized in that The step S17, calculating the uniformity index of stress distribution by using a quantitative model to determine the structural stability of the aluminum alloy door frame in different usage scenarios, includes: Step S171, obtaining original stress distribution data of the aluminum alloy door frame in different usage scenarios, wherein the original stress distribution data is obtained by performing mechanical simulation on the aluminum alloy door frame using a finite element analysis tool; Step S172, extracting distribution feature vectors from the original stress distribution data using principal component analysis; Step S173: input the distribution characteristic vector into a time series analysis tool, calculate the standard deviation of the stress distribution in each scenario, and use the standard deviation of the stress distribution as an indicator of stress distribution uniformity; Step S174, obtaining a 95th percentile threshold of historical data, and comparing the stress distribution uniformity index with the 95th percentile threshold; Step S175: If the stress distribution uniformity index exceeds the 95th percentile threshold, a gradient descent algorithm is called to adjust the weight coefficient of the distribution feature vector to obtain a corrected distribution feature vector; Step S176, re-inputting the corrected distribution characteristic vector into the finite element analysis tool to generate new stress distribution data; Step S177, calculating an updated stress distribution uniformity index based on the new stress distribution data; Step S178, using an allowable stress comparison method to determine structural stability, obtaining a maximum stress value of the new stress distribution data, and comparing the maximum stress value with a yield strength threshold; Step S179: If the maximum stress value exceeds the yield strength threshold, it is determined that the structure is unstable; Step S1710: for unstable structures, extract factors that are strongly correlated with the stress peak from the finite element analysis parameters using the Pearson correlation coefficient; Step S1711: Based on the strongly correlated factors, a stress weighting algorithm is used to locally enhance the distribution characteristics of the key areas; Step S1712: outputting the optimized stress distribution uniformity index and stress distribution stability determination result.

4. The method according to claim 3, characterized in that The yield strength threshold is 70% of the material's yield strength.

5. The method according to claim 1, characterized in that Step S5, by comparing the surface consistency index with the stress distribution model, obtains the distribution characteristics of local strength and deformation resistance. If the deformation resistance of the high stress area is lower than the functional requirement, the welding density is increased by a local repair welding algorithm to obtain optimized strength distribution data, including: Step S51, obtaining workpiece surface roughness data, inputting a pre-established elastic-plastic analysis model, and obtaining stress distribution data; Step S52, comparing the stress distribution data with a stress distribution threshold, and marking areas exceeding the stress distribution threshold as high stress areas; Step S53, for the high stress area, using a welding simulation module to calculate the anti-deformation capability under the current welding density; Step S54, if the anti-deformation capability is lower than a preset value, calculating the welding density increment according to a preset material strength formula to generate an adjusted welding density distribution; Step S55, inputting the adjusted welding density distribution into the elastic-plastic analysis model to recalculate the stress distribution; Step S56, extracting the stress value in the recalculated stress distribution and the adjusted welding density as features, and performing classification using a preset classification algorithm, wherein the classification label is whether the deformation resistance meets the standard; Step S57: if the classification result is not up to standard, the welding density of the high stress area is updated with a preset step size, and the welding simulation is repeated until the number of iterations reaches a preset value or the deformation resistance of the high stress area is all up to standard; Step S58: draw a curve showing how the intensity of each region changes with the number of iterations, and determine whether the final intensity of each region reaches the regional intensity threshold.

6. The method according to claim 5, characterized in that The preset step size is 10% of the welding density increment.

7. The method according to claim 1, characterized in that The step S6, based on the optimized strength distribution data, uses a resource allocation algorithm to analyze the impact of welding depth and density on production efficiency, extracts parameter redundancy values ​​of redundant welding areas, and determines the quantitative degree of resource waste, including: Step S61, loading a preset analysis module with the intensity distribution data, and using principal component analysis to decompose and obtain optimized data and distribution data; Step S62, using the simplex method to calculate the influence coefficients of welding depth and welding density based on the optimization data and distribution data, and obtaining the change trend of production efficiency; Step S63: extracting parameter redundancy values ​​based on the changing trend of production efficiency and determining the boundary range of the redundant area; Step S64, calculating the initial quantitative value of resource waste based on the boundary range of the redundant area, and obtaining the adjustment requirements for resource allocation; Step S65: If the initial quantified value of resource waste exceeds the waste quantification threshold, the least square method is used to optimize resource allocation to obtain an adjusted allocation plan; Step S66: updating the distribution data of welding depth and welding density according to the adjusted allocation scheme to determine the quantitative degree of resource waste; Step S67 : Determine the final resource allocation result by comparing the quantified degree of resource waste with the initial quantified value.

8. The method according to claim 1, characterized in that In step S7, after obtaining the quantitative degree of resource waste, the welding depth and density of the low-stress area are adjusted through a parameter optimization algorithm, and an iterative calculation is performed on the balance point between appearance quality and production efficiency to obtain the final differentiated welding solution, including: Step S71: obtaining quantitative data on resource waste from production data, calculating the degree of waste using descriptive statistics, and obtaining an initial assessment result; Step S72, based on the initial evaluation results, using a gradient descent algorithm to adjust the welding depth and welding density in the low stress area to obtain a preliminary adjustment plan; Step S73: Calculate the weld surface roughness and weld width as the appearance quality index values ​​for the preliminary adjustment plan, determine whether they meet the preset appearance index threshold, and obtain quality assessment data; Step S74: Calculate the welding length per unit time as the production efficiency index value based on the quality assessment data. If it is lower than the production efficiency threshold, adjust the welding density to obtain the efficiency optimization result. Step S75: Obtain the efficiency optimization result, combine it with the appearance quality index value, use the Newton iteration method to determine the parameter combination of the balance point, and obtain a set of candidate solutions; Step S76: Using the K-means clustering algorithm to analyze the differential features from the candidate solution set to obtain the final differentiated welding solution; In step S77 , the resource waste quantitative data before and after optimization are compared using the final differentiated welding solution to verify the downward trend and obtain the optimized data distribution.

9. The method according to claim 8, characterized in that The step S76, using the K-means clustering algorithm to analyze the differential features from the candidate solution set to obtain the final differentiated welding solution, includes: Step S761, obtaining a differentiated feature data set including current, voltage, and speed; Step S762, using a Z-score standardization method to standardize the differential feature dataset to obtain a standardized feature dataset; Step S763, clustering the standardized feature data set using a K-means clustering algorithm to obtain a preliminary grouping result, wherein the preliminary grouping result includes at least two clusters; Step S764, obtaining the mean feature vector of each cluster in the preliminary grouping result; Step S765, calculating the Euclidean distance between the mean eigenvectors. If the Euclidean distance is greater than the Euclidean distance threshold, adjusting the cluster center and re-clustering until the Euclidean distance is less than or equal to the Euclidean distance threshold, thereby obtaining an optimized grouping result. Step S766, obtaining a welding scheme parameter set corresponding to each cluster in the optimization grouping result; Step S767, using a Pearson correlation coefficient matrix to analyze the correlation between the welding scheme parameter set and the welding quality, and obtaining a welding scheme parameter subset having a correlation coefficient greater than a correlation threshold; Step S768, generating a differentiated welding plan according to the welding plan parameter subset; Step S769, obtaining the current, voltage, and speed parameter distribution data of each of the differentiated welding schemes; Step S7610: Use a box plot to analyze the discreteness of the parameter distribution data to obtain the solution distribution characteristics.

10. A differentiated welding system for aluminum alloy door frames, characterized in that: The aluminum alloy door frame is differentially welded using the method according to any one of claims 1 to 9, the system comprising: The stress analysis module is used to obtain stress distribution data of aluminum alloy door frames under different usage scenarios through finite element analysis. The module determines the initial thresholds of welding depth and welding density for high-stress areas and low-stress areas, respectively, to obtain a quantitative model of stress distribution. The parameter calculation module is used to extract the specific requirements of local strength from high-stress areas based on the quantitative model of stress distribution, calculate the matching values ​​of welding depth and width using the laser welding parameter adjustment algorithm, and determine the welding parameter configuration table for each part; The real-time control module is used to obtain the welding parameter configuration table and detect the actual values ​​of welding depth and density during the welding process through the real-time monitoring system. If the actual values ​​deviate from the set values ​​in the configuration table, the laser power and speed are adjusted through the feedback control algorithm to obtain a stable welding output result; The quality inspection module is used to detect the deviation of weld width and appearance quality based on the stable welding output results using optical scanning technology, extract surface consistency indicators from the scan data, and determine whether the preset appearance quality threshold is met; The strength optimization module is used to obtain the distribution characteristics of local strength and deformation resistance by comparing the surface consistency index with the stress distribution model. If the deformation resistance of the high-stress area is lower than the functional requirement, the welding density is increased through the local repair welding algorithm to obtain the optimized strength distribution data; The resource assessment module is used to analyze the impact of welding depth and density on production efficiency based on the optimized strength distribution data using a resource allocation algorithm, extract parameter redundancy values ​​of excess welding areas, and determine the quantitative degree of resource waste; The parameter balancing module is used to obtain the quantified degree of resource waste, then adjust the welding depth and density in low-stress areas through a parameter optimization algorithm. It iteratively calculates the balance point between appearance quality and production efficiency to obtain the final differentiated welding solution. The performance verification module uses digital twin technology to simulate the door frame's deformation resistance and structural stability under complex stress environments for the final differentiated welding solution. It then extracts performance distribution verification data from the simulation results to determine the solution's applicability. The dynamic correction module is used to obtain the deviation characteristics that may occur in the actual welding process through the verification data of the performance distribution, and dynamically correct the welding parameter configuration table using an adaptive adjustment algorithm to obtain a set of process parameters suitable for diverse scenarios.

Images