Welding seam morphology prediction and optimization method for laser arc overhead welding of medium-thickness plate
A mathematical model of weld morphology for medium-thick plate laser arc overhead welding was established using response surface methodology and analysis of variance. Significant influencing factors were screened and visualized. Welding parameters were optimized by combining non-dominated sorting genetic algorithm, which solved the stability and accuracy problems of weld morphology prediction and optimization in existing technologies, and improved welding efficiency and quality.
Patent Information
- Application Number
- CN202511830603.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-12-06
- Publication Date
- 2026-02-24
AI Technical Summary
Existing technologies make it difficult to establish stable and high-precision mathematical models to predict and optimize the weld morphology of medium and thick plates using laser arc overhead welding. This leads to the reliance on manual experience in setting welding parameters, making it impossible to achieve stable weld morphology. Furthermore, traditional methods are computationally complex and costly.
A mathematical model of weld morphology with respect to welding parameters was established using response surface methodology and analysis of variance. Significant influencing factors were screened, and visualization analysis was performed. Welding parameters were optimized using stepwise regression and multi-objective optimization was carried out by combining non-dominated sorting genetic algorithm. A high-precision method for predicting and optimizing weld morphology was established.
It enables high-precision weld morphology prediction and optimization with fewer experiments and lower costs, improving welding efficiency and quality, and guiding the automated and intelligent production of large-scale equipment manufacturing.
Smart Images
Figure CN121551843A_ABST
Abstract
Description
Technical Field
[0001] This invention belongs to the field of structural welding technology and relates to a method for predicting and optimizing the weld morphology of laser arc overhead welding, specifically a method for predicting and optimizing the weld morphology of medium and thick plates using laser arc overhead welding. Background Technology
[0002] Large-scale equipment manufacturing, including aerospace equipment and marine engineering equipment, is a key development area. Medium and heavy plates are widely used in this industry, with usage exceeding 80 million tons in 2020, accounting for over 60% of usage in the engineering machinery sector. Medium and heavy plates are primarily used in the production of load-bearing structural components such as ribs. However, due to their large size, complex structure, and various limitations imposed by construction conditions, actual operations inevitably face numerous non-flat welding requirements, such as overhead welding connections. In non-flat welding positions, the gravity of the molten pool can cause instability, leading to welding defects such as incomplete penetration. Laser-arc hybrid welding technology is a welding method with stable process parameters and strong process adaptability. However, there are coupling effects between welding parameters, and related research focuses on the influence of welding parameters on weld mechanical properties and molten pool flow patterns, failing to establish a mathematical model of the relationship between welding parameters and weld morphology. This results in welding parameter determination relying mainly on manual experience, lacking relevant mathematical models for guidance, and the inability to adaptively adjust welding parameters according to working conditions during the welding process to achieve a stable and ideal weld morphology.
[0003] Patent CN115048882A relates to a laser weld morphology prediction method. Although it can solve the weld morphology function at the next moment well based on classical physics theory and has good universality, the mechanism function expression has simplification assumptions in the derivation process, which cannot truly reflect the online real-time state of the weld. Moreover, the amount of computation is huge, and the calculation process is cumbersome and complicated, which is not conducive to engineering practice and application.
[0004] For example, patent CN117655605A proposes a conceptual method for weld control based on intelligent optimization algorithms. This method improves prediction accuracy by adjusting algorithm weights based on the comparison between sample results and predicted values. However, it only focuses on process design and does not propose specific prediction or optimization methods. Furthermore, it is not conducive to analyzing the influence of various welding parameters on weld morphology, nor does it consider the impact of assembly conditions on weld morphology, making it unsuitable for practical engineering applications.
[0005] Many existing technical solutions employ machine learning or orthogonal experimental design to establish mathematical models. However, due to the numerous parameters and coupling effects inherent in laser-arc hybrid welding technology, applying these methods significantly increases experimental costs. Therefore, obtaining a stable and highly accurate mathematical prediction optimization model while reducing model construction costs has significant practical guiding significance for the field of large-scale equipment manufacturing. Summary of the Invention
[0006] To address the aforementioned issues, this invention provides a method for predicting and optimizing the weld morphology of medium-thick plates using laser arc overhead welding. This method establishes a high-precision mathematical model of the weld morphology with a small sample size, enabling visualization of the influence of welding parameters and weld morphology, and optimization of welding parameters. This guides specific production processes, thereby improving manufacturing efficiency and product quality of medium-thick plates in the field of automated and intelligent manufacturing of large-scale equipment, significantly increasing overhead welding production efficiency and improving the quality of overhead welds.
[0007] The technical solution of the present invention is as follows: A method for predicting the weld morphology of laser arc overhead welding in medium and thick plates includes the following steps: S1, input welding parameters, including welding current, welding speed, laser peak power and butt gap, and determine the output parameters as back weld width and front weld height; S2, by controlling variables, experimentally determine the range of variation of a single input welding parameter; S3, Establish the experimental parameter matrix and establish a mathematical model of weld morphology with respect to welding parameters based on the response surface methodology; S4 verifies the mathematical model established in S3 based on the analysis of variance method, and selects the terms that have a significant impact on the weld morphology from single welding parameter terms and multiple coupled welding parameter terms. S5: After converting the items with significant influence obtained in S4 into coded values, visualize the degree of influence of the items, and analyze the influence weight of individual welding parameter items and the influence law of welding parameter coupling items.
[0008] Furthermore, in S2, the input welding parameter values, i.e., the range of input values, are divided into 5 levels, namely... -β , -1 , 0 , 1 and β ,in β The value depends on the type and number of input values.
[0009] in, k The number of input value variables.
[0010] Furthermore, in S3, the center point values and all other point values of the experimental parameter matrix are obtained according to the normalization formula of the response surface methodology. The formula is as follows:
[0011] in X For standard coded values, U The variable is the actual input value; U max The maximum value of the actual input value; U min This is the minimum value of the actual input.
[0012] Furthermore, in S3, a quadratic polynomial is used to fit the functional relationship between multiple input values and multiple output values, including:
[0013] in For the output value, x i For input values, ε This is the pure error of the system; Among them The transformed second-order polynomial regression equation is obtained by performing a power transformation. Y The result is:
[0014] in b o , b i , b ij , b ii In order, they are the constant term, the coefficient of the linear term, the coefficient of the interaction term, and the coefficient of the quadratic term.
[0015] Furthermore, in S3, based on fitting the functional relationship between multiple input values and multiple output values, a stepwise regression method is used to establish a regression model, and insignificant terms are eliminated until the significance and lack of fit of the model meet the requirements, thereby obtaining a mathematical model between the weld morphology and process parameters of laser arc overhead welding, including regression equations with coded values as input values and regression equations with actual values as input values.
[0016] Furthermore, in S5, a single welding parameter is used as the horizontal axis, and the back width and front clearance are used as the vertical axes to form a two-dimensional analysis diagram.
[0017] Furthermore, in S5, the welding parameter coupling terms are fitted into variation curves, with the back width and front clearance as the ordinates, respectively, to form a three-dimensional analysis diagram.
[0018] A method for optimizing the weld morphology of medium-thick plate laser arc overhead welding includes the following steps: [The method describes a method for predicting the weld morphology of medium-thick plate laser arc overhead welding, which optimizes the weld morphology obtained from the aforementioned method.] S6. With the maximum back weld width and minimum excess height as optimization objectives, within the range of parameter terms and parameter coupling terms obtained from the analysis in step S5, the mathematical model verified in S4 is optimized to obtain the optimal solution set after optimization. S7. Based on the optimized optimal solution set obtained in step S6, a verification experiment is conducted to analyze and compare the prediction accuracy and weld morphology of the optimized model.
[0019] The advantages of this invention are as follows: 1. The method for predicting and optimizing the weld morphology of medium-thick plate laser arc overhead welding proposed in this invention can establish a mathematical model of weld morphology with high accuracy and good stability regarding welding parameters.
[0020] 2. The overhead welding morphology prediction and optimization method proposed in this invention can visualize the influence of individual parameters and parameter coupling on weld morphology, and compare the influence weight of individual parameters on weld morphology and the influence of parameter coupling terms on weld morphology, so as to find the parameter optimization range.
[0021] 3. The overhead welding morphology prediction and optimization method proposed in this invention can obtain a high-precision prediction model with fewer experiments and lower experimental costs, effectively overcoming the problems of large experimental volume and high experimental cost of traditional modeling methods. Attached Figure Description
[0022] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. The drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0023] Figure 1 A flowchart illustrating the laser arc overhead welding weld morphology prediction and optimization method of the present invention.
[0024] Figure 2 This is a schematic diagram illustrating the stress analysis on the back side of the molten pool during laser arc overhead welding.
[0025] Figure 3 This is a comparison chart of the predicted and actual values of the response surface model obtained by applying the present invention in the implementation case.
[0026] Figure 4 The residual normal distribution diagram of the response surface model obtained by applying the present invention in the implementation case.
[0027] Figure 5 The diagram illustrates the influence of a single parameter based on the encoding value on the back-side weld width, obtained from the implementation examples of this invention.
[0028] Figure 6 The diagram illustrates the influence of a single parameter based on the encoded value on the remaining height, obtained from the implementation examples of this invention.
[0029] Figure 7 This is a schematic diagram illustrating the coupled effect of welding speed and butt joint gap on the back weld width obtained from the implementation examples of this invention.
[0030] Figure 8 The coupling effect of welding current and laser power on back weld width obtained from the implementation examples of this invention.
[0031] Figure 9 This is a schematic diagram illustrating the coupling effect of welding current and butt joint gap on the back weld width obtained from the implementation examples of this invention.
[0032] Figure 10 This is a schematic diagram illustrating the coupling effect of welding current and welding speed on the weld height obtained from the implementation examples of this invention.
[0033] Figure 11 This is a schematic diagram illustrating the coupling effect of welding current and butt gap on the reinforcement height obtained from the implementation examples of this invention.
[0034] Figure 12 A schematic diagram of the Pareto front of the optimized back weld width and excess height obtained by applying the embodiments of the present invention.
[0035] Figure 13 A schematic diagram of the cross-sectional morphology of the weld obtained from the implementation of this invention. Detailed Implementation
[0036] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.
[0037] The features and illustrative embodiments of various aspects of the present invention will now be described in detail. Numerous specific details are set forth in the following detailed description to provide a thorough understanding of the invention. However, it will be apparent to those skilled in the art that the invention may be practiced without requiring some of these specific details. The following description of embodiments is merely intended to provide a better understanding of the invention by illustrating examples of the invention. The invention is by no means limited to any specific setups and methods set forth below, but covers any improvements, substitutions, and modifications to structures, methods, and devices without departing from the spirit of the invention. Well-known structures and techniques are not shown in the drawings and the following description to avoid unnecessarily obscuring the invention.
[0038] In the description of this invention, it should be noted that the directions or positional relationships indicated by terms such as "center," "upper," "lower," "left," "right," "vertical," "horizontal," "inner," and "outer" are based on the directions or positional relationships shown in the accompanying drawings and are only for the convenience of describing and simplifying the invention, and should not be construed as limiting the invention. Furthermore, the use of ordinal numbers (e.g., "first and second," etc.) is for distinguishing objects and is not limited to this order, and should not be construed as indicating or implying relative importance.
[0039] In the description of this invention, it should be noted that, unless otherwise explicitly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly, encompassing both direct connection and indirect connection via an intermediate medium. Those skilled in the art can understand the specific meaning of these terms in this invention based on the specific circumstances.
[0040] It should be noted that, unless otherwise specified, the embodiments of the present invention and the features thereof can be combined with each other, and the various embodiments can be referenced and cited in each other. The present invention will now be described in detail with reference to the accompanying drawings and embodiments.
[0041] Example 1: A method for predicting the weld morphology of laser arc overhead welding in medium and thick plates includes the following steps: S1, input welding parameters, including welding current, welding speed, laser peak power and butt gap, and determine the output parameters as back weld width and front weld height; S2, by controlling variables, experimentally determine the range of variation of a single input welding parameter; S3, Establish the experimental parameter matrix and establish a mathematical model of weld morphology with respect to welding parameters based on the response surface methodology; S4 verifies the mathematical model established in S3 based on the analysis of variance method, and selects the terms that have a significant impact on the weld morphology from single welding parameter terms and multiple coupled welding parameter terms. S5: After converting the items with significant influence obtained in S4 into coded values, visualize the degree of influence of the items, and analyze the influence weight of individual welding parameter items and the influence law of welding parameter coupling items.
[0042] In S2, the input welding parameter values, i.e., the range of input values, are divided into 5 levels, namely... -β , -1 , 0 , 1 and β ,in β The value depends on the type and number of input values.
[0043] in, k The number of input value variables.
[0044] In S3, the center point values and all other point values of the experimental parameter matrix are obtained by applying the standardization formula of the response surface methodology. The formula is as follows:
[0045] in X For standard coded values, U The variable is the actual input value; U max The maximum value of the actual input value; U min This is the minimum value of the actual input.
[0046] In S3, a quadratic polynomial is used to fit the functional relationship between multiple input values and multiple output values, including:
[0047] in For the output value, x i For input values, ε This is the pure error of the system; Among them The transformed second-order polynomial regression equation is obtained by performing a power transformation. Y The result is:
[0048] in b o , b i , b ij , b ii In order, they are the constant term, the coefficient of the linear term, the coefficient of the interaction term, and the coefficient of the quadratic term.
[0049] In S3, based on fitting the functional relationship between multiple input values and multiple output values, a regression model is established using the stepwise regression method. Insignificant terms are removed until the significance and lack of fit of the model meet the requirements, thus obtaining a mathematical model between the weld morphology and process parameters of laser arc overhead welding, including regression equations with coded values as input values and regression equations with actual values as input values.
[0050] In S5, a single welding parameter is used as the horizontal axis, and the back width and front clearance are used as the vertical axes to form a two-dimensional analysis diagram.
[0051] In S5, the welding parameter coupling terms are fitted into a variation curve, and the back width and front clearance are used as the ordinates to form a three-dimensional analysis diagram.
[0052] A method for optimizing the weld morphology of medium-thick plate laser arc overhead welding includes the following steps: [The method describes a method for predicting the weld morphology of medium-thick plate laser arc overhead welding, which optimizes the weld morphology obtained from the aforementioned method.] S6. With the maximum back weld width and minimum excess height as optimization objectives, within the range of parameter terms and parameter coupling terms obtained from the analysis in step S5, the mathematical model verified in S4 is optimized to obtain the optimal solution set after optimization. S7. Based on the optimized optimal solution set obtained in step S6, conduct verification experiments to analyze and compare the prediction accuracy and weld morphology of the optimized model. Example 2: S1: Based on the stress analysis results of the molten pool during the weld formation process and relevant literature, the input parameters include welding current, welding speed, laser peak power, and butt gap. Based on relevant literature on the mechanical properties of weld morphology, the output parameters include back weld width and front weld reinforcement. S2: Based on the premise that the weld formation is good and defect-free, a preliminary experiment is conducted using the controlled variable method to determine the variation range of a single input parameter; S3: Establish the experimental parameter matrix based on the central composite design method to build a mathematical model of weld morphology with respect to welding parameters based on the response surface methodology; S4: The established mathematical model was verified based on the analysis of variance method, and the terms that have a significant impact on the weld morphology were selected from the single parameter terms and the parameter coupling terms. S5: Visualize the influence of individual parameter items and parameter coupling items based on the encoded values on the weld morphology, analyze the influence weight of individual parameter items and the influence of parameter coupling items, and analyze the range of variation of the parameter to be optimized based on the visualization results of the influence of parameters on the weld morphology. S6: Taking the maximum back weld width and minimum excess height as optimization objectives, within the parameter range obtained from the analysis in step S5, the second-generation non-dominated sorting genetic algorithm is used to perform multi-objective optimization on the mathematical model established in step S4 to obtain the optimal solution set after optimization.
[0053] S7: Based on the optimized optimal solution set obtained in step S6, conduct verification experiments to analyze and compare the prediction accuracy and weld morphology of the optimized model.
[0054] Example 3: This invention proposes a method for predicting and optimizing the weld morphology of laser-arc hybrid overhead welding for medium-thick plates. Taking the prediction and optimization of process parameters for laser-arc hybrid overhead welding of 4mm thick X70 pipeline steel as an example, the specific steps of this example are further explained below: Step 1: Based on the process described in step S1, analyze the weld formation process. During the laser-arc hybrid overhead welding process, the stress on the back molten pool is as follows: Figure 1 As shown, the molten pool only exists under the influence of the electric arc force. F a ,gravity F g Solid-liquid surface tension F α Laser impact force F l Achieving proper forming requires a balance of forces. Among these, laser impact force... F l Positively correlated with peak laser power; gravity F g Determined by the volume of the molten pool; solid-liquid surface tension F α It is a function of temperature; electric arc force F a It is mainly affected by the welding current and the volume of the molten pool, as shown in the following formula.
[0055] in I It is the welding current. R It is the volume of the molten pool. μ The value is constant. The volume of the molten pool is mainly affected by the butt joint gap and welding speed. According to the stress analysis results of the molten pool during the welding process, the welding current, welding speed, laser power, and butt joint gap welding parameters all have a significant impact on the weld morphology.
[0056] Based on relevant literature, two weld morphology features that can reflect weld quality were selected: back width (BW) and front reinforcement (WR). Based on the stress analysis of the molten pool during weld formation, four welding process parameters that significantly affect weld morphology were selected: welding current (…). I ), welding speed ( V ), laser power ( P ) and assembly clearance ( C ).
[0057] Step 2: Based on the process described in step S2, a preliminary experiment is conducted using the single-factor control variable method to determine the range of single-factor variations for each welding parameter.
[0058] To ensure that the variances of the input and output values are the same across all experimental points, the range of input values is divided into five levels, namely: -β , -1 , 0 , 1 , β , β The value depends on the number of types of input values, as shown in the formula:
[0059] in k Regarding the number of input value variables, this method has a total of four input value variables, namely... k = 4, so the upper and lower limits of the code value are 2 and -2, respectively. The parameter variation range and its corresponding encoding value are shown in the table below:
[0060] Step 3: Following the process described in step S3, obtain the center point values and all other point values of the experimental parameter matrix using the standardization formula of the response surface methodology. The formula is as follows:
[0061] in X For standard coded values, U For the actual input variables; U max This represents the actual maximum input value. U min This represents the actual minimum input value. The established experimental design matrix contains 30 experimental points, including 16 full-factor experimental points, 8 axis points, and 6 center points. Experiments were conducted according to the experimental design matrix, and the weld cross-sectional response values were recorded to complete the experimental design matrix, as shown in the table below:
[0062] A quadratic polynomial is used to fit the functional relationship between four input values and two output values. This function can be expressed as:
[0063] in For the output value, x i For input values, ε This represents the pure error of the system. To make the function more accurate, a power transformation needs to be performed on the function y. y Second-order polynomial regression equation after power transformation Y The result is:
[0064] in b o , b i , b ij , b ii These are the coefficients of the constant term, linear term, interaction term, and quadratic term, respectively. A stepwise regression method was used to establish a regression model, removing insignificant terms until the model's significance and lack-of-fit values met the requirements. A mathematical model relating the weld morphology and process parameters in laser arc overhead welding was obtained. The regression equation with the encoded value as the input value is:
[0065]
[0066] The regression equation with actual values as input is:
[0067]
[0068] Step 4: Following the process described in Step S4, the established model is validated using analysis of variance (ANOVA). According to the principles of ANOVA, a p-value less than 0.05 is considered significant, less than 0.01 is considered highly significant, and greater than 0.1 is considered insignificant. In this case, the p-values for the back weld width and weld reinforcement models are both less than 0.001, indicating high prediction accuracy. The p-values for the lack-of-fit terms are 0.1237 and 0.2236, both greater than 0.1, suggesting a good fit and high confidence level. A comparison of the distributions of the model's predicted and actual values is shown below. Figure 3 As shown, the actual values are evenly distributed around the predicted values; Figure 4The residuals of the model follow a normal distribution, which is approximately linear. All results indicate that the model fits the actual values well. Based on the variance analysis of the model's coefficients, the welding speed, a single factor in the back weld width model, is... V Welding current I Laser power P , docking gap C Coupling terms V×C , I×P , I×C and quadratic terms I 2 , C 2 The term is significant; in the residual height model, the single-factor term welding speed is significant. V Welding current I , docking gap C Coupling terms V×I , V×C quadratic term P 2 For significant terms Step 5: Based on the process described in step S5, and using the established mathematical model and the results of the model coefficient variance analysis, the results of the weld back weld width and reinforcement height with respect to the coded value parameters are as follows: Figure 5 , Figure 6 As shown, the effect of welding speed and butt gap coupling terms on the back weld width is as follows: Figure 7 As shown, the effect of the coupling term between welding current and laser power on the back weld width is as follows: Figure 8 As shown, the effect of welding current and butt gap coupling term on back weld width is as follows: Figure 9 As shown, the effect of the coupling term of welding current and welding speed on the reinforcement height is as follows: Figure 10 As shown, the effect of welding current and butt gap coupling term on the reinforcement height is as follows: Figure 11 As shown. Step 6: Based on the process described in Step S6 and the established mathematical model, a second-generation non-dominated sorting genetic algorithm is used for multi-objective optimization to obtain the optimal weld morphology characterized by the maximum back weld width and the minimum reinforcement height. The resulting Pareto solution set for the back weld width and reinforcement height is as follows: Figure 12 As shown.
[0069] Step 7: Following the process described in Step S7, arbitrarily select 5 sets of data from different frontiers in the solution set for verification experiments. The comparison results are shown in the table below. According to the comparison results, the error of each set of verification experiments is less than 5%. The weld morphology obtained from the optimized parameters is as follows: Figure 13 As shown, the Pareto optimal solutions with different frontiers can meet personalized needs and have high accuracy and effectiveness.
[0070] The above detailed embodiments are a description of the present invention. It should not be considered that the specific embodiments of the present invention are limited to these descriptions. For those skilled in the art, several simple deductions and substitutions can be made without departing from the concept of the present invention, and all of these should be considered to fall within the protection scope of the present invention.
Claims
1. A method for predicting the weld morphology of laser arc overhead welding in medium-thick plates, characterized in that, Includes the following steps: S1, input welding parameters, including welding current, welding speed, laser peak power and butt gap, and determine the output parameters as back weld width and front weld height; S2, by controlling variables, experimentally determine the range of variation of a single input welding parameter; S3, Establish the experimental parameter matrix and establish a mathematical model of weld morphology with respect to welding parameters based on the response surface methodology; S4 verifies the mathematical model established in S3 based on the analysis of variance method, and selects the terms that have a significant impact on the weld morphology from single welding parameter terms and multiple coupled welding parameter terms. S5: After converting the items with significant influence obtained in S4 into coded values, visualize the degree of influence of the items, and analyze the influence weight of individual welding parameter items and the influence law of welding parameter coupling items.
2. The method for predicting the weld morphology of medium-thick plate laser arc overhead welding according to claim 1, characterized in that, In S2, the input welding parameter values, i.e., the range of input values, are divided into 5 levels, namely... -β , -1 , 0 , 1 and β ,in β The value depends on the type and number of input values. in, k The number of input value variables.
3. The method for predicting the weld morphology of medium-thick plate laser arc overhead welding according to claim 2, characterized in that, In S3, the center point values and all other point values of the experimental parameter matrix are obtained by applying the standardization formula of the response surface methodology. The formula is as follows: in X For standard coded values, U The variable is the actual input value; U max The maximum value of the actual input value; U min This is the minimum value of the actual input.
4. The method for predicting the weld morphology of medium-thick plate laser arc overhead welding according to claim 1, characterized in that, In S3, a quadratic polynomial is used to fit the functional relationship between multiple input values and multiple output values, including: in For the output value, x i For input values, ε This is the pure error of the system; Among them The transformed second-order polynomial regression equation is obtained by performing a power transformation. Y The result is: in b o , b i , b ij , b ii In order, they are the constant term, the coefficient of the linear term, the coefficient of the interaction term, and the coefficient of the quadratic term.
5. The method for predicting the weld morphology of medium-thick plate laser arc overhead welding according to claim 4, characterized in that, In S3, based on fitting the functional relationship between multiple input values and multiple output values, a regression model is established using the stepwise regression method. Insignificant terms are removed until the significance and lack of fit of the model meet the requirements, thus obtaining a mathematical model between the weld morphology and process parameters of laser arc overhead welding, including regression equations with coded values as input values and regression equations with actual values as input values.
6. The method for predicting the weld morphology of medium-thick plate laser arc overhead welding according to claim 1, characterized in that, In S5, a single welding parameter is used as the horizontal axis, and the back width and front clearance are used as the vertical axes to form a two-dimensional analysis diagram.
7. The method for predicting the weld morphology of medium-thick plate laser arc overhead welding according to claim 1, characterized in that, In S5, the welding parameter coupling terms are fitted into a variation curve, and the back width and front clearance are used as the ordinates to form a three-dimensional analysis diagram.
8. A method for optimizing the weld morphology of medium-thick plate laser arc overhead welding, comprising optimizing the weld morphology of medium-thick plate laser arc overhead welding obtained by the weld morphology prediction method of medium-thick plate laser arc overhead welding as described in claim 1, characterized in that... Includes the following steps: S6. With the maximum back weld width and minimum excess height as optimization objectives, within the range of parameter terms and parameter coupling terms obtained from the analysis in step S5, the mathematical model verified in S4 is optimized to obtain the optimal solution set after optimization. S7. Based on the optimized optimal solution set obtained in step S6, a verification experiment is conducted to analyze and compare the prediction accuracy and weld morphology of the optimized model.