Process optimization method for laser-arc hybrid welding seam forming

Abstract

The invention discloses a process optimization method for laser-arc hybrid welding seam forming, which comprises the following steps: S1, acquiring a plurality of parameters of a hybrid welding experiment, taking fusion depth and fusion width in the parameters as experiment output parameters, and taking other parameters as experiment input parameters; s2, parameters are analyzed through Spearman correlation coefficients, and then key parameters influencing welding seam forming are obtained; s3, obtaining a range level of a single parameter variable value in the key parameters through a single variable method; s4, a response curved surface mathematical model is established, and more parameter matrix schemes meeting the welding seam forming conditions are screened out through the response curved surface mathematical model; and S5, taking the two input parameter variables, namely the fusion depth and the fusion width, as optimization targets, and performing multi-target optimization on the parameter matrix obtained in the step 4 by using a second-generation non-dominated sorting genetic algorithm to obtain a final optimization solution of the optimization targets. Compared with an existing method for improving welding seam forming based on process parameters, the problems of insufficient efficiency, high experiment cost and the like are solved.

Description

Technical Field

[0001] The invention relates to the technical field of welding, and in particular to a process optimization method for laser arc hybrid welding weld formation. Background Art

[0002] Laser arc hybrid welding is a technology that combines two heat sources with different physical properties, laser heat source and arc heat source. This welding technology involves laser guidance and stabilization of the arc. At the same time, the arc improves the metal's absorption rate of the laser, enhances the droplet transfer and bridging capabilities, thereby achieving a larger welding depth and an efficient and high-quality welding process. Significant welding efficiency has been achieved in key components in industries such as ships, bridges, and aviation, especially in thick plate processing.

[0003] The energy coupling mechanism of laser and arc in the hybrid welding process is complex and involves a variety of process parameters, including welding process parameters and welding condition parameters. Welding process parameters include laser power, welding speed, current and voltage, etc. Welding condition parameters include laser type and laser wavelength, welding angle, environmental factors, etc. Depth of penetration and width of penetration are key factors in determining the quality of weld formation. Different process parameters will directly affect weld formation. For example, if the laser power is too low, it cannot provide enough energy to fully melt the material, resulting in poor weld formation and incomplete penetration; if it is too high, it will cause excessive melting and produce weld depression; if the welding speed is too fast, the cooling speed of the molten pool will be accelerated, the depth of penetration will be shallower, and the weld width will be insufficient; if it is too slow, the heat affected zone will be enhanced, resulting in overheating of the weld, and the grain size will increase, affecting the mechanical properties of the welded joint. The welding process is the core step in processing key components. The quality of weld formation is the most intuitive indicator for evaluating the quality of welding. Therefore, the current research on the development of hybrid welding is mainly focused on producing an ideal weld shape without any visible defects.

[0004] The experimental process of the hybrid welding process is complex, with many process parameters and each parameter is not independent of each other. Many technical solutions use simulation or orthogonal experiments to explore different parameter combinations to obtain ideal weld quality, which usually consumes a lot of time and economic costs. The method of using machine learning to fit parameters for modeling is not suitable for hybrid welding, because hybrid welding has little experimental data, long experimental time, and high experimental cost, and machine learning models usually require a large amount of data for training. It can be seen that how to obtain well-formed welds by optimizing process parameters on the basis of improving efficiency and reducing costs is an urgent problem to be solved. Summary of the invention

[0005] In view of the shortcomings of the prior art, the purpose of the present invention is to provide a process optimization method for laser arc hybrid welding weld formation, in order to solve the problem that the process parameters in laser arc hybrid welding are complex, the welding experiment cost is high, and the overall efficiency of exploring the influence of parameters on weld formation is low. In order to achieve the above-mentioned purpose and other advantages according to the present invention, a process optimization method for laser arc hybrid welding weld formation is provided, comprising:

[0006] S1. Acquire multiple parameters of a hybrid welding experiment, wherein the penetration depth and the penetration width are used as experimental output parameters, and the other parameters are used as experimental input parameters;

[0007] S2. Analyze the parameters by using the Spearman correlation coefficient, and then obtain the key parameters affecting the weld formation, wherein the key parameters include laser power, defocus, filament spacing, current, and welding speed;

[0008] S3, obtaining the range level of the single parameter variable value in the key parameter by the single variable method;

[0009] S4, establishing a response surface mathematical model, and screening out more parameter matrix solutions that meet the weld formation conditions through the response surface mathematical model;

[0010] S5. Taking the two input parameter variables, melt depth and melt width, as optimization targets, the second-generation non-dominated sorting genetic algorithm is used to perform multi-objective optimization on the parameter matrix obtained in step 4 to obtain the final optimized solution of the optimization target.

[0011] Preferably, the parameters in step S1 include laser power, welding speed, wire feeding speed, filament spacing, defocusing amount, shielding gas flow, arc voltage, arc current, groove spacing, laser wavelength, guiding method, melting depth and melting width.

[0012] Preferably, in step S2, the correlation coefficient between the input parameter and the output parameter is denoted as r. When the absolute value of r is greater than or equal to 0.8, the input parameter is strongly correlated with the output parameter.

[0013] Preferably, when the absolute value of r is less than 0.8, the input parameter is weakly correlated with the output parameter, and the influence of the input parameter on the output parameter is ignored.

[0014] Compared with the prior art, the present invention has the following beneficial effects: by analyzing the correlation strength of each parameter through Spearman correlation, laser power, defocus, filament spacing, current, and welding speed are selected as key parameters affecting weld formation by comparing each correlation coefficient. By single-factor experiment, the absence of weld defects is taken as the experimental target, and the horizontal value ranges of 5 parameters are obtained respectively. Then, the functional relationship between the 5 key input parameters and the output parameters of penetration depth and width is obtained by response surface method, so as to construct a response surface and obtain more parameter matrices within the horizontal range of composite weld formation quality. Finally, the optimal solution for weld formation is obtained by multi-objective genetic algorithm, so that the optimized process parameters can meet a certain weld quality in welding processing. The present invention has important significance for significantly improving efficiency and reducing economic cost in the processing of important components in the fields of aerospace, shipbuilding, rail, etc. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 A flow chart of a process optimization method for laser arc hybrid welding weld formation according to the process optimization method for laser arc hybrid welding weld formation of the present invention;

[0016] Figure 2 A flow chart of parameter level ranges obtained for a single factor experiment of a process optimization method for laser arc hybrid welding weld formation according to the present invention;

[0017] Figure 3 A normal probability distribution diagram of residuals of a response surface model for a process optimization method for laser arc hybrid welding weld formation according to the present invention;

[0018] Figure 4 A response surface diagram showing the influence of laser power parameters and filament spacing parameters on the penetration depth in the process optimization method for laser arc hybrid welding weld formation according to the present invention;

[0019] Figure 5 A response surface diagram showing the influence of laser power parameters and defocus parameters on the penetration depth in the process optimization method for laser arc hybrid welding weld formation according to the present invention;

[0020] Figure 6 A response surface diagram showing the influence of laser power parameters and current parameters on the penetration depth in the process optimization method for laser arc hybrid welding weld formation according to the present invention;

[0021] Figure 7 A response surface diagram showing the influence of laser parameters and welding speed on the penetration depth in the process optimization method for laser arc hybrid welding weld formation according to the present invention;

[0022] Figure 8The fast non-dominated genetic algorithm flow chart of the process optimization method for laser arc hybrid welding weld formation according to the present invention. DETAILED DESCRIPTION

[0023] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0024] Example 1

[0025] The invention discloses an optimization method for laser arc hybrid welding weld formation, and takes the optimization process parameters of EH36 marine high-strength steel as an example to further illustrate the optimization method of the invention. Figure 1 As shown, the specific steps include:

[0026] Step 1: Obtain all parameters of the laser arc hybrid welding experiment, with penetration depth and penetration width as output parameters and other parameters as input parameters.

[0027] Step 2: Use the Spearman correlation coefficient to calculate the correlation of each parameter, and analyze and evaluate the key parameters that have the greatest impact on the depth of penetration and the width of the weld based on the correlation coefficient. The Spearman correlation coefficient evaluates the monotonic relationship between two variables, ignoring the linear relationship between the variables. The correlation coefficient is recorded as r. If |r|≥0.8, the input parameter is strongly correlated with the output parameter; if |r|<0.8, the input parameter is weakly correlated with the output parameter, and the influence of the input parameter on the output parameter is ignored. Spearman analysis was performed based on the existing hybrid welding experimental data, and laser power, defocus, filament spacing, current, and welding speed were used as key parameters affecting the depth of penetration and the width of the weld. Table 1 shows the Spearman correlation coefficient analysis results of the commonly used parameters in the hybrid welding process.

[0028] Table 1. Spearman correlation coefficients of common process parameters for hybrid welding

[0029]

[0030] Step 3: Use single variable method to study the influence of five key parameters on weld formation through single factor experiment. The specific process is as follows: Figure 2As shown. The purpose of this step is to obtain the upper and lower boundaries of the five key parameters for defect-free weld formation. The experimentally determined parameter level ranges that meet defect-free weld formation are: laser power 14kW-15kW, filament spacing 4mm-5mm, welding speed 2.0m / min-2.2m / min, current 400A-420A, and defocus -4mm-0mm.

[0031] Step 4: Use the response surface method to establish a response surface model of 5 input parameters and 2 output parameters. The purpose of this step is to explore the influence of each parameter and the interaction between parameters on the weld morphology. Since step 3 has obtained the parameter range of the weld morphology without defects through single factor experiments, this step is just to further improve and narrow the range to obtain the optimization interval, so as to obtain the optimal weld morphology.

[0032] In the response surface method, the P value is used to judge whether the model is significant and whether the lack of fit term is significant. If the model is significant and the lack of fit term is not significant, it means that the model can be used for subsequent optimization design. According to the statistical results of experimental factors and response values, mathematical models based on melt width and melt depth were established respectively.

[0033] y1=-858.2+0.19*A-2.09*B+6.18*C+0.67*D-2.72*E-2.47*A*B-1.99*A*C+1.24*A *D+5.31*A*E-4.97*B*C+5.62*B*D-0.2*B*E+0.22*C*D+0.1*C*E+9.02*D*E-1.29*A 2 -0.63*B 2 -0.75*C 2 -1.55*D 2 -0.99*E 2

[0034] y2=76.91+0.21*A-3.21*B+8.12*C-0.31*D+2.31*E-0.18*A*B-2.21*A*C+3.21*A*D -4.53*A*E+3.21*B*C+2.77*B*D+0.41*B*E-0.49*C*D+0.24*C*E+8.31*D*E-2.03*A 2 +0.52*B 2 +0.67*C 2 -2.71*D 2 -0.83*E 2

[0035] Among them, y1 and y2 represent the depth and width of the weld respectively, and A, B, C, D, and E represent the laser power, defocus, filament spacing, current, and welding speed respectively. Figure 3 It is the normal probability distribution plot of the residual, which shows that the model has good adaptability. Figure 4 , Figure 5 , Figure 6 and Figure 7 The following are the response surface diagrams of the effects of laser power parameters, filament spacing, defocus, current, and welding speed on the depth of penetration. The purpose of studying the response surface is to determine the two response indicators of depth of penetration and width of penetration to obtain the parameter matrix for preliminary optimization of process parameters.

[0036] Embodiment 2

[0037] The difference between the second embodiment and the first embodiment is that the first embodiment determines the range of factors affecting weld formation in composite welding to the level of factors without weld defects, which is a preparation for the second embodiment. In this embodiment, the parameter matrix of step 3 in the first embodiment is finally optimized using the second generation non-dominated genetic algorithm.

[0038] The second generation non-dominated genetic algorithm is a genetic algorithm for solving multi-objective optimization problems. When dealing with multi-objective optimization problems, it can find a set of Pareto optimal solutions and maintain the diversity and uniformity of the solution set. The present invention is a method for optimizing the weld formation of composite welding. The weld formation is usually related to the weld width and the weld depth. The algorithm flow is as follows Figure 8 As shown, the specific steps are as follows:

[0039] Step 1: Map the parameter matrix screened by the response surface methodology to the data space as the initial population.

[0040] Step 2: Perform non-dominated sorting on each individual in the population, where N(i) is used to represent the number of solution individuals that dominate individual i in the population, and S(i) is used to represent the set of solution individuals dominated by individual i. The calculation steps of the dominance relationship are as follows:

[0041] Step 2.1: For each individual i in the population, traverse all other individuals j in the population.

[0042] Step 2.2: If individual j is not worse than individual i in all objectives and is better than individual i in at least one objective, then individual i is dominated by individual j.

[0043] Step 2.3: Count N(i) and S(i). Select individuals with N(i) = 0, which are not dominated by any other individuals and are therefore non-dominated. Put all non-dominated individuals into the first layer and set the Pareto level to 1.

[0044] Step 2.4: For the individuals in the first layer, rescan the remaining individuals in the population and check whether they dominate each individual j in the remaining population. If the first-layer individual i dominates individual j, increase the domination count N(i) of individual j.

[0045] Step 2.5: After updating the dominance count, find individuals with N(i) = 0 to form the second layer, and set the Pareto level to 2. This process is repeated until all individuals are assigned to the corresponding non-dominated layers and the corresponding Pareto levels are set.

[0046] Step 2.6: In order to make the individuals more evenly distributed in each layer, crowding is introduced for each individual. The larger the crowding value, the less crowded the individual is in the solution space around it.

[0047] The calculation expression of congestion is:

[0048]

[0049] where i d represents the crowding degree of the ith individual, and Represent the jth objective function value of the i+1th point and the i-1th point respectively, and They represent the maximum and minimum values ​​of the j-th objective function respectively.

[0050] Step 3: According to the tournament selection method, select individuals from the layer with the highest Pareto level of 1 and the same level of crowding for crossover and mutation, and then generate the first generation subgroup Q t .

[0051] Step 4: Set the parent population P t and the offspring population Q t Merge to get a population R of size 2N t , and then for the population R t The individuals in the population are sorted non-dominatedly and the crowding degree is calculated. According to the non-dominated relationship and the crowding degree of the individuals, suitable individuals are selected to form a new parent population P. t+1 .

[0052] Step 5: Use the new parent population P t+1 Repeat the selection, crossover, and mutation steps until the termination number of iterations is met, and obtain a set of Pareto optimal solutions.

[0053] Repeating all the above steps and taking the optimized process parameters as actual parameters for experimental verification can obtain excellent weld morphology, and the weld joint has no defects such as pores and lack of fusion.

[0054] The equipment numbers and processing sizes described here are for simplicity and clarity.

[0055] Although the embodiments of the present invention have been disclosed as above, they are not limited to the applications listed in the specification and implementation modes. They can be fully applied to various fields suitable for the present invention. For those familiar with the art, additional modifications can be easily implemented. Therefore, without departing from the description of the claims of the present invention, the application, modification and variation of the present invention are the general concepts defined by the scope of equivalents for those skilled in the art. The present invention is not limited to the specific details and the illustrations shown and described herein.

Claims

1. A process optimization method for laser arc hybrid welding weld formation, characterized in that: include: S1. Acquire multiple parameters of a hybrid welding experiment, wherein the penetration depth and the penetration width are used as experimental output parameters, and the other parameters are used as experimental input parameters; S2. Analyze the parameters by using the Spearman correlation coefficient, and then obtain the key parameters affecting the weld formation, wherein the key parameters include laser power, defocus, filament spacing, current, and welding speed; S3, obtaining the range level of the single parameter variable value in the key parameter by the single variable method; S4, establishing a response surface mathematical model, and screening out more parameter matrix solutions that meet the weld formation conditions through the response surface mathematical model; S5. Taking the two input parameter variables, melt depth and melt width, as optimization targets, the second-generation non-dominated sorting genetic algorithm is used to perform multi-objective optimization on the parameter matrix obtained in step 4 to obtain the final optimized solution of the optimization target.

2. A process optimization method for laser arc hybrid welding weld formation according to claim 1, characterized in that: The parameters in step S1 include laser power, welding speed, wire feeding speed, wire spacing, defocusing amount, shielding gas flow, arc voltage, arc current, groove spacing, laser wavelength, guiding method, melting depth and melting width.

3. The process optimization method for laser arc hybrid welding weld formation according to claim 1, characterized in that: In step S2, the correlation coefficient between the input parameter and the output parameter is recorded as r. When the absolute value of r is greater than or equal to 0.8, the input parameter is strongly correlated with the output parameter.

4. A process optimization method for laser arc hybrid welding weld formation as claimed in claim 3, characterized in that: When the absolute value of r is less than 0.8, the input parameter is weakly correlated with the output parameter, and the influence of the input parameter on the output parameter is ignored.

5. The process optimization method for laser arc hybrid welding weld formation according to claim 1, characterized in that: The genetic algorithm in step S5 specifically includes: (1) The parameter matrix screened by response surface methodology is mapped to the data space and initialized as the initial population; (2) Perform non-dominated sorting on each individual in the population and calculate the crowding degree. After sorting, use the tournament method to select individuals for crossover and mutation to obtain the first-generation subpopulation; (3) The first-generation subpopulation is merged with the parent population to perform non-dominated sorting and crowding calculations, and new individuals are selected to form a new parent generation. (4) Repeat steps 2 and 3 until the termination number of iterations is met and a set of Pareto optimal solutions is obtained.

6. A process optimization method for laser arc hybrid welding weld formation as claimed in claim 5, characterized in that: Step 2 The calculation method of congestion is: where i d represents the crowding degree of the ith individual, and Represent the jth objective function value of the i+1th point and the i-1th point respectively, and They represent the maximum and minimum values ​​of the j-th objective function respectively.

Images