A multi-objective optimization method for steel / aluminum dissimilar metal laser welding
By optimizing laser welding parameters using the CatBoost-MOAVOA algorithm and CRITIC evaluation method, the problems of high cost and low prediction accuracy in existing technologies are solved. This achieves low-cost and high-efficiency optimization of steel/aluminum dissimilar metal welding, significantly improving the warping deformation and maximum tensile force of welded parts, and enhancing the performance of welded joints.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- XUZHOU NORMAL UNIVERSITY
- Filing Date
- 2024-09-30
- Publication Date
- 2026-04-14
AI Technical Summary
Existing methods for optimizing laser welding process parameters rely on experiments and experience, which are costly and have low predictive accuracy. They are also difficult to effectively suppress the formation of hard and brittle intermetallic compounds in steel/aluminum dissimilar metal welding, thus affecting the mechanical properties of the welded joint.
By employing the CatBoost-MOAVOA algorithm combined with the CRITIC evaluation method, and through the design of orthogonal experiments and the construction of the RIME-CatBoost model, laser power, welding speed, defocusing amount, and CeO2 addition amount are optimized to achieve multi-objective optimization, reduce the warping deformation of welded parts, increase the maximum tensile force, and reduce welding costs.
It achieves low-cost improvement in the reliability of steel/aluminum laser welding, reduces warpage deformation of welded parts by 37.7%, increases maximum tensile force by 20.6%, reduces welding costs by 18%, and significantly improves the quality and reliability of welded joints.
Smart Images
Figure CN119475964B_ABST
Abstract
Description
Technical Field
[0001] This invention relates to a multi-objective optimization method for laser welding, specifically a multi-objective optimization method for laser welding parameters and CeO2 addition amount of steel / aluminum dissimilar metals based on Catboost-MOAVOA, belonging to the field of laser welding technology. Background Technology
[0002] Automotive lightweighting aims to reduce a vehicle's curb weight as much as possible while maintaining its strength and safety performance, thereby improving its power, reducing fuel consumption, and lowering emissions. Steel / aluminum dissimilar metal composite structures are characterized by high strength and light weight. While ensuring safety, they effectively reduce the weight of the vehicle. Therefore, using steel / aluminum welded structures instead of pure steel welded structures is currently recognized as one of the most effective methods for automotive lightweighting.
[0003] However, due to the differences in chemical and physical properties between steel and aluminum, hard and brittle intermetallic compounds (IMCs) will inevitably be generated during the welding process. The presence of IMCs will seriously reduce the mechanical properties of the welded joint. Therefore, how to suppress the generation of IMCs has become the core issue in obtaining high-quality steel / aluminum welded joints.
[0004] Laser welding (LBW) is a common method for welding dissimilar metals such as steel and aluminum. By adjusting process parameters such as laser power, welding speed, and defocusing amount, the size of the heat-affected zone in the weld area can be controlled, reducing the formation of Fe-Al IMCs. However, the relationship between laser welding parameters and welding results is non-linear. Traditional methods for optimizing laser welding process parameters rely on experiments and experience, which require a large amount of human and material resources and are costly.
[0005] Existing methods for optimizing laser welding process parameters generally include two parts: laser welding process parameter prediction and parameter optimization. The laser welding parameter prediction model, as a method for establishing a nonlinear mapping relationship between laser welding process parameters and welding quality evaluation parameters, directly affects the optimization effect of the laser welding process parameters. Commonly used welding parameter prediction methods include empirical calculation, experimental summary, and numerical simulation. Empirical calculation and experimental summary are often based on experience, and the calculated results may have significant errors compared to the actual situation. Numerical simulation requires a significant investment of time and resources from professionals. Response surface methodology and machine learning methods are widely used due to their ability to quickly and accurately predict welding deformation; however, polynomial response surface methodology requires a large number of experimental samples when dealing with complex problems, resulting in high costs. Machine learning methods are gradually being promoted due to their advantages of requiring small sample sizes and high prediction accuracy; however, because different machine learning models perform differently under different conditions, existing welding process parameter prediction methods using a single prediction model may not achieve optimal prediction results for specific datasets. Summary of the Invention
[0006] To address the aforementioned problems, this invention provides a multi-objective optimization method for laser welding of dissimilar metals such as steel and aluminum. This method can reduce the warping deformation of the welded parts after laser welding and increase the maximum tensile force that the welded parts can withstand, while reducing welding costs. This results in improving the reliability of laser welding of steel and aluminum at low cost and provides theoretical basis and data support for obtaining the optimal combination of process parameters for laser welding of steel and aluminum.
[0007] To achieve the above objectives, the multi-objective optimization method for laser welding of dissimilar metals such as steel and aluminum specifically includes the following steps:
[0008] Step 1: The warping deformation amount y1 of the welded part, the maximum tensile force that the welded part can withstand y2, and the welding cost y3 are used as optimization objectives. The laser power x1, welding speed x2, defocusing amount x3, and CeO2 addition amount x4 are used as optimization design variables. The range of values for the optimization design variables is determined, and an orthogonal experiment is designed to conduct laser welding and obtain the experimental results.
[0009] The formula for calculating welding costs is as follows:
[0010]
[0011] In the formula: y3 is the welding cost, in ¥; x1 is the laser power, in W; x2 is the welding speed, in mm / min; x4 is the CeO2 addition amount, in g;
[0012] Step 2: Based on the experimental results, the laser power x1, welding speed x2, defocusing amount x3 and CeO2 addition amount x4 are used as inputs, and the warping deformation amount y1 and the maximum tensile force that the weldment can withstand y2 are used as outputs to construct a RIME-CatBoost model that optimizes CatBoost with RIME.
[0013] Step 3: Use the MOAVOA algorithm to perform multi-objective optimization on the RIME-CatBoost model to obtain the Pareto front of the optimization objective;
[0014] Step 4: Determine the optimal combination of process parameters in the Pareto front of the obtained optimization objective using the CRITIC evaluation method.
[0015] Furthermore, in Step 2, the reciprocal of the maximum tensile force that the weldment can withstand, the amount of warpage deformation of the weldment, and the welding cost are used as the objective functions of the optimization algorithm. The mathematical model for welding process optimization based on the RIME-CatBoost model is expressed as follows:
[0016]
[0017] In the formula: x1 is the laser power, x2 is the welding speed, x3 is the defocusing amount, x4 is the CeO2 addition amount, y1 is the warping deformation of the welded part, y2 is the maximum tensile force that the welded part can withstand, and y3 is the welding cost.
[0018] Furthermore, in Step 2, the specific steps for optimizing CatBoost using RIME are as follows:
[0019] ①The learning rate and maximum depth of the CatBoost model are used as hyperparameters to be optimized;
[0020] ② Set the range of values for hyperparameters and the initial parameters for the RIME algorithm;
[0021] ③ The CatBoost model is trained by updating the hyperparameters in each iteration of the RIME algorithm;
[0022] ④ Calculate the fitness value of the corresponding hyperparameter combination based on the model performance index, and select the hyperparameter combination with the best fitness.
[0023] Furthermore, in Step 3, the RIME-CatBoost model is optimized using the MOAVOA algorithm for multiple objectives, as detailed below:
[0024] ① Initialize MOAVOA parameters: Parameters include maximum number of iterations, population size, maximum archive capacity, and number of objective functions;
[0025] ② Determine if the current iteration count has reached the maximum iteration count: The algorithm starts from the initial iteration count and continues until the preset maximum iteration count is reached. If the current iteration count is less than the maximum iteration count, the algorithm continues to iterate; otherwise, it outputs the Pareto front and terminates.
[0026] ③ Fitness calculation: Each vulture represents a solution, and the fitness of each vulture is obtained according to the RIME-CatBoost model and the welding cost calculation formula;
[0027] ④ Select the best vulture: In each iteration, select the best-performing vulture as the leader based on its fitness value and save it to the archive;
[0028] ⑤ Update Vulture Locations: Update the locations of all vultures based on the leader vulture's behavior and current strategy;
[0029] ⑥ Determine file size: Check if the current solution set size exceeds the file size setting. If the solution set size is less than the file size, continue the iteration. If the solution set size is greater than the file size, delete solutions in dense areas through the grid mechanism to preserve the diversity of solutions.
[0030] ⑦ Update Archive: Update the archive according to the Pareto frontier criteria, retaining non-dominated solutions;
[0031] ⑧ Perform the next iteration: Increment the iteration count by 1, and proceed to step ② to iterate again.
[0032] Furthermore, in Step 4, when determining the optimal combination of process parameters from the obtained Pareto fronts of the optimization objective using the CRITIC evaluation method, a weighted aggregation method is applied to the obtained Pareto fronts, assigning weights to each Pareto front. The expression of the comprehensive evaluation model is shown below:
[0033]
[0034] In the formula: Z is the comprehensive evaluation value of each experimental design, W j Let y be the weight value of the j-th indicator. ij Let y be the j-th evaluation index value of the i-th scheme. i1 y represents the warpage deformation of the welded parts in each experimental design. i2 y represents the maximum tensile force that the welded component can withstand in each experimental design. i3 The welding cost in each experimental design.
[0035] Compared with existing technologies, the multi-objective optimization method for laser welding of dissimilar metals such as steel and aluminum first uses laser power, welding speed, defocusing amount, and CeO2 addition amount as inputs, and the warpage deformation of the weldment and the maximum tensile force that the weldment can withstand as outputs. Orthogonal experiments are designed and welding experiments are conducted to obtain experimental results. Then, based on the experimental results and the data measured from the welding samples, a RIME-CatBoost model is constructed to optimize CatBoost using RIME. Next, the warpage deformation of the weldment, the maximum tensile force that the weldment can withstand, and the calculated welding cost are used as optimization objectives, and the RIME-CatBoost model is optimized using the MOAVOA algorithm. Finally, the CRITIC comprehensive evaluation method is used to determine the optimal combination of welding parameters. The experimental results show that the error between the welding test results predicted by the optimization method used in this invention and the actual experimental values is controlled within 5%, the model prediction accuracy is good, and the warping deformation and welding cost of the optimized welded parts are significantly reduced compared with the results before optimization. The maximum tensile force that the optimized welded parts can withstand is significantly increased compared with the results before optimization. Welding using the optimized laser welding parameters can significantly reduce welding costs and improve product quality, thereby achieving low-cost improvement of the reliability of steel / aluminum laser welding. It can provide theoretical basis and data support for obtaining the optimal combination of process parameters for steel / aluminum laser welding. Attached Figure Description
[0036] Figure 1 These are comparison diagrams of the effects of the model before and after optimization of the present invention, where (a) is a comparison diagram of the predicted and actual values of the deformation of the welded parts before and after optimization, and (b) is a comparison diagram of the predicted and actual values of the maximum tensile force that the welded parts can withstand before and after optimization.
[0037] Figure 2 This is the Pareto front plot obtained by performing MOAVOA multi-objective optimization on the RIME-CatBoost model according to the present invention;
[0038] Figure 3 This is an optimized welding sample diagram of the present invention;
[0039] Figure 4 These are microscopic images of the tensile fracture surfaces of the sample before and after optimization according to the present invention, wherein (a) is a microscopic image of the tensile fracture surface of the sample before optimization and (b) is a microscopic image of the tensile fracture surface of the sample after optimization.
[0040] Figure 5 This is a flowchart of the multi-objective optimization of the RIME-CatBoost model using the MOAVOA algorithm in this invention. Detailed Implementation
[0041] In the laser welding of dissimilar metals like steel and aluminum, the metallurgical reaction of the molten pool can be altered by adding elements such as Ti, Cu, Zn, Ni, and rare earth elements to form an intermediate layer. The use of this intermediate layer not only reduces the mixing of Fe and Al elements in the molten pool and decreases the thickness of brittle Fe-Al IMCs, but also allows the added elements to combine with Fe or Al to form new binary or ternary IMCs with good elasticity, improving the mechanical properties of the steel / aluminum weld joint. Therefore, introducing an intermediate layer for alloying metallurgical control during the laser welding of dissimilar metals like steel and aluminum can effectively suppress the formation of Fe-Al IMCs. This invention, based on the addition of rare earth element CeO2 to form an intermediate layer for alloying metallurgical control during the laser welding of dissimilar metals like steel and aluminum to suppress IMC formation, aims to obtain high-quality steel / aluminum weld joints through multi-objective optimization of the CeO2 addition amount and laser welding parameters.
[0042] The multi-objective optimization method for laser welding of dissimilar metals such as steel and aluminum first uses laser power, welding speed, defocusing amount, and CeO2 addition amount as inputs, and weld warpage and maximum tensile force that the weldment can withstand as outputs, to conduct welding experiments. Then, a CatBoost mathematical model is established based on the data measured from the welding samples, and its hyperparameters are optimized using the robust iterative multiview embedding (RIME) algorithm. Next, the weld warpage, maximum tensile force that the weldment can withstand, and the calculated welding cost are used as optimization objectives, and multi-objective optimization is performed using the multi-objective African vultures optimization algorithm (MOAVOA). Finally, the optimal welding parameters are determined using the CRITIC comprehensive evaluation method.
[0043] The following describes the invention in detail using laser welding of DP780 duplex steel and 6016 aluminum alloy, which are widely used in the automotive industry.
[0044] Step 1: Determine the optimization objective, select the laser welding process parameters and their value ranges, use the value ranges of each process parameter as the experimental design space, design orthogonal experiments based on the value ranges of each process parameter, and perform laser welding to obtain the experimental results.
[0045] A laser lap welding test scheme with an intermediate CeO2 layer was designed. The CeO2 powder had a particle size of 20 nm. The dimensions of the DP780 duplex steel and 6016 aluminum alloy specimens were both 40 mm × 10 mm × 1 mm. The chemical composition of the specimens is shown in Table 1 below. Acetone was used to clean the surface oil and oxide film before the welding test.
[0046] Table 1 Chemical composition of the specimens (wt.%)
[0047]
[0048] Four parameters that significantly affect welding quality were selected for the experiment: laser power (x1, unit W), welding speed (x2, unit mm / min), defocusing amount (x3, unit mm), and CeO2 addition amount (x4, unit g) as influencing factors. Based on the recommended range for laser welding, trial welding was conducted to adjust and define the value ranges of the influencing factors. Within the adjusted range, no obvious defects were observed in the joint. Laser power (x1), welding speed (x2), and defocusing amount (x3) were divided into 6 levels, and CeO2 addition amount (x4) was divided into 3 levels, as shown in Table 2 below.
[0049] Table 2 Experimental factors and level parameters
[0050]
[0051] Using the four factors shown in Table 2, 36 different combinations of welding process parameters were generated based on the principle of orthogonal experimental design. Laser welding experiments were conducted on steel / aluminum dissimilar metal specimens using a JHM-1GX-1000F FBER laser welding machine. Argon gas was used as the shielding gas during the welding process at a rate of 10 L / min.
[0052] The deformation of the welded parts after laser welding of dissimilar metals such as steel and aluminum has a significant impact on the installation accuracy and appearance quality of the welded components. Meanwhile, the maximum tensile force that the welded parts can withstand is an important mechanical property indicator of the metallic material and a crucial factor affecting the overall mechanical performance of automobiles. Therefore, the warpage deformation (y1, in mm) and the maximum tensile force (y2, in N) of the welded parts are used as optimization targets.
[0053] Furthermore, welding cost (y3, in ¥) is one of the most important factors in industrial production. To achieve low-cost welding, welding cost (y3, in ¥) is also considered as an optimization objective. The total welding cost is the product of the hourly welding cost and the welding time (assuming a utilization rate of 0.8) plus the cost of consuming the intermediate CeO2 layer. The formula for calculating welding cost (y3) is as follows:
[0054]
[0055] In the formula: x1 is the welding power, x2 is the welding speed, and x4 is the amount of CeO2 added.
[0056] After welding, the deformation of 36 welded samples was measured using a 3D laser scanner, and the maximum tensile force that the samples could withstand was tested using the room-temperature tensile module of a dynamic thermal simulation testing machine. The measured deformation, maximum tensile force, and calculated welding cost are shown in Table 3 below.
[0057] Table 3 Test Results
[0058]
[0059] Step 2: Based on the experimental results, establish a CatBoost model optimized using the robust iterative multiviewembedding (RIME) algorithm.
[0060] CatBoost is a machine learning algorithm based on gradient boosting decision trees, specifically optimized for categorical feature processing and overfitting prevention to achieve efficient and robust model performance. Compared to traditional models, CatBoost models exhibit superior performance and effectively handle categorical features and improve model performance through strategies such as ranking learning and goal-oriented encoding.
[0061] Two CatBoost models were constructed using laser power (x1), welding speed (x2), defocusing amount (x3), and CeO2 addition amount (x4) as inputs, and weld warpage deformation (y1) and maximum tensile force that the weld can withstand (y2) as outputs. To demonstrate the predictive accuracy and effectiveness of the models, 80% of the data from 36 randomly selected orthogonal experiments were used as the training set for the CatBoost models, and the remaining 20% was used as the test set. The trained CatBoost models were then used for predictive analysis.
[0062] Although the CatBoost model has excellent performance, its hyperparameters have a significant impact on the model's performance. If the hyperparameters are not set properly, overfitting can easily occur. Therefore, it is necessary to optimize the hyperparameters of the CatBoost model.
[0063] The Robust Iterative Multiview Embedding (RIME) algorithm is an optimization algorithm based on the simulation of physical phenomena. It is inspired by the process of frost and ice crystals gradually forming on a cooled surface. The algorithm draws on the growth characteristics of soft frost and utilizes the strong randomness and coverage of frost particles to enable the algorithm to quickly cover the entire search space and find the optimal solution to complex optimization problems.
[0064] The RIME algorithm is used to optimize the CatBoost hyperparameters. The specific steps are as follows:
[0065] ①The learning rate and maximum depth of the CatBoost model are used as hyperparameters to be optimized;
[0066] ② Set the range of values for hyperparameters and the initial parameters for the RIME algorithm;
[0067] ③ The CatBoost model is trained by updating the hyperparameters in each iteration of the RIME algorithm;
[0068] ④ Calculate the fitness value of the corresponding hyperparameter combination based on the model performance index, and select the hyperparameter combination with the best fitness.
[0069] The RIME algorithm performs a global search and local fine-tuning in the parameter space through iteration, and continuously updates individuals and the optimal solution using a greedy selection mechanism, so that the depth and learning rate of the CatBoost model are optimally configured, thereby improving model performance and prediction accuracy.
[0070] The prediction errors of the Catboost model before and after optimization for the warping deformation of the weldment and the maximum tensile force that the weldment can withstand are shown in Tables 4 and 5 below.
[0071] Table 4 Prediction error values of warpage deformation of welded parts
[0072]
[0073] Table 5 Predicted error values of the maximum tensile force that welded parts can withstand
[0074]
[0075] Tables 4 and 5 show that the RIME-CatBoost model is more accurate in predicting the warpage deformation and the maximum tensile force that the weldment can withstand. The predicted data before and after optimization are compared with the actual welding test results. Figure 1 As shown, where Figure 1 (a) shows a comparison between the predicted warpage deformation of the weldment by the CatBoost model before and after optimization and the actual value. Figure 1 (b) shows a comparison between the maximum tensile force predicted by the CatBoost model before and after optimization and the actual value.
[0076] Step 3: Use the multi-objective African vultures optimization algorithm (MOAVOA) to perform multi-objective optimization and obtain the Pareto front of the optimization objective.
[0077] The optimization objective of this invention is to minimize the warpage deformation of the weldment (y1) and welding cost (y3), while maximizing the maximum tensile force that the weldment can withstand (y2). Considering the significant differences in magnitude and optimization direction among these three optimization objectives, and their equal importance, the reciprocal of the maximum tensile force that the weldment can withstand, the warpage deformation of the weldment, and the welding cost are used as the objective functions of the optimization algorithm. To balance the magnitudes of the factors in the objective function and facilitate comprehensive evaluation, the weighting coefficient for the warpage deformation of the weldment is set to 1, the weighting coefficient for the reciprocal of the maximum tensile force that the weldment can withstand is set to 1000, and the weighting coefficient for the welding cost is set to 20. Therefore, the mathematical model for welding process optimization based on the RIME-CatBoost model can be summarized as follows:
[0078]
[0079] In the formula: x1 is the laser power, x2 is the welding speed, x3 is the defocusing amount, x4 is the CeO2 addition amount, y1 is the warping deformation of the weldment, y2 is the maximum tensile force that the weldment can withstand, and y3 is the welding cost. y1 and y2 are predicted outputs from the RIME-CatBoost model, and y3 is calculated using the aforementioned formula.
[0080] The African Vultures Optimization Algorithm (AVOA) simulates the foraging behavior of African vultures to find the optimal solution to an optimization problem. The Multi-Objective African Vultures Optimization Algorithm (MOAVOA) builds upon the single-objective AVOA by adding three key mechanisms: Archive, Grid, and Leader Selection, to enhance its performance in multi-objective optimization. MOAVOA finds a uniformly distributed Pareto front in the search space through processes such as population initialization, fitness calculation, hunger level calculation, exploration phase, development phase, archive updating, and iterative updates, thus effectively solving multi-objective optimization problems.
[0081] Using the aforementioned mathematical model for welding process optimization based on the RIME-CatBoost model as the objective function of MOAVOA, such as... Figure 5 As shown, the MOAVOA algorithm is used to perform multi-objective optimization on the RIME-CatBoost model, as detailed below:
[0082] ① Initialize MOAVOA parameters: Initialize the parameters of the algorithm, including the maximum number of iterations, population size, maximum archive capacity, and number of objective functions.
[0083] ② Determine if the current iteration count has reached the maximum iteration count: The algorithm starts from the initial generation and continues until the preset maximum iteration count is reached. If the current iteration count is less than the maximum iteration count, the algorithm continues to iterate; otherwise, it outputs the Pareto front and terminates.
[0084] ③ Calculate the fitness of each vulture: Each vulture represents a set of solutions. The fitness of each vulture (i.e., solution) is obtained according to the RIME-CatBoost model and the welding cost calculation formula.
[0085] ④ Select the best vulture: In each iteration, select the best-performing vulture as the leader (i.e., the non-dominant solution) based on its fitness value and save it in the archive.
[0086] ⑤ Update Vulture Locations: Update the locations of all vultures based on the leader vulture's behavior and current strategy.
[0087] ⑥ Determine file size: Check if the current solution set size exceeds the file's set size. If the solution set size is less than the file size, continue iteration; if the solution set size is greater than the file size, delete solutions in dense regions using a grid mechanism to preserve solution diversity.
[0088] ⑦ Update Archive: Update the archive according to the Pareto Frontier guidelines, retaining non-dominated solutions.
[0089] ⑧ Perform the next iteration: Increment the iteration count by 1, and proceed to step ② to iterate again.
[0090] The final Pareto front of CatBoost-MOAVOA output is as follows: Figure 2 As shown, a total of 23 Pareto fronts were obtained.
[0091] Step 4: Determine the optimal combination of process parameters using the CRITIC evaluation method.
[0092] Figure 2 Twenty-three Pareto fronts were obtained through MOAVOA optimization, and a comprehensive evaluation method is needed to select the optimal solution. The 23 Pareto fronts are weighted and aggregated, with appropriate weights assigned to each front. The expression of the comprehensive evaluation model is shown below:
[0093]
[0094] In the formula: Z is the comprehensive evaluation value of each experimental design, W j Let y be the weight value of the j-th indicator. ij Let y be the j-th evaluation index value of the i-th scheme. i1 y represents the warpage deformation of the welded parts in each experimental design.i2 y represents the maximum tensile force that the welded component can withstand in each experimental design. i3 The welding cost in each experimental design.
[0095] The comprehensive evaluation model fully considers the importance of each evaluation criterion by weighted aggregation. This method generates comprehensive evaluation values covering each Pareto front, thus enabling a comprehensive evaluation and ranking of different solutions. For the 23 Pareto fronts obtained using the MOAVOA optimization method, the weights of each objective were objectively determined using the CRITIC method. The objective weights for the maximum tensile force that the weldment can withstand, the amount of warpage deformation of the weldment, and the welding cost are 0.58, 0.29, and 0.13, respectively. Introducing the calculated objective weights into the above formula yields the comprehensive evaluation score of the Pareto front. Figure 2 The circled solution's evaluation score is the minimum of the 23 Pareto fronts, corresponding to a maximum tensile force of 751.2 N, a warpage of 0.676 mm, and a welding cost of ¥0.046. The corresponding laser welding parameter combination is: laser power of 593.20 W, welding speed of 816.78 mm / min, defocusing distance of 11.62 mm, and CeO2 addition of 0.0202 g.
[0096] Based on the actual accuracy of the JHM-1GX-1000F FBER fiber laser welding machine, the optimal laser welding parameters were determined to be: laser power 593W, welding speed 816.8mm / min, defocusing distance 11.60mm, and CeO2 addition amount 0.02g. Using these optimal laser welding parameters, steel / aluminum dissimilar metal laser welding was performed on the JHM-1GX-1000F FBER fiber laser welding machine. Argon gas was used as the shielding gas during the welding process, with a flow rate of 10L / min. To verify the accuracy of the optimization results, three laser welding experiments were conducted using the optimized laser welding parameters. The welded samples are shown below. Figure 3 As shown in the figure. A 3D laser scanner was used to measure the deformation of the samples, and the room-temperature tensile module of a dynamic thermal simulation testing machine was used to measure the maximum tensile force of the samples after welding. The average value of the test values for the three welded samples was taken to obtain the optimized test results. The comparison of the results before and after optimization is shown in Table 6 below.
[0097] Table 6 Comparison of Optimization Results and Model Prediction Results
[0098]
[0099] Table 6 shows that the errors between the experimental and predicted values after CatBoost-MOAVOA optimization are 3.7% and 4.8%, respectively, both less than 5%, indicating that the optimization method has high accuracy. The optimized weldment warpage is 0.650 mm, the maximum tensile force the weldment can withstand is 715 N, and the welding cost is ¥0.046. Compared with the experimental values before optimization, the optimized sample shows a 37.7% reduction in weldment warpage, a 20.6% increase in the maximum tensile force the weldment can withstand, and an 18% reduction in welding cost. This demonstrates that using the optimized parameters not only significantly improves the performance of the welded joint but also reduces processing costs, validating the feasibility of this optimization scheme.
[0100] Electron microscopy (SEM) was used to compare and analyze the tensile fracture surfaces of samples before and after welding parameter optimization. The fracture morphology is shown in the figure. Figure 4 As shown. Figure 4 As shown in (a), the specimen welded using the parameters before optimization exhibited numerous cracks at the fracture interface. These cracks not only increased the brittleness of the material but also led to localized stress concentration, significantly reducing the overall reliability and load-bearing capacity of the welded joint. Figure 4 As shown in (b), welding tests were conducted using optimized laser welding parameters. It was observed that the tensile fracture surface of the sample was significantly improved. The optimized parameters reduced the number of cracks and tear ridges appeared on the fracture surface. This indicates that the toughness and fracture mode of the welded joint have been improved. This improvement can reduce the stress concentration effect, thereby enhancing the overall strength and durability of the welded joint and reducing the risk of sudden fracture.
[0101] The reasons why optimized laser welding parameters can improve fracture performance are analyzed as follows: ① By optimizing laser welding parameters, the size of the heat-affected zone in the welding area can be controlled, stress concentration and thermal stress can be reduced, thereby reducing the formation of Fe-Al IMCs. ② The addition of CeO2 can promote the formation and flow of the molten pool, thereby reducing the accumulation of thermal stress and lowering the hot cracking susceptibility of the weld joint. ③ The addition of CeO2 can effectively reduce the formation of intermetallic compounds (IMCs) during the welding process. By promoting alloying reactions, inhibiting the growth rate of IMCs, and reducing the presence of oxide inclusions, it improves the quality and corrosion resistance of the weld joint.
[0102] The above results verify the significant impact of optimized parameters on fracture performance, and can provide a reliable theoretical basis and data support for the application of laser welding of dissimilar metals such as steel and aluminum.
Claims
1. A multi-objective optimization method for laser welding of dissimilar metals such as steel and aluminum, characterized in that, Specifically, the following steps are included: Step 1: The warping deformation amount y1 of the welded part, the maximum tensile force that the welded part can withstand y2, and the welding cost y3 are used as optimization objectives. The laser power x1, welding speed x2, defocusing amount x3, and CeO2 addition amount x4 are used as optimization design variables. The range of values for the optimization design variables is determined, and an orthogonal experiment is designed to conduct laser welding and obtain the experimental results. The formula for calculating welding costs is as follows: In the formula: y3 is the welding cost, in ¥; x1 is the laser power, in W; x2 is the welding speed, in mm / min; x4 is the CeO2 addition amount, in g; Step 2: Based on the experimental results, the laser power x1, welding speed x2, defocusing amount x3 and CeO2 addition amount x4 are used as inputs, and the warping deformation amount y1 and the maximum tensile force that the weldment can withstand y2 are used as outputs to construct a RIME-CatBoost model that optimizes CatBoost with RIME. The specific steps for optimizing CatBoost using RIME are as follows: ①The learning rate and maximum depth of the CatBoost model are used as hyperparameters to be optimized; ② Set the range of values for hyperparameters and the initial parameters for the RIME algorithm; ③ The CatBoost model is trained by updating the hyperparameters in each iteration of the RIME algorithm; ④ Calculate the fitness value of the corresponding hyperparameter combination based on the model performance index, and select the hyperparameter combination with the best fitness. Using the reciprocal of the maximum tensile force that the weldment can withstand, the amount of warpage deformation of the weldment, and the welding cost as the objective functions of the optimization algorithm, the mathematical model for welding process optimization based on the RIME-CatBoost model is expressed as follows: In the formula: x1 is the laser power, x2 is the welding speed, x3 is the defocusing amount, x4 is the CeO2 addition amount, y1 is the warping deformation of the welded part, y2 is the maximum tensile force that the welded part can withstand, y3 is the welding cost, 1000 is the reciprocal coefficient of the maximum tensile force that the welded part can withstand, and 20 is the weighting coefficient of the welding cost. Step 3: Use the MOAVOA algorithm to perform multi-objective optimization on the RIME-CatBoost model to obtain the Pareto front of the optimization objective; Step 4: Determine the optimal combination of process parameters in the Pareto front of the obtained optimization objective using the CRITIC evaluation method.
2. The multi-objective optimization method for laser welding of dissimilar metals such as steel and aluminum according to claim 1, characterized in that, In Step 3, the RIME-CatBoost model is optimized using the MOAVOA algorithm for multiple objectives, as detailed below: ① Initialize MOAVOA parameters: Parameters include maximum number of iterations, population size, maximum archive capacity, and number of objective functions; ② Determine if the current iteration count has reached the maximum iteration count: The algorithm starts from the initial iteration count and continues until the preset maximum iteration count is reached. If the current iteration count is less than the maximum iteration count, the algorithm continues to iterate; otherwise, it outputs the Pareto front and terminates. ③ Fitness calculation: Each vulture represents a solution, and the fitness of each vulture is obtained according to the RIME-CatBoost model and the welding cost calculation formula; ④ Select the best vulture: In each iteration, select the best-performing vulture as the leader based on its fitness value and save it to the archive; ⑤ Update Vulture Locations: Update the locations of all vultures based on the leader vulture's behavior and current strategy; ⑥ Determine file size: Check if the current solution set size exceeds the file size setting. If the solution set size is less than the file size, continue the iteration. If the solution set size is greater than the file size, delete solutions in dense areas through the grid mechanism to preserve the diversity of solutions. ⑦ Update Archive: Update the archive according to the Pareto frontier criteria, retaining non-dominated solutions; ⑧ Perform the next iteration: Increment the iteration count by 1, and proceed to step ② to iterate again.
3. The multi-objective optimization method for laser welding of dissimilar metals such as steel and aluminum according to claim 1, characterized in that, In Step 4, when determining the optimal combination of process parameters from the Pareto fronts of the obtained optimization objectives using the CRITIC evaluation method, a weighted aggregation method is applied to each Pareto front, assigning weights to them. The expression of the comprehensive evaluation model is shown below: In the formula: Z is the comprehensive evaluation value of each experimental design, W j Let y be the weight value of the j-th indicator. ij Let y be the j-th evaluation index value of the i-th scheme. i1 y represents the warpage deformation of the welded parts in each experimental design. i2 y represents the maximum tensile force that the welded component can withstand in each experimental design. i3 The welding cost in each experimental design.