White vehicle body sheet welding positioning layout method based on response surface and genetic algorithm

Abstract

The invention relates to the technical field of automobile body-in-white thin plate welding process analysis, in particular to a body-in-white thin plate welding, positioning and layout method based on a response surface and a genetic algorithm, which comprises the following steps of: performing three-dimensional assembly modeling on body-in-white thin plates; carrying out rigid assembly deviation modeling by utilizing three-dimensional tolerance simulation analysis software; finite element pretreatment is carried out based on rigid modeling, and grid division is carried out by adopting quadrilateral shell unit grid fragments; a body-in-white thin plate positioning design parameter space and an optimization target are determined through flexible assembly deviation modeling; designing a test by adopting a response surface method, and analyzing and calculating all test results by utilizing a three-dimensional flexible assembly deviation simulation model; analyzing a response surface to confirm the advantages and disadvantages of a fitting effect; and carrying out positioning parameter optimization by using an improved genetic algorithm, and obtaining an optimal parameter combination of positioning coordinates on the basis of the established second-order response surface model by taking the change of the positioning coordinate values as a constraint condition and taking the minimum value of the assembly gap size deviation of the key measurement point as a target.

Description

Technical Field

[0001] The invention relates to the technical field of automobile body-in-white thin plate welding process analysis, and in particular to a body-in-white thin plate welding positioning layout method based on response surface and genetic algorithm. Background Art

[0002] The automobile body in white is assembled and welded from flexible thin plates. The assembly quality of the body in white directly affects the appearance and performance of the whole vehicle. Factors such as the design and positioning layout of the body in white thin plate welding fixture have an important impact on the assembly accuracy. The body in white thin plate welding positioning is mainly achieved by fixtures. The change of the positioning point position will cause the fluctuation of assembly deviation. Therefore, it is necessary to select a reasonable positioning point position to reduce production costs.

[0003] At present, the optimization of the BIW thin plate welding fixture positioning points usually adopts the comparison method, which is difficult to obtain the global optimum and ignores the elastic deformation of the parts. Designers use the empirical analysis method to repeatedly modify the position of the fixture positioning points, and determine the optimal layout of the positioning points after multiple numerical simulations. The traditional assembly analysis method considers the BIW thin plate as a rigid part, and cannot consider the gravity and welding deformation of the parts, making it difficult to ensure the optimal layout of the positioning points. A lot of time and energy are spent on selecting parameter combinations and conducting numerical simulation analysis, and the optimization efficiency is low, which cannot meet the design and production requirements of enterprises.

[0004] Therefore, a body-in-white thin plate welding positioning parameter optimization design method based on response surface model and genetic algorithm is proposed to guide the design of tooling fixtures, which not only solves the problems of low optimization efficiency and insufficient precision, but also saves design costs. Summary of the invention

[0005] In view of the problems existing in the prior art, the object of the present invention is to provide a body-in-white thin plate welding positioning layout method based on response surface and genetic algorithm.

[0006] The technical solution adopted by the present invention to solve the technical problem is: a body-in-white thin plate welding positioning layout method based on response surface and genetic algorithm, comprising the following steps: S1: Construct a 3D assembly model of the body-in-white sheet metal, determine the assembly process of parts according to the assembly positioning principle and the flexible sheet metal assembly idea, determine the shape and position tolerances of parts according to standard requirements and experience, and establish a 3D assembly model; S2: 3D assembly model is used to perform rigid assembly deviation analysis modeling, and a rigid assembly deviation analysis model for assembly, measurement, and tolerance analysis is created through 3D tolerance simulation analysis software; S3: Perform finite element pre-processing on the three-dimensional assembly model, use FEA pre-processing software to divide the mesh, set the assembly component materials and unit properties, and use shell unit mesh to divide the assembly model; S4: The three-dimensional assembly model is used for flexible assembly deviation analysis modeling. The FEA solver is selected to import the mesh with physical properties in S3 into the three-dimensional simulation analysis software of S2 to perform flexible assembly point coupling association and obtain the FEA file required for assembly; S5: The three-dimensional flexible simulation model is used to calculate assembly deviations, analyze the coordinate parameters of the positioning points in the assembly process, and determine the positioning design parameter space and optimization targets; S6: Experimental design and sample data set construction. According to the design parameter space and optimization objectives determined in S5, the response surface method is used to design the experiment, and the experimental results are calculated based on the flexible assembly deviation analysis model established in S4. S7: response surface analysis, based on the sample data set obtained in S6, a second-order response surface image reflecting the positioning design parameters and assembly deviation is established, and the image fitting and parameter sensitivity analysis are performed; S8: Establish a second-order response surface model, perform regression processing on the test data based on the response surface method, establish a multivariate regression model by testing the significance of the first-order term, the second-order term and the constant term, fit the second-order polynomial formula, and correct the coefficients of each term in the polynomial; S9: Positioning parameter optimization. According to the response surface image analysis results, the optimization constraints are determined, and the response surface model established in S8 is optimized using the improved genetic algorithm to obtain the positioning design parameters under the optimal parameter combination.

[0007] Preferably, the simulation verification of the optimal parameter combination substitutes the positioning design parameters in S9 into the second-order response surface model in S8 to obtain the predicted value of the target under the optimal parameter combination; at the same time, the coordinate values ​​of the positioning points in the three-dimensional assembly model in S4 are adjusted to the optimal combination for simulation calculation to obtain the simulated value under the optimal parameters, and the effectiveness of the established second-order response surface model is verified by comparing and analyzing the predicted value and the simulated value.

[0008] Preferably, the positions of the three-dimensional assembly model include clamping points, welding points and control measuring points; the clamping points, welding points, pressure points and control measuring points are determined in the rigid assembly deviation analysis modeling.

[0009] Preferably, during the response surface analysis in S7, a response surface is generated based on Design Expert, and the response surface is analyzed and tested for accuracy. The significance of the model is determined by analyzing the variance of the fitted model; the response fitness diagram of the test sample points and the model prediction values ​​is established for analysis, and the closer the test sample points are to the fitted straight line, the higher the model fit; the fitting of the response surface model is tested by the determination coefficient, and the closer its value is to 1, the higher the accuracy of the response surface model.

[0010] Preferably, in S8, Design Expert is used to establish a second-order response surface model, and the general expression of the second-order response surface model is: ; Where: is the calculated value of the response, is the design variable, is the number of independent variables, , , , are the coefficients of the constant term, linear term, quadratic term, and cross term, respectively.

[0011] Preferably, in S9, an improved genetic algorithm is used to optimize the established response surface model, and a genetic algorithm optimization model is established by combining constraints and objective functions: ; Where: is the design variable, is the number of independent variables, is the design variable space, Calculate the value for the response quantity.

[0012] The present invention has the following beneficial effects: The present invention provides a new idea for optimizing the fixture positioning strategy of the body-in-white sheet structure, avoiding repeated and tedious trial calculation processes; The present invention adopts a three-dimensional tolerance simulation analysis method to simulate and model the flexible assembly deviation of the body-in-white sheet, taking into account the material properties and assembly process (positioning, clamping, welding and release) of the body-in-white sheet, and combining the finite element method to accurately simulate the assembly deviation of the flexible sheet, and can fully reflect the deformation of the flexible parts in the actual production and assembly process, as well as the influence of clamping, welding and external force on the deformation of the parts assembly. The flexible assembly deviation analysis method improves the simulation calculation accuracy and efficiency; The present invention establishes a response surface test within the range of the positioning design parameters, and obtains a relatively accurate second-order response surface model by fitting, which intuitively reflects the relationship between the positioning parameters and the objective function. By analyzing the variance, response fitness diagram and determination coefficient of the response surface model, the significance of the model is determined and the quality of the fitting effect is judged; Taking the positioning coordinate parameters satisfying the positioning design parameter space as the constraint condition and the minimum assembly deviation as the goal, the improved genetic algorithm is used for parameter optimization to iterate and obtain the optimal process parameters; Finally, simulation and experimental verification of the optimal parameter combination were carried out, which confirmed the practicability of the optimization design method and can be used to guide the design of tooling and fixtures, while shortening the development cycle, improving work efficiency and saving design costs. BRIEF DESCRIPTION OF THE DRAWINGS

[0013] Figure 1 It is an overall flow chart of the analysis method described in the embodiment of the present invention.

[0014] Figure 2 It is a schematic diagram of the structure of the front windshield upper crossbeam assembly according to an embodiment of the present invention.

[0015] Figure 3 It is a grid unit diagram of the upper crossbeam of the front windshield described in an embodiment of the present invention.

[0016] Figure 4 This is a diagram of the positioning points and measuring points of the crossbeam on the front windshield according to an embodiment of the present invention.

[0017] Figure 5 It is the actual value and predicted value distribution response fitness diagram described in the embodiment of the present invention.

[0018] Figure 6 It is the residual and predicted value distribution response fitness diagram described in the embodiment of the present invention.

[0019] Figure 7 This is a diagram of the measurement points of the assembly gap of the upper crossbeam of the front windshield described in an embodiment of the present invention.

[0020] Figure 8 It is a comparison diagram of assembly results of the original positioning scheme and the optimized scheme of the front windshield upper crossbeam described in the embodiment of the present invention.

[0021] In the figure: 1-lower body of the upper crossbeam of the front windshield; 2-upper body of the upper crossbeam of the front windshield; L1, L2, L3, L4-positioning points; - location of measuring points; , , -Three gap measurement points of the assembly; a-assembly result of the original design; b-assembly result of the optimized design. DETAILED DESCRIPTION

[0022] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely further describe the technical solutions in the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0023] A body-in-white thin plate welding positioning layout method based on response surface and genetic algorithm comprises the following steps: (1) Constructing a three-dimensional assembly model for the body-in-white thin panel: Determine the component assembly process based on the assembly positioning principle and the flexible thin panel assembly idea, determine the component shape and position tolerances based on standard requirements and experience, and establish a three-dimensional assembly model.

[0024] (2) Perform rigid assembly deviation analysis modeling on the constructed 3D assembly model: Use 3D tolerance simulation analysis software to create a rigid assembly deviation analysis model for assembly, measurement, and tolerance analysis, identify the flexible analysis parts in the model, and create key flexible assembly points on the flexible parts.

[0025] (3) Perform finite element pre-processing on the constructed three-dimensional assembly model: Use FEA pre-processing software to divide the mesh, set the assembly component materials and unit properties, and use shell unit mesh to divide the assembly model into meshes to obtain a mesh file that meets the requirements.

[0026] (4) Conduct flexible assembly deviation analysis modeling on the constructed three-dimensional assembly model: First, select a suitable FEA solver to import the mesh unit model with physical properties in step (3) into the three-dimensional simulation analysis software in step (2) to perform flexible assembly point coupling association to obtain an FEA file that meets the assembly requirements. Secondly, based on the rigid assembly modeling in step (2), perform flexible modeling in accordance with the assembly process requirements in step (1).

[0027] (5) Determine the positioning design parameters and optimization targets based on the body-in-white sheet metal assembly process: Calculate the assembly deviation of the three-dimensional flexible simulation model constructed in step (4), analyze the influence trend of the assembly process positioning point coordinate design parameters on the performance evaluation index, and determine the positioning design parameter space and optimization targets.

[0028] (6) Experimental design and sample data set construction: Based on the design parameter space and optimization objectives determined in step (5), the response surface methodology is used to design the experiment, and all experimental results are calculated based on the flexible assembly deviation analysis model established in step (4).

[0029] (7) Response surface analysis: Based on the sample data set obtained in step (6), a second-order response surface image reflecting the positioning design parameters and assembly deviation is established, and the image fit and parameter sensitivity analysis are performed to determine the impact of the positioning design parameters on the assembly deviation.

[0030] (8) Establishing a second-order response surface model: Based on the response surface methodology, the experimental data were regressed and processed. The significance of the linear, quadratic and constant terms of the influencing equation was tested, and a multivariate regression model was established. The quadratic polynomial formula was fitted and the coefficients of each term in the polynomial were corrected.

[0031] (9) Positioning parameter optimization: According to the response surface image analysis results, the optimization constraints are determined, and the improved genetic algorithm is used to optimize the response surface model established in step (8) to obtain the positioning design parameters under the optimal parameter combination.

[0032] (10) Simulation verification of optimal parameter combination: Substitute the optimized positioning design parameters in step (9) into the second-order response surface model in step (8) to obtain the predicted value of the target under the optimal parameter combination; at the same time, adjust the coordinate value of the positioning point in the three-dimensional assembly model in step (4) to the optimal combination for simulation calculation to obtain the simulated value under the optimal parameters. Compare and analyze the predicted value and the simulated value to verify the effectiveness of the established second-order response surface model.

[0033] (11) Trial production after improving the tooling: Use the positioning point target parameters optimized in step (9) to guide the design of the body-in-white sheet metal tooling fixture, and conduct welding process verification to determine the practicality of the optimization method.

[0034] In step (1), the positions of the clamping points, welding points, control measuring points, etc. of the 3D assembly model are clearly defined.

[0035] In step (2), the assembly is rigidly modeled based on 3DCS, and the key flexible assembly points created include clamping points, welding points, pressure points, etc., which serve as the basic points for flexible modeling.

[0036] In step (3), the improved three-dimensional model in step (2) is geometrically pre-processed based on FEA pre-processing software (such as Presys, Hypermesh, etc.). The model is divided using a quadrilateral shell unit mesh independent slicing algorithm. The mesh contains points, elements, mesh type, thickness (shell unit), unit size, material type, and defines the physical properties of each component of the model (elastic modulus, Poisson's ratio, expansion coefficient, etc.).

[0037] In step (4), compliance modeling is performed based on the 3DCS FEA Compliant Modeler, which specifically includes: 1) Based on the rigid model established in step (2), the Autobend assembly method is used to solve the over-constrained positioning problem of the main positioning surface, and the Bend point is selected for geometric bending deformation processing.

[0038] 2) Use FEA solution tools (such as Abaqus, Nastran, etc.) to obtain the stiffness matrix, mass matrix and mesh file that meet the requirements by coupling the finite element model nodes with the three-dimensional deviation analysis model points.

[0039] 3) Load the mesh, stiffness matrix, and mass matrix into the software and perform nominal assembly so that the node set in the mesh of step (3) is associated with the flexible points established on the model in step (2).

[0040] 4) Create flexible assemblies based on assembly process (positioning, clamping, welding and releasing).

[0041] In step (7), when performing response surface analysis, response surfaces are generated based on 3DCS FEA and Design Expert respectively, and the response surface model is analyzed and the accuracy is tested to evaluate the quality of the fitting effect: 1) Determine the significance of the model by performing variance analysis on the fitted model; 2) Analyze by establishing a response fitness graph of the test sample points and the model prediction values. The closer the test sample points are to the fitting line, the better the model fitting effect. 3) The fitting of the response surface model was tested by the coefficient of determination. The closer its value was to 1, the higher the accuracy of the response surface model.

[0042] In step (8), Design Expert is used to establish a second-order response surface model. The general expression of the second-order response surface model is: ; Where: is the calculated value of the response, is the design variable, is the number of independent variables, , , , are the coefficients of the constant term, linear term, quadratic term, and cross term, respectively.

[0043] In step (9), the established response surface model is optimized using an improved genetic algorithm. Combining the constraints and the objective function, the genetic algorithm optimization model is established as follows: ; Where: is the design variable, is the number of independent variables, is the design variable space, Calculate the value for the response quantity.

[0044] Example A method for positioning and layout of thin-plate welding of a body-in-white based on response surface and genetic algorithm mainly includes three-dimensional modeling, rigid assembly deviation analysis modeling, finite element pre-processing, flexible assembly deviation analysis modeling, assembly deviation simulation calculation, response surface analysis, target parameter optimization, simulation verification, test scheme verification, etc., wherein the rigid assembly deviation analysis modeling and the flexible assembly deviation analysis modeling are completed on the same platform. The specific implementation steps of the present invention are as follows: (1) Constructing a 3D assembly model: This embodiment takes the upper crossbeam assembly of the front windshield of a heavy truck cab as an example, and performs 3D solid modeling using Catia software. Figure 2 The assembly mainly includes an upper body and a lower body, which are welded together by positioning, clamping, welding and releasing processes.

[0045] (2) Perform rigid assembly deviation analysis modeling on the assembly: Determine the geometric tolerances and performance evaluation indicators of the assembly parts based on the process design parameter statistical database. Perform 3DCS rigid assembly deviation modeling based on the assembly process of the assembly in step (1) to identify key flexible assembly points such as positioning points, clamping points, and welding points.

[0046] (3) Assembly FEA pre-processing: Perform Hypermesh geometry pre-processing on the assembly model established in step (2), use quadrilateral shell element mesh slices for mesh division, and focus on mesh refinement of the flexible assembly point area in step (2). Define the material and element properties of each component of the assembly, as shown in Table 1. Finally, a mesh file that meets the requirements is obtained. Figure 3 It is the grid unit of the crossbar on the front windshield.

[0047] Table 1 Materials and unit properties of the front windshield upper beam assembly ; (4) Modeling the flexible assembly deviation analysis of the assembly: Use the 3DCS finite element flexible module for modeling, call the Abaqus solver to couple the finite element model nodes in step (3) with the three-dimensional deviation analysis model points in step (2), and obtain the stiffness matrix, mass matrix and mesh file that meet the flexible assembly deviation requirements. Based on the modeling in step (2), use the 3DCS_Autobend module and FEA module to complete the main positioning surface assembly and other flexible assembly modeling in step (1).

[0048] (5) Determine the positioning design parameters and optimization targets based on the assembly process of the front windshield upper crossbeam: The coordinate values ​​of the positioning points on the main positioning surface of the crossbeam are used as parameter variables, and the assembly deviation of the target measuring point of the assembly is calculated. The 3DCS flexible assembly deviation simulation analyzes the influence of the coordinate parameter changes of the positioning point on the assembly deviation of the measuring point, and finally determines the positioning point (L1~L4) The change of the coordinate value has a significant impact on the assembly deviation of the measuring point. The range of the coordinates is shown in Table 2. The arrangement of the crossbeam positioning points and measuring points on the front windshield is shown in Table 2. Figure 4 shown.

[0049] Table 2 Design parameters of the front windshield upper beam positioning ; Experimental design and sample data set construction: According to the established three-dimensional flexible assembly deviation simulation model, the Box-Behnken method was used to design the experiment. The experiment had 4 factors and 3 levels, as shown in Table 3, and the number of experiments was 29. The process parameters and levels of the Box-Behnken Design (BBD) experiment were determined, among which the positioning point The coordinate value change is the process parameter variable, and each parameter variable corresponds to the coordinate of a specific point on the model. Based on the Monte Carlo simulation assembly of the front windshield upper beam in step (4), all test results are obtained. A set of data is calculated each time the parameter changes, and then a sample data set is constructed.

[0050] Table 3 BBD test factors and levels ; (7) Response surface analysis: Based on the experimental data set constructed in step (6), the response surface graph is generated using the Design Expert software, as shown in Figure 5 , Figure 6 As shown. Perform image fitting analysis and precision test to evaluate the quality of the fitting effect: 1) Through the variance analysis of the fitted model, the variance analysis results show that the model The value is 44.37, The value is less than 0.0001, so the model is significant.

[0051] 2) By establishing a response fitness graph of the test sample points and the model prediction values, the closer the actual value of the sample point is to the predicted value, the closer the points in the image are to a straight line distribution, and the more random the residual values ​​of the data points are and the better the fitting effect is. Figure 5 , Figure 6 The goodness of fit graph of the response surface shows that the fitting effect is good.

[0052] 3) The response surface model fitting was tested by the coefficient of determination. The closer the value is to 1, the higher the accuracy of the response surface model. As shown in Table 4, the coefficient of determination and the modified coefficient of determination of the model were obtained by using the Design Expert software. As shown in Table 4, the coefficients of determination of the response surface model are all greater than 0.95, indicating that the model has a high accuracy.

[0053] Table 4 Determination coefficient of response surface model ; (8) Establish a second-order response surface model: Based on Design Expert, the sample data set constructed in step (6) is regressed and fitted to obtain a quadratic formula. The coefficients of each formula are modified to establish a second-order response surface model. The model contains a constant term, a linear term, and a quadratic term. The quadratic term contains a coupling term for any two input parameters. The coefficients and signs indicate the significance and direction of the influence of each input parameter on the shell characteristics. The model clearly shows the degree of influence of the input parameters on the output parameters.

[0054] The general expression of the second-order response surface model is: ; Where: is the calculated value of the response, is the design variable, is the number of independent variables, , , , are the coefficients of the constant term, linear term, quadratic term, and cross term, respectively.

[0055] Fitting to obtain the target value of the deviation of key measuring points and (L1~L4) positioning points Direction coordinates ( , , , ) is as follows: ; (9) Positioning parameter optimization: Improved genetic algorithm is used for parameter optimization, that is, the improved genetic algorithm with good local search function is used to achieve model optimization. The algorithm uses global and local hybrid search methods to perform optimization iterations to achieve the purpose of rapid optimization. The program is compiled through MATLAB. The sample data comes from step (6) and the positioning points (L1~L4) are used. The coordinate value change is the constraint condition, and the assembly gap size deviation of the key measuring point is The goal is to minimize the value and finally find the optimal parameter combination of the front windshield beam positioning coordinates.

[0056] The optimization model is: ; When optimizing the genetic algorithm, the parameters are set as follows: population size is 100, termination generation is 100, global crossover and mutation probabilities are 0.50 and 0.10, and local crossover and mutation probabilities are 0.80 and 0.15. The optimal parameter values ​​obtained are: is -0.5979, is 0.7754, is 0.9943, is 0.1243, which is converted into (L1~L4) positioning points To the real coordinate value: is 761.565, 801.110, is -739.010, The true coordinates of the positioning points under the optimal parameters are determined by the grid model established in step (3), as shown in Table 5.

[0057] Table 5 Coordinate values ​​of positioning points under optimal parameters ; (10) Simulation verification of optimal parameter combination: Substitute the optimized positioning design parameters in step (9) into the second-order response surface model in step (8) to obtain the predicted value of the target under the optimal parameter combination; at the same time, adjust the positioning point coordinate values ​​in the three-dimensional assembly model in step (4) to the optimal combination for 3DCS simulation calculation, and obtain the predicted value and simulated value of the target value of the assembly gap size deviation of the key measuring point, as shown in Table 6. The error between the predicted value and the simulated value is 0.1%, which proves the effectiveness of the second-order response surface approximation model and its high prediction accuracy.

[0058] Table 6 Prediction and simulation results of optimal parameter combination ; The optimized positioning parameters are used to conduct test simulation verification and evaluate the assembly quality. Figure 7 The original positioning scheme and the optimized design scheme of the front windshield upper crossbeam are simulated and analyzed in 3DCS to obtain comparative analysis results, as shown in Table 7 and Figure 8 shown.

[0059] Table 7 Comparison of numerical results between original design and optimized design ; From Table 6 and Figure 8 It can be seen that under the optimized design scheme of the front windshield beam positioning layout, the average measurement value of the assembly gap of the key measuring points is reduced by 3.2% compared with the original scheme, the average value of the standard deviation is reduced by 16.5%, and the average deviation rate is reduced by 64.6%.

[0060] (11) Trial production after improving the tooling: In actual production, the positioning point target parameters optimized in step (9) are used to guide the design of the tooling fixture, and trial welding verification is carried out. Compared with the original positioning design scheme, the qualified rate of the measuring points is increased by 11.7%.

[0061] The present invention is not limited to the above-mentioned implementation modes, and anyone should be aware of the structural changes made under the enlightenment of the present invention, and all those having the same or similar technical solutions as the present invention fall within the protection scope of the present invention.

[0062] The techniques, shapes, and structural parts not described in detail in the present invention are all well-known techniques.

Claims

1. A body-in-white sheet metal welding positioning layout method based on response surface and genetic algorithm, characterized in that: The following steps are involved: S1: Construct a 3D assembly model of the body-in-white sheet metal, determine the assembly process of parts based on the assembly positioning principle and the flexible sheet metal assembly idea, determine the shape and position tolerances of parts based on standard requirements and experience, and establish a 3D assembly model; S2: 3D assembly model is used to perform rigid assembly deviation analysis modeling, and a rigid assembly deviation analysis model for assembly, measurement, and tolerance analysis is created through 3D tolerance simulation analysis software; S3: Perform finite element pre-processing on the three-dimensional assembly model, use FEA pre-processing software to divide the mesh, set the assembly component materials and unit properties, and use shell unit mesh to divide the assembly model; S4: The three-dimensional assembly model is used for flexible assembly deviation analysis modeling. The FEA solver is selected to import the mesh with physical properties in S3 into the three-dimensional simulation analysis software of S2 to perform flexible assembly point coupling association and obtain the FEA file required for assembly; S5: The three-dimensional flexible simulation model is used to calculate assembly deviations, analyze the coordinate parameters of the positioning points in the assembly process, and determine the positioning design parameter space and optimization targets; S6: Experimental design and sample data set construction. According to the design parameter space and optimization objectives determined in S5, the response surface method is used to design the experiment, and the experimental results are calculated based on the flexible assembly deviation analysis model established in S4. S7: response surface analysis, based on the sample data set obtained in S6, a second-order response surface image reflecting the positioning design parameters and assembly deviation is established, and the image fitting and parameter sensitivity analysis are performed; S8: Establish a second-order response surface model, perform regression processing on the test data based on the response surface method, establish a multivariate regression model by testing the significance of the first-order term, the second-order term and the constant term, fit the second-order polynomial formula, and correct the coefficients of each term in the polynomial; S9: Positioning parameter optimization. According to the response surface image analysis results, the optimization constraints are determined, and the response surface model established in S8 is optimized using the improved genetic algorithm to obtain the positioning design parameters under the optimal parameter combination.

2. The body-in-white sheet metal welding positioning layout method based on response surface and genetic algorithm according to claim 1 is characterized in that: The simulation verification of the optimal parameter combination is to substitute the positioning design parameters in S9 into the second-order response surface model in S8 to obtain the predicted value of the target under the optimal parameter combination; at the same time, the coordinate values ​​of the positioning points in the three-dimensional assembly model in S4 are adjusted to the optimal combination for simulation calculation to obtain the simulated value under the optimal parameters, and the effectiveness of the established second-order response surface model is verified by comparing and analyzing the predicted value and the simulated value.

3. The body-in-white sheet metal welding positioning layout method based on response surface and genetic algorithm according to claim 1 is characterized in that: The positions of the three-dimensional assembly model include clamping points, welding points and control measuring points; the clamping points, welding points, pressure points and control measuring points are determined in the rigid assembly deviation analysis modeling.

4. The body-in-white sheet metal welding positioning layout method based on response surface and genetic algorithm according to claim 1 is characterized in that: In the response surface analysis in S7, the response surface is generated based on Design Expert, the response surface is analyzed and the precision is tested, and the significance of the model is determined by variance analysis of the fitted model; By establishing a response fitness graph of the test sample points and the model prediction values, the closer the test sample points are to the fitting line, the higher the model fit; The fitting of the response surface model was tested by the determination coefficient. The closer its value was to 1, the higher the accuracy of the response surface model.

5. The body-in-white sheet metal welding positioning layout method based on response surface and genetic algorithm according to claim 1 is characterized in that: In S8, Design Expert is used to establish a second-order response surface model, and the general expression of the second-order response surface model is: ； Where: is the calculated value of the response, is the design variable, is the number of independent variables, , , , are the coefficients of the constant term, linear term, quadratic term, and cross term, respectively.

6. The body-in-white sheet metal welding positioning layout method based on response surface and genetic algorithm according to claim 1 is characterized in that: In S9, an improved genetic algorithm is used to optimize the established response surface model, and a genetic algorithm optimization model is established by combining constraints and objective functions: ； Where: is the design variable, is the number of independent variables, is the design variable space, Calculate the value for the response quantity.

Images