Method and system for predicting and compensating stamping springback of automobile lengthened longitudinal beam

Abstract

The invention discloses an automobile lengthened longitudinal beam stamping springback prediction and compensation method, which comprises the following steps: constructing an initial data set containing an initial prediction value according to the influence of preset process parameters on stamping springback; constructing a springback prediction model by adopting an incremental plasticity theory, training the springback prediction model by adopting a gradient boosting decision tree, and adjusting process parameters by adopting a Bayesian optimization method; according to the optimal process parameters, finite element simulation is adopted to manufacture a longitudinal beam, simulation stamping springback data is generated, and a genetic algorithm and a variational auto-encoder are adopted to optimize a die compensation shape; and calculating compensation data of auxiliary stamping according to the die compensation shape, generating longitudinal beam processing data, and taking the longitudinal beam processing data as real stamping process parameters of the longitudinal beam. According to the stamping springback prediction and compensation method for the lengthened longitudinal beam of the automobile, the springback prediction accuracy can be improved, and the forming quality of the longitudinal beam is improved based on the optimized mold compensation shape and the auxiliary stamping compensation data.

Description

Technical Field

[0001] The present invention relates to the technical field of automotive component stamping, and particularly relates to a method and system for predicting and compensating springback in the stamping of automotive extended longitudinal beams. Background Art

[0002] Automotive longitudinal beams are key load-bearing components in the body structure, and their processing accuracy directly affects the safety and stability of the entire vehicle. Due to the elastic recovery effect of materials during the stamping forming process, the geometric shape of longitudinal beam parts often produces springback errors after stamping, resulting in the actual shape deviating from the design requirements.

[0003] Currently, stamping process designers usually try to minimize springback by optimizing stamping process parameters and adjusting die structures during the process design and die design stages; however, due to the complex geometric shapes and diverse material properties of longitudinal beam products, and multiple processes involved in the forming process, it is difficult to completely eliminate stamping springback through simple design adjustments.

[0004] Therefore, designers often need to rely on experience and optimize the shape of stamping dies through repeated tests, die repair, and trial molding to control the springback error of the product within a reasonable range. However, this test-based method not only has a long cycle and high cost but may also lead to unstable compensation effects due to human experience deviations.

[0005] Therefore, there is an urgent need for a method for predicting and compensating stamping springback that can optimize process parameters and the shape of stamping dies in the early stage of production and manufacturing, reduce the number of die repair tests, and lower manufacturing costs. Summary of the Invention

[0006] Object of the Invention: To overcome the above deficiencies, the object of the present invention is to provide a method and system for predicting and compensating springback in the stamping of automotive extended longitudinal beams. By combining incremental plasticity theory, gradient boosting decision trees, Bayesian optimization, genetic algorithms, and variational autoencoders, it can improve the accuracy of springback prediction and, based on the optimized die compensation shape and auxiliary stamping compensation data, improve the forming quality of longitudinal beams.

[0007] To solve the above technical problems, the present invention provides a method for predicting and compensating springback in the stamping of automotive extended longitudinal beams, including:

[0008] Step S1: Construct an initial data set containing initial prediction values according to the influence of preset process parameters on stamping springback;

[0009] Step S2: Construct a springback prediction model using incremental plasticity theory, then train the springback prediction model using gradient boosting decision trees and adjust process parameters using Bayesian optimization method to generate optimal process parameters;

[0010] Step S3: Manufacture the longitudinal beam by finite element simulation according to the optimal process parameters, generate simulation stamping springback data, and optimize the die compensation shape by using genetic algorithm and variational autoencoder according to the simulation stamping springback data;

[0011] Step S4: Calculate the longitudinal beam compensation data for assisted stamping according to the die compensation shape, generate the longitudinal beam compensation data, and use the longitudinal beam processing data as the true stamping process parameters of the longitudinal beam.

[0012] As a preferred embodiment of the present invention, in step S1, the method includes:

[0013] Step S11: Select a material model and define the elastic modulus, Poisson's ratio, and yield criterion;

[0014] Step S12: Perform mesh division using four-node shell elements or eight-node solid elements;

[0015] Step S13: Set preset process parameters;

[0016] Step S14: Simulate the stamping process of the longitudinal beam by finite element simulation, including: applying stamping load to simulate the stress and deformation of the sheet metal, calculating the stress field and strain field after deformation; applying boundary conditions to remove the stamping load, and calculating the residual stress and the maximum springback amount and springback angle after deformation;

[0017] Step S15: Extract the maximum springback amount, residual stress distribution, and springback angle change under the preset process parameters to generate an initial prediction value.

[0018] As a preferred embodiment of the present invention, in step S2, the method includes:

[0019] Construct a springback prediction model using the incremental plasticity theory:

[0020]

[0021] where Δε is the predicted springback error, E is the material elastic modulus, σ res is the residual stress after stamping, and σ y is the material yield stress.

[0022] As a preferred embodiment of the present invention, in step S2, the method further includes:

[0023] Step S21: Use the preset process parameters as input features and train the springback prediction model using gradient boosting decision tree:

[0024]

[0025] where F t(x) is the decision tree in the t-th round, T is the number of decision trees, and R ptredicted is the springback error predicted by the model;

[0026] Step S22: Define the Bayesian optimization objective function:

[0027] f(x) = min|R real -R ptredicted (x)|

[0028] where R real is the springback error measured in the experiment, and x is the preset process parameter;

[0029] Step S23: Optimize the preset process parameters using the Gaussian process regression method:

[0030] p(y|x) = N(μ(x), k(x, x'))

[0031] where p(y|x) is the probability distribution of the springback error y after the given parameter x, μ(x) is the mean function representing the estimate of the springback error, and k(x, x') is the covariance function representing the correlation between parameters;

[0032] Step S24: Evaluate the generalization ability of the springback prediction model using five-fold cross-validation;

[0033] Step S25: Iteratively loop through steps S21 to S24 a preset number of times until the preset requirements are met and output the optimal process parameters.

[0034] As a preferred embodiment of the present invention, in step S21, the method includes:

[0035] Step S210: Initialize the springback error prediction value:

[0036]

[0037] where F0(x) is the mean of the springback error, N is the number of training samples, and R real,i is the springback error of the i-th sample;

[0038] Step S211: Calculate the deviation between the springback error and the predicted value:

[0039]

[0040] where is the deviation of the i-th sample in the t-th iteration, and F t-1 (x i ) is the predicted springback error of the i-th sample in the (t-1)-th iteration;

[0041] Step S212: Fit the deviation between the springback error and the predicted value:

[0042]

[0043] where F t (x)' is the decision tree for the t-th round of training, used to fit the residual, and F(x i ) is the predicted value of the trained decision tree model for the i-th sample, and argmin is the optimization objective that minimizes the sum of squared deviations;

[0044] Step S213: Update the predicted value and control the step size with the learning rate η:

[0045] F t (x) = F t-1 (x) + ηF t (x)

[0046] where F t (x) is the predicted value of the model after the t-th iteration, and F t-1 (x) is the predicted value after the (t - 1)-th iteration;

[0047] Step S214: Loop steps S211 to S213 for a preset number of times.

[0048] As a preferred embodiment of the present invention, in step S3, the method includes:

[0049] Step S31: According to the optimal process parameters, the selected material model, the defined elastic modulus, Poisson's ratio, and yield criterion, use finite element simulation to manufacture the longitudinal beam and output simulation stamping springback data including the maximum springback amount, residual stress distribution, springback angle change, and stress distribution in the plate thickness direction;

[0050] Step S32: Use a variational autoencoder to learn the relationship between the die shape and the stamping springback data, and generate a parameterized initial compensation shape;

[0051] Step S33: Sample the initial die shape from the initial compensation shape generated by the variational autoencoder and optimize the initial die shape using a genetic algorithm to generate a die compensation shape;

[0052] Step S34: Use the die compensation shape to perform finite element simulation again, and determine whether the springback error is reduced. If so, it is determined that the optimization is completed. If not, return to step S33 and use the genetic algorithm to optimize the die compensation shape again.

[0053] As a preferred embodiment of the present invention, in step S33, the method includes:

[0054] Step S331: Sample the initial die shape from the initial compensation shape generated by the variational autoencoder;

[0055] Step S332: Calculate the fitness function;

[0056] Step S333: Select individuals with high fitness to enter the next generation using a preset selection method;

[0057] Step S334: Perform crossover and mutation operations;

[0058] Step S335: Repeat Steps S332 to S334 until a preset number of iterations or convergence criteria are met.

[0059] As a preferred embodiment of the present invention, in Step S4, the method includes:

[0060] Step S41: Calculate the local compensation offset based on the optimized die compensation shape:

[0061] Δh = h real -h compensated

[0062] where h real is the die shape after uncompensated stamping, and h compensated is the ideal shape of the die after compensation;

[0063] Step S42: Calculate the local stress distribution:

[0064] σ res (x,y) = σ plastic (x,y) - σ elastic (x,y)

[0065] where σ plastic (x,y) is the plastic stress, and σ elastic (x,y) is the elastic stress;

[0066] Step S43: Set the springback critical threshold and select the area meeting the first preset condition as the local heating area:

[0067] A critical = {(x,y)|σ res (x,y) > σ crit}

[0068] where σ crit is the springback critical stress;

[0069] Step S44: Use finite element heat conduction analysis to calculate the temperature field of the local heating area:

[0070]

[0071] where ρ is the material density, c p is the specific heat capacity, k is the thermal conductivity, and Q is the heating source term;

[0072] Step S45: Calculate the locally optimal heating temperature:

[0073]

[0074] where σ0 is the room temperature yield stress, T0 is the initial temperature, T m is the melting point temperature, and m is the material parameter;

[0075] Step S46: Calculate the locally optimal heating time:

[0076]

[0077] where ΔT is the desired temperature rise and Q is the heating power.

[0078] As a preferred embodiment of the present invention, in step S43, the method further includes:

[0079] Step S431: Select the area satisfying the second preset condition as the ultrasonic vibration action area:

[0080] A UV ={(x,y)|σ res (x,y)>σ crit}

[0081] Step S432: Calculate the flow stress reduced by ultrasonic vibration:

[0082] σ eff =σ flow -αf US

[0083] where σ flow is the material resistance without ultrasonic vibration, α is the ultrasonic softening coefficient, and f US is the ultrasonic frequency; Step S433: Calculate the optimal ultrasonic parameters:

[0084] A US =k·σ max

[0085] where k is the empirical coefficient and σ max is the local maximum offset.

[0086] The present application also provides an automobile lengthened longitudinal beam stamping springback prediction and compensation system using the above method, including:

[0087] A data construction module, configured to calculate the influence of preset process parameters on stamping springback and construct an initial data set including an initial prediction result;

[0088] A model optimization module, which is used to construct a springback prediction model by using the incremental plasticity theory, and then adjust process parameters by using the Bayesian optimization method and train the springback prediction model by using the gradient boosting decision tree method to minimize the springback error;

[0089] A simulation compensation module, which is used to manufacture a longitudinal beam through preset process parameters by using finite element simulation, generate simulation stamping springback data, and optimize the die compensation shape by using the genetic algorithm and variational autoencoder according to the simulation stamping springback data;

[0090] A real feedback module, which is used to calculate the longitudinal beam compensation data for assisted stamping according to the die compensation shape, generate the longitudinal beam compensation data, and use the longitudinal beam processing data as the real stamping process parameters of the longitudinal beam.

[0091] In some embodiments, the present application also relates to a computer medium, on which a computer program is stored, and the computer program is executed by a processor to implement the method for predicting and compensating stamping springback of an extended longitudinal beam of an automobile.

[0092] In some embodiments, the present application also relates to a computer, including the above-mentioned computer medium.

[0093] The above technical solution of the present application has the following advantages compared with the prior art:

[0094] 1. The present application calculates the springback error by using the incremental plasticity theory, combines finite element simulation, establishes a high-precision springback prediction model, effectively avoids the problem of large predicted springback error, and at the same time uses the gradient boosting decision tree to model the springback error, adjusts process parameters by using the Bayesian optimization method, improves the adaptability and generalization ability of the springback prediction model, and then uses five-fold cross-validation to evaluate the springback prediction model to ensure accurate prediction of springback error under different materials and different stamping process parameters.

[0095] 2. The present application uses a variational autoencoder to learn the mapping relationship between the die shape and stamping springback data, generates a parameterized initial compensation shape, improves the compensation shape optimization efficiency, and then uses the genetic algorithm to further optimize the die compensation shape, avoids local optimum, improves the accuracy of the die compensation shape, and then verifies the effectiveness of the die compensation shape through finite element simulation to ensure that the shape of the compensated longitudinal beam meets the design requirements.

[0096] 3. Calculate the local compensation offset based on the optimized die compensation shape, accurately adjust the die shape, improve the machining accuracy, and then use the method of combining local heating and ultrasonic-assisted stamping to calculate and adjust the assisted stamping compensation data, optimize the stamping springback control strategy, and ensure that the error between the size of the longitudinal beam after forming and the target size does not exceed 1%. Description of the Drawings

[0097] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the drawings in the following description are only the embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on the provided drawings.

[0098] Figure 1 It is a flowchart of a method for predicting and compensating the springback of a lengthened longitudinal beam of an automobile provided by an embodiment of the present invention.

[0099] Figure 2 It is a flowchart of a method for constructing an initial data set provided by an embodiment of the present invention.

[0100] Figure 3 It is a flowchart of a method for generating optimal process parameters provided by an embodiment of the present invention.

[0101] Figure 4 It is a flowchart of a method for optimizing the die compensation shape provided by an embodiment of the present invention.

[0102] Figure 5 It is a flowchart of a local heating-assisted stamping method provided by an embodiment of the present invention.

[0103] Figure 6 It is a flowchart of an ultrasonic vibration-assisted stamping method provided by an embodiment of the present invention.

[0104] Figure 7 It is a module connection diagram of a system for predicting and compensating the springback of a lengthened longitudinal beam of an automobile provided by an embodiment of the present invention.

[0105] Explanation of the reference numerals in the drawings of the specification:

[0106] 100. Data construction module, 101. Model optimization module, 102. Simulation compensation module, 103. Real feedback module. Detailed implementation manners

[0107] The following will describe in detail the embodiments of the present invention. The examples of the embodiments are shown in the drawings, where the same or similar reference numerals represent the same or similar elements or elements with the same or similar functions throughout. The embodiments described below with reference to the drawings are exemplary and are intended to explain the present invention, and should not be construed as a limitation of the present invention.

[0108] Refer to Figure 1 As shown, in some embodiments, a method for predicting and compensating the springback of a lengthened longitudinal beam of an automobile involves the following steps:

[0109] Step S1: Construct an initial data set containing initial predicted values according to the influence of preset process parameters on stamping springback.

[0110] Specifically, referring to Figure 2 As shown, in step S1, the method includes:

[0111] Step S11: Select a material model and define the elastic modulus, Poisson's ratio, and yield criterion.

[0112] Among them, the material model includes but is not limited to bilinear hardening model, multilinear isotropic hardening model, Johnson-Cook Strength plastic material constitutive model, etc., and the yield criterion includes but is not limited to VonMises, Hill48, etc.; specifically, the bilinear hardening model is usually applicable to low-carbon steel, and the elastic-plastic behavior of the material is simulated by the yield stress and hardening modulus. The linear isotropic hardening model is usually applicable to high-strength steel, allowing multiple segments of non-linear stress-strain relationships. The J-C plastic material constitutive model is usually applicable to thermally sensitive materials under high strain rate conditions, such as aluminum alloys, stainless steels, etc. Specifically, the VonMises yield criterion is usually applicable to isotropic materials, and the Hill48 yield criterion is usually applicable to anisotropic materials.

[0113] Exemplarily, for example, if DP980 high-strength steel is used in this application, the multilinear isotropic hardening model and VonMises yield criterion are selected.

[0114] Step S12: Perform mesh division using four-node shell elements or eight-node solid elements.

[0115] Among them, the four-node shell elements are usually applicable to thin plate stamping, and the eight-node solid elements are usually applicable to complex deformation regions; the size of the specific mesh is set according to the size of the longitudinal beam workpiece.

[0116] Step S13: Set preset process parameters, and the preset process parameters include but are not limited to parameters related to the stamping process such as punching force, punching time, punching temperature, etc.

[0117] Step S14: Use finite element simulation to simulate the longitudinal beam stamping process, including: applying stamping load to simulate the stress and deformation of the sheet metal, calculating the stress field and strain field after deformation; applying boundary conditions to remove the stamping load, and calculating the residual stress, the maximum springback amount and springback angle after deformation.

[0118] Among them, the simulation process is:

[0119] Apply stamping load, calculate the stress and strain fields, and obtain the stress distribution of the workpiece in the maximum deformation state;

[0120] Apply boundary conditions, release the stamping load, and calculate the residual stress and the maximum springback amount after deformation;

[0121] Calculate the springback angle, which is calculated from the residual bending moment.

[0122] Among them, the maximum springback amount is obtained by subtracting the length of the target workpiece from the length of the workpiece after unloading; the springback angle is:

[0123]

[0124] Among them, M res is the residual bending moment, L is the length of the sheet, E is the elastic modulus, h is the thickness of the sheet, and b is the width of the sheet.

[0125] Step S15: Extract the maximum springback amount, residual stress distribution, and springback angle change under the preset process parameters to generate an initial prediction value.

[0126] Step S2: Construct a springback prediction model using the incremental plasticity theory, then train the springback prediction model using the gradient boosting decision tree and adjust the process parameters using the Bayesian optimization method to generate the optimal process parameters.

[0127] Specifically, a brief introduction to the incremental plasticity theory:

[0128] During the stamping process, the material undergoes a loading-unloading process:

[0129] In the loading stage, the punching force causes plastic deformation and the stress state changes;

[0130] In the unloading stage, the sheet metal rebounds, and part of the plastic strain cannot be recovered, resulting in residual stress and shape error.

[0131] Thus, the basic equation of the incremental plasticity theory:

[0132] dε = dε e + dε p

[0133] Among them, dε is the total strain increment, dε e is the elastic strain increment, and dε p is the plastic strain increment.

[0134] In the unloading stage, the material mainly shows elastic recovery, and its plastic strain no longer increases. Therefore, the incremental plasticity theory is used to construct a springback prediction model:

[0135]

[0136] Among them, Δε is the springback error, E is the material elastic modulus, σ res is the residual stress after stamping, σy is the yield stress of the material.

[0137] Specifically, referring to Figure 3 as shown, in step S2, the method further includes:

[0138] Step S21: Using the preset process parameters as input features, training the springback prediction model with a gradient boosting decision tree:

[0139]

[0140] where, F t (x) is the t-th decision tree, T is the number of decision trees, and R ptredicted is the springback error predicted by the model.

[0141] Step S22: Defining the Bayesian optimization objective function:

[0142] f(x) = min|R real - R ptredicted (x)|

[0143] where, R real is the springback error measured experimentally, and x is the preset process parameter.

[0144] where, the optimization objective is to adjust the preset process parameters to make the predicted springback error close to the true error.

[0145] Step S23: Optimizing the preset process parameters using the Gaussian process regression method:

[0146] p(y|x) = N(μ(x), k(x,x'))

[0147] where, p(y|x) is the probability distribution of the springback error y after given the parameter x, μ(x) is the mean function representing the estimation of the springback error, and k(x,x') is the covariance function representing the correlation between parameters.

[0148] where, the covariance function (kernel function) preferably uses a radial basis kernel function:

[0149]

[0150] where, σ 2 is the variance of the kernel function to control the prediction range; l is the length scale to control the similarity between different inputs; thus, the optimization process of the Gaussian process regression method is: predicting μ(x) and the variance k(x,x') based on the current springback error, and then selecting the optimal parameter x to minimize the objective function.

[0151] Step S24: Evaluating the generalization ability of the springback prediction model using five-fold cross-validation.

[0152] Among them, five-fold cross-validation means dividing the dataset into 5 parts, using 4 parts for training and 1 part for testing each time, repeating 5 times, and then ensuring the generalization ability of the model and preventing overfitting based on the mean squared error and coefficient of determination.

[0153] Step S25: Iteratively loop through steps S21 to S24 for a preset number of times until the preset requirements are met and the optimal process parameters are output.

[0154] Among them, the preset number of times in this application is set by the designer according to actual needs, and the preset requirements refer to the springback error converging to a set threshold, for example, |R real -R ptredicted | < 0.05mm.

[0155] In some examples of this application, as shown in Figure 3 In step S21, the method further includes:

[0156] Step S210: Initialize the predicted value of the springback error:

[0157]

[0158] Among them, F0(x) is the mean of the springback error, N is the number of training samples, and R real,i is the springback error of the i-th sample; since the gradient boosting decision tree gradually improves the prediction accuracy by learning the residuals, the initial predicted value uses the mean of the springback errors of all samples as the benchmark.

[0159] Step S211: Calculate the deviation between the springback error and the predicted value:

[0160]

[0161] Among them, is the deviation of the i-th sample in the t-th iteration, and F t-1 (x i ) is the predicted springback error of the i-th sample in the (t - 1)-th iteration; thus, the error between the current model's predicted value and the true value is measured by this deviation.

[0162] Step S212: Fit the deviation between the springback error and the predicted value:

[0163]

[0164] Among them, F t (x)' is the decision tree trained in the t-th round, used to fit the residuals, and F(x i) is the predicted value of the $i$-th sample by the trained decision tree model, and argmin is the optimization objective that minimizes the sum of squared deviations; that is, train a new decision tree to predict the deviation to find the relationship between the input $x$ and the residual, so that the decision tree can minimize the error.

[0165] Step S213: Update the predicted value and control the step size with the learning rate $\eta$:

[0166] F t (x) = F t-1 (x) + $\eta$F t (x)

[0167] where, F t (x) is the predicted value of the model after the $t$-th iteration, and F t-1 (x) is the predicted value after the $(t - 1)$-th iteration; the learning rate $\eta$ is usually greater than 0 and less than or equal to 1, which controls the contribution of the new decision tree to the final predicted value.

[0168] Step S214: Loop steps S211 to S213 for a preset number of times.

[0169] where, the preset number of times in this application is set by the designer according to actual needs.

[0170] Step S3: Manufacture the longitudinal beam by finite element simulation according to the optimal process parameters, generate simulation stamping springback data, and optimize the die compensation shape by using the genetic algorithm and variational autoencoder according to the simulation stamping springback data.

[0171] Specifically, as shown in Figure 4 , in step S3, the method includes:

[0172] Step S31: Manufacture the longitudinal beam by finite element simulation according to the optimal process parameters, the selected material model, the defined elastic modulus, Poisson's ratio and yield criterion, and output simulation stamping springback data including the maximum springback amount, residual stress distribution, springback angle change, and stress distribution in the plate thickness direction.

[0173] In this way, based on the optimal process parameters, perform stamping simulation and extract the springback data after stamping.

[0174] Step S32: Use the variational autoencoder to learn the relationship between the die shape and the stamping springback data, and generate a parameterized initial compensation shape.

[0175] where, by learning the relationship between the die shape $M$ and the stamping springback data, generate a possible distribution of compensation shapes; the structure of the variational autoencoder includes an encoder and a decoder, where the encoder is:

[0176]

[0177] Among them, the initial mold shape M is input, and the mean μ of the latent variable z is output. M and variance The decoder is:

[0178]

[0179] Among them, the latent variable z is input, and the reconstructed mold shape is output.

[0180] Step S33: Based on the initial compensation shape generated by the variational autoencoder, sample the initial mold shape and optimize the initial mold shape using the genetic algorithm to generate the mold compensation shape.

[0181] Step S34: Use the mold compensation shape to perform finite element simulation again to determine whether the springback error is reduced. If so, it is judged that the optimization is completed. If not, return to step S33 and use the genetic algorithm to optimize the mold compensation shape again.

[0182] Among them, use the optimized mold compensation shape to perform finite element simulation stamping again to verify whether the springback error is reduced. If |R new -R real | ≤ ∈, it is judged that the optimization is successful, otherwise continue to optimize; where the R new is the springback error of the optimized mold compensation shape, and ∈ is the error tolerance.

[0183] In some examples of this application, as shown in Figure 4 In step S33, the method further includes:

[0184] Step S331: Sample the initial mold shape M0 from the latent distribution p(M|z) generated by the variational autoencoder.

[0185] Step S332: Calculate the fitness function:

[0186] F fitness (M) = -|R real -R ptredicted (M)|

[0187] Step S333: Use the preset selection method to select individuals with high fitness to enter the next generation.

[0188] Among them, the preset selection method includes but is not limited to at least one of roulette wheel selection, tournament selection, Monte Carlo selection, probability selection, stochastic universal sampling, Boltzmann selection, etc. in the genetic algorithm to select individuals with high fitness to enter the next generation.

[0189] Step S334: Perform crossover:

[0190] M new = αM1+(1 + α)M2

[0191] where M1 and M2 are two selected parent die shapes, and α here is the crossover rate;

[0192] Perform mutation processing:

[0193] M mut = M + βN(0, σ)

[0194] where β is the mutation rate and N(0, σ) is a random perturbation following a normal distribution.

[0195] Step S335: Repeat steps S332 to S334 until a preset number of iterations or a convergence criterion is reached.

[0196] where the preset number of iterations or the preset convergence criterion is set by the designer according to actual requirements.

[0197] Step S4: Calculate the longitudinal beam compensation data for assisted stamping based on the die compensation shape, generate the longitudinal beam compensation data, and use the longitudinal beam processing data as the true stamping process parameters of the longitudinal beam.

[0198] Specifically, referring to Figure 5 as shown, in step S4, the method includes:

[0199] Step S41: Calculate the local compensation offset based on the optimized die compensation shape:

[0200] Δh = h real - h compensated

[0201] where h real is the die shape after uncompensated stamping, and h compensated is the ideal shape of the die after compensation.

[0202] Step S42: Calculate the local stress distribution:

[0203] σ res (x, y) = σ plastic (x, y) - σ elastic (x, y)

[0204] where σ plastic (x, y) is the plastic stress and σ elastic (x, y) is the elastic stress;

[0205] Step S43: Set the springback critical threshold and select the area meeting the first preset condition as the local heating area:

[0206] A critical = {(x, y)|σ res (x, y) > σ crit}

[0207] where σ crit is the critical stress for springback; the region of the first preset condition is the region where σ res (x, y) > σ crit .

[0208] Step S44: Calculate the temperature field of the local heating area by using finite element heat conduction analysis:

[0209]

[0210] where ρ is the material density, c p is the specific heat capacity, k is the thermal conductivity, and Q is the heat source term.

[0211] Step S45: Calculate the locally optimal heating temperature:

[0212]

[0213] where σ0 is the room temperature yield stress, T0 is the initial temperature, T m is the melting point temperature, and m is the material parameter.

[0214] Step S46: Calculate the locally optimal heating time:

[0215]

[0216] where ΔT is the desired temperature rise and Q is the heating power.

[0217] Thus, based on the local heating area, the temperature field of the local heating area, the locally optimal heating temperature, and the locally optimal heating time, the longitudinal beam processing data is generated and used as the actual stamping process parameters of the longitudinal beam.

[0218] Thus, in the above example, local heating is performed in the key deformation areas (such as the bending or bulging areas) to improve the material ductility and reduce the springback amount; the specific heating mode can be selected as resistance heating, induction heating, laser heating, etc., to locally heat the metal before or during stamping; according to the actual material of the longitudinal beam workpiece, an appropriate temperature range is selected. For example, for DP980, Q&P steel, etc., a temperature range between 350°C and 600°C can be adopted. Thus, it is ensured that the longitudinal beam workpiece material is stamped in a heated and softened state to obtain the optimal plastic deformation effect. Further, after stamping, a liquid nitrogen cooling spray cooling device, a compressed air spray cooling device, or a spray cooling device can be used to quickly cool the material immediately to lock the deformation form in the mold and reduce the springback.

[0219] According to the above local heating assisted stamping technical solution, an example is given as follows:

[0220] By using FEM simulation and experimental measurement, stress data at different positions after stamping are extracted. Selecting a certain key bending area, the stress data are as follows:

[0221] Position (x, y) <![CDATA[σ plastic (MPa)]]> <![CDATA[σ elastic (MPa)]]> <![CDATA[σ res (MPa)]]> (10,20) 450 200 250 (15,25) 420 190 230 (20,30) 380 200 180 (25,35) 300 150 150

[0222] Among them, through experimental and simulation analysis, σ crit is set to 200 MPa. Thus, the area where σ res (x, y)>200 MPa is screened, that is, A critical ={(10, 20), (15, 25)}. The residual stress at these two points is relatively high. Therefore, local heating assisted stamping is required. Due to the existing parameters of DP980 high-strength steel, the heating temperature is obtained as 400 °C through thermo-mechanical coupling finite element analysis, and then the heating time is obtained as 10 seconds.

[0223] In some embodiments of the present application, electromagnetic assisted stamping can also be used. By applying a dynamic magnetic field or pulsed current, the metal flow characteristics can be optimized to reduce stamping springback.

[0224] Specifically, for applying the dynamic magnetic field: coils are arranged on the die or stamping platform, an alternating magnetic field is applied, the frequency is set to 10 - 100 kHz, and the magnetic field intensity is set to 0.5 - 2 T. In this way, eddy currents are induced inside the metal through the alternating magnetic field, the local temperature rises, the material is softened and the plastic deformation ability is improved, and the grain orientation is changed by the magnetic field to reduce the springback error.

[0225] Specifically, for applying the pulsed current: before or during stamping, a high-frequency pulsed current is applied to the material to reduce the yield strength of the material, and the high-frequency pulsed current is set to 1 - 10 kA. The current passes through the metal to generate the Joule heat effect, causing the local temperature of the material to rise, softening the metal, improving the plastic deformation ability, and using the current to affect the dislocation movement inside the metal, optimizing the stress distribution, and reducing the springback error.

[0226] By adopting the above technical solution, non-contact control can be adopted, avoiding affecting the die structure, and being compatible with existing stamping equipment, improving the forming accuracy.

[0227] In some embodiments of the present application, as shown in Figure 6 , in step S43, the method further includes:

[0228] Step S431: Select the area that meets the second preset condition as the ultrasonic vibration action area:

[0229] A UV ={(x, y)|σres (x, y) > σ crit}

[0230] The second preset condition is |σ res (x, y)| > σ crit area.

[0231] Step S432: Calculate the flow stress reduced by ultrasonic vibration:

[0232] σ eff = σ flow - αf US

[0233] where σ flow is the material resistance without the action of ultrasonic vibration, α is the ultrasonic softening coefficient, and f US is the ultrasonic frequency;

[0234] Step S433: Calculate the optimal ultrasonic parameters:

[0235] A US = k·σ max

[0236] where k is the empirical coefficient and σ max is the local maximum offset.

[0237] Thus, by applying high-frequency ultrasonic vibration on the surface of the mold or stamping sheet, the deformation resistance can be reduced and the forming accuracy can be improved; the ultrasonic vibrator can be installed at the positions of the mold, punch, fixture or longitudinal beam workpiece sheet and the ultrasonic vibration frequency can be set in the range of 20 - 60 kHz and the amplitude in the range of 5 - 20 μm to utilize the ultrasonic vibration to break the slip resistance inside the metal and reduce the friction between the metal and the mold, making the material easier to deform and reducing the springback error.

[0238] According to the above ultrasonic vibration-assisted stamping technical solution, an example is given as follows:

[0239] By using FEM simulation and experimental measurement, the deformation errors in some areas are obtained:

[0240] Position (x, y) <![CDATA[h real (mm)]]> <![CDATA[h compensated (mm)]]> Δh (mm) (10,20) 12.5 12.0 0.5 (15,25) 13.8 13.0 0.8 (20,30) 12.2 12.0 0.2

[0241] Set the error threshold to 0.4 mm, A UV = {(10, 20), (15, 25)}, that is, ultrasonic vibration needs to be applied at the positions of (10, 20) and (15, 25); then calculate the ultrasonic softening coefficient α = 0.2 MPa / kHz, and use 20 kHz ultrasonic wave: σ eff = σ flow - 0.2 × 20, if the original flow stress of the material σ flow = 400 MPa, then σ eff= 400 - 4 = 396 MPa. The softening effect can reduce the flow stress by 1% and reduce springback. Thus, the ultrasonic amplitude A is calculated. US = 0.2 × 0.8 = 0.16 mm, that is, ultrasonic vibration with an amplitude of 0.16 mm is applied at positions (10, 20) and (15, 25).

[0242] In this application, the designer can also combine the incremental plasticity theory with crystal plasticity, starting from the microscale, considering grain orientation, twinning deformation, and dislocation evolution, constructing a high-precision material constitutive model and optimizing springback prediction by combining finite element simulation.

[0243] Specifically, first, at the microscale, a three-dimensional crystal plasticity finite element method can be used to define the polycrystalline microstructure model and assume that each grain has a random orientation. Then, based on the orientation distribution function, the initial grain orientation is defined and its Schmid factor is calculated to analyze the influence of different grain orientations on plastic deformation behavior. Furthermore, different stamping process parameters (stamping speed, temperature, stamping time, etc.) are applied to analyze the deformation, orientation evolution, and twinning behavior of each grain. Then, the molecular dynamics method is used to simulate the micro-deformation mechanism of the metal during stamping and study the dislocation movement, dislocation climb, and dislocation cross-slip processes of the material under different strain rates, temperatures, and pressures. Furthermore, based on the dislocation density evolution during stress relaxation, the Orowan equation is used to characterize micro-plastic flow, so as to obtain the dislocation structure evolution data under different stamping process parameters through simulation and provide micro-mechanism support for the constitutive equation. Thus, based on the results of the above three-dimensional crystal plasticity finite element method and molecular dynamics method, a material hardening characteristic equation is established by combining the Taylor model and the Voce hardening model. Furthermore, uniaxial tensile experiments, springback tests, and nano-indentation experiments are used to measure the stress-strain relationship of the material, and the parameters of the Taylor-Voce hardening model are calibrated by numerical optimization methods (such as least squares fitting or Bayesian optimization) to improve the prediction accuracy of the constitutive model. Combining digital image correlation technology to monitor the micro-strain distribution and verify the influence of grain orientation and twinning behavior on the rheological properties of the material to predict the dynamic yield behavior of the metal sheet under the action of various stamping process parameters.

[0244] Secondly, at the macroscopic scale, the constitutive equation of incremental plasticity theory is used to predict the stamping springback behavior, and the springback prediction model is optimized through finite element simulation; then, a variational autoencoder is used to learn the mapping relationship between the die shape and the stamping springback data to generate a parameterized compensation shape, and the genetic algorithm is used to optimize the compensation shape with the minimum springback error as the fitness objective to ensure that the final shape can effectively reduce the springback error; then, the finite element simulation is used to verify the optimized die compensation shape, and the springback errors of the longitudinal beam before and after compensation are compared. If the accuracy requirements are still not met, return to the optimization link to adjust the parameters until the accuracy requirements are met; finally, the generated die compensation shape and the longitudinal beam compensation data for assisting stamping are used as the true stamping process parameters of the longitudinal beam.

[0245] Specifically, the above microscopic modeling process is based on Schmid's law:

[0246]

[0247] where τ ω is the shear stress on the slip system ω, σ ij is the stress tensor, is the direction cosine of the slip system; its slip rate is:

[0248]

[0249] where is the shear strain rate on the slip system ω, is the reference strain rate, is the critical shear stress, and n is the strain rate sensitivity index. Thus, the activation of different grain slip systems during the stamping process of the metal sheet is described, which in turn affects the overall plastic deformation characteristics.

[0250] In some embodiments of the present application, when the stamping equipment performs stamping processing according to the longitudinal beam compensation data, laser measurement or three-dimensional vision detection technology can also be used to monitor the springback of the stamped longitudinal beam workpiece in real time, so as to dynamically adjust the stamping process parameters (such as pressure, speed, local heating temperature, local heating time, local heating position, ultrasonic-assisted stamping power, ultrasonic-assisted stamping time, and ultrasonic-assisted stamping position, etc.) for feedback compensation; or through 3D scanning and digital twin technology, real-time stamping prediction and compensation are performed.

[0251] Refer to Figure 7 As shown, in some embodiments, an automotive extended longitudinal beam stamping springback prediction and compensation system involved includes:

[0252] A data construction module 100, configured to calculate the influence of preset process parameters on stamping springback and construct an initial data set including initial prediction results;

[0253] The model optimization module 101 is configured to construct a springback prediction model using the incremental plasticity theory, and then adjust the process parameters using the Bayesian optimization method and train the springback prediction model using the gradient boosting decision tree method to minimize the springback error;

[0254] The simulation compensation module 102 is configured to manufacture a longitudinal beam with preset process parameters through finite element simulation, generate simulation stamping springback data, and optimize the die compensation shape using a genetic algorithm and a variational autoencoder according to the simulation stamping springback data;

[0255] The real feedback module 103 is configured to calculate the longitudinal beam compensation data for assisted stamping according to the die compensation shape, generate the longitudinal beam compensation data, and use the longitudinal beam processing data as the real stamping process parameters of the longitudinal beam.

[0256] In some embodiments, the present application also relates to a computer medium having a computer program stored thereon, and the computer program is executed by a processor to implement the method for predicting and compensating the stamping springback of an extended longitudinal beam of an automobile.

[0257] In some embodiments, the present application also relates to a computer including the computer medium described above.

[0258] In the description of this specification, the descriptions with reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples", etc., mean that the specific features, structures, materials, or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representations of the above terms are not necessarily directed to the same embodiment or example. Moreover, the specific features, structures, materials, or characteristics described may be combined in any one or more embodiments or examples in a suitable manner. In addition, without conflict, those skilled in the art may combine and combine the different embodiments or examples described in this specification and the features of the different embodiments or examples.

[0259] Although the embodiments of the present invention have been shown and described above, it can be understood that the above embodiments are exemplary and should not be construed as limiting the present invention. Those of ordinary skill in the art can make changes, modifications, substitutions, and variations to the above embodiments within the scope of the present invention.

Claims

1. A method for predicting and compensating springback in stamping of automotive extended longitudinal beams, characterized in that, It includes the following steps: Step S1: According to the influence of preset process parameters on stamping springback, construct an initial data set containing initial predicted values; Step S2: Construct a springback prediction model using the incremental plasticity theory, then train the springback prediction model using the gradient boosting decision tree and adjust the process parameters using the Bayesian optimization method to generate optimal process parameters; Step S3: Manufacture a longitudinal beam using finite element simulation according to the optimal process parameters, generate simulation stamping springback data, and optimize the die compensation shape using the genetic algorithm and variational autoencoder according to the simulation stamping springback data; Step S4: Calculate the longitudinal beam compensation data for auxiliary stamping according to the die compensation shape, generate the longitudinal beam compensation data, and use the longitudinal beam processing data as the true stamping process parameters of the longitudinal beam.

2. A method for predicting and compensating springback in stamping of an extended longitudinal beam of an automobile according to claim 1, characterized in that In step S1, the method includes: Step S11: Select a material model and define the elastic modulus, Poisson's ratio, and yield criterion; Step S12: Perform mesh division using four-node shell elements or eight-node solid elements; Step S13: Set preset process parameters; Step S14: Simulate the stamping process of the longitudinal beam using finite element simulation, including: applying stamping loads to simulate the stress and deformation of the sheet metal, calculating the stress field and strain field after deformation; applying boundary conditions to remove the stamping loads, and calculating the residual stress and the maximum springback amount and springback angle after deformation; Step S15: Extract the maximum springback amount, residual stress distribution, and springback angle change under the preset process parameters to generate initial predicted values.

3. A method for predicting and compensating springback in stamping of an extended longitudinal beam of an automobile according to claim 1 or 2, characterized in that, In step S2, the method includes: Construct a springback prediction model using the incremental plasticity theory: Among them, Δε is the predicted springback error, E is the material elastic modulus, and σ res is the residual stress after stamping, and σ y is the material yield stress.

4. A method for predicting and compensating the springback of a lengthened longitudinal beam of an automobile according to claim 3, characterized in that, In step S2, the method further includes: Step S21: Use the preset process parameters as input features and train the springback prediction model using the gradient boosting decision tree: Among them, F t (x) is the decision tree in the t-th round, T is the number of decision trees, and R ptredicted is the rebound error predicted by the model; Step S22: Define the Bayesian optimization objective function: f(x) = min|R real -R ptredicted (x)| where R real is the rebound error measured experimentally, and x is the preset process parameter; Step S23: Optimize the preset process parameters using the Gaussian process regression method: p(y|x) = N(μ(x), k(x,x')) where p(y|x) is the probability distribution of the springback error y after given the parameter x, μ(x) is the mean function representing the estimation of the springback error, and k(x,x') is the covariance function representing the correlation between parameters; Step S24: Use five-fold cross-validation to evaluate the generalization ability of the springback prediction model; Step S25: Iteratively loop through steps S21 to S24 a preset number of times until the preset requirements are met and output the optimal process parameters.

5. A method for predicting and compensating springback in stamping of an extended longitudinal beam of an automobile according to claim 1 or 4, characterized in that, In step S21, the method includes: Step S210: Initialize the springback error prediction value: Among them, F0(x) is the mean value of the springback error, N is the number of training samples, and R real,i is the springback error of the i-th sample; Step S211: Calculate the deviation between the springback error and the predicted value: Among them, is the deviation of the i-th sample in the t-th iteration, F t-1 (x i ) is the predicted springback error of the i-th sample in the (t - 1)-th iteration; Step S212: Fit the deviation between the springback error and the predicted value: Among them, F t (x)' is the decision tree for the t-th round of training, used to fit the residuals, and F(x i ) is the predicted value of the trained decision tree model for the i-th sample, and argmin is the optimization objective that minimizes the sum of squared deviations; Step S213: Update the predicted value and control the step size with the learning rate η: F t F(x) = t-1 F(x)+ηF t (x) Among them, F t (x) is the predicted value of the model after the t-th iteration, and F t-1 (x) is the predicted value after the (t - 1)-th iteration; Step S214: Loop through steps S211 to S213 a preset number of times.

6. A method for predicting and compensating the springback of a lengthened longitudinal beam of an automobile according to claim 2, characterized in that In step S3, the method includes: Step S31: According to the optimal process parameters, the selected material model, the defined elastic modulus, Poisson's ratio, and yield criterion, use finite element simulation to manufacture the longitudinal beam and output simulation stamping springback data including the maximum springback amount, residual stress distribution, springback angle change, and stress distribution in the plate thickness direction; Step S32: Use a variational autoencoder to learn the relationship between the die shape and the stamping springback data, and generate a parameterized initial compensation shape; Step S33: Sample the initial die shape from the initial compensation shape generated by the variational autoencoder and optimize the initial die shape using a genetic algorithm to generate a die compensation shape; Step S34: Use the die compensation shape to perform finite element simulation again to determine whether the springback error is reduced. If so, it is determined that the optimization is completed. If not, return to Step S33 and use the genetic algorithm to optimize the die compensation shape again.

7. A method for predicting and compensating the springback of a lengthened longitudinal beam of an automobile according to claim 6, characterized in that, In Step S33, the method includes: Step S331: Sample the initial die shape from the initial compensation shape generated by the variational autoencoder; Step S332: Calculate the fitness function; Step S333: Use a preset selection method to select individuals with high fitness to enter the next generation; Step S334: Perform crossover and mutation processing; Step S335: Repeat Steps S332 to S334 until a preset number of iterations or convergence criteria are reached.

8. A method for predicting and compensating springback in stamping of an extended longitudinal beam of an automobile according to claim 6 or 7, characterized in that, In Step S4, the method includes: Step S41: Calculate the local compensation offset based on the optimized die compensation shape: Δh = h real -h compensated where h real is the die shape after uncompensated stamping, and h compensated is the ideal shape of the die after compensation; Step S42: Calculate the local stress distribution: σ res (x,y) = σ plastic (x,y) - σ elastic (x,y) Among them, σ plastic (x, y) is the plastic stress, and σ elastic (x, y) is the elastic stress; Step S43: Set a springback critical threshold and select the area that meets the first preset condition as the local heating area: A critical = {(x, y)|σ res (x, y) > σ crit} Among them, σ crit is the critical stress for springback; Step S44: Use finite element heat conduction analysis to calculate the temperature field of the local heating area: where ρ is the material density, c p is the specific heat capacity, k is the thermal conductivity, and Q is the heat source term; Step S45: Calculate the local optimal heating temperature: Among them, σ0 is the yield stress at room temperature, T0 is the initial temperature, T m is the melting point temperature, and m is the material parameter; Step S46: Calculate the local optimal heating time: where ΔT is the desired temperature rise and Q is the heating power.

9. A method for predicting and compensating the springback of a lengthened longitudinal beam of an automobile according to claim 8, characterized in that In Step S43, the method further includes: Step S431: Select the area that meets the second preset condition as the ultrasonic vibration action area: A UV = {(x,y)|σ res (x,y) > σ crit} Step S432: Calculate the flow stress reduced by ultrasonic vibration: σ eff = σ flow - αf US Among them, σ flow is the material resistance without ultrasonic vibration, α is the ultrasonic softening coefficient, and f US is the ultrasonic frequency; Step S433: Calculate the optimal ultrasonic parameters: A US = k·σ max where k is an empirical coefficient, and σ max is the local maximum offset.

10. An automobile lengthened longitudinal beam stamping springback prediction and compensation system using the method according to any one of claims 1-9, characterized in that including: A data construction module for calculating the influence of preset process parameters on stamping springback and constructing an initial data set containing initial prediction results; A model optimization module for constructing a springback prediction model using incremental plasticity theory, then adjusting process parameters using the Bayesian optimization method and training the springback prediction model using the gradient boosting decision tree method to minimize the springback error; A simulation compensation module for using finite element simulation to manufacture the longitudinal beam through preset process parameters, generating simulation stamping springback data, and optimizing the die compensation shape using a genetic algorithm and a variational autoencoder according to the simulation stamping springback data; A real feedback module for calculating the longitudinal beam compensation data for assisted stamping according to the die compensation shape, generating the longitudinal beam compensation data, and using the longitudinal beam processing data as the real stamping process parameters of the longitudinal beam.

Images