MMC fracture curve optimization method and device based on bayesian optimization

Abstract

The application discloses a MMC fracture curve optimization method and device based on Bayesian optimization, inputs the adjusted MMC curve minimum point and end point after constructing a black box function, outputs a single comprehensive error index, then carries out Bayesian optimization on the black box function to obtain optimization process information, and finally finds the minimum value point and end point which can minimize the single comprehensive error index according to the optimization process information. In the method, the Bayesian optimization takes the single comprehensive error index as an algorithm optimization target, takes the adjusted MMC curve minimum value point and end point as parameters to be optimized, carries out automatic iterative optimization, and finally obtains optimal parameters, i.e. the optimal values of the adjusted MMC fracture curve minimum value point and end point. The method can realize automatic adjustment and optimization of the MMC fracture curve, no longer relies on manual adjustment, and can greatly improve the efficiency and accuracy of the correction of the MMC fracture curve.

Description

Technical Field

[0001] This application relates to the field of materials testing technology, and more specifically, to a method and apparatus for optimizing MMC fracture curves based on Bayesian optimization. Background Technology

[0002] Increasingly stringent automotive crash regulations and the demand for vehicle lightweighting have presented various technical challenges to the selection, design, and optimization of automotive components. In response, computer technology has been continuously upgraded, with tools such as Abaqus and LS_DYNA playing a crucial role in crash simulation. However, considering the complex stresses experienced by automotive structural components in actual operation, conventional prediction methods are no longer sufficient to meet the growing demands. Therefore, more advanced plasticity and fracture models are needed for in-depth research. Based on this, the MMC fracture model is proposed. This advanced model, based on the Mohr-Coulomb fracture mechanism, not only includes traditional Mohr-Coulomb parameters but also incorporates other key factors, such as equivalent plastic strain. After parameter fitting, the MMC model can successfully represent the fracture behavior of materials under various complex stress states.

[0003] When performing calculations using the MCC fracture model, it is possible to generate the force-displacement curve of the specimen under specific simulation conditions. However, these simulation-generated curves may not perfectly match the experimental data. This requires correction of the MCC fracture curve, but determining the optimal adjustment range is currently a difficult problem to solve.

[0004] In the process of matching force-displacement curves, manually adjusting the MMC fracture curve requires repeatedly trying different adjustment ranges to achieve a match between the simulated and experimental force-displacement curves. This approach is not only time-consuming and labor-intensive but also demands a high level of experience from the operators. Therefore, exploring a method that can automatically perform curve linear optimization to improve the efficiency and accuracy of curve matching has become an important research direction for experts in this field. Summary of the Invention

[0005] The purpose of this application is to provide a Bayesian optimization-based method and apparatus for optimizing the MMC break curve. The core of this method is the construction of a black-box function. The input to this function is the minimum and end points of the adjusted MMC break curve, and the output is a single comprehensive error index. This single comprehensive error index is set as the target of Bayesian optimization, and the minimum and end points that minimize the single comprehensive error index are obtained through multiple iterations. This method can automatically adjust and optimize the MMC break curve, eliminating the need for manual adjustment and significantly improving the efficiency and accuracy of MMC break curve correction. This objective is achieved through the following technical solution:

[0006] Firstly, this application proposes a Bayesian optimization-based method for finding the MMC fracture curve, the method comprising:

[0007] Construct a black-box function, taking the minimum and end points of the adjusted MMC curve as inputs and a single comprehensive error index as output.

[0008] Bayesian optimization is performed on the black-box function to obtain optimization process information;

[0009] Based on the optimization process information, find the minimum point and the endpoint that minimize the single comprehensive error index.

[0010] In one possible implementation, the steps of using the adjusted minimum and end points of the MMC curve as inputs to a black-box function and using a single comprehensive error index as the output of the black-box function include:

[0011] The corrected MMC fracture curve is obtained by adjusting the minimum and end points in conjunction with the data deviation rate.

[0012] Input the corrected MMC fracture curve into the dynamic simulation solver to obtain various force-displacement simulation curves;

[0013] The force-displacement error is calculated from various force-displacement simulation curves and force-displacement experimental curves to obtain an error array;

[0014] The error array is transformed into a single comprehensive error index based on a custom logic algorithm.

[0015] In one possible implementation, the modified MMC fracture curve includes an adjusted first curve and an adjusted second curve, wherein the adjusted first curve... The formula is: Where y0 is the turning point before the minimum point, y c1 The minimum point, The minimum point after adjustment, y i1 To adjust from y0 to y c1 Each point on the curve;

[0016] Adjusted second curve Where y c1 The minimum point is y c2 For the endpoint, The minimum point after adjustment, y i2 To adjust y c1 To y c2 Each curve point, This is the adjusted endpoint.

[0017] In one possible implementation, the single comprehensive error index result is: result = 5 × max(fra_err_list) + mean(fra_err_list), where max(fra_err_list) is the maximum value in the error array, and mean(fra_err_list) is the average value of the error array. The calculation formula for obtaining the single comprehensive error index result is a custom logic algorithm.

[0018] In one possible implementation, the various force-displacement simulation curves correspond to the force-displacement simulation curves for pure shear, R5 notch tension, R10 notch tension, center hole tension, and cup protrusion, respectively.

[0019] Secondly, a Bayesian optimization-based MMC fracture curve optimization device is proposed, the device comprising:

[0020] The function construction module is used to construct black-box functions, taking the minimum and end points of the adjusted MMC curve as inputs to the black-box function and a single comprehensive error index as the output of the black-box function.

[0021] The optimization module is used to perform Bayesian optimization on the black-box function to obtain optimization process information;

[0022] The minimization module is used to find the minimum point and the endpoint that minimize a single comprehensive error index based on the optimization process information.

[0023] In one possible implementation, the function building module includes:

[0024] The optimization submodule is used to adjust the minimum and end points in conjunction with the data deviation rate to obtain the corrected MMC fracture curve;

[0025] The simulation submodule is used to input the corrected MMC fracture curve into the dynamic simulation solver to obtain various force-displacement simulation curves.

[0026] The calculation submodule is used to calculate the force-displacement error of various force-displacement simulation curves and force-displacement experimental curves to obtain an error array;

[0027] The transformation submodule converts the error array into a single comprehensive error index based on a custom logic algorithm.

[0028] In one possible implementation, the modified MMC fracture curve includes an adjusted first curve and an adjusted second curve, wherein the adjusted first curve... The formula is: Where y0 is the turning point before the minimum point, y c1 The minimum point, The minimum point after adjustment, y i1 To adjust from y0 to y c1 Each point on the curve;

[0029] Adjusted second curve Where y c1 The minimum point is y c2 For the endpoint, The minimum point after adjustment, y i2 To adjust y c1 To y c2 Each curve point, This is the adjusted endpoint.

[0030] In one possible implementation, the single comprehensive error index result is: result = 5 × max(fra_err_list) + mean(fra_err_list), where max(fra_err_list) is the maximum value in the error array, and mean(fra_err_list) is the average value of the error array. The calculation formula for obtaining the single comprehensive error index result is a custom logic algorithm.

[0031] In one possible implementation, the various force-displacement simulation curves correspond to the force-displacement simulation curves for pure shear, R5 notch tension, R10 notch tension, center hole tension, and cup protrusion, respectively.

[0032] The main solution and its various further alternatives described above can be freely combined to form multiple solutions, all of which are solutions that can be adopted and are claimed in this application; furthermore, the (non-conflicting alternatives) can also be freely combined with each other and with other alternatives. Those skilled in the art, after understanding the solution of this application, will realize from the prior art and common general knowledge that there are many combinations, all of which are technical solutions to be protected by this application, and will not be exhaustively listed here.

[0033] This application discloses a method and apparatus for optimizing the MMC break curve based on Bayesian optimization. First, a black-box function is constructed, using the minimum and terminal points of the adjusted MMC curve as inputs and a single comprehensive error index as output. Then, Bayesian optimization is performed on the black-box function to obtain optimization process information. Finally, based on this information, the minimum and terminal points that minimize the single comprehensive error index are found. In this method, Bayesian optimization uses the single comprehensive error index as the algorithm's optimization objective, and the minimum and terminal points of the adjusted MMC curve as parameters to be optimized. Automatic iterative optimization is then performed to obtain the optimal parameters, i.e., the optimal values ​​of the minimum and terminal points of the adjusted MMC break curve. This method enables automatic adjustment and optimization of the MMC break curve, eliminating the need for manual adjustment and significantly improving the efficiency and accuracy of MMC break curve correction. Attached Figure Description

[0034] Figure 1 The diagram shows a flowchart of a Bayesian optimization-based MMC fracture curve optimization method proposed in an embodiment of this application.

[0035] Figure 2 A schematic diagram of the modified MMC fracture curve proposed in the embodiments of this application is shown.

[0036] Figure 3 A convergence diagram of the Bayesian optimization iterative process proposed in an embodiment of this application is shown. Detailed Implementation

[0037] The following specific examples illustrate the implementation of this application. Those skilled in the art can easily understand other advantages and effects of this application from the content disclosed in this specification. This application can also be implemented or applied through other different specific embodiments, and various details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of this application. It should be noted that, unless otherwise specified, the following embodiments and features in the embodiments can be combined with each other.

[0038] Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without inventive effort are within the scope of protection of this application.

[0039] In the existing technology, if the MMC fracture curve is adjusted manually, it is necessary to try different adjustment ranges to achieve the match between the simulation and experimental curves. This approach is not only time-consuming and labor-intensive, but also requires a high level of experience from the operators.

[0040] Therefore, to address the aforementioned problems, this application provides a Bayesian optimization-based method and apparatus for optimizing MMC break curves. The core of this method is the construction of a black-box function. The input to this function is the minimum and end points of the adjusted MMC break curve, and the output is a single comprehensive error index. This single comprehensive error index is set as the target of Bayesian optimization, and the minimum and end points that minimize the single comprehensive error index are obtained through multiple iterations. This method enables automatic adjustment and optimization of the MMC break curve, eliminating the need for manual adjustment and significantly improving the efficiency and accuracy of MMC break curve correction. A detailed explanation follows.

[0041] Please refer to Figure 1 , Figure 1 This paper presents a flowchart illustrating a Bayesian optimization-based MMC fracture curve optimization method proposed in an embodiment of this application, including the following steps:

[0042] Construct a black-box function, taking the minimum and end points of the adjusted MMC curve as inputs and a single comprehensive error index as output.

[0043] A black-box function is constructed to calculate the force-displacement error between the simulation curve and the experimental curve, resulting in a single comprehensive error index. The function definition of the black-box function is as follows:

[0044] def objective(m):

[0045] fra_name,fra_err_list,fra_dis_list,fra_reforc_list=cal(path_doc,solverpath,testdata,k4,k5,x,y,m[0],m[1]);

[0046] print(fra_err_list);

[0047] fra_err_list=abs(np.array(fra_err_list)),tolist();

[0048] result=max(fra_err_list)*5+np.mean(fra_err_list);

[0049] Return result.

[0050] During the optimization process, a custom `cal` function is defined to implement the key functionality of the black-box function. The `cal` function performs the following steps:

[0051] MMC curve correction: The corrected MMC breakage curve is obtained by adjusting the minimum and end points in conjunction with the data deviation rate.

[0052] Simulation solution and error calculation: The simulation curves are analyzed according to the specified path and the parameters of the simulation solver, where fra_err_list is the error array of the force-displacement curve.

[0053] The transmission of variable node values: In the black-box function, the parameters m[0] and m[1] represent the values ​​of the variable nodes (the minimum and end points of the adjusted MMC curve).

[0054] Organization of error data: fra_err_list contains force-displacement errors for five different specimens, reflecting the deviation between model simulation and experimental results.

[0055] The input values ​​of the black-box function are the parameters corresponding to Bayesian optimization, and the output values ​​of the black-box function are the function output values ​​corresponding to Bayesian optimization.

[0056] The minimum and end points of the initial MMC fracture curve are randomly adjusted upwards to obtain different minimum and end points. The different minimum and end points obtained by adjustment are then corrected to obtain the corrected MMC fracture curve.

[0057] In one possible implementation, the steps of using the adjusted minimum and end points of the MMC curve as inputs to a black-box function and using a single comprehensive error index as the output of the black-box function include:

[0058] The corrected MMC fracture curve is obtained by adjusting the minimum and end points in conjunction with the data deviation rate.

[0059] Input the corrected MMC fracture curve into the dynamic simulation solver to obtain various force-displacement simulation curves;

[0060] The force-displacement error is calculated from various force-displacement simulation curves and force-displacement experimental curves to obtain an error array;

[0061] The error array is transformed into a single comprehensive error index based on a custom logic algorithm.

[0062] Please refer to Figure 2 , Figure 2 The diagram shows a modified MMC fracture curve proposed in the embodiments of this application. The solid line in the diagram is the initial MMC fracture curve, and the dashed line is the modified MMC fracture curve. To adjust the initial MMC fracture curve, it is necessary to first determine the key variable nodes. In this application, the minimum point and the end point in the initial MMC fracture curve are taken as variable nodes, and the two are adjusted and the adjusted nodes are calculated.

[0063] In one possible implementation, the modified MMC fracture curve includes an adjusted first curve and an adjusted second curve, wherein the adjusted first curve... The formula is: Where y0 is the turning point before the minimum point, y c1 The minimum point, The minimum point after adjustment, y i1 To adjust from y0 to y c1 Each point on the curve;

[0064] Adjusted second curve Where y c1 The minimum point is y c2 For the endpoint, The minimum point after adjustment, y i2 To adjust y c1 To y c2 Each curve point, This is the adjusted endpoint.

[0065] Select the turning point y0 before the smallest node, y c1 and y c2 Let y be the minimum and end points to be adjusted, respectively. c1 and y c2 They are respectively denoted as and The original y0 to y c1 Each point on the curve is denoted as y. i1 , will the original y c1 To y c2 Each curve point on the curve in the equation is denoted as y. i2 y is calculated using the following formula. c1 Adjusted to After that, for each data point y i Adjusted to Value: Then the calculations were performed. The fitting is now complete. The adjusted data points are then plotted as a curve.

[0066] By vertically moving the minimum and end points, we obtain the adjusted minimum and end points. With the adjusted minimum and end points being adjusted upwards simultaneously, and by adjusting the curve guided by the data deviation rate, we can obtain a corrected MMC curve, which shows the correction result of adjusting according to the data deviation.

[0067] The corrected MMC fracture curve is input into a dynamic simulation solver to obtain updated force-displacement simulation curves. The dynamic simulation solver can be LS-DYNA, which is a fully functional program for geometric nonlinearity, material nonlinearity, and contact nonlinearity.

[0068] Multiple force-displacement simulation curves correspond to the force-displacement simulation curves of pure shear, R5 notch tension, R10 notch tension, center hole tension, and cupping. Pure shear, R5 notch tension, R10 notch tension, center hole tension, and cupping are five tests of the MMC fracture model. The error value between the force-displacement experimental curve corresponding to each tensile test and the actual measured force-displacement experimental curve is calculated to obtain a set of error arrays. According to a custom logic algorithm, the error array is transformed into a single comprehensive error index. The single comprehensive error index result is: result = 5 × max(fra_err_list) + mean(fra_err_list), where max(fra_err_list) is the maximum value in the error array, and mean(fra_err_list) is the average value of the error array. The calculation formula for obtaining the single comprehensive error index result is a custom logic algorithm.

[0069] The entire process described above is constructed as a black-box function. The input of this black-box function is the adjusted minimum point and terminal point, and the output is a defined single comprehensive error index. The single comprehensive error index is regarded as the optimization objective (minimization optimization) through the Bayesian optimization algorithm, and automatic iterative optimization is performed until the best result of adjusting the minimum point and terminal point of the MMC fracture curve is found.

[0070] The single comprehensive error index reflects the focus of the optimization process on reducing the maximum error among all sample error parameters, while ensuring that the overall average error is as low as possible. This ensures that Bayesian optimization does not only focus on a single worst case, but also reduces the error uniformly across all samples.

[0071] When inputting parameters into the aforementioned black-box function, it is first necessary to determine the adjustment range of the variable nodes, which defines the explorable space for the variable nodes of the fracture model during optimization. Setting y1 as the ordinate corresponding to the lowest point of the initial MMC curve and y2 as the ordinate corresponding to the end point of the initial MMC curve establishes a parameter space to define the optimization adjustment range of these two variable nodes. This parameter space is as follows: the upper limit of y1 and y2 is set to 1.4 times the initial ordinate value to allow sufficient adjustment space to accommodate possible data changes; the lower limit of y1 and y2 is set to 0.8 times the initial ordinate value. This ensures that the adjustment will not fall below a certain threshold, maintaining the basic stability and practicality of the model. This approach provides sufficient search space for the optimization algorithm while avoiding exceeding the range of actual physical meaning or prior experience.

[0072] Bayesian optimization of the black-box function yields information about the optimization process.

[0073] Using Bayesian optimization algorithms to optimize black-box functions, the corresponding parameters are adjusted with the goal of minimizing the function output value. This approach can find the global optimum with fewer function evaluations. The optimization process information includes the horizontal and vertical coordinates of the optimized MMC curve, the number of iterations, and other information.

[0074] Based on a single comprehensive error index, the preset minimum point and preset end point are searched from the optimization process information.

[0075] The preset minimum point and preset end point are the optimal minimum point and optimal end point that satisfy the preset conditions. During the search process, the parameter values ​​and corresponding function outputs, optimal parameter values, and optimal objective function outputs are included at each step.

[0076] The Bayesian optimization algorithm uses a single comprehensive error index as the optimization target, and the minimum point and end point of the adjusted MMC curve as the parameters to be optimized. It automatically iterates to find the optimal parameters, thereby realizing the automatic adjustment and optimization of the MMC break curve. It no longer relies on manual adjustment, which greatly improves the efficiency and accuracy of MCC break curve correction.

[0077] The entire process combines advanced computing technology with a deep understanding of material behavior, enabling automated and precise model optimization.

[0078] To verify the aforementioned Bayesian optimization-based MMC fracture curve optimization method, Bayesian optimization training was initiated. This training was implemented using the `gp_minimize` function from the `scikit-optimize` library, with a maximum iteration count of 30 to ensure sufficient optimization. DH780 material was selected as the research object to test the Bayesian optimization scheme.

[0079] When the training entered the initial optimization stage, parameters were obtained by constitutive model simulation based on DH780 material. The initial MMC model curve was constructed using the parameters, and the error array between the simulated force-displacement curve and the experimental data was obtained: [4.56%, 7.17%, 0.44%, 4.08%, 14.31%]. The above array shows that the maximum error reached 14.31% before Bayesian optimization was applied.

[0080] Next, the Bayesian optimization algorithm was used to guide the iterative update of the black-box function input parameters in order to minimize the overall prediction error of the force-displacement curve. As the optimization process progressed, the trajectory of the target approaching the optimal solution was more intuitively observed through visualization of the iterative process, as well as its continuous improvement in the model's prediction accuracy.

[0081] Figure 3 The diagram illustrates the convergence of the Bayesian optimization iterative process proposed in this application. The horizontal axis represents the number of iterations, and the vertical axis represents the output value of the black-box function. After 30 rounds of Bayesian optimization adjustments to the simulation model of the DH780 material, the optimization process tended to stabilize after the 7th iteration. The output value of the black-box function (the objective function that the optimization algorithm attempts to minimize) showed slight fluctuations in the range of 25 to 26. Notably, the optimal result was achieved in the 7th iteration. The error between the simulation curve and the experimental data curve obtained in this iteration was significantly reduced, with the force-displacement error data presented as an array of [4.56%, 4.45%, 3.52%, 0.94%, 0.59%]. After Bayesian optimization, the maximum error of the model was significantly reduced from 14.31% to 4.56%, a significant improvement that greatly enhanced the agreement between the simulation model and the experimental results.

[0082] Overall, Bayesian optimization enables automatic adjustment of the MMC fracture curve. The corrected simulation results show a high degree of agreement with experimental results, indicating that the corrected simulation model can accurately simulate the fracture behavior of materials under real working conditions. Furthermore, the intelligent tuning characteristics of Bayesian optimization reduce the need for reliance on intuition or traditional trial-and-error methods. By constructing a probabilistic model to guide the search process, it can effectively find the optimal parameters, significantly improving optimization efficiency.

[0083] The following is a possible implementation of a Bayesian optimization-based MMC fracture curve optimization device, which is used to execute the various execution steps and corresponding technical effects of the Bayesian optimization-based MMC fracture curve optimization device method shown in the above embodiments and possible implementations. The device includes:

[0084] The function construction module is used to construct black-box functions, taking the minimum and end points of the adjusted MMC curve as inputs to the black-box function and a single comprehensive error index as the output of the black-box function.

[0085] The optimization module is used to perform Bayesian optimization on black-box functions to obtain optimization process information;

[0086] The minimization module is used to find the minimum point and the endpoint that minimize a single comprehensive error index based on the optimization process information.

[0087] In one possible implementation, the function building module includes:

[0088] The optimization submodule is used to adjust the minimum and end points in conjunction with the data deviation rate to obtain the corrected MMC fracture curve;

[0089] The simulation submodule is used to input the corrected MMC fracture curve into the dynamic simulation solver to obtain various force-displacement simulation curves.

[0090] The calculation submodule is used to calculate the force-displacement error of various force-displacement simulation curves and force-displacement experimental curves to obtain an error array;

[0091] The transformation submodule converts the error array into a single comprehensive error index based on a custom logic algorithm.

[0092] In one possible implementation, the modified MMC fracture curve includes an adjusted first curve and an adjusted second curve, wherein the adjusted first curve... The formula is: Where y0 is the turning point before the minimum point, y c1 The minimum point, The minimum point after adjustment, y i1 To adjust from y0 to y c1 Each point on the curve;

[0093] Adjusted second curve Where y c1 The minimum point is y c2 For the endpoint, The minimum point after adjustment, y i2 To adjust y c1 To y c2 Each curve point, This is the adjusted endpoint.

[0094] In one possible implementation, the single comprehensive error index result is: result = 5 × max(fra_err_list) + mean(fra_err_list), where max(fra_err_list) is the maximum value in the error array, and mean(fra_err_list) is the average value of the error array. The calculation formula for obtaining the single comprehensive error index result is a custom logic algorithm.

[0095] In one possible implementation, the various force-displacement simulation curves correspond to the force-displacement simulation curves for pure shear, R5 notch tension, R10 notch tension, center hole tension, and cup protrusion, respectively.

[0096] The above description is merely a preferred embodiment of this application and is not intended to limit this application. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this application should be included within the protection scope of this application.

Claims

1. A method for MMC fracture curve optimization based on Bayesian optimization, characterized in that, The method includes: A black-box function is constructed, with the minimum and end points of the adjusted MMC curve as inputs and a single comprehensive error index as output. The adjusted minimum and end points are then combined with the data deviation rate to obtain the corrected MMC breakage curve. The corrected MMC fracture curve is input into the dynamic simulation solver to obtain various force-displacement simulation curves; the corrected MMC fracture curve includes an adjusted first curve and an adjusted second curve, wherein the adjusted first curve The formula is: ,in The turning point before the minimum point. The minimum point, This is the adjusted minimum point. Before adjustment Each curve point; Adjusted second curve ,in The minimum point, For the endpoint, This is the adjusted minimum point. Before adjustment Each curve point, The adjusted endpoint; Bayesian optimization is performed on the black-box function to obtain optimization process information; Based on the optimization process information, find the minimum point and the endpoint that minimize the single comprehensive error index.

2. The MMC fracture curve optimization method as described in claim 1, characterized in that, The steps of using the adjusted minimum and end points of the MMC curve as inputs to a black-box function and a single comprehensive error index as the output of the black-box function include: The force-displacement error is calculated from various force-displacement simulation curves and force-displacement experimental curves to obtain an error array; The error array is transformed into a single comprehensive error index based on a custom logic algorithm.

3. The MMC fracture curve optimization method as described in claim 1, characterized in that, Single comprehensive error index for: ,in The maximum value in the error array. To obtain a single comprehensive error index from the average of the error array. The calculation formula is a custom logic algorithm.

4. The MMC fracture curve optimization method as described in claim 1, characterized in that, The various force-displacement simulation curves correspond to the force-displacement simulation curves for pure shear, R5 notch tension, R10 notch tension, center hole tension, and cup protrusion, respectively.

5. A Bayesian optimization-based MMC fracture curve optimization device, characterized in that, The device includes: The function construction module is used to construct black-box functions, taking the minimum and end points of the adjusted MMC curve as inputs to the black-box function and a single comprehensive error index as the output of the black-box function. The function construction module includes: The optimization submodule is used to adjust the minimum and end points in conjunction with the data deviation rate to obtain the corrected MMC fracture curve; The simulation submodule is used to input the corrected MMC fracture curve into the dynamic simulation solver to obtain various force-displacement simulation curves. The corrected MMC fracture curve includes an adjusted first curve and an adjusted second curve, wherein the adjusted first curve... The formula is: ,in The turning point before the minimum point. The minimum point, This is the adjusted minimum point. Before adjustment Each curve point; Adjusted second curve ,in The minimum point, For the endpoint, This is the adjusted minimum point. Before adjustment Each curve point, The adjusted endpoint; The optimization module is used to perform Bayesian optimization on the black-box function to obtain optimization process information; The minimization module is used to find the minimum point and the endpoint that minimize a single comprehensive error index based on the optimization process information.

6. The MMC fracture curve optimization device as described in claim 5, characterized in that, The function construction module includes: The calculation submodule is used to calculate the force-displacement error of various force-displacement simulation curves and force-displacement experimental curves to obtain an error array; The transformation submodule converts the error array into a single comprehensive error index based on a custom logic algorithm.

7. The MMC fracture curve optimization device as described in claim 5, characterized in that, Single comprehensive error index for: ,in The maximum value in the error array. To obtain a single comprehensive error index from the average of the error array. The calculation formula is a custom logic algorithm.

8. The MMC fracture curve optimization device as described in claim 5, characterized in that, The various force-displacement simulation curves correspond to the force-displacement simulation curves for pure shear, R5 notch tension, R10 notch tension, center hole tension, and cup protrusion, respectively.

Images