3D analogue simulation analysis method for mold

Abstract

The invention relates to a mold 3D analogue simulation analysis method, which comprises the following steps: acquiring a three-dimensional CAD model of a mold, carrying out preprocessing and grid division, encrypting grids in a key area, calculating a stress value by utilizing finite element analysis, optimizing the grids through iterative calculation until a convergence requirement is met, and obtaining a stress distribution condition. A support vector machine regression algorithm is adopted to establish a nonlinear prediction model of stress characteristics and fatigue life, mold structure optimization is carried out, an optimization scheme is converted into a CAD model, a machining program is generated, accurate prediction and automatic optimization design of the fatigue life of the mold are achieved, the service life of the mold is prolonged, and the reliability of the mold is improved. And important technical support is provided for 3D analogue simulation analysis of mold design and manufacturing.

Description

Technical Field

[0001] This application relates to the technical field of image data processing, and specifically relates to a method for 3D simulation analysis of molds. Background Art

[0002] The prediction and optimization of the fatigue life of molds is a complex engineering problem, involving multiple interrelated technical challenges. During the use of molds, they are subjected to complex cyclic loads, and the internal stress distribution shows high nonlinearity and non-uniformity. Traditional simplified analysis methods are difficult to capture such complex stress states, which easily cause large deviations in life prediction. Even when using the finite element method for refined modeling, there are still challenges in ensuring the rationality of mesh division and the accuracy of boundary conditions; the fatigue failure mechanism of molds is very complex, involving multiple stages such as crack initiation and propagation. A single stress index is difficult to comprehensively reflect the fatigue damage process. It is urgent to extract the most representative characteristic parameters from a large amount of stress analysis data and establish a quantitative relationship between the stress state and the fatigue life; the optimal design of the mold structure faces many challenges. While ensuring the strength and stiffness of the mold, specifically improving local high-stress areas, it is necessary to avoid material waste caused by over-design and ensure that the optimized structure still has good processability and economy. It is necessary to establish an organic connection between simulation analysis, life prediction, and structural optimization.

[0003] The solution proposed by the present invention for the above-mentioned drawbacks is as follows: Through a refined 3D CAD model and encrypted mesh division, accurately capture the complex stress inside the mold, extract key stress characteristic parameters, establish a non-linear prediction model to comprehensively reflect fatigue damage. Through simulation analysis and structural optimization, improve high-stress areas, avoid over-design, and ensure the processability and economy of the optimized structure. At the same time, iteratively optimize to reduce stress concentration, improve fatigue life, and verify the prediction accuracy through actual tests, optimize cutting parameters, and improve machining accuracy and surface quality, thereby effectively solving the defects in the prior art. Summary of the Invention

[0004] In order to solve the problems existing in the above-mentioned prior art, the purpose of this application is to provide a method for 3D simulation analysis of molds.

[0005] A method for 3D simulation analysis of molds according to this application includes the following steps:

[0006] S1. According to the 3D CAD model of the mold, tetrahedral meshes are used to discretize the space of the mold. During the mesh generation process in the sharp corners and thin-wall parts of the stress concentration area, the meshes are refined. The stress concentration area is selected as the key area, and the relative error of the stress value in the key area is calculated. The formula for the relative error is: |(σ_n - σ_n-1) / σ_n-1|, where σ_n is the stress value under the current mesh density, and σ_n-1 is the stress value under the previous-level mesh density;

[0007] S2. For the 3D CAD model of the mold after mesh generation in step S1, input the elastic modulus and Poisson's ratio parameters of the mold material, apply the working load and direction of the mold, set the displacement constraints and contact conditions, and solve for the displacement and stress components of each node of the mold through finite element analysis software;

[0008] S3. Select a color scale according to the yield strength of the mold material in step S2, generate a stress distribution contour map, adjust the viewing angle to highlight the distribution characteristics of the stress concentration area, and add dimension markings;

[0009] S4. Extract the features of the stress distribution contour map in step S3, obtain the numerical values and spatial coordinates of the maximum principal stress value and von Mises stress value, calculate the stress gradient direction and amplitude, and quantitatively evaluate the uniformity of the stress distribution;

[0010] S5. Combining the fatigue performance data of the mold material, a non-linear prediction model of stress characteristic parameters and mold fatigue life is established using the support vector machine regression algorithm. Select the maximum principal stress value, von Mises stress value, and stress gradient amplitude in step S4 as the characteristic parameters, and the fatigue life as the label data. Cross-validation is performed through the grid search method to determine the kernel function type, regularization parameter, and kernel function parameter of the support vector machine;

[0011] S6. Use the non-linear prediction model in step S5 to determine whether the stress concentration degree of the current mold design scheme exceeds the material fatigue strength limit. If it exceeds the material fatigue strength limit, chamfer the sharp corners in the mold structure and add stiffeners to the thin-wall parts. If it does not exceed the material fatigue strength limit, no treatment is performed, and an optimized mold design scheme is obtained;

[0012] S7. Re-perform finite element analysis on the optimized mold design scheme in step S6, extract the stress characteristic parameters, and calculate the reduction amplitude of the maximum stress value and the improvement degree of the stress distribution uniformity to predict the increase in fatigue life;

[0013] S8. Convert the optimized die design solution in step S6 into a 3D CAD model, import it into computer-aided manufacturing software for tool path planning and cutting parameter selection, simulate the NC machining process and estimate the machining time, generate an NC machining program, and upload it to the NC machine tool for machining.

[0014] Preferably, in step S1, obtain the 3D CAD model data of the die, preprocess the model, and perform mesh generation; determine the stress concentration areas according to the structural characteristics and stress distribution of the die, where the stress concentration areas include sharp corners and thin-wall parts, and mark them as key areas; discretize the die in space using tetrahedral meshes to generate an initial mesh model; during the mesh generation process, refine the meshes in the key areas; perform finite element analysis to calculate the stress values of each element at the current mesh density, and obtain the stress distribution contour map of the entire die; select the elements in the key areas, extract their stress values, and calculate the relative error between the stress value at the current mesh density and the stress value at the previous mesh density.

[0015] Among them, the calculation formula for the relative error is: |(σ_n - σ_n-1) / σ_n-1|, where σ_n is the stress value at the current mesh density, and σ_n-1 is the stress value at the previous mesh density.

[0016] Judge whether the calculated relative error is less than the preset threshold: if so, it is considered that the current mesh size meets the convergence requirement, and stop the calculation; if not, return to the key areas to further refine the meshes and recalculate.

[0017] According to the final mesh model, perform finite element analysis to obtain the stress distribution of the die under actual working conditions, which is used to identify the stress concentration areas and weak links.

[0018] Preferably, in step S2, according to the geometric shape and structural characteristics of the mold, a grid division method is used to discretize the three-dimensional CAD model of the mold to generate a finite element model; the mechanical property parameters of the mold material, including the elastic modulus and Poisson's ratio, are obtained and input into the finite element analysis software as material properties, and the material properties are input into each element of the mold finite element model; the magnitude and direction of the external load applied to the mold are determined and applied as boundary conditions to the corresponding positions of the finite element model, and at the same time, the displacement constraint conditions of the mold are set, including restricting the movement of degrees of freedom; the contact relationship and contact properties between the various components of the mold are defined, including the contact type, contact surface, and friction coefficient, and a contact pair is established in the finite element model; the solution module of the finite element analysis software is called, the solution method and convergence criterion are selected, and the force analysis of the mold finite element model is carried out to obtain the deformation field and stress field distribution results of the mold under the working load; the displacement component and stress component data of each node of the mold are extracted from the finite element analysis results, and according to the magnitude and distribution characteristics of the data, it is judged whether the mold meets the strength and stiffness requirements under the working load, and the stress concentration area is identified;

[0019] If the deformation and stress levels of the mold exceed the preset range, then according to the optimized mold design scheme, a finite element analysis model is constructed, the material properties, boundary conditions, and load conditions are set, and finite element solution calculations are carried out;

[0020] Stress characteristic parameters are extracted from the finite element analysis results, including the maximum stress value and the stress distribution uniformity index. Using a fatigue life prediction model and combining the stress characteristic parameters, the fatigue life of the optimized mold is calculated to obtain the fatigue life increase amount; through the stress distribution contour map and the stress uniformity index, the stress distribution uniformity of the optimized mold is evaluated;

[0021] If both the reduction amplitude of the maximum stress value and the fatigue life increase amount reach the preset thresholds, and the stress distribution uniformity meets the requirements, then the optimized mold design scheme is output.

[0022] Preferably, in step S3, according to the yield strength of the die material, a mapping relationship between the color scale and the stress value is established for generating the stress distribution contour map; by using the finite element analysis method, calculate the stress distribution of the die under the stress state to obtain the stress value data of each region; according to the stress value data and the color scale mapping relationship, generate the stress distribution contour map, where different colors represent different stress value ranges; perform image processing on the generated stress distribution contour map by adjusting the viewing angle and rendering parameters; add dimension markings to the stress distribution contour map to mark the dimension information of the key regions; use an image segmentation algorithm to divide the stress distribution contour map into regions, identify the position and range of the stress concentration region; according to the region division result, extract the stress value data of the stress concentration region, perform statistical analysis to obtain the characteristic parameters of the stress concentration degree and distribution; calculate the mean square deviation of the local stress gradient amplitude to quantitatively evaluate the uniformity of the overall stress distribution; according to the quantitative evaluation result of the stress distribution uniformity, use a clustering algorithm to partition the stress distribution contour map to identify the high-risk regions with uneven stress distribution.

[0023] Preferably, in step S4, according to the yield strength of the die material, a mapping relationship between the color scale and the stress value is established for generating the stress distribution contour map; by using the finite element analysis method, calculate the stress distribution of the die under the stress state to obtain the stress value data of each region; according to the stress value data and the color scale mapping relationship, generate the stress distribution contour map; perform image processing on the generated stress distribution contour map, use an image processing algorithm to extract the features of the contour map to obtain a digital representation of the stress distribution; for the extracted digital representation of the stress distribution, determine the numerical values and spatial coordinates corresponding to the maximum principal stress value and the von Mises stress value; according to the spatial coordinates of the maximum principal stress value and the von Mises stress value, calculate the gradient direction and gradient amplitude of the stress in space; if the gradient amplitude exceeds the preset threshold, it is determined that the stress distribution is uneven, and a machine learning algorithm is used to mark this region; perform a uniformity evaluation on the entire stress distribution contour map, calculate the mean square deviation of the local stress gradient amplitude to quantitatively evaluate the uniformity of the overall stress distribution; according to the quantitative evaluation result of the stress distribution uniformity, use a clustering algorithm to partition the stress distribution contour map to identify the high-risk regions with uneven stress distribution.

[0024] Preferably, in the step S5, fatigue performance data of the die material are obtained, including stress characteristic parameters such as the maximum principal stress value, von Mises stress value, and stress gradient amplitude value, and corresponding fatigue life data; the obtained stress characteristic parameters and fatigue life data are preprocessed to remove outliers and noise data and perform data normalization; based on the preprocessed data, a nonlinear prediction model between the stress characteristic parameters and fatigue life is established using the support vector machine regression algorithm; through the grid search method, different combinations of kernel function types, regularization parameters, and kernel function parameters are traversed for cross-validation to evaluate the prediction performance of the nonlinear prediction model; according to the cross-validation results, the optimal support vector machine model parameters are determined, including the kernel function type, regularization parameter, and kernel function parameter; using the optimized support vector machine model, the stress characteristic parameters of the new die material are predicted to obtain the corresponding predicted fatigue life value; the predicted fatigue life is compared with the actual test results to evaluate the prediction accuracy of the nonlinear prediction model; the finite element analysis method is used to calculate the stress distribution of the die under the working condition to obtain the stress concentration area and stress concentration coefficient; the calculated stress concentration coefficient is used as the input and substituted into the pre-established fatigue life prediction model to predict the fatigue life of the die under the current stress level;

[0025] If the fatigue life of the die is lower than the preset threshold corresponding to the fatigue strength limit of the material, it is determined that the stress concentration degree of the current die design scheme is high and structural optimization is required; for the stress concentration area, the chamfering treatment method is adopted, and an arc transition with a certain radius is set at the sharp corners; for the thin-walled part, by setting ribs in the thin-walled area, the cross-sectional area is increased;

[0026] The optimized die is re-analyzed by finite element analysis to calculate the stress distribution and predict the fatigue life.

[0027] Preferably, in step S6, fatigue performance data of the die material are obtained, including stress characteristic parameters such as the maximum principal stress value, von Mises stress value, and stress gradient amplitude value, and corresponding fatigue life data; the obtained stress characteristic parameters and fatigue life data are preprocessed to remove outliers and noise data and perform data normalization; according to the preprocessed data, a nonlinear prediction model between stress characteristic parameters and fatigue life is established using the support vector machine regression algorithm; through the grid search method, different combinations of kernel function types, regularization parameters, and kernel function parameters are traversed for cross-validation to evaluate the prediction performance of the model; according to the cross-validation results, the optimal support vector machine model parameters are determined, including kernel function type, regularization parameter, and kernel function parameter; using the optimized support vector machine model, the stress characteristic parameters of the new die material are predicted to obtain the corresponding predicted fatigue life value; the finite element analysis method is used to calculate the stress distribution of the die under the working condition to obtain the stress concentration area and stress concentration coefficient; the calculated stress concentration coefficient is used as the input and substituted into the pre-established fatigue life prediction model to predict the fatigue life of the die at the current stress level; the predicted die fatigue life is compared with the fatigue strength limit value of the material to determine whether the stress concentration degree of the current die design scheme exceeds the material fatigue strength limit;

[0028] If the die fatigue life is lower than the preset threshold corresponding to the material fatigue strength limit, it is determined that the stress concentration degree of the current die design scheme is high and structural optimization is required; for the stress concentration area, the chamfering treatment method is adopted, and an arc transition with a certain radius is set at the sharp corners; for the thin-wall part, by presetting stiffeners in the thin-wall area to increase the cross-sectional area, the optimized die is re-analyzed by finite element method to calculate the stress distribution and predict the fatigue life;

[0029] If the fatigue life of the optimized die is still lower than the preset threshold corresponding to the material fatigue strength limit, continue to iteratively optimize the die structure design until the fatigue life requirement is met;

[0030] Through iterative optimization, gradually reduce the stress concentration degree of the die and improve the fatigue life of the die; compare the optimized die design scheme with the original design scheme to evaluate the optimization effect; quantify the optimization effect by comparing the stress distribution, stress concentration coefficient, and fatigue life prediction value before and after optimization, determine the final die design scheme, convert the finally determined die design scheme into a three-dimensional model, and generate the numerical control machining code required for die machining; according to the generated numerical control machining code, perform die machining and manufacturing, and conduct actual fatigue tests on the machined die to verify the accuracy of the fatigue life prediction model and the effectiveness of the optimized design.

[0031] Preferably, in step S7, according to the optimized die design scheme, a finite element analysis model is constructed, material properties, boundary conditions, and load conditions are set, and the finite element solution method is used to calculate the stress distribution of the die under the working state, and the stress concentration area and stress concentration coefficient are obtained; stress characteristic parameters are extracted from the finite element analysis results, including the maximum stress value and the stress distribution uniformity index, and they are compared with the stress characteristic parameters before optimization, the reduction amplitude of the maximum stress value is calculated, and it is judged whether the preset threshold is reached. If not, return to continue the die design optimization iteration; take the calculated stress concentration coefficient as the input and substitute it into the pre-established fatigue life prediction model to predict the fatigue life of the die at the current stress level, and obtain the fatigue life increase amount; judge whether the fatigue life increase amount reaches the preset percentage. If not, return to continue the die design optimization iteration until the fatigue life increase amount is not less than 20% of the original life; through the stress distribution cloud map and the stress uniformity index, evaluate the stress distribution uniformity of the optimized die. If the uniformity does not meet the requirements, for the stress concentration area, adopt the chamfering treatment method, set an arc transition with a certain radius at the sharp corners to improve the stress distribution uniformity; for the thin-walled part, by presetting stiffeners in the thin-walled area, increasing the cross-sectional area, and reducing the stress concentration degree, an optimized die design scheme is obtained;

[0032] Re-perform finite element analysis on the optimized die design scheme in step S6, calculate the stress distribution, extract the maximum stress value and the stress uniformity index, and judge whether it meets the requirements of the preset threshold;

[0033] If both the reduction amplitude of the maximum stress value and the increase amount of the fatigue life reach the preset threshold, and the stress distribution uniformity meets the requirements, then output the optimized die design scheme and determine the final die structure parameters;

[0034] Transfer the optimized die design scheme to the next link for die processing and manufacturing to complete the die design optimization process.

[0035] Preferably, in step S8, according to the optimized mold design scheme, a 3D CAD model is constructed, and the CAD model is imported into CAM software through a model conversion interface for subsequent processing; in the CAM software, select the tool type, cutting parameters such as cutting amount and feed rate, plan the tool movement path, and generate a numerical control processing program; utilize the machining simulation function provided by the CAM software to simulate the numerical control machining process, analyze the interference between the tool and the workpiece and the cutting force situation, estimate the machining time, optimize the cutting parameters and the tool path; convert the generated numerical control processing program into a code format recognizable by the numerical control machine tool through a post-processor, upload it to the numerical control system, and the numerical control machine tool controls the tool movement according to the processing program instructions. During the machining process, the cutting parameters are dynamically adjusted by on-line monitoring of tool wear and workpiece dimensions; after machining, use the in-machine probe to detect the key dimensions of the workpiece, compare and analyze the detected data with the CAD model, evaluate the machining accuracy and surface quality, and form a quality feedback report; according to the quality feedback report, optimize and adjust the numerical control processing program and the cutting parameters, and apply the optimized processing scheme to subsequent batch production;

[0036] If the machining accuracy and surface quality do not meet the requirements, then return to adjust the cutting parameters and the tool path, regenerate the numerical control processing program until the machining quality meets the standard;

[0037] Save the optimized numerical control processing program and cutting parameters to the process database.

[0038] The advantages of a mold 3D simulation analysis method described in this application are as follows: First, perform finite element analysis on the mold, ensure the calculation accuracy through mesh encryption and convergence analysis, then extract stress distribution characteristics, including parameters such as the maximum principal stress, von Mises stress, and stress gradient, establish a non-linear prediction model of these stress characteristics and fatigue life using the support vector machine regression algorithm, based on the prediction results, perform structural optimization on the areas exceeding the fatigue strength limit, including chamfers and ribs, through iterative analysis and optimization until the stress reduction amplitude and life increase amount reach the preset goals, and finally convert the optimization scheme into a CAD model and generate a processing program;

[0039] The present invention realizes the accurate prediction of the mold fatigue life and the automated optimization design, effectively improves the service life and reliability of the mold, and provides important technical support for the 3D simulation analysis of mold design and manufacturing. Brief Description of the Drawings

[0040] Figure 1 is the flow of a mold 3D simulation analysis method described in this application Figure One ;

[0041] Figure 2 is the flow of a mold 3D simulation analysis method described in this applicationFigure Two ;

[0042] Figure 3 is the flow of a 3D simulation analysis method for a mold described in this application Figure Three . Detailed implementation manner

[0043] As Figures 1 - 3 shown, a 3D simulation analysis method for a mold described in this application includes the following steps:

[0044] S1. According to the three-dimensional CAD model of the mold, the mold is spatially discretized using tetrahedral meshes, and mesh refinement is performed during the mesh generation process at the sharp corners and thin-walled parts in the stress concentration area. The stress concentration area is selected as the key area, and the relative error of the stress value in the key area is calculated. Among them, the calculation formula for the relative error is: |(σ_n - σ_n-1) / σ_n-1|, where σ_n is the stress value under the current mesh density, and σ_n-1 is the stress value under the previous-level mesh density;

[0045] S2. For the three-dimensional CAD model of the mold after mesh generation in step S1, the elastic modulus and Poisson's ratio parameters of the mold material are input, the mold working load and direction are applied, the displacement constraints and contact conditions are set, and the displacement and stress components of each node of the mold are solved through finite element analysis software;

[0046] S3. According to the yield strength of the mold material in step S2, a color scale is selected to generate a stress distribution contour map, and the viewing angle is adjusted to highlight the distribution characteristics of the stress concentration area, and dimension labels are added;

[0047] S4. Feature extraction is performed on the stress distribution contour map in step S3 to obtain the numerical values and spatial coordinates of the maximum principal stress value and von Mises stress value, calculate the stress gradient direction and amplitude, and quantitatively evaluate the uniformity of the stress distribution;

[0048] S5. Combining the fatigue performance data of the mold material, a non-linear prediction model of stress characteristic parameters and mold fatigue life is established using the support vector machine regression algorithm. The maximum principal stress value, von Mises stress value, and stress gradient amplitude in step S4 are selected as feature parameters, and the fatigue life is used as label data. Cross-validation is performed through the grid search method to determine the kernel function type, regularization parameter, and kernel function parameter of the support vector machine;

[0049] S6. Using the non-linear prediction model in step S5, it is judged whether the stress concentration degree of the current mold design scheme exceeds the material fatigue strength limit. If it exceeds the material fatigue strength limit, chamfering is performed on the sharp corners in the mold structure, and ribbing is performed on the thin-walled parts. If it does not exceed the material fatigue strength limit, no treatment is performed, and an optimized mold design scheme is obtained;

[0050] S7. Re - conduct finite - element analysis on the optimized die design scheme in step S6, extract stress characteristic parameters, calculate the reduction amplitude of the maximum stress value and the improvement degree of the stress distribution uniformity, and use them to predict the increase in fatigue life;

[0051] S8. Convert the optimized die design scheme in step S6 into a 3D CAD model, import it into computer - aided manufacturing software for tool - path planning and cutting - parameter selection, simulate the NC machining process and estimate the machining time, generate an NC machining program, and upload it to the NC machine tool for machining.

[0052] As Figures 1 - 3 shown, in step S1, according to the 3D CAD model of the die, the die is spatially discretized using tetrahedral meshes. During the mesh - generation process in the sharp - corner and thin - wall parts of the stress - concentration area, the mesh is refined, and these areas are selected as key areas to calculate the relative error of the stress values in the key areas. The relative - error calculation formula is: |(σ_n - σ_n - 1) / σ_n - 1|, where σ_n is the stress value under the current mesh density and σ_n - 1 is the stress value under the previous - level mesh density. When the relative error is less than the preset threshold, it is considered that the mesh size meets the convergence requirement.

[0053] Furthermore, in step S1, obtain the 3D CAD model data of the die, pre - process the model, check the integrity and rationality of the model to ensure that the model can be meshed;

[0054] According to the structural characteristics and stress - distribution of the die, determine the stress - concentration areas, mainly including sharp - corners and thin - wall parts, mark these areas as key areas, and spatially discretize the die using tetrahedral meshes to generate an initial mesh model;

[0055] During the mesh - generation process, refine the mesh in the key areas to increase the mesh density;

[0056] Conduct finite - element analysis, calculate the stress values of each element under the current mesh density, and obtain the stress - distribution contour map of the entire die; select the elements in the key areas, extract their stress values, and calculate the relative error between the stress value under the current mesh density and the stress value under the previous - level mesh density. The relative - error calculation formula is: |(σ_n - σ_n - 1) / σ_n - 1|, where σ_n is the stress value under the current mesh density and σ_n - 1 is the stress value under the previous - level mesh density;

[0057] Judge whether the calculated relative error is less than the preset threshold,

[0058] If so, it is considered that the current mesh size meets the convergence requirement and stop the calculation;

[0059] If not, return to further encrypt the grid in the key area, increase the grid density, and recalculate.

[0060] Through iterative calculations, continuously optimize the grid model until the relative error meets the preset threshold requirements, and obtain the final grid model that meets the convergence requirements; according to the final grid model, perform finite element analysis to obtain the stress distribution of the mold under actual working conditions, and identify the stress concentration areas and weak links; output the grid model that meets the convergence requirements and the corresponding stress distribution nephogram to provide data support for subsequent mold structure optimization and manufacturing process optimization.

[0061] Specifically, in step S1, first obtain the three-dimensional CAD model data of the mold, import it into the CAD software in the IGES and STEP general formats, and repair and simplify the model to eliminate factors such as small facets and sharp corners that affect mesh generation.

[0062] Then, according to the structural characteristics and stress distribution of the mold, use the curvature analysis method to determine the stress concentration areas, and mark the areas with a curvature greater than 0.5 mm^-1 as the key areas.

[0063] Next, use tetrahedral meshes to spatially discretize the mold to generate an initial grid model, set the grid size to 5 mm, and encrypt the grid size to 0.5 mm in the key areas.

[0064] Perform finite element analysis, use the von Mises stress criterion to calculate the stress values of each element, and obtain the stress distribution nephogram of the entire mold.

[0065] Select the elements in the key areas, extract their stress values, calculate the relative error. If the relative error is greater than 5%, return to the mesh generation step, reduce the grid size in the key areas to 50% of the original, and recalculate.

[0066] Through iterative calculations, continuously optimize the grid model until the relative error is less than 5%, and obtain the final grid model that meets the convergence requirements.

[0067] According to the final grid model, perform finite element analysis, use the static analysis method to calculate the stress distribution of the mold under actual working conditions, and identify the areas where the stress is greater than 80% of the material yield strength as the stress concentration areas and weak links.

[0068] Finally, output the grid model file and the stress distribution nephogram to provide data support for subsequent mold structure optimization and manufacturing process optimization.

[0069] Such as Figures 1 - 3As shown, in step S2, for the three-dimensional CAD model of the mold after meshing in step S1, input the elastic modulus and Poisson's ratio parameters of the mold material, apply the working load and direction of the mold, set the displacement constraints and contact conditions, and solve for the displacement and stress components of each node of the mold through finite element analysis software.

[0070] Further, in step S2, according to the geometric shape and structural characteristics of the mold, use a meshing method to discretize the three-dimensional CAD model of the mold to generate a finite element model;

[0071] Obtain the mechanical property parameters of the mold material, including elastic modulus and Poisson's ratio, input them as material properties into the finite element analysis software, and assign the material properties to each element of the mold finite element model;

[0072] Determine the magnitude and direction of the external load on the mold, apply it as a boundary condition to the corresponding position of the finite element model, and at the same time set the displacement constraint conditions of the mold, including restricting the movement of degrees of freedom;

[0073] Define the contact relationship and contact properties between each component of the mold, including contact type, contact surface, friction coefficient, and establish contact pairs in the finite element model;

[0074] Call the solution module of the finite element analysis software, select the solution method and convergence criterion, perform a force analysis on the mold finite element model, and obtain the distribution results of the deformation field and stress field of the mold under the working load; Extract the displacement component and stress component data of each node of the mold from the finite element analysis results, and judge whether the mold meets the strength and stiffness requirements under the working load according to the magnitude and distribution characteristics of the data, and identify the stress concentration areas;

[0075] If the deformation and stress levels of the mold exceed the allowable range, then according to the optimized mold design scheme, construct a finite element analysis model, set the material properties, boundary conditions, and load conditions, and perform finite element solution calculations;

[0076] Extract stress characteristic parameters from the finite element analysis results, including the maximum stress value and the stress distribution uniformity index, use a fatigue life prediction model, combine with the stress characteristic parameters, calculate the fatigue life of the optimized mold, and obtain the increase in fatigue life; Evaluate the uniformity of the stress distribution of the optimized mold through the stress distribution contour map and the stress uniformity index. If both the reduction amplitude of the maximum stress value and the increase in fatigue life reach the preset thresholds, and the uniformity of the stress distribution meets the requirements, then output the optimized mold design scheme to complete the design optimization process.

[0077] Specifically, in step S2, first, a hexahedral mesh is used to divide the die solid model, with the mesh size set to 5 mm, generating a finite element model containing approximately 500,000 nodes and 300,000 elements;

[0078] The die material is set as H13 hot work die steel, with an elastic modulus of 210 GPa and a Poisson's ratio of 0.3, and these material properties are assigned to all elements of the finite element model;

[0079] According to the actual working conditions, a pressure load of 3 MPa is applied to the die cavity surface, and a full constraint is applied to the bottom surface of the die;

[0080] The contact relationship between the various components of the die is defined as surface-to-surface contact, with a friction coefficient of 0.1, and contact pairs are established at the die parting surface and the slider surface;

[0081] Using ANSYS finite element analysis software, select the static analysis type, adopt the direct sparse matrix solution method, set the residual convergence criterion to 0.001, and perform a stress analysis calculation on the die;

[0082] Extract the displacement and stress data of each node of the die from the calculation results. It is found that the maximum deformation is 0.15 mm and the maximum stress is 650 MPa, which is located at the sharp corners of the die cavity;

[0083] Since the maximum stress exceeds the material yield strength, the die structure needs to be optimized;

[0084] According to the optimization plan, fillets are added in the stress concentration area, the slider guiding method is adjusted, and finite element analysis is performed again;

[0085] The maximum stress after optimization is reduced to 520 MPa, the maximum deformation is reduced to 0.12 mm, and the fatigue life prediction shows that the number of uses can reach more than 500,000 times, meeting the design requirements;

[0086] Finally, through the stress nephogram and the stress uniformity coefficient (the uniformity coefficient is increased from 1.8 to 1.5), it is verified that the stress distribution of the optimized die is more reasonable, and the design scheme can be finalized and the process preparation can be carried out.

[0087] As Figures 1 - 3 shown, in step S3, according to the material yield strength of the die, a color scale is selected to generate a stress distribution nephogram, and the viewing angle is adjusted to highlight the distribution characteristics of the stress concentration area, and dimension labels are added.

[0088] Furthermore, in step S3, according to the yield strength of the die material, a mapping relationship between the color scale and the stress value is established for the generation of the stress distribution nephogram;

[0089] By using the finite element analysis method, calculate the stress distribution of the mold under the stress state to obtain the stress value data of each region;

[0090] According to the stress value data and the color scale mapping relationship, generate a stress distribution contour map, where different colors represent different stress value ranges;

[0091] Perform image processing on the generated stress distribution contour map. By adjusting the viewing angle and rendering parameters, highlight the distribution characteristics of the stress concentration region;

[0092] Add dimension markings to the stress distribution contour map, mark the dimension information of the key regions, and use an image segmentation algorithm to divide the stress distribution contour map into regions to identify the location and range of the stress concentration regions;

[0093] According to the region division results, extract the stress value data of the stress concentration regions, conduct statistical analysis, and obtain the characteristic parameters of the stress concentration degree and distribution;

[0094] Quantitatively evaluate the uniformity of the overall stress distribution by calculating the mean square deviation of the local stress gradient amplitude;

[0095] According to the quantitative evaluation results of the stress distribution uniformity, use a clustering algorithm to partition the stress distribution contour map to identify the high-risk regions with non-uniform stress distribution.

[0096] Specifically, in step S3, establish the mapping relationship between the color scale and the stress value according to the yield strength of the mold material. For example, if the yield strength is 500 MPa, map the stress value range of 0 - 500 MPa to the color scale from blue to red;

[0097] By using the finite element analysis method, including using ANSYS software, perform mesh division and boundary condition setting on the mold, and calculate the stress value data of each region of the mold under the stress state;

[0098] According to the stress value data and the color scale mapping relationship, use the matplotlib library in Python to generate a stress distribution contour map, where different colors represent different stress value ranges;

[0099] Perform image processing on the generated stress distribution contour map, including using the OpenCV library. By adjusting the viewing angle and rendering parameters, including light intensity and material properties, highlight the distribution characteristics of the stress concentration region;

[0100] Add dimension markings to the stress distribution contour map, use the PIL library to add text markings and arrow indicators in the key regions to mark the key dimension information;

[0101] Adopt an image segmentation algorithm, including a segmentation algorithm based on region growing, to divide the stress distribution contour map into regions and identify the location and scope of the stress concentration region;

[0102] According to the region division result, extract the stress value data of the stress concentration region, use the NumPy library for statistical analysis, and calculate the average stress value of the stress concentration region and the distribution characteristic parameters of the stress gradient;

[0103] By calculating the mean square deviation of the local stress gradient amplitude, including using the sliding window method, sliding a window of 20×20 pixels on the stress distribution contour map, and calculating the mean square deviation of the stress gradient amplitude within each window, quantitatively evaluate the overall stress distribution uniformity;

[0104] According to the quantitative evaluation result of the stress distribution uniformity, adopt the K-means clustering algorithm to partition the stress distribution contour map, identify the high-risk regions with non-uniform stress distribution, and provide a basis for subsequent die optimization design.

[0105] Specifically, it also includes: in step S3, through the establishment of a mapping relationship, preset the yield strength to 500 MPa, and establish the mapping relationship between the color scale and the stress value. The mapping range is preset to 0 - 500 MPa, corresponding to the color scale from blue to red, including 0 MPa corresponding to blue, 500 MPa corresponding to red, and the intermediate stress values are distributed to the corresponding color gradients according to the linear interpolation method. The color corresponding to 250 MPa is purple, which is between blue and red;

[0106] Use ANSYS software for finite element analysis, preset the mesh size to 0.5 mm, and set the fixed constraints and actual working loads. The fixed constraints are realized by simulating the die support surface, while the actual working loads are obtained through the equivalent conversion and distribution calculation of forces according to the force conditions during die operation;

[0107] Generate the stress distribution contour map through the matplotlib library of Python, and use the OpenCV library to adjust the viewing angle and rendering parameters, including presetting the light intensity to 50%, the material property to metallic luster, and at the same time using the PIL library for dimension marking;

[0108] Adopt a segmentation algorithm based on region growing, preset the growth threshold to 150 MPa, identify the stress concentration region, and use the NumPy library to extract relevant data, including starting from the seed point, comparing the stress values of adjacent pixels. If the stress value is greater than 150 MPa, include it in the growth region until no new pixels meet the conditions;

[0109] Perform statistical analysis, calculate the average stress value and stress gradient parameters. Preset the sample size to 100 data points, calculate the mean square deviation of the local stress gradient amplitude by the sliding window method, and preset the window moving step size to 10 pixels. Evaluate the stress distribution uniformity, including the stress gradient, which is obtained by calculating the change rate of the stress values of adjacent data points.

[0110] Finally, use the K-means clustering algorithm, preset the number of cluster centers to 3, perform clustering analysis, identify the high-risk areas with non-uniform stress distribution, and preset the threshold to 3 times the standard deviation of the mean square deviation. This includes calculating the mean and standard deviation of the stress gradient within each 10-pixel window, and then averaging the standard deviations of all windows to obtain the mean square deviation.

[0111] As Figures 1 - 3 shown, in step S4, feature extraction is performed on the stress distribution contour map in step S3 to obtain the numerical values and spatial coordinates of the maximum principal stress value and von Mises stress value, calculate the stress gradient direction and amplitude, and quantitatively evaluate the uniformity of the stress distribution.

[0112] Furthermore, in step S4, according to the yield strength of the mold material, establish the mapping relationship between the color scale and the stress value for the generation of the stress distribution contour map.

[0113] By the finite element analysis method, calculate the stress distribution of the mold under the stress state to obtain the stress value data of each region.

[0114] According to the stress value data and the color scale mapping relationship, generate the stress distribution contour map, where different colors represent different stress value ranges.

[0115] Perform image processing on the generated stress distribution contour map, use image processing algorithms to extract features from the contour map, and obtain the digital representation of the stress distribution.

[0116] For the extracted digital representation of the stress distribution, determine the numerical values and spatial coordinates corresponding to the maximum principal stress value and von Mises stress value.

[0117] According to the spatial coordinates of the maximum principal stress value and von Mises stress value, calculate the gradient direction and gradient amplitude of the stress in space.

[0118] If the gradient amplitude exceeds the preset threshold, it is judged that the stress distribution is non-uniform, and a machine learning algorithm is used to mark this region.

[0119] Evaluate the uniformity of the entire stress distribution contour map, and quantitatively evaluate the uniformity of the overall stress distribution by calculating the mean square deviation of the local stress gradient amplitude.

[0120] Based on the quantitative evaluation results of the stress distribution uniformity, a clustering algorithm is used to partition the stress distribution contour map to identify high-risk areas with non-uniform stress distribution.

[0121] Specifically, in step S4, according to the yield strength of the die material, a mapping relationship between the color scale and the stress value is established. For example, if the yield strength is 500 MPa, the stress value range of 0 - 500 MPa is mapped to the color scale from blue to red.

[0122] Through the finite element analysis method, including using ANSYS software, the stress distribution of the die under a force of 100 kN is calculated to obtain the stress value data of each grid element.

[0123] According to the stress value data and the color scale mapping relationship, a stress distribution contour map is generated. For example, the area with a stress value of 0 - 100 MPa is shown in blue, the area of 100 - 200 MPa is shown in green, and so on.

[0124] Image processing is performed on the generated stress distribution contour map. The Canny edge detection algorithm is used to extract the features of the contour map to obtain a digital representation of the stress distribution, including extracting the gradient information of the stress value.

[0125] For the extracted digital representation of the stress distribution, through the search algorithm, it is determined that the maximum principal stress value is 450 MPa, located at the coordinates (100, 200) of the contour map, and the von Mises stress value is 380 MPa, located at the coordinates (150, 250).

[0126] According to the spatial coordinates of the maximum principal stress value and the von Mises stress value, the gradient direction of the stress in space is calculated to be 45 degrees and the gradient amplitude is 20 MPa / mm through the Sobel operator.

[0127] If the gradient amplitude exceeds the preset threshold of 15 MPa / mm, it is determined that the stress distribution is non-uniform, and the support vector machine algorithm is used to mark this area.

[0128] The uniformity of the entire stress distribution contour map is evaluated. By calculating the mean square deviation of the local stress gradient amplitude, the overall stress distribution uniformity evaluation value is obtained as 0.85.

[0129] According to the stress distribution uniformity evaluation value, the K-means clustering algorithm is used to partition the stress distribution contour map to identify high-risk areas with non-uniform stress distribution, including dividing the area with a stress gradient amplitude greater than 18 MPa / mm into high-risk areas.

[0130] Such as Figures 1 - 3As shown, in step S5, a non-linear prediction model of stress characteristic parameters and die fatigue life is established by using the support vector machine regression algorithm in combination with die material fatigue performance data. The maximum principal stress value, von Mises stress value, and stress gradient amplitude in step S4 are selected as characteristic parameters, and the fatigue life is used as label data. Cross-validation is performed through the grid search method to determine the kernel function type, regularization parameter, and kernel function parameter of the support vector machine.

[0131] Further, in step S5, the fatigue performance data of the die material is obtained, including the maximum principal stress value, von Mises stress value, stress gradient amplitude stress characteristic parameters, and corresponding fatigue life data.

[0132] The obtained stress characteristic parameters and fatigue life data are preprocessed to remove outliers and noise data, and data normalization is performed.

[0133] According to the preprocessed data, a non-linear prediction model between stress characteristic parameters and fatigue life is established by using the support vector machine regression algorithm.

[0134] Through the grid search method, different combinations of kernel function types, regularization parameters, and kernel function parameters are traversed for cross-validation to evaluate the prediction performance of the model.

[0135] According to the cross-validation results, the optimal support vector machine model parameters are determined, including the kernel function type, regularization parameter, and kernel function parameter.

[0136] Using the optimized support vector machine model, the stress characteristic parameters of the new die material are predicted to obtain the corresponding predicted fatigue life value.

[0137] The predicted fatigue life is compared with the actual test results to evaluate the prediction accuracy of the model. The finite element analysis method is used to calculate the stress distribution of the die in the working state to obtain the stress concentration area and stress concentration coefficient.

[0138] The calculated stress concentration coefficient is used as input and substituted into the pre-established fatigue life prediction model to predict the fatigue life of the die at the current stress level.

[0139] If the die fatigue life is lower than the preset threshold corresponding to the material fatigue strength limit, it is determined that the stress concentration degree of the current die design scheme is high and structural optimization is required.

[0140] For the stress concentration area, the chamfering treatment method is adopted, and an arc transition with a certain radius is set at the sharp edges and corners.

[0141] For the thin-wall part, the cross-sectional area is increased by adding ribs in the thin-wall area.

[0142] Perform finite element analysis on the optimized mold again, calculate the stress distribution, and predict the fatigue life.

[0143] Specifically, in step S5, first obtain the fatigue performance data of the mold material, including stress characteristic parameters such as the maximum principal stress value, von Mises stress value, stress gradient amplitude, and the corresponding fatigue life data. The preset dataset contains 1000 samples, including 700 training sets, 200 validation sets, and 100 test sets. Preprocess the obtained data to remove outliers and noise data, and normalize the data to the [0, 1] interval using the min-max normalization method. The preset minimum value is 0, and the maximum value is 1.

[0144] Then, based on the preprocessed data, establish a model using the support vector machine regression algorithm. Through the grid search method, the preset kernel function types include linear kernel, polynomial kernel (degree preset to 3), and radial basis kernel; the range of the regularization parameter is {0.1, 1, 10}; the range of the kernel function parameter gamma value is {0.001, 0.01, 0.1, 1}. Perform 5-fold cross-validation, calculate the mean squared error (MSE) of each cross-validation, preset the MSE threshold to 0.001, and select the parameter combination with the smallest MSE as the optimal model parameter.

[0145] Use the optimized support vector machine model to predict the stress characteristic parameters of the new mold material, obtain the corresponding predicted fatigue life value, and compare the predicted value with the actual test result. Calculate the average absolute error and root mean square error evaluation indicators to evaluate the prediction accuracy of the model.

[0146] Include calculating the average absolute error (MAE) and root mean square error (RMSE) between the predicted value and the actual test result. Preset the error thresholds of MAE and RMSE to 0.05. If the error exceeds the threshold, adjust the model parameters and retrain.

[0147] Next, adopt the finite element analysis method. Import the 3D CAD model of the mold into the finite element analysis software, set the material properties, mesh division, load, and boundary condition parameters, calculate the stress distribution of the mold under the working condition, and obtain the stress concentration area and stress concentration coefficient.

[0148] Include using the finite element analysis method, preset the mesh size to 0.5 mm, perform mesh division on the 3D CAD model of the mold, set the material properties (elastic modulus, Poisson's ratio), the load condition is 1.5 times the actual working load, and the boundary conditions are fixed constraints and force loads. Calculate the stress distribution, identify the stress concentration area, and preset the stress concentration coefficient threshold to 3.

[0149] Substitute the calculated stress concentration factor into the pre-established fatigue life prediction model to predict the fatigue life of the mold under the current stress level. If the fatigue life of the mold is lower than the preset threshold (100,000 times) corresponding to the fatigue strength limit of the material, it is determined that the stress concentration degree of the current mold design scheme is high and structural optimization is required;

[0150] Adopt a chamfering treatment method for the stress concentration area and set an arc transition with a radius of 2 mm at the sharp corners;

[0151] Preset stiffeners with a thickness of 3 mm in the thin-wall area for the thin-wall part. Re-perform finite element analysis on the optimized mold, calculate the stress distribution, and predict the fatigue life until the fatigue life requirement is met.

[0152] As Figures 1 - 3 shown, in step S6, use the trained fatigue life prediction model to judge whether the stress concentration degree of the current mold design scheme exceeds the fatigue strength limit of the material. If it exceeds the preset threshold, chamfer the sharp corners in the mold structure and design and process stiffeners for the thin-wall part.

[0153] Furthermore, in step S6, obtain the fatigue performance data of the mold material, including the maximum principal stress value, von Mises stress value, stress gradient amplitude stress characteristic parameters, and the corresponding fatigue life data;

[0154] Preprocess the obtained stress characteristic parameters and fatigue life data, remove outliers and noise data, and perform data normalization;

[0155] According to the preprocessed data, use the support vector machine regression algorithm to establish a non-linear prediction model between the stress characteristic parameters and the fatigue life;

[0156] Through the grid search method, traverse different combinations of kernel function types, regularization parameters, and kernel function parameters, perform cross-validation, and evaluate the prediction performance of the model;

[0157] According to the cross-validation results, determine the optimal support vector machine model parameters, including the kernel function type, regularization parameter, and kernel function parameter;

[0158] Use the optimized support vector machine model to predict the stress characteristic parameters of the new mold material and obtain the corresponding predicted fatigue life value;

[0159] Adopt the finite element analysis method to calculate the stress distribution of the mold under the working condition, and obtain the stress concentration area and the stress concentration factor;

[0160] Take the calculated stress concentration factor as the input and substitute it into the pre-established fatigue life prediction model to predict the fatigue life of the mold under the current stress level;

[0161] Compare the predicted die fatigue life with the fatigue strength limit value of the material to determine whether the stress concentration degree of the current die design scheme exceeds the fatigue strength limit of the material;

[0162] If the die fatigue life is lower than the preset threshold corresponding to the fatigue strength limit of the material, it is determined that the stress concentration degree of the current die design scheme is high and structural optimization is required;

[0163] For the stress concentration area, adopt the chamfering treatment method and set a circular arc transition with a certain radius at the sharp corners;

[0164] For the thin-wall part, increase the cross-sectional area by presetting stiffeners in the thin-wall area;

[0165] Re-perform finite element analysis on the optimized die, calculate the stress distribution, and predict the fatigue life;

[0166] If the fatigue life of the optimized die is still lower than the preset threshold corresponding to the fatigue strength limit of the material, continue to iteratively optimize the die structure design until the fatigue life requirement is met;

[0167] Through iterative optimization, gradually reduce the stress concentration degree of the die and improve the fatigue life of the die;

[0168] Compare the optimized die design scheme with the original design scheme to evaluate the optimization effect;

[0169] Quantify the optimization effect by comparing the stress distribution, stress concentration coefficient, and fatigue life prediction values before and after optimization, and determine the final die design scheme;

[0170] Convert the finally determined die design scheme into a 3D model and generate the numerical control machining code required for die machining;

[0171] According to the generated numerical control machining code, carry out the machining and manufacturing of the die, and conduct actual fatigue tests on the machined die to verify the accuracy of the fatigue life prediction model and the effectiveness of the optimized design.

[0172] Specifically, in step S6, by conducting fatigue tests on the die material, obtain fatigue life data under different stress levels, including stress characteristic parameters such as the maximum principal stress value, von Mises stress value, and stress gradient amplitude;

[0173] Use the 3σ criterion to remove outliers and adopt the min-max normalization method to normalize the data to the [0,1] interval;

[0174] Based on the normalized data, a Gaussian kernel function is selected. Different combinations of the regularization parameter C and the kernel function parameter γ are traversed by the grid search method. The prediction performance of the support vector machine regression model is evaluated using 5-fold cross-validation, and the optimal parameter combination is obtained as C = 10, γ = 0.1;

[0175] Using the optimized support vector machine model, a finite element analysis of the stress distribution of the mold in the working state is carried out. The stress concentration coefficient in the stress concentration area is obtained as 2.5, and it is substituted into the fatigue life prediction model. The predicted fatigue life of the mold at the current stress level is 10,000 times;

[0176] The predicted life is compared with the fatigue strength limit of the material, which is 50,000 times, and it is determined that the stress concentration degree of the current mold design scheme is high;

[0177] For the stress concentration area, a fillet treatment is carried out with an arc transition of a radius of 2 mm, and a reinforcing rib with a preset thickness of 3 mm is provided in the thin-wall area;

[0178] After the finite element analysis of the optimized mold, the stress concentration coefficient is reduced to 1.8, and the fatigue life is increased to 30,000 times, meeting the requirements of the material fatigue strength limit;

[0179] The stress concentration coefficients and fatigue lives before and after optimization are compared. The stress concentration coefficient is reduced by 28%, and the fatigue life is increased by 2 times, verifying the effectiveness of the optimized design;

[0180] The optimized mold design scheme is converted into a three-dimensional model, and numerical control machining codes are generated to complete the machining and manufacturing of the mold and the actual fatigue test. The error between the test result and the predicted life is within 10%, verifying the accuracy of the fatigue life prediction model.

[0181] As Figures 1 - 2 shown, in step S7, the optimized mold design scheme in step S6 is re-analyzed by finite element method, stress characteristic parameters are extracted, the reduction amplitude of the maximum stress value and the improvement degree of stress distribution uniformity are calculated, and the increase amount of fatigue life is predicted. If the reduction amplitude of the maximum stress value is less than the preset threshold or the increase amount of fatigue life is less than the preset percentage, return to the previous step to continue optimization. The preset percentage can be set as the increase amount of fatigue life is not less than 20% of the original life.

[0182] Furthermore, in step S7, according to the optimized mold design scheme, a finite element analysis model is constructed, material properties, boundary conditions and load conditions are set, and the finite element solution method is used to calculate the stress distribution of the mold in the working state, and the stress concentration area and the stress concentration coefficient are obtained;

[0183] Extract stress characteristic parameters from the finite element analysis results, including the maximum stress value and the stress distribution uniformity index. Compare them with the stress characteristic parameters before optimization, calculate the reduction amplitude of the maximum stress value, and determine whether the preset threshold is reached. If not, return to continue the die design optimization iteration.

[0184] Take the calculated stress concentration factor as the input and substitute it into the pre-established fatigue life prediction model to predict the fatigue life of the die under the current stress level and obtain the increase in fatigue life.

[0185] Determine whether the increase in fatigue life reaches the preset percentage. If not, return to continue the die design optimization iteration until the increase in fatigue life is not less than 20% of the original life.

[0186] Evaluate the stress distribution uniformity of the optimized die through the stress distribution nephogram and the stress uniformity index. If the uniformity does not meet the requirements, for the stress concentration area, adopt the chamfering treatment method, set a circular arc with a certain radius at the sharp edges and corners to improve the stress distribution uniformity.

[0187] For the thin-walled part, by presetting ribs in the thin-walled area, increasing the cross-sectional area, reducing the stress concentration degree, and obtaining the optimized die design scheme.

[0188] Re-perform finite element analysis on the optimized die design scheme in step S6, calculate the stress distribution, extract the maximum stress value and the stress uniformity index, and determine whether it meets the preset threshold requirements.

[0189] If both the reduction amplitude of the maximum stress value and the increase in fatigue life reach the preset thresholds, and the stress distribution uniformity meets the requirements, then output the optimized die design scheme, determine the final die structure parameters, transfer the optimized die design scheme to the next link for die processing and manufacturing, and complete the die design optimization process.

[0190] Specifically, in step S7, first, according to the optimized die design scheme, construct a finite element analysis model in ANSYS software. Set the die material as Cr12MoV, the elastic modulus as 210 GPa, the Poisson's ratio as 0.3, apply the temperature load of 80 °C and the pressure load of 150 MPa during the die working process, use tetrahedral mesh division, with a mesh size of 2 mm. Through solution calculation, obtain the stress distribution nephogram of the die and the maximum stress value of 86 MPa, which is located at the fillet of the die cavity.

[0191] Extract the maximum stress values before and after optimization, calculate the reduction amplitude as (95 - 86) / 95 = 9.5%, which does not reach the preset threshold of 10%, and return to continue optimization.

[0192] Substitute the stress concentration factor of 1.8 into the Morrow fatigue life prediction model logσ = logC - 1 / m·logN. With material parameters C = 2030 and m = 9.18, the calculated fatigue life is 82,000 cycles, which is a 17.1% increase compared to 70,000 cycles before optimization. Since it does not reach the preset percentage of 20%, return to continue the optimization.

[0193] Through the analysis of the stress distribution contour map, the stress gradient is relatively large at the fillet of the concave die. A chamfer radius of R5mm is used for transition, and the stress value is relatively high at the thin-wall part, so a reinforcing rib with a preset thickness of 5mm is set.

[0194] Perform finite element analysis on the optimized die again. The maximum stress value is reduced to 78MPa, a decrease of 17.9%, the fatigue life is increased to 96,000 cycles, an increase of 37.1%, and the stress distribution uniformity is improved, meeting the preset requirements. Output the optimized die design scheme to complete the design optimization process.

[0195] As Figures 1 - 3 shown, in step S8, convert the optimized die design scheme in step S6 into a 3D CAD model, import it into the computer-aided manufacturing software for tool path planning and cutting parameter selection, simulate the NC machining process and estimate the machining time, generate the NC machining program, and upload it to the NC machine tool for machining.

[0196] Furthermore, in step S8, according to the optimized die design scheme, construct a 3D CAD model, and import the CAD model into the CAM software through the model conversion interface for subsequent machining processing;

[0197] In the CAM software, select cutting tool types, cutting parameters such as cutting depth, feed rate, and plan the tool movement path to generate the NC machining program;

[0198] Utilize the machining simulation function provided by the CAM software to simulate the NC machining process, analyze the interference between the tool and the workpiece and the cutting force situation, estimate the machining time, and optimize the cutting parameters and tool path;

[0199] Convert the generated NC machining program into a code format that can be recognized by the NC machine tool through the post-processor and upload it to the NC system;

[0200] The NC machine tool controls the tool movement according to the machining program instructions. During the machining process, dynamically adjust the cutting parameters by on-line monitoring of tool wear and workpiece dimensions;

[0201] After machining, use the in-machine probe to detect the key dimensions of the workpiece, compare the detected data with the CAD model for analysis, evaluate the machining accuracy and surface quality, and form a quality feedback report;

[0202] Optimize and adjust the numerical control machining program and cutting parameters according to the quality feedback report, and apply the optimized machining plan to subsequent batch production;

[0203] If the machining accuracy and surface quality do not meet the requirements, return to step 2, adjust the cutting parameters and tool path, regenerate the numerical control machining program until the machining quality meets the standard;

[0204] Save the optimized numerical control machining program and cutting parameters to the process database to provide reference for the numerical control machining of subsequent similar molds, and improve the consistency of production efficiency and machining quality.

[0205] Specifically, in S8, according to the optimized mold design scheme, use CATIA software to construct a 3D CAD model, and import the model into UG CAM software in STEP format;

[0206] In UG CAM, select a ball-end mill with a diameter of 10 mm, a spindle speed of 8000 rpm, a feed rate of 800 mm / min, a cutting depth of 0.5 mm, and a lateral cutting depth of 0.3 mm. When planning the tool path, adopt the equal-height Z-axis machining strategy to generate the numerical control machining G code;

[0207] Use the machining simulation function of UG CAM to simulate the numerical control machining process, calculate the tool cutting force through finite element analysis, optimize the cutting parameters, adjust the feed rate to 1000 mm / min, and estimate the machining time to be 2.5 hours; upload the G code generated by post-processing to the FANUC numerical control system, and the numerical control machine tool controls the tool movement according to the G code instructions. Use a Renishaw OMP60 probe to perform on-line monitoring of tool wear and workpiece dimensions. When the tool wear exceeds 0.2 mm, dynamically adjust the compensation parameters; after machining, use a Hexagon coordinate measuring machine to detect the key dimensions of the workpiece, and compare and analyze the measurement data with the CAD model through reverse engineering software Geomagic Control. The machining accuracy is 0.02 mm, and the surface roughness Ra is 0.8 μm, meeting the mold machining requirements; optimize the machining parameters and tool path, compensate for the dimensional error caused by tool wear, and apply it to batch production; save the optimized machining parameters and G code to the PLM system to realize the accumulation and reuse of process knowledge.

[0208] In the description of the present application, the technical features of the above embodiments can be combined arbitrarily. For the sake of brevity of description, not all possible combinations of the technical features in the above embodiments are described. However, as long as the combination of these technical features does not conflict, it should be considered as the scope recorded in this specification.

[0209] For those skilled in the art, various corresponding changes and deformations can be made according to the technical solutions and concepts described above, and all such changes and deformations should fall within the protection scope of the claims of this application.

Claims

1. A mold 3D simulation analysis method, characterized in that: The following steps are involved: S1. According to the three-dimensional CAD model of the mold, the mold is spatially discretized using tetrahedral meshes, mesh encryption is performed during mesh division of the sharp corners and thin-walled parts of the stress concentration area, the stress concentration area is selected as the key area, and the relative error of the stress value of the key area is calculated, wherein the calculation formula of the relative error is: |(σ_n-σ_n-1) / σ_n-1|, σ_n is the stress value under the current mesh density, and σ_n-1 is the stress value under the previous mesh density; S2, for the three-dimensional CAD model of the mold after meshing in step S1, input the elastic modulus and Poisson's ratio parameters of the mold material, apply the mold working load and direction, set the displacement constraint and contact conditions, and solve the displacement and stress components of each node of the mold through finite element analysis software; S3, selecting a color scale according to the yield strength of the mold material in step S2, generating a stress distribution cloud map, adjusting the viewing angle to highlight the distribution characteristics of the stress concentration area, and adding dimension annotations; S4, extracting features from the stress distribution cloud map of step S3, obtaining the numerical values ​​and spatial coordinates of the maximum principal stress value and the von Mises stress value, calculating the stress gradient direction and amplitude, and quantitatively evaluating the uniformity of the stress distribution; S5. Combined with the fatigue performance data of the mold material, a nonlinear prediction model of stress characteristic parameters and mold fatigue life is established by using a support vector machine regression algorithm. The maximum principal stress value, von Mises stress value and stress gradient amplitude in step S4 are selected as characteristic parameters, and fatigue life is used as label data. A cross-validation is performed by a grid search method to determine the kernel function type, regularization parameter and kernel function parameter of the support vector machine. S6, using the nonlinear prediction model in step S5, determine whether the stress concentration of the current mold design exceeds the material fatigue strength limit. If it exceeds the material fatigue strength limit, chamfer the sharp corners in the mold structure and strengthen the thin-walled parts. If it does not exceed the material fatigue strength limit, no treatment is performed to obtain an optimized mold design; S7, re-performing finite element analysis on the mold design optimized in step S6, extracting stress characteristic parameters, calculating the maximum stress value reduction and the degree of improvement in the uniformity of stress distribution, so as to predict the increase in fatigue life; S8. Convert the optimized mold design in step S6 into a three-dimensional CAD model, import it into computer-aided manufacturing software for tool path planning and cutting parameter selection, simulate the CNC machining process and estimate the machining time, generate a CNC machining program, and upload it to the CNC machine tool for machining.

2. A mold 3D simulation analysis method according to claim 1, characterized in that: In the step S1, the three-dimensional CAD model data of the mold is obtained, the model is preprocessed, and the stress concentration area is determined; The stress concentration area includes sharp corners and thin-walled parts. The stress concentration area is marked as a key area, an initial mesh model is obtained, and mesh encryption is performed on the key area; Calculating the stress value of each unit under the current grid density after mesh encryption to obtain a stress distribution cloud map of the entire mold; The relative error between the stress value at the current grid density and the stress value at the previous grid density is calculated based on the cells in the key area in the stress distribution cloud map.

3. The mold 3D simulation analysis method according to claim 1, characterized in that: In the step S2, the three-dimensional CAD model of the mold is discretized by using a meshing method to generate a finite element model, and the mechanical property parameters of the mold material are obtained. The mechanical property parameters include elastic modulus and Poisson's ratio. The mechanical property parameters are input into finite element analysis software, and then the material properties are input into each unit of the finite element model. Applying the magnitude and direction of the external load as boundary conditions to the finite element model, setting displacement constraints of the mold, the displacement constraints including limiting the movement of the degrees of freedom, and establishing contact pairs in the finite element model; The solution module of the finite element analysis software is called to perform force analysis on the finite element model to obtain the deformation field and stress field distribution results of the mold under the working load, and extract the displacement component and stress component data of each node of the mold.

4. The mold 3D simulation analysis method according to claim 1, characterized in that: In the step S3, according to the yield strength of the mold material in step S2, a mapping relationship between the color scale and the stress value is established for generating a stress distribution cloud map, and the stress distribution of the mold under the stress state is calculated by a finite element analysis method to obtain stress value data of each area; Generate the stress distribution cloud map according to the stress value data and the color scale mapping relationship; Performing image processing on the generated stress distribution cloud map, and adding dimension annotations to the stress distribution cloud map to annotate dimension information of key areas; Using an image segmentation algorithm to divide the stress distribution cloud map into regions to identify the location and range of the stress concentration region; After the area division, the stress value data of the stress concentration area is extracted to obtain the characteristic parameters of the stress concentration degree and distribution.

5. The mold 3D simulation analysis method according to claim 1, characterized in that: In step S4, an image processing algorithm is used to extract features from the cloud image to obtain a digital representation of the stress distribution; The extracted digital representation of the stress distribution is used to determine the numerical values ​​and spatial coordinates corresponding to the maximum principal stress value and the von Mises stress value; According to the spatial coordinates of the maximum principal stress value and the von Mises stress value, the gradient direction and gradient amplitude of the stress in space are calculated to determine the stress distribution.

6. The mold 3D simulation analysis method according to claim 1, characterized in that: In the step S5, fatigue performance data of the mold material in step S4 is obtained, wherein the fatigue performance data includes a maximum principal stress value, a von Mises stress value, stress characteristic parameters of a stress gradient amplitude, and corresponding fatigue life data; Preprocessing the acquired stress characteristic parameters and fatigue life data; Based on the preprocessed data, a nonlinear prediction model between stress characteristic parameters and fatigue life is established using a support vector machine regression algorithm, and cross-validation is performed to evaluate the prediction performance of the model; Determine optimal support vector machine model parameters according to the cross-validation result to obtain an optimized support vector machine model, wherein the optimal support vector machine model parameters include a kernel function type, a regularization parameter, and a kernel function parameter; Using the optimized support vector machine model, the stress characteristic parameters of the new mold material are predicted to obtain the corresponding fatigue life prediction value; The predicted fatigue life is compared with the actual test results to evaluate the prediction accuracy of the model, and the stress distribution of the mold under working conditions is calculated to obtain the stress concentration area and stress concentration factor; The calculated stress concentration factor is used as input and substituted into the nonlinear prediction model to determine the degree of stress concentration.

7. The mold 3D simulation analysis method according to claim 1, characterized in that: In step S6, the stress distribution of the mold under the working state is calculated by the finite element analysis method to obtain the stress concentration area and the stress concentration coefficient. The stress concentration coefficient is substituted into the fatigue life prediction model to predict the fatigue life of the mold, and the predicted mold fatigue life is compared with the material fatigue strength limit value to determine whether structural optimization is needed.

8. The mold 3D simulation analysis method according to claim 1, characterized in that: In the step S7, according to the mold design scheme optimized in step S6, a finite element analysis model is constructed, and the stress distribution of the mold in the working state is calculated by a finite element solution method to obtain the stress concentration area and the stress concentration coefficient; Extracting stress characteristic parameters from the results of the finite element analysis, and comparing them with the stress characteristic parameters before optimization, calculating the reduction amplitude of the maximum stress value, and determining whether a preset threshold value has been reached; The stress concentration factor obtained by calculating the reduction amplitude of the maximum stress value is used as input and substituted into the nonlinear prediction model to predict the fatigue life of the mold under the current stress level and obtain the fatigue life increase, which is used to determine whether the fatigue life increase reaches a preset percentage; Evaluate the stress distribution uniformity of the optimized mold through stress distribution cloud diagram and stress uniformity index; The optimized mold design scheme is re-analyzed by finite element analysis to calculate the stress distribution, extract the maximum stress value and stress uniformity index, and determine whether the preset threshold requirements are met.

9. The mold 3D simulation analysis method according to claim 1, characterized in that: In the step S8, a three-dimensional CAD model is constructed according to the mold design scheme optimized in step S6, and the CAD model is imported into the CAM software through the model conversion interface for processing; In the CAM software, select the cutting parameters of tool type, cutting amount, and feed speed, plan the tool movement path, and generate the CNC machining program; The machining simulation function provided by CAM software is used to simulate the CNC machining process, analyze the interference between the tool and the workpiece and the cutting force, estimate the machining time, and optimize the cutting parameters and tool path.

10. A mold 3D simulation analysis method according to claim 9, characterized in that: The CNC machining program is converted into a G-code format recognized by a CNC machine tool by a post-processor and uploaded to a FANUC CNC system; The CNC machine tool controls the movement of the tool according to the machining program instructions in the G-code format, and dynamically adjusts the cutting parameters by online monitoring of tool wear and workpiece size; After machining is completed, the key dimensions of the workpiece are inspected using an on-machine probe, and the inspection data is compared and analyzed with the CAD model to evaluate the machining accuracy and surface quality and obtain a quality feedback report.

Images