A data-driven design method, system, device, and medium for heat treating a part

Abstract

The application discloses a data-driven design method, system, device and medium for heat treatment parts, and relates to the field of heat treatment design; the method comprises the following steps: acquiring parameter information of a heat treatment part; constructing a finite element analysis model based on material parameters; determining boundary conditions and a design space of heat treatment according to the parameter information and the finite element analysis model; determining a response relationship between heat treatment process parameters and target physical quantities of the finite element analysis model by using a surrogate model based on the boundary conditions and the design space; determining optimal heat treatment process parameters by using a target optimization method and taking the design space as a constraint condition; and determining a heat treatment data-driven model according to the optimal heat treatment process parameters and the finite element analysis model, so as to analyze, simulate and predict a heat treatment process; and the application can quickly realize simulation and prediction of the heat treatment process.

Description

Technical Field

[0001] This invention relates to the field of heat treatment design, and in particular to a data-driven design method, system, device and medium for heat-treated parts. Background Technology

[0002] Heat treatment, as a key process in manufacturing, plays a crucial role in improving material properties, extending component lifespan, and enhancing product quality. However, traditional experience-based heat treatment design methods are increasingly unable to meet the demands of increasingly complex applications. These methods, often based on trial and error and experience, struggle to accurately predict the microstructural evolution, residual stress distribution, dimensional deformation, and property changes of complex materials under different heat treatment conditions. Therefore, a more precise and efficient heat treatment process design method is urgently needed to address this challenge. Summary of the Invention

[0003] The purpose of this invention is to provide a data-driven design method, system, device, and medium for heat-treated parts, which can quickly realize the simulation and prediction of heat treatment processes.

[0004] To achieve the above objectives, the present invention provides the following solution:

[0005] A data-driven design method for heat-treated parts, the method comprising:

[0006] Obtain parameter information of heat-treated parts; the parameter information includes: material parameters and heat treatment process parameters; the material parameters include: thermophysical property parameters, mechanical property parameters and phase transformation kinetic parameters; the heat treatment process parameters include: heating temperature, heating time, cooling medium, medium stirring rate and tooling fixture dimensions;

[0007] Based on the material parameters, a finite element analysis model is constructed. The finite element analysis model is obtained by using modeling software to perform geometric modeling on the heat-treated part, obtaining a geometric model, and then meshing the geometric model.

[0008] Based on the parameter information and the finite element analysis model, the boundary conditions and design space of the heat treatment are determined; the boundary conditions include: heat transfer coefficient and ambient temperature; the design space is the value range corresponding to the heat treatment process parameters.

[0009] Based on the boundary conditions and the design space, a surrogate model is used to determine the response relationship between the heat treatment process parameters and the target physical quantities of the finite element analysis model; the target physical quantities include: microstructure, residual stress, dimensional deformation, and property distribution.

[0010] The optimal heat treatment process parameters are determined by using the design space as a constraint, based on the target optimization method. The optimal heat treatment process parameters are the heat treatment process parameters corresponding to the minimum distribution value of the target physical quantity, based on the response relationship.

[0011] The heat treatment data-driven model is determined based on the optimal heat treatment process parameters and the finite element analysis model; the heat treatment data-driven model is used to analyze, simulate, and predict the heat treatment process.

[0012] Optionally, based on the boundary conditions and the design space, a surrogate model is used to determine the response relationship between the heat treatment process parameters and the target physical quantities of the finite element analysis model, specifically including:

[0013] A sample point set is determined; the sample point set is a collection of multiple sample points determined according to the design space; each sample point corresponds to a parameter design combination; the parameter design combination is obtained by randomly combining the heat treatment process parameters.

[0014] Based on the boundary conditions and the sample point set, the simulation results of the target physical quantity are determined by numerical simulation.

[0015] Based on the simulation results, a surrogate model is used to determine the response relationship between the heat treatment process parameters and the target physical quantity.

[0016] Optionally, the sample point set can be determined using the Latin hypercube method or the orthogonal experimental method.

[0017] Optionally, the surrogate model is a multinomial model, radial basis function, Kriging model, or neural network model.

[0018] Optionally, the functional expression of the polynomial model is:

[0019]

[0020] Where y(X) is a function of the polynomial model; x i and x j For the i-th and j-th heat treatment process parameters; β0, β ii β i β i All are undetermined coefficients of a quadratic polynomial; ε is the error between the fitted value and the actual value; N is the total number of heat treatment process parameters.

[0021] Optionally, an objective optimization method is employed, using the design space as a constraint, to determine the optimal heat treatment process parameters, specifically including:

[0022] For any sample point in the sample point set, the first difference is determined based on the response relationship and the set target value, using the design space as a constraint.

[0023] For any sample point in the sample point set, the function value is determined based on the response relationship and the set variable values, using the design space as a constraint.

[0024] For any of the function values, a second difference is determined based on the function value and the set target value;

[0025] Determine the target function value; the target function value is the distribution value of the target physical quantity that has the smallest value after comparing all the first differences and all the second differences.

[0026] The optimal heat treatment process parameters are determined by the heat treatment process parameters corresponding to the objective function value.

[0027] A data-driven design system for heat-treated parts, the system comprising:

[0028] The parameter information acquisition module is used to acquire parameter information of heat-treated parts; the parameter information includes: material parameters and heat treatment process parameters; the material parameters include: thermophysical property parameters, mechanical property parameters and phase transformation kinetic parameters; the heat treatment process parameters include: heating temperature, heating time, cooling medium, medium stirring rate and tooling fixture dimensions;

[0029] The model building module is used to construct a finite element analysis model based on the material parameters. The finite element analysis model is obtained by using modeling software to perform geometric modeling on the heat-treated part, obtaining a geometric model, and then meshing the geometric model.

[0030] The determination module is used to determine the boundary conditions and design space of the heat treatment based on the parameter information and the finite element analysis model; the boundary conditions include: heat transfer coefficient and ambient temperature; the design space is the value range corresponding to the heat treatment process parameters;

[0031] The response module is used to determine the response relationship between the heat treatment process parameters and the target physical quantities of the finite element analysis model based on the boundary conditions and the design space, using a surrogate model; the target physical quantities include: microstructure, residual stress, dimensional deformation, and property distribution.

[0032] The optimization module is used to determine the optimal heat treatment process parameters by adopting the target optimization method and taking the design space as a constraint. The optimal heat treatment process parameters are the heat treatment process parameters corresponding to the minimum distribution value of the target physical quantity based on the response relationship.

[0033] The processing module is used to determine the heat treatment data-driven model based on the optimal heat treatment process parameters and the finite element analysis model; the heat treatment data-driven model is used to analyze, simulate and predict the heat treatment process.

[0034] An electronic device includes a memory and a processor, the memory storing a computer program, and the processor running the computer program to cause the electronic device to perform the data-driven design method for heat-treated parts described above.

[0035] A computer-readable storage medium storing a computer program that, when executed by a processor, implements the data-driven design method for heat-treated parts described above.

[0036] According to specific embodiments provided by the present invention, the present invention discloses the following technical effects:

[0037] This invention provides a data-driven design method, system, device, and medium for heat-treated parts. It acquires parameter information of the heat-treated parts; constructs a finite element analysis model based on material parameters; determines the boundary conditions and design space of the heat treatment based on the parameter information and the finite element analysis model; uses a surrogate model to determine the response relationship between the heat treatment process parameters and the target physical quantities of the finite element analysis model based on the boundary conditions and the design space; uses an objective optimization method with the design space as a constraint to determine the optimal heat treatment process parameters; and determines a data-driven heat treatment model based on the optimal heat treatment process parameters and the finite element analysis model to analyze, simulate, and predict the heat treatment process. Therefore, this invention can quickly achieve the simulation and prediction of heat treatment processes. Attached Figure Description

[0038] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0039] Figure 1 A flowchart of a data-driven design method for heat-treated parts provided in an embodiment of the present invention;

[0040] Figure 2 A structural diagram of a data-driven design system for heat-treated parts provided in an embodiment of the present invention.

[0041] Symbol explanation:

[0042] Parameter information acquisition module-1, model construction module-2, determination module-3, response module-4, optimization module-5, processing module-6. Detailed Implementation

[0043] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0044] To overcome the limitations of traditional methods, this invention integrates advanced data analysis techniques with materials science expertise. Through in-depth analysis of vast amounts of heat treatment and simulation data, it provides a new paradigm for the design and optimization of heat treatment processes. The core idea of ​​the data-driven approach lies in using machine learning and artificial intelligence technologies to build surrogate models from large datasets, thereby enabling real-time, rapid, and automatic optimization of the heat treatment target, achieving accurate prediction and intelligent design of the heat treatment process. This invention provides a data-driven design method, system, equipment, and medium for heat-treated parts, which boasts advantages such as high computational speed, high intelligence, and low cost, offering a new paradigm for heat treatment process design.

[0045] The purpose of this invention is to provide a data-driven design method, system, device, and medium for heat-treated parts, which can quickly realize the simulation and prediction of heat treatment processes.

[0046] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0047] Example 1

[0048] like Figure 1 As shown, this embodiment of the invention provides a data-driven design method for heat-treated parts, the method comprising:

[0049] Step 100: Obtain parameter information for the heat-treated part. This parameter information includes: material parameters and heat treatment process parameters; material parameters include: thermophysical properties, mechanical properties, and phase transformation kinetics; heat treatment process parameters include: heating temperature, heating time, cooling medium, medium stirring rate, and tooling fixture dimensions.

[0050] Step 200: Construct a finite element analysis model based on material parameters. The finite element analysis model is obtained by using modeling software to geometrically model the heat-treated part, and then meshing the geometric model.

[0051] Step 300: Based on the parameter information and the finite element analysis model, determine the boundary conditions and design space for the heat treatment. The boundary conditions include the heat transfer coefficient and ambient temperature; the design space is the range of values ​​corresponding to the heat treatment process parameters.

[0052] Step 400: Based on boundary conditions and design space, a surrogate model is used to determine the response relationship between the heat treatment process parameters and the target physical quantities of the finite element analysis model. The target physical quantities include: microstructure, residual stress, dimensional deformation, and property distribution.

[0053] Based on boundary conditions and design space, a surrogate model is used to determine the response relationship between heat treatment process parameters and target physical quantities in the finite element analysis model, specifically including:

[0054] The sample point set is determined; the sample point set is a collection of multiple sample points determined based on the design space; each sample point corresponds to a parameter design combination; the parameter design combination is obtained by randomly combining heat treatment process parameters. Specifically, the sample point set is determined using the Latin hypercube method or orthogonal experimental method.

[0055] Based on the boundary conditions and sample point set, numerical simulation methods are used to determine the simulation results of the target physical quantity.

[0056] Based on the simulation results, a surrogate model is used to determine the response relationship between heat treatment process parameters and target physical quantities.

[0057] The surrogate model can be a multinomial model, radial basis function, Kriging model, or neural network model.

[0058] The functional expression of the polynomial model is:

[0059]

[0060] Where y(X) is a function of the polynomial model; x i and x j For the i-th and j-th heat treatment process parameters; β0, β ii β i β i All are undetermined coefficients of a quadratic polynomial; ε is the error between the fitted value and the actual value; N is the total number of heat treatment process parameters.

[0061] Step 500: Using an objective optimization method, and with the design space as a constraint, determine the optimal heat treatment process parameters. The optimal heat treatment process parameters are those that minimize the distribution value of the objective physical quantity based on the response relationship.

[0062] Using an objective optimization method with design space as a constraint, the optimal heat treatment process parameters are determined, specifically including:

[0063] For any sample point in the sample point set, the first difference is determined based on the response relationship and the set target value, with the design space as a constraint.

[0064] For any sample point in the sample point set, the function value is determined based on the response relationship and the set variable values, with the design space as the constraint.

[0065] For any function value, determine the second difference based on the function value and the set target value.

[0066] Determine the objective function value; the objective function value is the distribution value of the objective physical quantity that is the smallest after comparing all the first differences and all the second differences.

[0067] The optimal heat treatment process parameters are determined by identifying the heat treatment process parameters corresponding to the objective function value.

[0068] Step 600: Determine the heat treatment data-driven model based on the optimal heat treatment process parameters and the finite element analysis model. The heat treatment data-driven model is used to analyze, simulate, and predict the heat treatment process.

[0069] Example 2

[0070] like Figure 2 As shown, this embodiment of the invention provides a data-driven design system for heat-treated parts. The system includes: a parameter information acquisition module 1, a model construction module 2, a determination module 3, a response module 4, an optimization module 5, and a processing module 6.

[0071] The parameter information acquisition module 1 is used to acquire parameter information of heat-treated parts. The parameter information includes: material parameters and heat treatment process parameters. The material parameters include: thermophysical property parameters, mechanical property parameters and phase transformation kinetic parameters. The heat treatment process parameters include: heating temperature, heating time, cooling medium, medium stirring rate and tooling fixture dimensions.

[0072] Model building module 2 is used to build a finite element analysis model based on material parameters. The finite element analysis model is obtained by using modeling software to perform geometric modeling on the heat-treated parts, obtaining a geometric model, and then meshing the geometric model.

[0073] Module 3 is used to determine the boundary conditions and design space of heat treatment based on parameter information and finite element analysis model. The boundary conditions include heat transfer coefficient and ambient temperature. The design space is the range of values ​​corresponding to the heat treatment process parameters.

[0074] Response module 4 is used to determine the response relationship between heat treatment process parameters and target physical quantities of the finite element analysis model based on boundary conditions and design space using a surrogate model; the target physical quantities include: microstructure, residual stress, dimensional deformation and property distribution.

[0075] The optimization module 5 is used to determine the optimal heat treatment process parameters by using the target optimization method with the design space as a constraint. The optimal heat treatment process parameters are the heat treatment process parameters corresponding to the minimum distribution value of the target physical quantity based on the response relationship.

[0076] Processing module 6 is used to determine the heat treatment data-driven model based on the optimal heat treatment process parameters and the finite element analysis model; the heat treatment data-driven model is used to analyze, simulate and predict the heat treatment process.

[0077] Example 3

[0078] This invention provides an electronic device, including a memory and a processor. The memory stores a computer program, and the processor runs the computer program to enable the electronic device to execute the data-driven design method for heat-treated parts in Embodiment 1.

[0079] As an alternative implementation, the present invention also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the data-driven design method for heat-treated parts in Embodiment 1.

[0080] In practical applications, the specific operational steps of the data-driven design method for heat-treated parts are as follows:

[0081] Step 1: Establish a finite element analysis model of the heat treatment process of the part, set the material parameters and boundary conditions of the model, carry out finite element simulation calculations of heat treatment, and conduct experiments to optimize the model and parameters, so as to achieve accurate simulation and prediction of physical quantities such as microstructure, residual stress, dimensional deformation, and performance distribution of the part after heat treatment.

[0082] Step 1.1: The finite element analysis model of the heat treatment process includes the geometric model and mesh model of the heat-treated part; specifically, the geometric model is first established using CAD modeling software, and then the geometric model is meshed.

[0083] Step 1.2: Set the material parameters and boundary conditions of the model. The basic material parameters include thermophysical properties, mechanical properties, and phase transformation kinetics. The boundary conditions are constraints and heat transfer boundary conditions set based on the heat treatment process parameters before optimization.

[0084] Specifically, setting boundary conditions refers to the conditions on the surface boundary of a part, mainly heat transfer boundary conditions, which are the heat transfer coefficient between the part and the environment and the ambient temperature.

[0085] For example, if the heat treatment heating temperature is 1000℃, then the ambient temperature in the heat exchange boundary is set to 1000℃.

[0086] Step 1.3: Conducting finite element simulation calculations for heat treatment and performing experiments to optimize the model and parameters refers to performing finite element simulations of the heat treatment process before optimization and conducting corresponding experiments to verify the results of the simulation calculations, such as microstructure, residual stress, dimensional deformation, and property distribution. Based on the results, the relevant parameters in the model, material parameters, and boundary conditions are further optimized.

[0087] Step 2: Determine the design variables and design space of the heat treatment process parameters. The design variables are the heat treatment process parameters to be optimized, which are usually heating temperature, heating time, cooling medium, medium stirring rate, tooling fixture size, etc. The design space is the range of values ​​for the design variables.

[0088] Step 2.1: The design variables for heat treatment process parameters refer to the heat treatment process parameters to be optimized, typically including heating temperature, heating time, cooling medium, medium stirring rate, and tooling fixture dimensions. The design space refers to the range of values ​​for the above heat treatment process parameters. For example, if the heating temperature is optimized within the range of 800-1000℃, then the heating temperature is a design variable, and 800-1000℃ is the design space corresponding to this design variable.

[0089] Step 3: Determine sample points in the design space established in Step 2 based on the Latin hypercube method or orthogonal experimental design. Each sample point corresponds to a set of heat treatment process parameter design combinations.

[0090] Step 3.1: A sample point refers to a heat treatment process. For example, if the process to be optimized is heating temperature and heating time, then multiple values ​​are taken and combined within the design space of these two design variables. That is, sampling is performed using the Latin hypercube method or orthogonal experimental method, and each combination is a sample point.

[0091] Step 4: Perform numerical simulation based on the sample points from Step 3 to obtain physical quantities such as microstructure, residual stress, dimensional deformation, and performance distribution at all nodes of the model under different design combinations.

[0092] Step 4.1: Numerical simulation based on the sample points in Step 3 means setting the parameters of the design variables in the sample points in Step 3 to the corresponding parameters on the model boundary conditions in Step 1. For example, if 30 sample points are designed in Step 3, each sample point has a different combination of heating temperature and heating time, then 30 sets of heat transfer boundary conditions and heating times are set in Step 1, and numerical simulations are performed separately to obtain 30 sets of simulation results.

[0093] Step 5: Determine the target physical quantity, i.e., which physical quantity to optimize. This physical quantity is usually microstructure, residual stress, dimensional deformation, property distribution, etc. Based on the simulation results of sample points, a surrogate model is used to establish the response relationship between the process parameters to be optimized and the target physical quantity at all nodes of the model. The surrogate model here includes polynomial models, radial basis functions, Kriging models, neural network models, etc.

[0094] Step 5.1: Establishing the response relationship between the process parameters to be optimized and the target physical quantities at all nodes of the model using a surrogate model refers to establishing the correspondence between design variables and results based on the simulation results of all sample points in Step 4. For example, a polynomial model can be used for modeling.

[0095] After obtaining simulation results for 30 different combinations of heating temperature and heating time using numerical simulation, the 30 sets of sample parameters and nodal target physical quantity data were used as training data to fit a multinomial model. This resulted in different β0 and β values ​​at different nodes. ii β ij β i ε. This establishes the response relationship between the process parameters to be optimized and the physical quantities at all nodes of the model.

[0096] Step 6: Establish a visualization simulation platform. Import the finite element analysis model from Step 1 into the visualization platform. Set the design variable inputs in the visualization platform, calculate the data of all nodes based on the surrogate model from Step 5, and assign different colors to different nodes of the model in the visualization platform according to the numerical values, thereby realizing the three-dimensional visualization of the finite element analysis model and the target physical quantity.

[0097] Step 7: Establish the target optimization system. Let the distribution of the target physical quantities of the part after heat treatment be y, and the optimization target be y0. Then the objective function formula is F = min(y - y0). Using the adjustable range of the process parameter design variables as constraints, solve the optimal problem by: randomly generating a set of process parameter combinations; using the surrogate model from Step 5 to calculate the physical field distribution at the nodes, and then calculating the objective function y1; adding a small amount δ (5%-10% of the variable value) to the initial point and substituting it into the surrogate model to calculate the corresponding objective function y2; if |y2 - y0| < |y1 - y0|, continue repeating the above steps; the process parameter combination corresponding to the optimal objective function y2 obtained after traversing all process parameter combinations is the optimal process parameter combination.

[0098] Existing heat treatment process optimization methods are trial-and-error design approaches based on experience or numerical simulation. This involves setting different process parameters to obtain multiple results and then selecting the optimal parameters. This method is costly, time-consuming, and inefficient, and it cannot definitively determine if the final process parameters represent the optimal solution. This invention establishes a surrogate model based on high-throughput numerical simulation results and constructs an automatic optimization method for heat treatment objectives. It is a data-driven heat treatment process design method with advantages such as fast computation speed, high intelligence, and low cost.

[0099] The high computational speed stems from the surrogate model established in step 5. Traditional methods for performing a single heat treatment finite element numerical simulation require solving the thermo-elastic-plastic constitutive equations, resulting in low efficiency. However, the surrogate model can rapidly calculate the response to the input parameters; it is a data-driven model with a computational response speed in the millisecond range.

[0100] The high level of intelligence stems from the establishment of the visualization platform and the automatic optimization algorithm in steps 6 and 7. The visualization platform can intuitively display the physical field results under different inputs, and the automatic optimization algorithm can automatically traverse all parameter combinations and quickly calculate the objective function, thereby determining the optimal process parameters.

[0101] The low cost is based on obtaining training samples for the surrogate model through numerical simulation in steps 1-4, which has the advantage of low cost compared to high-throughput experimental trial and error.

[0102] Example 4

[0103] When the surrogate model uses radial basis functions, this invention is illustrated using the die-pressing and quenching heat treatment process design of a spiral bevel gear as an example. The spiral bevel gear employs an 815℃ die-pressing and quenching process. The process parameters need to be optimized to reduce the axial and tooth profile cone angle deformation after die-pressing and quenching. The heat treatment process parameters to be optimized include the die pressure angle x1, the combined dimension x2, the expansion ring pressure x3, and the inner ring pressure x4. Therefore, in this embodiment, the design variables are x1, x2, x3, and x4; the target physical quantity is the dimensional deformation.

[0104] Step 1: Establish a finite element analysis model of the heat treatment process of spiral bevel gears, set a set of parameters x1, x2, x3, and x4, carry out numerical simulation calculations of die-forming and quenching, and conduct corresponding experiments to verify the simulation results, and simulate the dimensional deformation of the parts after heat treatment.

[0105] Step 2: Design variables are mold pressure angle x1, combined dimension x2, expansion ring pressure x3, and inner ring pressure x4. Confirm the design space for each of the four parameters.

[0106] Step 3: Design 50 sets of design variable combinations based on the Latin hypersolution method, which will serve as 50 sample points to be calculated.

[0107] Step 4: Using the 50 sets of design variable combinations obtained in Step 3 as input parameters, conduct numerical simulations to obtain deformation data of all nodes in the x, y, and z directions under 50 different design combinations. The deformation of the model is magnified by 50 times.

[0108] Step 5: Using the 50 sets of design variable combinations of the sample as input and the deformation data of all nodes in the x, y, and z directions as output, the surrogate model is trained using the radial basis function model to establish the surrogate model for input and output.

[0109] Step 6: Establish a visualization simulation platform based on the VTK image library and Qt interface design library. This platform can input four design variables and display the deformation results as contour plots. To make the deformation results more obvious, the model deformation is magnified 50 times for display, and the contour plot data is displayed according to the actual deformation values.

[0110] Step 7: Using the minimization of axial deformation and tooth profile cone angle deformation of the part as the optimization objective, and the design space of four design variables as constraints, the optimal problem is solved. First, a set of design variable combinations is randomly generated, and then the axial deformation data and tooth profile cone angle data are calculated. Then, the four design variables are continuously adjusted so that they iterate through all design combinations at 5% of their values, resulting in a total of 194,481 combinations. Since it is difficult to simultaneously minimize the axial deformation and tooth profile cone angle deformation of the part under the same set of design variables, the optimal objective range is appropriately increased to obtain the best optimal combination.

[0111] The time required to calculate nodal deformation data under any combination of design variables using the surrogate model is approximately 1e-4 seconds, and the calculation time required to traverse the deformation amount under 194,481 combinations is approximately 20 seconds. Compared with the traditional method for designing the die-cutting and quenching process, the heat treatment deformation amount is reduced by 60%, and the development time is reduced from tens of days to several hours, which greatly improves the efficiency of heat treatment process development.

[0112] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the systems disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the descriptions are relatively simple; relevant parts can be referred to the method section.

[0113] This document uses specific examples to illustrate the principles and implementation methods of the present invention. The descriptions of the above embodiments are only for the purpose of helping to understand the method and core ideas of the present invention. Furthermore, those skilled in the art will recognize that, based on the ideas of the present invention, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of the present invention.

Claims

1. A data-driven analysis method for heat-treated parts, characterized in that, The method includes: Obtain parameter information of heat-treated parts; the parameter information includes: material parameters and heat treatment process parameters; the material parameters include: thermophysical property parameters, mechanical property parameters and phase transformation kinetic parameters; the heat treatment process parameters include: heating temperature, heating time, cooling medium, medium stirring rate and tooling fixture dimensions; Based on the material parameters, a finite element analysis model is constructed. The finite element analysis model is obtained by using modeling software to perform geometric modeling on the heat-treated part, obtaining a geometric model, and then meshing the geometric model. Based on the parameter information and the finite element analysis model, the boundary conditions and design space of the heat treatment are determined; the boundary conditions include: heat transfer coefficient and ambient temperature; the design space is the value range corresponding to the heat treatment process parameters. Based on the boundary conditions and the design space, a surrogate model is used to determine the response relationship between the heat treatment process parameters and the target physical quantities of the finite element analysis model; the target physical quantities include: microstructure, residual stress, dimensional deformation, and property distribution. The optimal heat treatment process parameters are determined by using the design space as a constraint, based on the target optimization method. The optimal heat treatment process parameters are the heat treatment process parameters corresponding to the minimum distribution value of the target physical quantity, based on the response relationship. A heat treatment data-driven model is determined based on the optimal heat treatment process parameters and the finite element analysis model; the heat treatment data-driven model is used to analyze, simulate, and predict the heat treatment process. Using an objective optimization method, with the design space as a constraint, the optimal heat treatment process parameters are determined, specifically including: For any sample point in the sample point set, the first difference is determined based on the design space as a constraint, the response relationship, and the set target value. For any sample point in the sample point set, the function value is determined based on the response relationship and the set variable values, using the design space as a constraint. For any of the function values, a second difference is determined based on the function value and the set target value; Determine the target function value; the target function value is the distribution value of the target physical quantity that has the smallest value after comparing all the first differences and all the second differences. The optimal heat treatment process parameters are determined by the heat treatment process parameters corresponding to the objective function value.

2. The data-driven analysis method for heat-treated parts according to claim 1, characterized in that, Based on the boundary conditions and the design space, a surrogate model is used to determine the response relationship between the heat treatment process parameters and the target physical quantities of the finite element analysis model, specifically including: A sample point set is determined; the sample point set is a collection of multiple sample points determined according to the design space; each sample point corresponds to a parameter design combination; the parameter design combination is obtained by randomly combining the heat treatment process parameters. Based on the boundary conditions and the sample point set, the simulation results of the target physical quantity are determined by numerical simulation. Based on the simulation results, a surrogate model is used to determine the response relationship between the heat treatment process parameters and the target physical quantity.

3. The data-driven analysis method for heat-treated parts according to claim 2, characterized in that, The sample point set is determined using the Latin hypercube method or orthogonal experimental method.

4. The data-driven analysis method for heat-treated parts according to claim 2, characterized in that, The surrogate model can be a multinomial model, radial basis function, Kriging model, or neural network model.

5. The data-driven analysis method for heat-treated parts according to claim 4, characterized in that, The functional expression of the polynomial model is: ； in, y ( X ) is a function of a polynomial model; x i and x j For the first i and j One heat treatment process parameter; β 0 , β ii , β ij , β i All are undetermined coefficients of a quadratic polynomial; ε This represents the error between the fitted value and the actual value. N This represents the total number of heat treatment process parameters.

6. A data-driven analysis system for heat-treated parts, characterized in that, The system employs the data-driven analysis method for heat-treated parts according to any one of claims 1 to 5, wherein the system comprises: The parameter information acquisition module is used to acquire parameter information of heat-treated parts; the parameter information includes: material parameters and heat treatment process parameters; the material parameters include: thermophysical property parameters, mechanical property parameters and phase transformation kinetic parameters; the heat treatment process parameters include: heating temperature, heating time, cooling medium, medium stirring rate and tooling fixture dimensions; The model building module is used to construct a finite element analysis model based on the material parameters. The finite element analysis model is obtained by using modeling software to perform geometric modeling on the heat-treated part, obtaining a geometric model, and then meshing the geometric model. The determination module is used to determine the boundary conditions and design space of the heat treatment based on the parameter information and the finite element analysis model; the boundary conditions include: heat transfer coefficient and ambient temperature; the design space is the value range corresponding to the heat treatment process parameters; The response module is used to determine the response relationship between the heat treatment process parameters and the target physical quantities of the finite element analysis model based on the boundary conditions and the design space, using a surrogate model; the target physical quantities include: microstructure, residual stress, dimensional deformation, and property distribution. The optimization module is used to determine the optimal heat treatment process parameters by adopting the target optimization method and taking the design space as a constraint. The optimal heat treatment process parameters are the heat treatment process parameters corresponding to the minimum distribution value of the target physical quantity based on the response relationship. The processing module is used to determine the heat treatment data-driven model based on the optimal heat treatment process parameters and the finite element analysis model; the heat treatment data-driven model is used to analyze, simulate and predict the heat treatment process.

7. An electronic device, characterized in that, The device includes a memory and a processor, the memory being used to store a computer program, and the processor running the computer program to cause the electronic device to perform the data-driven analysis method for heat-treated parts as described in any one of claims 1 to 5.

8. A computer-readable storage medium, characterized in that, It stores a computer program that, when executed by a processor, implements the data-driven analysis method for heat-treated parts as described in any one of claims 1 to 5.

Images