Multi-parameter and multi-field intelligent optimization method and system for press free forging large-scale die casting

Abstract

The invention relates to a multi-parameter and multi-field intelligent optimization method and system for a press free forging large-scale falling die, and the method achieves the collaborative optimization of technological parameters, die geometry and microstructure through a heat-force-microstructure three-field coupling modeling, a deep kernel learning agent model and a digital twinning guided multi-objective evolutionary algorithm, improves the optimization efficiency, and improves the optimization precision. The mold testing times are reduced; and meanwhile, a real-time closed-loop control system is constructed, the grain size uniformity of forgings is improved, the forming load is reduced, the production period is shortened, the process stability is improved, and the die service life is prolonged.

Description

Technical Field

[0001] This invention relates to the field of design of large-scale die for free forging presses, and in particular to a multi-parameter, multi-field intelligent optimization method and system for large-scale die for free forging presses. Background Technology

[0002] With the rapid development of high-end equipment manufacturing, large forgings are increasingly widely used in nuclear power, aerospace and other fields. Press free forging, as a core forming process, directly affects the performance and production efficiency of forgings due to the design quality of its die. In the field of press free forging large die design, traditional methods often rely on numerical simulation combined with manual experience for fine-tuning. For example, patent CN119203652A discloses an online optimization method and device for hot forging process parameters. It optimizes process parameters through multi-scale simulation of hot forging, but its dynamic recrystallization model relies on empirical formulas, resulting in large errors in grain size prediction and the lack of hard constraint optimization. The problems include: 1. Lack of multi-field coupling: Only considering the thermo-mechanical dual fields, without constructing a dynamic feedback mechanism between the microstructure field and macroscopic stress, leading to grain coarsening (standard deviation ≥ 20%) and low pass rate of forging mechanical properties. 2. Low optimization efficiency: Relying on empirical formulas and simple biomimetic algorithms, it lacks adaptive variation strategies for temperature gradients. Isolated optimization of process parameters, die geometry, and microstructure parameters is inefficient and prolongs the design cycle. 3. Control lag: Lack of real-time data acquisition and closed-loop feedback results in anomaly response times exceeding 2 seconds and a yield rate of less than 80%. Furthermore, in actual production, the high coefficient of variation in grain size uniformity significantly shortens mold life. Therefore, a mold design method integrating multi-physics coupled modeling, intelligent optimization algorithms, and real-time feedback control is urgently needed. Summary of the Invention

[0003] (a) Technical problems to be solved

[0004] In view of the problems of incomplete multi-field coupling, low efficiency of intelligent optimization, and lack of closed-loop control in existing large-scale die-casting design methods, this invention provides a multi-parameter, multi-field intelligent optimization method and system for large-scale die-casting in press free forging. It breaks through the limitations of traditional thermodynamic simulation, accurately predicts grain evolution through three-field coupling modeling, realizes the synergistic optimization of process parameters, die geometry and microstructure, improves optimization efficiency and reduces the number of trial runs, and constructs a real-time closed-loop control system to improve the uniformity of forging grain size, reduce forming load and shorten production cycle, thereby improving process stability and die life.

[0005] (II) Technical Solution

[0006] To achieve the above objectives, the present invention provides the following technical solution: a multi-parameter, multi-field intelligent optimization method for large-scale free forging dies in presses, comprising the following steps:

[0007] Multi-field modeling: Input the target shape of the forging, material properties and process constraints, construct a multi-physics coupled simulation model, and realize real-time interaction of multiple fields through a unified time step;

[0008] Parameter pre-screening: An initial parameter set is generated using Latin hypercube sampling. A deep kernel learning surrogate model is then used to simultaneously predict macroscopic molding accuracy, microscopic grain size, and residual stress. The variables covered by Latin hypercube sampling include the temperature gradient coefficient, strain rate threshold, and mold heat transfer correction term. Specifically, the temperature gradient coefficient ranges from 0.5 to 2.0 °C / mm, and the strain rate threshold ranges from 0.1 to 5.0 s⁻¹. -1 The mold heat exchange correction term is 0.8 to 1.2.

[0009] Objective optimization: A Pareto optimal solution set satisfying dynamic recrystallization volume fraction > 85% is generated using a digital twin-guided multi-objective evolutionary algorithm (QT-GuidedEA).

[0010] Mold generation: Based on optimal parameters, the parametric CAD template is used to generate the mold geometry, and the grain size distribution map and thermal parameter mapping database are called to automatically verify the design compliance. The parametric CAD generation is based on B-spline surface reconstruction of the mold cavity, outputting a STEP format model, and then automatically generating a compliance report and triggering CAD template updates.

[0011] Processing monitoring: Temperature and strain field data are collected in real time using a high frame rate infrared thermal imager and an embedded strain sensor, and dynamic recrystallization calculations are performed at the edge computing node;

[0012] Closed-loop feedback: Based on LSTM network analysis of actual production data and simulation errors, the digital twin and surrogate model are updated, and cross-production line knowledge transfer is achieved through a federated learning framework; LSTM error modeling takes time-series data (temperature, strain rate, grain size) as input and outputs the recrystallization activation energy correction amount.

[0013]

[0014] The multiphysics coupling simulation model is a three-field coupling model of thermo-mechanical-microstructure, in which:

[0015] Thermodynamic field modeling: Establish unsteady-state heat transfer equations, simultaneously establish the mold-workpiece interface heat transfer model, and integrate plastic work-heat source terms.

[0016]

[0017] Where η is the plastic work-to-heat efficiency, Q friction For the interfacial frictional heat generation term, Q contact For contact heat dissipation; here η = 0.9; Q friction=μ·p·v, further, μ = 0.3 is the coefficient of friction; Q contact =h·(T) workpiece -T die Furthermore, h = 5000 W / m 2 K.

[0018] Stress-strain field modeling: Based on the modified Johnson-Cook constitutive equation, piecewise correction of flow stress:

[0019]

[0020] Where A is the initial yield strength (unit: MPa), B is the strain hardening coefficient (unit: MPa), n is the hardening exponent (dimensionless), ε is the equivalent plastic strain (dimensionless), and ε recr Let ε be the volume fraction of dynamic recrystallization (dimensionless), Q be the activation energy (unit: J / mol), R be the gas constant (unit: J / mol·K), and T be the absolute temperature (unit: K). drex The correction is activated when the percentage is >85%.

[0021] Microstructure field modeling: Dynamic recrystallization and grain evolution are simulated using cellular automata. Nucleation density is correlated with strain excess, and grain boundary migration is determined by curvature k and storage energy gradient. Co-driving:

[0022]

[0023] Where M is the grain boundary mobility, M = 1.2 × 10⁻⁶ -3 m 2 / s, where λ is the storage energy weighting coefficient, λ = 0.5.

[0024] The three fields interact in real time through a unified time step: the temperature field drives recrystallization to activate energy updates, the strain field triggers grain evolution, and the grain size inversely corrects the macroscopic flow stress, with a unified time step Δt = 0.01s.

[0025] The implementation steps of the deep kernel learning surrogate model in parameter pre-screening include:

[0026] The input variables are defined as process parameters (temperature gradient coefficient, strain rate threshold, mold heat transfer correction term) and three-field coupled state variables;

[0027] The network structure is designed to use a 3-layer fully connected neural network, where the output layer embeds the Matern kernel function.

[0028] The joint probability distribution of the three indicators was established through Bayesian optimization to screen constraint parameters, including the dynamic recrystallization initiation threshold (ε). c≥0.15) and grain size standard deviation (<15%);

[0029] The model is trained using historical data and cross-validated to evaluate its generalization ability.

[0030] Digital twin-guided multi-objective evolutionary algorithms (QT-Guided EA) in objective optimization include:

[0031] Digital twin pre-verification: A reduced-order model is used to quickly simulate candidate solutions, retaining only solutions with a dynamic recrystallization volume fraction > 85%; that is, a lightweight digital twin is constructed, integrating a thermo-mechanical-microstructure reduced-order model; candidate solutions with a dynamic recrystallization volume fraction > 85% are screened, eliminating invalid solutions, improving efficiency by 3 times.

[0032] Gradient-sensitive variation: based on temperature gradient field The mutation probability is dynamically adjusted, and the mutation intensity coefficient is calculated using the following formula:

[0033]

[0034] Where α = 0.4, β = 0.3, γ = 0.3; for high gradient regions Apply Gaussian variation (standard deviation σ = 0.2).

[0035] Target hard constraint optimization: The NSGA-III algorithm is adopted, with the goal of minimizing forming load and average grain size <50μm, and the hard constraint grain size standard deviation <15%.

[0036] Automatic verification of design compliance during mold generation includes:

[0037] Access the historical thermal parameter database. Match the grain size variation coefficient of the current mold cavity;

[0038] For regions where the coefficient of variation exceeds the limit (>15%), adjust the fillet radius of the mold based on the strain gradient distribution (compensation amount). Where k = 0.1 - 0.3).

[0039] The federated learning framework in closed-loop feedback includes:

[0040] The central server aggregates the model parameters of each production line and uses the FedAvg algorithm to allocate weights.

[0041] The client only uploads encrypted model parameters (SHA-256 checksum), local data is not shared, and the model update cycle is 24 hours.

[0042] A system for a multi-parameter, multi-field intelligent optimization method for large-scale free forging dies in presses includes:

[0043] Multi-field coupling modeling module: used to construct a three-field coupling model of thermo-mechanical-microstructure field, including temperature field modeling unit, stress field modeling unit and microstructure evolution unit, integrating plastic work heat source and dynamic recrystallization evolution equation;

[0044] Parameter pre-screening module: Deploys Latin hypercube sampling and deep kernel learning models, and outputs the initial parameter set;

[0045] Intelligent optimization module: Runs the DT-Guided EA algorithm to generate the Pareto optimal solution set;

[0046] Mold generation module: Generates mold geometry based on parametric CAD templates and triggers local geometric compensation;

[0047] Processing monitoring module: Connects an infrared thermal imager and a strain sensor to calculate the dynamic recrystallization volume fraction in real time;

[0048] Closed-loop feedback module: The digital twin is updated through an LSTM network, and the federated learning framework enables cross-production line collaboration.

[0049] The system runs on a distributed architecture, including:

[0050] High-performance simulation cluster: Employs MPI parallel computing for three-field coupled models, with load difference <5%;

[0051] Edge computing node: Data is acquired at a frequency of 1kHz, noise is reduced by an FIR filter, and the response time is <2ms;

[0052] Cloud platform: Stores thermal parameter mapping database (MongoDB sharded cluster) and federated learning model library (AES-256 encryption).

[0053] (III) Beneficial Effects

[0054] (1) By using a three-field coupled modeling approach involving thermo-mechanical-microstructure, the limitations of traditional thermo-mechanical simulation are overcome. Through real-time interactive feedback of temperature, stress, and grain evolution, the grain size distribution and mechanical properties are accurately predicted. Simultaneously, deep kernel learning and the DT-Guided EA algorithm are combined with multi-objective optimization to simultaneously minimize forming load and grain size while ensuring the dynamic recrystallization volume fraction. This achieves dual optimization of macroscopic forming accuracy and microstructure quality, improving grain size uniformity, reducing residual stress, and increasing the pass rate of forging mechanical properties. Experimental data show that:

[0055] (2) The LSTM-Federated learning mechanism enables dynamic updating of digital twin parameters and breaks down data silos through cross-device knowledge transfer, effectively reducing errors. The entire system achieves a closed-loop control and feedback of "perception-analysis-decision-execution" from design to production, promoting the forging industry to upgrade to intelligent manufacturing and improving the consistency of the entire industrial chain process. Attached Figure Description

[0056] Various other advantages and benefits will become apparent to those skilled in the art upon reading the following detailed description of preferred embodiments. The accompanying drawings are for illustrative purposes only and are not intended to limit the invention. Furthermore, the same reference numerals denote the same parts throughout the drawings. In the drawings:

[0057] Figure 1 This is a flowchart of a multi-parameter, multi-field intelligent optimization method and system for a large-scale free forging die in a press, as described in this application.

[0058] Figure 2 This is a schematic diagram of the multi-parameter, multi-field intelligent optimization method and system for large-scale free forging die of a press, as described in this application;

[0059] Figure 3 This is a schematic diagram illustrating the multi-parameter, multi-field intelligent optimization method and system for large-scale free forging dies in a press, based on this application; (arrows indicate the direction of data transmission).

[0060] Figure 4 This document presents a flowchart of the DT-Guided EA optimization algorithm within a multi-parameter, multi-field intelligent optimization method for large-scale free forging dies in a press, as described in this application (three stages: pre-validation, mutation, and optimization).

[0061] Figure 5 This application presents a multi-parameter, multi-field intelligent optimization method for a large-scale free forging die in a press and a timing diagram of closed-loop feedback in the system.

[0062] Figure 6 This is a compliance verification interface diagram for a multi-parameter, multi-field intelligent optimization method and system for large-scale free forging dies in a press, as described in this application; (showing grain size distribution and geometric compensation scheme).

[0063] Figure 7 This application presents a multi-parameter, multi-field intelligent optimization method and system federated learning architecture diagram for a large-scale free forging die in a press; (data interaction between the central server and multiple production line clients).

[0064] Figure 8 This is a diagram showing the module composition under the distributed architecture support of the intelligent optimization method and system for large-scale free forging die of a press according to this application; Detailed Implementation

[0065] The following will refer to the appendix in the examples of this invention. Figure 1 - Appendix Figure 8 The technical solutions in the embodiments of the present invention are clearly and completely described. 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.

[0066] A multi-parameter, multi-field intelligent optimization method for large-scale free forging dies in presses includes the following steps:

[0067] Multi-field modeling: Input the target shape of the forging, material properties, and process constraints to construct a multi-physics coupled simulation model, and achieve real-time interaction of multiple fields through a unified time step; The target shape of the forging is input in STL format, and the material properties include but are not limited to elastic modulus, thermal conductivity, etc. The process constraints include but are not limited to press tonnage, allowable stress of mold, etc., to construct a thermo-mechanical-microstructure three-field coupled model, and achieve real-time interaction through a unified time step Δt = 0.01s.

[0068] Parameter pre-screening: Latin hypercube sampling is used to generate an initial parameter set, and a deep kernel learning surrogate model is used to simultaneously predict macroscopic forming accuracy, microscopic grain size and residual stress;

[0069] Objective optimization: A Pareto optimal solution set satisfying dynamic recrystallization volume fraction > 85% is generated using a digital twin-guided multi-objective evolutionary algorithm (QT-GuidedEA).

[0070] Mold generation: Based on optimal parameter-driven parametric CAD templates, mold geometry is generated, and the design compliance is automatically verified by calling the grain size distribution map and thermodynamic parameter mapping database; grain size mapping and compensation, that is, calling the historical database to match the grain size variation coefficient:

[0071]

[0072] Then, for CV grain For areas >15%, the fillet radius is adjusted based on the strain gradient using the following formula:

[0073]

[0074] Processing monitoring: Temperature and strain field data are collected in real time using a high frame rate infrared thermal imager and an embedded strain sensor, and dynamic recrystallization calculations are performed at the edge computing node;

[0075] Closed-loop feedback: Based on LSTM network analysis of actual production data and simulation errors, update digital twins and agent models, and realize cross-production line knowledge transfer through federated learning framework.

[0076] The multiphysics coupling simulation model is a three-field coupling model of thermo-mechanical-microstructure, in which:

[0077] Thermodynamic field modeling: Establish unsteady-state heat transfer equations, simultaneously establish the mold-workpiece interface heat transfer model, and integrate plastic work-heat source terms.

[0078]

[0079] Where k is the thermal conductivity of the material, η is the plastic work-to-heat conversion efficiency, and Q friction For the interfacial frictional heat generation term, Q contact For contact heat dissipation; here η = 0.9; Q friction =μ·p·v, where μ = 0.3 is the coefficient of friction, p is the contact pressure, and v is the relative sliding velocity; Q contact =h·(T) workpiece -T die Furthermore, h = 5000 W / m 2 K is the interfacial heat transfer coefficient.

[0080] Stress-strain field modeling: Based on the modified Johnson-Cook constitutive equation, piecewise correction of flow stress:

[0081]

[0082] Where A is the initial yield strength, B is the strain hardening coefficient, n is the hardening exponent, and X is the strain hardening coefficient. drex The dynamic recrystallization volume fraction is given by α = 0.3-0.4, which is the softening coefficient. This dynamic recrystallization volume fraction is calculated using cellular automata simulation. When X... drex The correction is activated when the percentage is >85%.

[0083] Microstructure field modeling: Dynamic recrystallization and grain evolution are simulated using cellular automata. Nucleation density is correlated with strain excess, and grain boundary migration is determined by curvature k and storage energy gradient. Co-driving:

[0084]

[0085] Where, M = 1.2 × 10 -3 m 2 / s represents the grain boundary mobility, and λ = 0.5 is the storage energy weighting coefficient.

[0086] The three fields interact in real time through a unified time step: the temperature field drives recrystallization to activate energy updates, the strain field triggers grain evolution, and the grain size inversely corrects the macroscopic flow stress, with a unified time step Δt = 0.01s.

[0087] The implementation steps of the deep kernel learning surrogate model in parameter pre-screening include:

[0088] The input variables are defined as process parameters (temperature gradient coefficient, strain rate threshold, and mold heat transfer correction term) and three-field coupled state variables. It should be further explained that the parameter pre-screening uses Latin hypercube sampling to generate an initial parameter set (covering temperature gradient coefficients of 0.5-2.0℃ / mm and strain rate thresholds of 0.1-5.0s). -1 The mold heat transfer correction term is 0.8-1.2), and it is predicted by a deep kernel learning surrogate model.

[0089] The network structure is designed to use a 3-layer fully connected neural network, where the output layer embeds the Matern kernel function.

[0090] By establishing the joint probability distribution of the three indices through Bayesian optimization, constraint parameters are selected, retaining those that satisfy the dynamic recrystallization initiation threshold (ε). c Parameters including ≥0.15 and grain size standard deviation (<15%);

[0091] The model is trained using historical data and cross-validated to evaluate its generalization ability.

[0092] Digital twin-guided multi-objective evolutionary algorithms (QT-Guided EA) in objective optimization include:

[0093] Digital twin pre-verification: A reduced-order model is used to quickly simulate candidate solutions, and only candidate solutions with a dynamic recrystallization volume fraction >85% are retained, with an elimination rate of ≥60%.

[0094] Gradient-sensitive variation: based on temperature gradient field The mutation probability is dynamically adjusted, and the mutation intensity coefficient is calculated using the following formula:

[0095]

[0096] Where α = 0.4, β = 0.3, γ = 0.3; Gaussian variation is applied to the high gradient region, with a standard deviation of σ = 0.2;

[0097] Target hard constraint optimization: The NSGA-III algorithm is adopted, with the goal of minimizing forming load and average grain size <50μm, and the hard constraint grain size standard deviation <15%.

[0098] Automatic verification of design compliance during mold generation includes:

[0099] Access the historical thermal parameter database. Match the grain size variation coefficient of the current mold cavity;

[0100] To further explain, mold generation is based on determining the parameterized CAD template using optimal parameters, i.e., B-spline surface reconstruction, with a tolerance of ±0.1mm. The grain size distribution map is checked against the thermal parameter database for compliance and regularity.

[0101] For regions where the coefficient of variation exceeds the limit (>15%), based on the strain gradient distribution ( Where k = 0.1-0.3) adjusts the fillet radius of the mold.

[0102] The federated learning framework in closed-loop feedback includes:

[0103] The central server aggregates the model parameters of each production line and uses the FedAvg algorithm to allocate weights.

[0104] The client only uploads encrypted model parameters (SHA-256 checksum), and local data is not shared.

[0105] A system for a multi-parameter, multi-field intelligent optimization method for large-scale free forging dies in presses includes:

[0106] Multi-field coupling modeling module: used to construct a three-field coupling model of thermo-mechanical-microstructure field, including temperature field modeling unit, stress field modeling unit and microstructure evolution unit, integrating plastic work heat source and dynamic recrystallization evolution equation;

[0107] Parameter pre-screening module: Deploys Latin hypercube sampling and deep kernel learning models, and outputs the initial parameter set;

[0108] Intelligent optimization module: Runs the DT-Guided EA algorithm to generate the Pareto optimal solution set;

[0109] Mold generation module: Generates mold geometry based on parametric CAD templates and triggers local geometric compensation;

[0110] Processing monitoring module: Connects an infrared thermal imager and a strain sensor to calculate the dynamic recrystallization volume fraction in real time; that is, it collects data in real time through a high frame rate infrared thermal imager (accuracy ±1℃) and an embedded strain sensor (sampling rate 1kHz), and the edge computing node performs dynamic recrystallization calculation at a frequency of 1kHz, with a response time of <2ms.

[0111] Closed-loop feedback module: Updates the digital twin through an LSTM network, and achieves cross-production line collaboration through a federated learning framework. Specifically, it inputs time-series data (temperature, strain rate, grain size) into the LSTM network, outputs recrystallization activation energy correction, and updates the digital twin. Then, it aggregates model parameters from each production line through a federated learning framework (FedAvg algorithm), with the client uploading SHA-256 encrypted parameters. Local data is not shared, and the model update cycle is 24 hours.

[0112] The system runs on a distributed architecture, including:

[0113] High-performance simulation cluster: Employs MPI parallel computing for three-field coupled models, with load difference <5%;

[0114] Edge computing node: Data is acquired at a frequency of 1kHz, noise is reduced by an FIR filter, and the response time is <2ms;

[0115] Cloud platform: Stores thermal parameter mapping database (MongoDB sharded cluster) and federated learning model library (AES-256 encryption).

[0116] Example 1: Referring to the accompanying drawings, this example aims to intelligently optimize a certain aero-engine turbine disk forging through a multi-parameter, multi-field intelligent optimization method and system. That is, to design a high-precision die to ensure the uniformity of grain size in the blade tenon groove area (target coefficient of variation ≤ 15%), and avoid fatigue fracture caused by grain coarsening during service.

[0117] The implementation steps are as follows: Figure 1 As shown, the parameter settings for the forging must first be clarified, including its material properties (Ti-6Al-4V alloy, initial temperature 950℃, strain rate 0.1s). -1 The parameters are: process constraints (maximum hydraulic press tonnage 10000 kN, allowable mold stress 800 MPa) and optimization objectives (forming load < 1000 kN, average grain size < 50 μm, dynamic recrystallization volume fraction ≥ 85%). In other words, parameters need to be input before constructing the three-field coupling model.

[0118] After the parameter settings are clarified, the multi-field coupling modeling module begins to construct a three-field coupling model, including the thermodynamic field, the stress-strain field, and the microstructure field. The thermodynamic field is constructed by establishing an unsteady heat transfer equation and setting the heat transfer coefficient at the mold-workpiece interface to h = 5000 W / m. 2 K, and through the plastic work heat source term Modeling was achieved; the stress-strain field was modeled using a modified Johnson-Cook model with the main parameters A = 850 MPa, B = 680 MPa, n = 0.4, C = 0.015, m = 1.2, and the dynamic recrystallization softening coefficient α = 0.35; the microstructure field was calculated using cellular automata with a mesh resolution of 10 μm × 10 μm and an initial grain size of 30 μm.

[0119] After the thermo-mechanical-microstructure three-field coupling modeling is completed, the module proceeds to parameter pre-screening for parameter pre-screening and surrogate model training. This involves generating 200 initial parameter sets using Latin hypercube sampling, covering temperature gradient coefficients (0.5-2.0℃ / mm) and strain rate thresholds (0.1-5.0s). -1 The dataset was then divided into a training set (70%), a validation set (15%), and a test set (15%) using a deep kernel learning model. After the dataset was divided, the error was predicted using a surrogate model, resulting in an MSE < 0.05 and a grain size standard deviation prediction accuracy > 90%.

[0120] Subsequently, the intelligent optimization module uses a digital twin-guided multi-objective evolutionary algorithm (QT-GuidedEA) to generate the optimal solution set. First, digital twin pre-validation is performed, initially with 100 candidate solutions. After screening using a lightweight twin, 35 solutions are retained. These 35 solutions simultaneously satisfy a single requirement, namely X. drex >85%; after that, for the region with a temperature gradient >50℃ / mm, the mutation probability increased to 0.4, and the standard deviation σ=0.2; then the NSGA-III algorithm was used to iterate 50 times to obtain the final Pareto solution set containing 20 sets of parameters, with a forming load range of 820-950kN and an average grain size of 45-48μm.

[0121] After the target intelligent optimization is completed, the module proceeds to the mold generation module. In the mold generation module, the mold cavity geometry is generated based on the optimal parameters (B-spline surface reconstruction, tolerance within ±0.1mm). Subsequently, the grain size variation coefficient is checked, that is, the out-of-limit region (CV = 17%) is compensated by adjusting the fillet radius.

[0122] The processing monitoring module collects data in real time during these processes. The collected data includes the temperature field collected by the infrared thermal imager (accuracy ±1℃) and the strain field collected by the embedded sensor (sampling rate 1kHz).

[0123] The closed-loop feedback module updates the digital twin parameters through the LSTM network, reducing the recrystallization activation energy prediction error from 12% to 3.2%; at the same time, federated learning aggregates three production line models, reducing the global grain size standard deviation prediction error from 15% to 7%.

[0124] The specific optimization results are shown in the table below:

[0125] index Before optimization After optimization Increase Forming load (kN) 1200 820 31.7％ Average grain size (μm) 65 47 27.7％ Grain size standard deviation (%) 18 7.5 58.3％ Dynamic recrystallization volume fraction (%) 72 89 23.6％

[0126] Example 2: Optimization of forging die for low-pressure rotor in nuclear power plants. First, input the parameters, specifically: material properties (SA508-3 steel, initial temperature 850℃, strain rate 0.05s). -1 The process constraints are defined as follows: mold preheating temperature 300℃, maximum pressing speed 5mm / s; optimization objectives are defined as follows: avoid abnormal grain growth, i.e., grain size > 100μm, mold life > 1000 cycles. Then, the multi-field coupling modeling module will perform modeling, starting with the thermodynamic field, where the mold-workpiece interface heat transfer coefficient h = 4500W / m. 2 K, frictional heat generation term Q plastic =0.3·p·v, where the pressure p = 50MPa and the velocity v = 2mm / s; followed by the stress-strain field, where the Johnson-Cook parameters A = 600MPa, B = 550MPa, n = 0.35, C = 0.02, m = 1.0, and the dynamic recrystallization softening coefficient α = 0.3; the initial grain size in the microstructure field is 50μm, and the stored energy is calculated based on the dislocation density model; after the thermo-mechanical-microstructure three-field coupling model is established, it enters the parameter pre-screening module, where Latin hypercube sampling generates 150 sets of parameters, and candidate solutions with a grain size standard deviation <15% are screened out, while the surrogate model predicts a maximum grain size error <8%; then it enters the intelligent optimization module, where solutions with a dynamic recrystallization volume fraction >80% are screened through digital twin pre-verification, with an elimination rate of 65%, and gradient-sensitive variation is used to optimize high strain rate regions. The mutation probability is increased to 0.5, and the objective function optimized by hard constraints minimizes the maximum grain size (d). max For grain sizes <100μm), 15 Pareto optimal solution sets are generated. After intelligent optimization, the model enters the mold generation module. In the mold generation module, the draft angle is increased by 2° for the region with a grain size >90μm through mold cavity geometry compensation, and a compliance report is generated, triggering CAD template updates. Finally, the model enters the closed-loop feedback template, and the recrystallization connection energy parameters are corrected through an LSTM network, reducing the prediction error from 18% to 5%. At the same time, the federated learning update cycle is shortened to 12 hours, and the cross-site model synchronization error is <3%. The processing monitoring module monitors the grain size online in real time through EBSD, with a monitoring rate of 10Hz.

[0127] The specific optimization results are shown in the table below:

[0128] index Before optimization After optimization Increase Maximum grain size (μm) 120 85 29.2％ Mold life (times) 500 1200 140％ Dynamic recrystallization volume fraction (%) 68 82 20.6％ Production yield (%) 88 96 9.1％

Claims

1. A multi-parameter, multi-field intelligent optimization method for large-scale free forging dies in presses, characterized in that, Includes the following steps: Multi-field modeling: Input the target shape of the forging, material properties and process constraints, and construct a three-field coupled simulation model of thermo-mechanical-microstructure. Real-time interaction is achieved through a unified time step. The temperature field drives recrystallization to activate energy updates, the strain field triggers grain evolution, and the grain size inversely corrects the macroscopic flow stress. Parameter pre-screening: Latin hypercube sampling is used to generate an initial parameter set, and a deep kernel learning surrogate model is used to simultaneously predict macroscopic forming accuracy, microscopic grain size and residual stress; Objective optimization: A multi-objective evolutionary algorithm guided by digital twins is used to generate a Pareto optimal solution set that satisfies a dynamic recrystallization volume fraction > 85%; Mold generation: The mold geometry is generated based on the optimal parameter-driven parametric CAD template, and the design compliance is automatically verified by calling the grain size distribution map and thermodynamic parameter mapping database; Processing monitoring: Temperature and strain field data are collected in real time using a high frame rate infrared thermal imager and an embedded strain sensor, and dynamic recrystallization calculations are performed at the edge computing node; Closed-loop feedback: Time-series data, including temperature, strain rate, and grain size, are input through an LSTM network, and the recrystallization activation energy correction is output to update the digital twin; the federated learning framework uses the FedAvg algorithm to aggregate parameters, and the client uploads SHA-256 encrypted model parameters, and local data is not shared.

2. The intelligent optimization method for large-scale free forging die of a press according to claim 1, characterized in that, The thermo-mechanical-microstructure three-field coupling modeling includes: Thermodynamic field modeling: Establish unsteady-state heat transfer equations, simultaneously establish the mold-workpiece interface heat transfer model, and integrate plastic work-heat source terms. Where T is temperature, γ is the thermal conductivity of the material (W / m·K), η is the plastic work-to-heat conversion efficiency, and ε represents stress. Q represents the strain rate. friction Q represents the interfacial frictional heat generation term. contact This refers to the contact heat dissipation item. Stress-strain field modeling: Based on the modified Johnson-Cook constitutive equation, piecewise correction of flow stress: Where A is the initial yield strength, B is the strain hardening coefficient, n is the hardening exponent, and ε is the equivalent plastic strain. recr Let Q be the volume fraction of dynamic recrystallization, R be the activation energy, and T be the gas constant. Microstructure field modeling: Dynamic recrystallization and grain evolution are simulated through cellular automata. Nucleation density is correlated with strain excess, and grain boundary migration is driven by curvature and storage energy gradient. The three fields achieve real-time data interaction through a unified time step: the temperature field drives recrystallization to activate energy updates, the strain field triggers grain evolution, and the grain size inversely corrects macroscopic flow stress.

3. The intelligent optimization method for large-scale free forging die of a press according to claim 2, characterized in that, The term for interfacial frictional heat generation is: Q friction =μ·p·v Wherein, the coefficient of friction μ=0.3, the contact pressure p=50MPa, and the relative sliding speed v=2mm / s; Contact heat dissipation items are: Q contact =h·(T workpiece -T die ) Interfacial heat transfer coefficient h = 5000 W / m 2 K.

4. The intelligent optimization method for large-scale free forging die of a press according to claim 1, characterized in that, The implementation steps of the deep kernel learning proxy model in the parameter pre-screening include: The input variables are defined as process parameters and three-field coupling state variables; The network structure is designed to use a 3-layer fully connected neural network, where the output layer embeds the Matern kernel function. By establishing the joint probability distribution of the three indices through Bayesian optimization, constraint parameters are screened, including the dynamic recrystallization initiation threshold and the standard deviation of grain size. The model is trained using historical data and cross-validated to evaluate its generalization ability.

5. The intelligent optimization method for large-scale free forging die of a press according to claim 1, characterized in that, The digital twin-guided multi-objective evolutionary algorithm (DT-Guided EA) in the objective optimization includes: Digital twin pre-verification: A reduced-order model is used to quickly simulate candidate solutions, and only solutions with a dynamic recrystallization volume fraction >85% are retained; Gradient-sensitive variability: The variability probability is dynamically adjusted based on the temperature gradient field. The formula for the variability intensity coefficient is: Where α = 0.4 represents the basic variability probability, β = 0.3 represents the temperature gradient influence coefficient, and γ = 0.3 represents the strain gradient influence coefficient. After normalization, these values ​​are dimensionless. The temperature gradient field and strain gradient field are represented by the normalized values, which are dimensionless. Target hard constraint optimization: The NSGA-III algorithm is adopted, with the goal of minimizing forming load and average grain size <50μm, and the hard constraint grain size standard deviation <15%.

6. The intelligent optimization method for large-scale free forging die of a press according to claim 1, characterized in that, The automatic verification of design compliance during mold generation includes: Access the historical thermal parameter database. Match the grain size variation coefficient of the current mold cavity; For regions where the coefficient of variation exceeds the limit (>15%), adjust the mold fillet radius based on the strain gradient distribution, and compensate for the excess amount.

7. The intelligent optimization method for large-scale impact molds in the free section of a press according to claim 6, characterized in that, The fillet radius compensation amount: Where k = 0.1 to 0.3, strain gradient Obtained through finite element analysis.

8. The intelligent optimization method for large-scale free forging die of a press according to claim 1, characterized in that, The federated learning framework in the closed-loop feedback includes: The central server aggregates the model parameters of each production line and uses the FedAvg algorithm to allocate weights. The client only uploads encrypted model parameters; local data is not shared.

9. A system for implementing the multi-parameter, multi-field intelligent optimization method for large-scale free forging die of a press according to any one of claims 1-8, characterized in that, include: Multi-field coupling modeling module: used to construct a three-field coupling model of thermo-mechanical-microstructure field, including temperature field modeling unit, stress field modeling unit and microstructure evolution unit, integrating plastic work heat source and dynamic recrystallization evolution equation; Parameter pre-screening module: Deploys Latin hypercube sampling and deep kernel learning models, and outputs the initial parameter set; Intelligent optimization module: Runs the DT-Guided EA algorithm to generate the Pareto optimal solution set; Mold generation module: Generates mold geometry based on parametric CAD templates and triggers local geometric compensation; Processing monitoring module: Connects an infrared thermal imager and a strain sensor to calculate the dynamic recrystallization volume fraction in real time; Closed-loop feedback module: The digital twin is updated through an LSTM network, and the federated learning framework enables cross-production line collaboration.

10. The system of a multi-parameter, multi-field intelligent optimization method for a large-scale free forging die in a press according to claim 9, characterized in that, The system operates in a distributed architecture, including: High-performance simulation cluster: Employs MPI parallel computing for three-field coupled models, with load difference <5%; Edge computing node: Data is acquired at a frequency of 1kHz, noise is reduced by an FIR filter, and the response time is <2ms; Cloud platform: MongoDB sharded cluster database for storing thermal parameter mappings, and AES-256 encrypted federated learning model library.

Images