Forging Process Control Method and System Based on Digital Twin

Abstract

This application relates to the field of processing control technology, and discloses a method and system for controlling the forging processing process based on digital twin. The method includes: collecting data streams of the temperature field and stress field of large hydraulic press equipment and forgings at multiple points in partitions; performing organization-defect correlation pattern recognition on the data streams to generate a forging process database; constructing a digital twin model using the constitutive unit nesting method, with macroscopic units embedding microscopic tissue evolution units; performing non-linear progressive simulation to obtain critical deformation conditions and defect thresholds; performing reverse solution of the process window to obtain the optimal process parameter combination. This application establishes a forging digital twin model through the constitutive unit nesting method, realizes the bidirectional coupling of macroscopic deformation and microscopic tissue evolution, and determines a robust process parameter combination through the reverse solution technology of the process window, thereby improving the quality stability and production efficiency of forgings.

Description

Technical Field

[0001] This application relates to the field of machining control technology, and in particular, to a forging part machining process control method and system based on digital twin. Background Art

[0002] In the traditional forging part machining process, process design mainly relies on engineers' experience and trial-and-error method for optimization, resulting in unstable forging quality, high cost, long cycle and low efficiency. With the development of computer technology, finite element simulation has been introduced into the forging field. However, these simulations often only focus on the macroscopic deformation process and are difficult to accurately predict the microscopic tissue evolution and defect formation. Most of the existing forging simulations in the prior art adopt single physical field simulation and cannot reflect the complex coupling relationship among the temperature field, stress field and tissue field during the forging process. In addition, it is difficult to collect forging processing data, and the data chain is incomplete, resulting in difficulty in establishing an accurate correlation between process parameters and product quality.

[0003] The main deficiencies of the prior art are as follows: there is a lack of a multi-scale modeling method that closely combines macroscopic deformation with microscopic tissue evolution, and it is impossible to achieve accurate prediction of the entire forging process. At the same time, the existing methods lack effective means to process multi-source heterogeneous data in the forging process and are difficult to extract valuable information from complex process data. Another key problem is the insufficient sensitivity analysis of process parameter perturbations, resulting in poor stability of the optimized process plan in actual production and large quality fluctuations between batches. Especially for large alloy steel components, the defect formation mechanism during their high-temperature large-deformation forming process is complex, and the existing methods are difficult to accurately predict and effectively control. Summary of the Invention

[0004] This application provides a forging part machining process control method and system based on digital twin, which is used to establish a forging digital twin model by the constitutive unit nesting method, realize the bidirectional coupling of macroscopic deformation and microscopic tissue evolution, and determine a robust process parameter combination through the process window reverse solution technology, so as to improve the quality stability and production efficiency of forging parts.

[0005] In a first aspect, the present application provides a method for controlling the forging process based on digital twins. The method for controlling the forging process based on digital twins includes: collecting the temperature field and stress field at multiple points in zones of large hydraulic press equipment and forgings during the forging process to obtain spatio-temporal process data streams; performing tissue-defect correlation pattern recognition on the spatio-temporal process data streams to generate a multi-dimensional forging process database; constructing a forging digital twin model based on the forging process database using the constitutive unit nesting method, where the forging material is divided into multiple macroscopic deformation units, and each macroscopic deformation unit embeds a microscopic tissue evolution unit to obtain a digital mapping body; performing non-linear progressive simulation on the digital mapping body to obtain critical deformation conditions and microstructural defect formation thresholds; and performing reverse solution of the process window for the critical deformation conditions and microstructural defect formation thresholds to obtain a process parameter combination.

[0006] In a second aspect, the present application provides a control system for the forging process based on digital twins. The control system for the forging process based on digital twins includes:

[0007] A collection module for collecting the temperature field and stress field at multiple points in zones of large hydraulic press equipment and forgings during the forging process to obtain spatio-temporal process data streams;

[0008] An identification module for performing tissue-defect correlation pattern recognition on the spatio-temporal process data streams to generate a multi-dimensional forging process database;

[0009] A division module for constructing a forging digital twin model based on the forging process database using the constitutive unit nesting method, where the forging material is divided into multiple macroscopic deformation units, and each macroscopic deformation unit embeds a microscopic tissue evolution unit to obtain a digital mapping body;

[0010] A simulation module for performing non-linear progressive simulation on the digital mapping body to obtain critical deformation conditions and microstructural defect formation thresholds;

[0011] A solution module for performing reverse solution of the process window for the critical deformation conditions and microstructural defect formation thresholds to obtain a process parameter combination.

[0012] In a third aspect, there is provided a control device for the forging process based on digital twins, including: a memory and at least one processor, where instructions are stored in the memory; the at least one processor calls the instructions in the memory so that the control device for the forging process based on digital twins executes the above-mentioned method for controlling the forging process based on digital twins.

[0013] In a fourth aspect, a computer-readable storage medium is provided, in which instructions are stored, and when the instructions are run on a computer, the computer is caused to execute the above-mentioned method for controlling the forging process based on digital twins.

[0014] In the technical solution provided by this application, by collecting the temperature field and stress field at multiple points in different regions of large hydraulic press equipment and forgings during the forging process, a process data stream with spatio-temporal characteristics is obtained, realizing the comprehensive monitoring and digital expression of key physical fields during the forging process, and providing a high-quality data basis for subsequent analysis; organizing-defect association pattern recognition is performed on the spatio-temporal process data stream to generate a multi-dimensional forging process database, establishing a mapping relationship between process parameters, material microstructure evolution, and defect formation, breaking through the limitations of traditional experience-based process design; based on the forging process database, a forging digital twin model is constructed using the constitutive unit nesting method, in which the forging material is divided into multiple macro-deformation units, and each macro-deformation unit embeds a microstructure evolution unit, realizing the two-way coupled calculation of macro-deformation and microstructure evolution, and solving the problem of macro-micro separation in traditional simulation methods; non-linear progressive simulation is performed on the digital mapping body to obtain the critical deformation conditions and the threshold for the formation of microstructural defects, accurately identifying the safety boundaries of materials under different deformation conditions, and providing a scientific basis for process design; reverse solving of the process window is performed on the critical deformation conditions and the threshold for the formation of microstructural defects to obtain a combination of process parameters that takes into account quality requirements and production stability, realizing a closed-loop control from defect prevention to process optimization. The hierarchical clustering algorithm used in the solution performs pattern recognition on the spatio-temporal process data stream, the constitutive unit nesting method realizes macro-micro coupled modeling, the non-linear progressive simulation algorithm predicts deformation behavior and microstructure evolution, and the process window reverse solving algorithm optimizes process parameters. These artificial intelligence algorithms and models together constitute a complete data-driven and intelligent decision-making forging process control system. Especially in the field of extreme manufacturing of large alloy steel components, it significantly improves the consistency and stability of forging quality, reduces the process parameter deviation rate and the over-tolerance rate of blank dimensions, increases the first-pass inspection qualification rate, shortens the manufacturing cycle, and reduces energy consumption and manpower requirements. Description of the Drawings

[0015] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the following drawings are some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0016] Figure 1 It is a schematic diagram of an embodiment of the method for controlling the forging process based on digital twins in the embodiments of this application;

[0017] Figure 2 This is a schematic diagram of an embodiment of the forging part processing process control system based on digital twin in the embodiment of the present application;

[0018] Figure 3 This is a structural schematic block diagram of the forging part processing process control device based on digital twin in the embodiment of the present invention. Specific embodiments

[0019] The embodiment of the present application provides a forging part processing process control method and system based on digital twin. The terms "first", "second", "third", "fourth", etc. (if any) in the specification, claims and above-mentioned drawings of the present application are used to distinguish similar objects, and do not have to be used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged under appropriate circumstances, so that the embodiments described here can be implemented in an order different from that shown or described here. In addition, the term "comprising" or "having" and any variation thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product or device comprising a series of steps or units does not have to be limited to those steps or units clearly listed, but may include other steps or units not clearly listed or inherent to these processes, methods, products or devices.

[0020] For ease of understanding, the specific process of the embodiment of the present application will be described below. Please refer to Figure 1 An embodiment of the forging part processing process control method based on digital twin in the embodiment of the present application includes:

[0021] Step S101: Collect the temperature field and stress field at multiple points in different zones of the large hydraulic press equipment and the forging during the forging process to obtain the spatio-temporal process data stream;

[0022] Step S102: Identify the organization-defect correlation pattern of the spatio-temporal process data stream to generate a multi-dimensional forging process database;

[0023] Step S103: Build a forging digital twin model based on the forging process database using the constitutive unit nesting method, where the forging material is divided into multiple macro deformation units, and each macro deformation unit embeds a microstructural evolution unit to obtain a digital mapping body;

[0024] Step S104: Perform non-linear progressive simulation on the digital mapping body to obtain the critical deformation condition and the threshold for the formation of microstructural defects;

[0025] Step S105: Perform reverse solution of the process window for the critical deformation condition and the threshold for the formation of microstructural defects to obtain the process parameter combination.

[0026] It can be understood that the execution entity of this application can be a forging process control system based on digital twins, or it can also be a terminal or a server. Specifically, it is not limited here. In this embodiment of the application, the server is taken as the execution entity for illustration.

[0027] Specifically, collecting the temperature field and stress field at multiple points in different regions of large hydraulic press equipment and forgings during the forging process is the data basis for the entire method. This step installs temperature sensors, thermal imagers, and stress and strain sensors on the hydraulic cylinders, sliders, and workbench surfaces of the hydraulic press to monitor the equipment status in real time. The thermal imager divides the surface of the forging into grid monitoring areas and collects temperature data at each grid point. At the same time, pressure sensors are arranged on the forging surface to collect pressure distribution data during the forging process. After these raw data are aligned with time stamps and denoised by wavelet transform, the spatio-temporal evolution characteristics of the temperature field and stress field are extracted through tensor decomposition methods, and finally a spatio-temporal process data stream is formed. The spatio-temporal process data stream is a four-dimensional data structure describing the dynamic characteristics of the forging process, containing complete information on the distribution of the temperature field and stress field over time and space. Perform organization-defect correlation pattern recognition on the spatio-temporal process data stream to generate a multi-dimensional forging process database. This step first segments the spatio-temporal process data stream according to three process stages: heating, deformation, and cooling to obtain process characteristic data for each stage. Based on these data, calculate the grain size, phase fraction, and segregation degree of different regions of the forging to form microstructural characteristic data. At the same time, spatially correspond the existing forging defect data with the spatio-temporal process data stream to establish a defect-process mapping relationship. By analyzing the correlation between process parameters and tissue evolution and defect formation, construct organization-defect correlation rules, and finally generate a forging process database containing multi-dimensional associations of materials, processes, tissues, and properties. This database systematically manages and stores multi-dimensional data through logical connections and causal relationship networks.

[0028] The constitutive unit nesting method is a multi-scale modeling method that divides the forging material into multiple macroscopic deformation units, and each macroscopic deformation unit embeds a microstructural evolution unit. In specific implementation, extract the geometric data of the forging from the forging process database, divide the forging into grids to form a grid structure containing multiple macroscopic deformation units. Then extract material parameters from the database and set thermodynamic and mechanical parameters for each macroscopic deformation unit. Extract tissue evolution data and defect formation data from the database, construct microstructural evolution units, and embed them into each macroscopic deformation unit to form a two-layer data structure. Set up a data transfer channel so that the deformation data of the macroscopic unit can be converted into the input of the microscopic unit, and at the same time, the tissue characteristic data of the microscopic unit can also correct the material properties of the macroscopic unit. Finally, obtain the equipment operation parameters from the database, generate a virtual equipment data template, establish a data mapping relationship with the two-layer data structure, form a data interaction network for the forging process, and obtain a digital mapping body.

[0029] Set the initial state and boundary conditions of the digital mapping body, apply incremental loads and thermal boundary conditions to it, conduct the first simulation calculation, and obtain the deformation amounts and temperature distributions of the macroscopic deformation units. Then transfer these data to the corresponding microstructure evolution units, calculate the microstructure changes, and obtain the grain size, phase fraction, and dislocation density data. Update the local mechanical property parameters of the material based on these data and back-transfer them to the macroscopic deformation units for the next progressive simulation. Set monitoring points on the deformation units to record the stress, strain, and temperature data at different deformation stages, and construct a deformation path spectrogram. By analyzing the correspondence between the deformation path and the microstructure evolution, identify the critical points where plastic instability or microstructure defects form under different deformation conditions, and obtain the critical deformation condition data. Conduct statistical analysis on these data to determine the formation thresholds of microstructure defects such as cracks, folds, and microstructure inhomogeneity.

[0030] Sort out the data of the critical deformation conditions and the formation thresholds of microstructure defects, determine the critical process conditions for various defects to occur, and establish the safe boundary ranges of process parameters. Based on these boundary ranges, set optimization goals such as deformation uniformity, microstructure integrity, and production efficiency to form multi-objective process parameter screening conditions. Calculate the influence weights of each process parameter on the forging quality, and screen out key process parameters such as temperature, deformation rate, and deformation amount. Then generate process parameter combination schemes within the safe boundary ranges and input them into the digital mapping body for verification to obtain the performance data of each parameter combination. Through multi-objective comprehensive analysis, screen out the set of process parameter combinations that meet multiple quality indicators. Finally, considering the production conditions and process stability, determine the final process parameter combination from them.

[0031] In the embodiment of the present application,

[0032] In a specific embodiment, the process of executing step S101 may specifically include the following steps:

[0033] Install temperature sensors, thermal imagers, and stress-strain sensors on the hydraulic cylinder, slider, and workbench surface of the hydraulic press equipment to monitor the equipment parameters and generate a parameter matrix containing temperature values, pressure values, and displacement values;

[0034] Divide the grid monitoring area on the surface of the forging, obtain the temperature data of each grid point through thermal imaging scanning, and perform time-frequency analysis on the temperature field using Fourier transform to obtain the temperature field distribution eigenvector;

[0035] Arrange pressure sensors on the forging surface, collect the pressure distribution data during the forging process, calculate the internal stress field distribution of the forging through the finite element inversion method, and construct a three-dimensional stress tensor field;

[0036] Synchronize the data and align the timestamps of the parameter matrix, the eigenvector of the temperature field distribution, and the three-dimensional stress tensor field to obtain multi-source data;

[0037] Input the multi-source data into the tensor decomposition algorithm. Decompose the temperature field and the stress field into a time mode matrix and a spatial mode matrix through the Tucker decomposition method to construct a four-dimensional time-space data model;

[0038] Perform singular value decomposition on the four-dimensional data model, calculate the eigenvectors corresponding to the singular values, and perform data compression through discrete cosine transform to finally form a spatio-temporal process data stream containing the spatio-temporal evolution characteristics of the temperature field and the stress field.

[0039] Specifically, install temperature sensors, thermal imagers, and stress-strain sensors on the hydraulic cylinder, slider, and workbench surface of the hydraulic press. The temperature sensor uses infrared temperature measurement technology and can measure the surface temperature of the equipment without contact; the thermal imager generates a temperature distribution image by detecting the infrared radiation emitted by the object; the stress-strain sensor measures the force based on the piezoresistive effect. These sensors jointly collect the temperature, pressure, and displacement data during the operation of the equipment to form a parameter matrix. The parameter matrix is a multi-dimensional data structure, where each row represents a time point, each column represents a sensor data, and the matrix elements are the corresponding measured values. Grid monitoring of the forging surface is a key step in obtaining comprehensive temperature distribution data. By virtually dividing the forging surface into regular grids, the thermal imager scans and measures each grid point to obtain the complete surface temperature distribution. These temperature data change over time to form a temperature field time series. Use the Fourier transform to process these series and convert the time-domain signal into a frequency-domain representation. The Fourier transform can identify the periodic change characteristics in the temperature field and separate the temperature fluctuation components of different frequencies. By selecting the main frequency components, construct the eigenvector of the temperature field distribution. The eigenvector contains the key characteristics of the spatial distribution and time evolution of the forging surface temperature and is a digital expression of the temperature characteristics during the forging process.

[0040] The acquisition of pressure distribution data during the forging process is achieved by arranging a pressure sensor array on the forging working surface. The sensor array records the pressure values at each point on the working surface during the forging process to form the surface pressure distribution data. However, it is extremely difficult to directly measure the internal stress state of the forging, so the finite element inversion method is used for calculation. Finite element inversion is a numerical calculation method. By establishing a finite element model and using the pressure data measured on the surface as boundary conditions, the internal stress distribution of the forging is solved. The calculation results form a three-dimensional stress tensor field, which describes the stress state of each point inside the forging in all directions. The three-dimensional stress tensor field provides an important basis for predicting the deformation behavior and defect formation of the forging.

[0041] The synchronous processing of multi-source data is the basis for ensuring the accuracy of data analysis. The parameter matrix, the eigenvector of the temperature field distribution, and the three-dimensional stress tensor field come from different sensors, and there may be differences in the acquisition frequency and timestamp. Through the timestamp alignment technology, the data from different sources are reorganized according to a unified time series to ensure the temporal consistency between the data. The timestamp alignment process includes data interpolation, resampling, and synchronization marking, and finally forms a multi-source dataset that is coordinated in the time dimension. The tensor decomposition algorithm is an effective tool for processing high-dimensional data. The temperature field and stress field formed by multi-source data are high-order tensors with time and space dimensions. Tucker decomposition is a multilinear algebra method that decomposes a high-order tensor into the product of several low-order tensors. In this method, the temperature field and stress field are decomposed into a time mode matrix and a space mode matrix through Tucker decomposition. The time mode matrix captures the variation law of the data over time, and the space mode matrix characterizes the spatial distribution characteristics of the data. This decomposition enables the extraction and expression of the spatio-temporal variation laws contained in the original data, and constructs a four-dimensional time-space data model.

[0042] The four-dimensional data model contains a large amount of information and requires further dimensionality reduction processing. Through the singular value decomposition technology, a high-dimensional matrix is decomposed into several singular values and their corresponding eigenvectors. The singular values represent the importance weights of the eigenvectors. By retaining the eigenvectors corresponding to the main singular values, the dimensionality reduction and feature extraction of the data are achieved. The discrete cosine transform is a signal processing technology that transforms a signal into a linear combination of cosine functions. The eigenvectors are compressed through the discrete cosine transform to reduce the data size while retaining the key information, and finally form a spatio-temporal process data stream containing the spatio-temporal evolution characteristics of the temperature field and stress field.

[0043] For example, 12 temperature sensors, 8 pressure sensors, and 6 displacement sensors are installed on an 18,000-ton hydraulic press, with a sampling frequency of 10 Hz, and the forging process is continuously monitored for 4 hours. The surface of the forging is divided into a 20×20 grid, and the temperature data of each grid point is collected once per second. The thermal imaging data is processed through Fourier transform to extract the variation characteristics of the surface temperature, and it is found that there are obvious temperature gradients at the head and tail of the forging. The data of the pressure sensors is combined with finite element calculations to reconstruct the internal stress distribution of the forging, and it is identified that there is a stress concentration area in the central region. After these data are aligned by timestamps, 3 main time modes and 4 main space modes are extracted through Tucker decomposition, capturing the main variation characteristics of the temperature and stress fields. After singular value decomposition and discrete cosine transform, the data volume is reduced by 85%, while retaining the key spatio-temporal evolution characteristics. The finally generated spatio-temporal process data stream accurately reflects the dynamic evolution law of the temperature field and stress field during the forging process.

[0044] In a specific embodiment, the process of executing step S102 may specifically include the following steps:

[0045] Perform data segmentation processing on the spatio-temporal process data stream, divide the forging process into three process stages: heating, deformation, and cooling, and obtain the process characteristic data of each stage;

[0046] Calculate the grain size, phase fraction, and segregation degree of the forging in different regions based on the process characteristic data of each stage, and generate the microstructure characteristic data;

[0047] Classify and organize the forging defect data, establish a defect feature library, and spatially correspond the defect positions with the spatio-temporal process data stream to form a defect-process mapping relationship;

[0048] Based on the microstructure characteristic data and the defect-process mapping relationship, analyze the correlation between process parameters and tissue evolution and defect formation, and establish tissue-defect association rules;

[0049] Integrate the tissue-defect association rules with the spatio-temporal process data stream, construct a data structure containing the relationships between materials, processes, tissues, and properties, and form a multi-dimensional data framework;

[0050] Establish a logical connection and causal relationship network between entities for the multi-dimensional data framework, store and manage it through a database, and generate a multi-dimensional forging process database.

[0051] Specifically, identify the key stages of the forging process. The forging process usually includes three main process stages: heating, deformation, and cooling, and each stage has different process characteristics and effects on the material structure. The data segmentation processing uses the feature point recognition method. By analyzing the change trends of temperature, pressure, and displacement data in the spatio-temporal process data stream, the stage transition points are identified. Specifically, the heating stage is characterized by a continuous increase in temperature and stabilization at the forging temperature, the deformation stage is characterized by significant changes in pressure and displacement data, and the cooling stage is characterized by a continuous decrease in temperature. Through the recognition of these characteristics, the continuous process data stream is segmented into three discrete process stages, and the temperature field, stress field, and deformation field data of each stage are obtained to form the process characteristic data of each stage.

[0052] The calculation of grain size is based on thermodynamics and kinetics principles, mainly considering the effects of temperature, strain, and strain rate. The specific calculation process includes determination of the initial grain size, calculation of the dynamically recrystallized grain size, and calculation of the statically recrystallized grain size. The calculation of phase fraction is based on the phase transformation kinetics theory, combined with heat treatment temperature, holding time, and cooling rate, to calculate the volume fractions of different phases (such as ferrite, pearlite, bainite, martensite, etc.). The calculation of segregation degree is based on the diffusion theory, analyzing the migration behavior of elements at high temperatures, and combined with solidification conditions and deformation history, to calculate the non-uniformity of element distribution in the material. Through these calculations, microstructure characteristic data including grain size, phase fraction, and segregation degree of each region of the forging are generated.

[0053] Sort out the historical production data and quality inspection records, classify the forging defects into surface defects (such as cracks, folds) and internal defects (such as porosity, segregation) according to the type, and establish a defect feature library including defect morphology, size, and distribution characteristics. Then, spatially match the positions of these defects with the process parameters at the corresponding positions in the spatio-temporal process data stream to establish a spatial correspondence relationship. This matching is based on coordinate transformation and spatial interpolation techniques to ensure the accurate correspondence between the defect positions and the process data. Through this spatial mapping, a direct association between the defect positions and the process parameters, that is, the defect-process mapping relationship, is formed.

[0054] Based on the microstructure characteristic data and the defect-process mapping relationship, data mining techniques are used to analyze the correlation between process parameters, material microstructure evolution, and defect formation. Association rule mining is a data mining technique for discovering association relationships between item sets. By calculating the support, confidence, and lift between process parameters (such as temperature, deformation amount, deformation rate) and microstructure characteristics and defect formation, strong correlation rules are identified. For example, when the forging temperature is too high and the cooling rate is too fast, it is easy to form coarse grains and quench cracks. Through this analysis, a series of organization-defect association rules describing the causal relationships between process parameters, microstructure changes, and defect formation are established.

[0055] Integrating the organization-defect association rules with the spatio-temporal process data stream requires constructing a unified data structure framework. Using a multi-dimensional data model, taking material composition, process parameters, microstructure characteristics, and performance indicators as different dimensions, and establishing the mapping relationships between them. Specifically, when implementing, design a data structure including material dimension, process dimension, microstructure dimension, and performance dimension, and each dimension contains the corresponding parameter set. In this multi-dimensional data framework, material composition and process parameters jointly affect microstructure characteristics, and microstructure characteristics in turn determine the final performance. Through this structured expression, discrete rule knowledge is transformed into a systematic data framework.

[0056] Establishing logical connections and causal relationship networks between entities in a multi-dimensional data framework is the key to constructing a complete knowledge system. Knowledge graph technology provides an effective way to express the relationships between entities. Entities such as materials, processes, structures, and properties and their attributes are regarded as nodes, and the influencing relationships between them are regarded as edges to construct a network structure containing causal logic. For example, forging temperature (process entity) affects grain size (structure entity), and grain size in turn affects strength (property entity). This network structure is stored and managed through a graph database, supporting complex relationship queries and inferences, and finally generating a forging process database containing multi-dimensional associations of material-process-structure-property.

[0057] In a specific embodiment, the process of executing step S103 may specifically include the following steps:

[0058] Extract the forging geometry data and material property data from the forging process database, perform mesh division processing on the forging to obtain a three-dimensional mesh data structure containing multiple macroscopic deformation units;

[0059] Screen the temperature field and stress field historical data from the forging process database, assign corresponding thermodynamic parameters and mechanical parameters to each macroscopic deformation unit to generate a macroscopic deformation data set;

[0060] Extract the microstructure evolution data and defect formation data from the forging process database, construct a microstructure evolution data set, and embed it into each macroscopic deformation unit to form a two-layer data structure;

[0061] Create a data transfer channel between the two-layer data structures, set data conversion rules, so that the deformation data of the macroscopic unit can be converted into the input data of the microscopic unit, and at the same time the tissue characteristic data of the microscopic unit is converted into the material property correction data of the macroscopic unit;

[0062] Obtain the equipment operation parameters from the forging process database, generate a virtual equipment data template, and establish a data mapping relationship with the two-layer data structure to form a data interaction network for the forging process;

[0063] Design a real-time data acquisition and update mechanism, input the collected process parameter data into the data interaction network, update the state values of each data node, and obtain a digital mapping body that can change dynamically with the process.

[0064] Specifically, the forging geometric data includes information such as the external dimensions, contour features, and internal structure of the forging. The material property data includes basic physical parameters such as the composition, density, heat capacity, and thermal conductivity of the material. The data extraction process uses database query technology to retrieve relevant records from the process database based on the forging model number and material grade, and extract complete geometric and material information. The extracted geometric data usually exists in the form of CAD format or point cloud data and needs to be preprocessed for subsequent mesh generation. Mesh generation for the forging is the process of discretizing a continuum into a finite number of computational elements. Using finite element mesh generation technology, according to the geometric characteristics of the forging and the expected analysis accuracy requirements, the forging is divided into tetrahedral or hexahedral elements of appropriate quantity and size. The mesh generation process needs to consider the balance between mesh quality, element quantity, and computational efficiency, and use a denser mesh distribution in key areas of the forging such as cross-section change points and stress concentration areas. After division, a three-dimensional mesh data structure containing geometric coordinates, element topological relationships, and material property assignments is formed, and these macroscopic deformation elements constitute the framework of the digital twin model.

[0065] Screening the temperature field and stress field historical data from the forging process database is to initialize the state of the digital twin model. Historical data refers to the actual temperature and stress distribution data collected during the production of previous similar forgings. Through database query and similarity matching algorithms, the historical case data closest to the current forging is found. During the data screening process, the similarities of the forging shape, material type, and process conditions are considered, and a weighted similarity calculation method is used to evaluate the reference value of the historical data. The screened historical data undergoes spatial interpolation processing and is mapped onto the three-dimensional mesh of the current forging to assign initial temperature values and stress states to each macroscopic deformation element. At the same time, the thermodynamic parameters (such as thermal conductivity, specific heat capacity, etc.) and mechanical parameters (such as elastic modulus, yield strength, strain hardening coefficient, etc.) of the material are extracted from the database. These parameters determine the behavior characteristics of the material during the deformation process. The parameter assignment process considers the influence of temperature and strain rate, and appropriate parameter values are assigned to each macroscopic deformation element through interpolation calculation, finally forming a macroscopic deformation dataset containing geometry, physical fields, and material properties.

[0066] Extracting the microstructure evolution data and defect formation data from the forging process database is the basis for constructing the micro-unit model. The microstructure evolution data describes the grain growth, recrystallization, and phase transformation behaviors of materials under different temperature, strain, and strain rate conditions, while the defect formation data records the critical conditions for the formation of defects such as cracks and voids. The data extraction process is also based on similarity matching to retrieve historical data records from the database that are similar to the current forging conditions. The extracted data is processed and normalized to construct a mathematical model and parameter set that describe the laws of microstructure evolution, forming a microstructure evolution data set. These micro-data are embedded into each macro-deformation unit, and a corresponding microstructure evolution model is configured for each macro unit, forming a macro-micro double-layer nested data structure. This constitutive unit nesting method enables the digital twin model to simultaneously reflect the macro-deformation behavior and microstructure evolution process, thus more accurately predicting the performance and quality of forgings.

[0067] Creating a data transfer channel between the double-layer data structures is the key to realizing macro-micro coupled calculation. The data transfer channel defines the data exchange rules and calculation order between the macro unit and the micro unit. In the specific implementation, a downward data flow from the macro to the micro and an upward data flow from the micro to the macro are set. The downward data flow converts the data such as temperature, stress, strain, and strain rate calculated by the macro unit into the input conditions of the micro unit, and the micro unit calculates the microstructure characteristics such as grain size, phase fraction, and defect probability based on these conditions. The upward data flow then converts the calculated microstructure characteristics into the material property correction data of the macro unit. For example, the strength coefficient is corrected according to the grain size, the elastic modulus is adjusted according to the phase fraction, and the fracture toughness is adjusted according to the defect probability. The data conversion process uses the principles of materials science and experimental empirical formulas to establish a quantitative relationship between the microstructure and the macro properties. Through this two-way data transfer mechanism, the macro-deformation and micro-evolution interact with each other and jointly determine the overall performance and quality of the forging.

[0068] Obtaining the equipment operation parameters from the forging process database, including information such as the geometric dimensions, motion characteristics, and control parameters of the equipment, is the basis for creating the virtual equipment model. The equipment parameter extraction uses a database query method similar to that of the forging data to retrieve relevant records according to the equipment model and configuration information. The extracted equipment parameters are used to construct the geometric model, kinematic model, and dynamic model of the virtual equipment, generating a virtual equipment data template. This template describes the working state and control logic of the equipment during the forging process, including the hydraulic system parameters, transmission system characteristics, and control system strategies. A data mapping relationship is established between the virtual equipment template and the previously constructed double-layer data structure to define how the equipment actions affect the deformation conditions of the forging and how the forging state feedback affects the equipment control. This mapping relationship constitutes a complete forging process data interaction network, describing the complex interaction between the equipment-workpiece-process.

[0069] The real-time data acquisition system obtains process parameter data from various sensors installed on forging equipment, such as real-time information like forging temperature, pressure, displacement, etc. After the collected data is preprocessed, it is input into the previously constructed data interaction network to update the state values of each node. The data update process adopts an iterative calculation method. First, the state of the macroscopic deformation unit is updated, then the change of the microstructure is calculated, and then the change of the microstructure is fed back to update the material properties of the macroscopic unit, forming a closed-loop calculation process. Through this continuous data update mechanism, the digital mapping body can dynamically evolve with the actual forging process, reflecting the physical state and tissue changes of the forging in real time, providing strong support for the monitoring and optimization of the forging process.

[0070] In a specific embodiment, the process of executing step S104 may specifically include the following steps:

[0071] Set the initial state and boundary conditions of the digital mapping body, apply incremental load and thermal boundary conditions to the digital mapping body, and perform the first simulation calculation to obtain the deformation amount and temperature distribution of the macroscopic deformation unit;

[0072] Transfer the deformation amount and temperature distribution data of the macroscopic deformation unit to the corresponding microstructure evolution unit, calculate the change amount of the microstructure, and obtain data such as grain size, phase fraction, and dislocation density;

[0073] Based on the data of grain size, phase fraction, and dislocation density, update the local mechanical property parameters of the material, and transmit the updated parameters back to the macroscopic deformation unit for the next progressive simulation;

[0074] Set deformation monitoring points on the macroscopic deformation unit, record the stress, strain, and temperature data at different deformation stages, construct a deformation path spectrum diagram, and analyze the corresponding relationship between the deformation path and the microstructure evolution;

[0075] According to the deformation path spectrum diagram and the microstructure evolution data, identify the critical points where the material undergoes plastic instability or the formation of tissue defects under different deformation conditions, and obtain the critical deformation condition data;

[0076] Conduct statistical analysis on the critical deformation condition data, establish the correlation mapping between the process parameters and the formation of defects, and determine the threshold values for the formation of microstructural defects such as cracks, folds, and tissue inhomogeneity in the material.

[0077] Specifically, set its initial state and boundary conditions. The initial state setting refers to assigning initial temperature, stress, and strain state values to each macroscopic deformation unit in the digital mapping body. These initial values can be extracted from the forging process database or set artificially according to process requirements. The boundary condition setting includes geometric boundary conditions and physical boundary conditions. The geometric boundary conditions specify the displacement constraints on the contact surface between the forging and the die, and the physical boundary conditions specify the heat exchange and friction conditions on the forging surface. Applying incremental loads to the digital mapping body means decomposing the pressure or displacement change amount during the forging process into multiple small incremental steps and gradually applying them to the forging. The thermal boundary conditions describe the heat exchange process between the forging and the environment or the die, including conduction, convection, and radiation heat transfer. After the settings are completed, perform the first simulation calculation, solve the thermo-mechanical coupling equations using the finite element method, and obtain the deformation amount and temperature distribution data of the macroscopic deformation units under specific loads and thermal boundary conditions. These data are stored in the form of displacement vectors and temperature scalar fields for each unit, describing the macroscopic physical state of the forging at that moment. Transferring the deformation amount and temperature distribution data of the macroscopic deformation units to the corresponding microstructural evolution units is the key link connecting the macroscopic and microscopic scales. The data transfer process first needs to convert the deformation amount of the macroscopic unit into strain tensors and strain rate tensors, and the temperature distribution into temperature and its gradient fields. These converted data are used as the input conditions for the microstructural evolution units to drive the microscale microstructure evolution calculation. The microstructural calculation is based on material science theories and empirical formulas to calculate the grain growth, recrystallization, and phase transformation processes under given strain, strain rate, and temperature conditions. The specific calculation process includes: calculating the occurrence degree of dynamic and static recrystallization according to temperature and strain conditions and updating the grain size; calculating the phase transformation kinetics according to temperature and cooling rate to determine the volume fraction of each phase (such as ferrite, pearlite, bainite, etc.); calculating the evolution of dislocation density according to deformation conditions. The calculation results form a microstructural characteristic data set, including information such as grain size, volume fraction of each phase, and dislocation density distribution, describing the structural state of the material at the microscale.

[0078] Mechanical property parameters include parameters describing the mechanical behavior of materials such as yield strength, flow stress, hardening coefficient, Poisson's ratio, etc. The parameter update is based on the structure-property relationship theory in material science. For example: calculate the yield strength of the material according to the grain size, following the Hall-Petch relationship; calculate the composite hardness and strength of the material according to the phase fraction, using the mixture rule; calculate the work hardening degree of the material according to the dislocation density. These updated mechanical property parameters are passed back from the microstructural evolution units to the corresponding macroscopic deformation units to correct the parameter values in their constitutive equations. The corrected constitutive equation more accurately reflects the mechanical behavior of the material in the current microstructural state, so as to calculate a more accurate deformation response in the next progressive simulation. In this way, through cyclic iteration, the two-way coupling calculation between macroscopic deformation and microstructures is achieved.

[0079] Deformation monitoring points are set for the macroscopic deformation unit to track the state changes of key positions during the forging process. The deformation monitoring points are usually selected in the characteristic areas of the forgings, such as key positions prone to problems in the central area, surface area, cross-section change area, etc. At these monitoring points, the change data of the stress tensor, strain tensor, and temperature scalar over time during the entire forging process are recorded. The data recording adopts a time series method, and the recording frequency is set according to the forging deformation step size to ensure that the state changes in the key deformation stages are captured. The recorded data is further processed to construct a deformation path spectrogram. The deformation path spectrogram is a multi-dimensional data visualization method that represents the change trajectories of multi-parameters such as stress, strain, and temperature over time in a coordinate system to form the deformation path lines of each monitoring point. By analyzing the correspondence between these deformation paths and the microscopic tissue evolution data of the corresponding monitoring points, it can be identified which deformation paths lead to the formation of favorable tissues and which paths may cause tissue defects.

[0080] Identifying the critical deformation conditions of the material based on the deformation path spectrogram and microscopic tissue evolution data is a key step in predicting defect formation. The critical deformation condition refers to the boundary state point where the material undergoes plastic instability or tissue defect formation. The identification process is based on data mining and pattern recognition techniques to analyze the correlation between the characteristic points in the deformation path spectrogram and microscopic tissue anomalies. Specific methods include: cluster analysis, grouping similar deformation paths and checking the tissue evolution results corresponding to each group of paths; threshold analysis, determining the critical strain, strain rate, and temperature values that lead to tissue defects; regression analysis, establishing a mathematical relationship between the deformation parameters and the probability of defect formation. Through these analyses, the critical state points of the material under different deformation conditions are identified, such as the critical point of abnormal grain growth caused by high-temperature and low-speed deformation, and the critical point of shear band formation caused by low-temperature and high-speed deformation. These critical points form a critical deformation condition data set, which describes the safety boundary of the material in the deformation space.

[0081] Statistical analysis of critical deformation condition data is an important step in quantifying the relationship between process parameters and defect formation. Statistical analysis methods include frequency analysis, correlation analysis, regression analysis, etc. Frequency analysis statistically analyzes the frequency distribution of various defects to identify the main defect types; correlation analysis calculates the correlation coefficient between process parameters and defect formation to quantify their association strength; regression analysis establishes a mathematical model between process parameters and defect probability to achieve defect prediction. Through these analyses, a quantitative mapping relationship is established between process parameters (such as temperature, deformation amount, deformation rate, etc.) and defect formation (such as cracks, folds, non-uniform microstructure, etc.), forming an association mapping database. Based on this database, critical thresholds for the formation of various defects are determined, such as: the critical strain rate threshold for crack formation, the critical reduction ratio threshold for fold formation, the critical temperature gradient threshold for non-uniform microstructure, etc. These thresholds provide clear boundary conditions for subsequent process optimization and guide the reasonable selection of forging process parameters.

[0082] In a specific embodiment, the process of executing step S105 may specifically include the following steps:

[0083] Sort out the data on critical deformation conditions and the formation thresholds of microstructural defects, determine the critical process conditions for the generation of various defects, and establish a safe boundary range for process parameters;

[0084] Set multiple process optimization goals based on the safe boundary range of process parameters, including deformation uniformity, tissue integrity, and production efficiency indicators, to form multi-objective process parameter screening conditions;

[0085] Rank the importance of the multi-objective process parameter screening conditions, calculate the influence weights of each process parameter on forging quality, and screen out the key process parameters of temperature, deformation rate, and deformation amount;

[0086] Generate process parameter combination plans within the safe boundary range according to the key process parameters, and input these plans into the digital mapping body for verification and evaluation to obtain the performance data of each parameter combination;

[0087] Conduct multi-objective comprehensive analysis on the performance data, and screen out the set of process parameter combinations that simultaneously meet the requirements of multiple quality indicators;

[0088] Screen from the set of process parameter combinations to determine the process parameter combination.

[0089] Specifically, the critical deformation condition data obtained from non-linear progressive simulation are classified according to defect types. Common defect types include cracks, folds, mixed crystals, inhomogeneous microstructure, etc. After classification, statistical processing is performed on the critical condition data of each defect, and the correlation coefficients between various process parameters (such as temperature, deformation amount, deformation rate) and defect formation are calculated. The correlation coefficient reflects the association strength between the parameter and the defect. By setting a correlation threshold, the process parameters that have a significant impact on the formation of each defect are screened out. Subsequently, boundary analysis is performed on the selected key parameters to determine the critical process conditions for the generation of various defects. For example, for crack defects, the critical condition of "easily generated when the temperature is lower than T1 and the deformation rate is higher than R1" may be determined; for fold defects, the critical condition of "easily generated when the reduction ratio is greater than P1 and the die temperature is lower than T2" may be determined. Through comprehensive analysis of the critical conditions of all defects, a safe boundary range of process parameters is established. This range stipulates the upper and lower limits that each process parameter should follow to avoid the formation of various defects. Clearly set the deformation uniformity index, which reflects the consistency of the deformation degree of each part of the forging, usually expressed by the standard deviation or coefficient of variation of the deformation amount. The higher the deformation uniformity, the more uniform the internal properties of the forging. Secondly, set the tissue integrity index, which reflects the ideal degree of the forging's microstructure, including aspects such as grain size uniformity, phase distribution rationality, and minimum defect content, usually expressed by quantifying the ideal deviation degree of these microscopic features. In addition, the production efficiency index includes aspects such as heating time, forging time, and energy consumption, which are directly related to production costs. These indicators often have competing and restrictive relationships. For example, improving deformation uniformity may require reducing production efficiency, and improving tissue integrity may require a more complex deformation path. By quantifying these indicators, screening conditions for process parameters containing multiple objectives are formed, and these conditions together constitute the target space for process optimization.

[0090] Establish an index importance comparison matrix. Through expert scoring or historical data analysis, the relative importance values are obtained by pairwise comparison of each index. Then, calculate the eigenvector of the comparison matrix to obtain the weight coefficients of each index. At the same time, calculate the influence coefficient of each process parameter on each index. This process uses the sensitivity analysis method. By changing a single process parameter and keeping other parameters unchanged, observe the ratio of the change in the index to the change in the parameter to obtain the influence coefficient. Multiply the weight coefficient of each index by the corresponding parameter influence coefficient and sum them up to obtain the comprehensive influence weight of each process parameter. According to the magnitude of the comprehensive influence weight, sort the process parameters and screen out several key process parameters that have the most significant impact on forging quality, such as temperature, deformation rate, deformation amount, etc. These key parameters will become the main control objects for subsequent process optimization.

[0091] The generation of parameter combination schemes adopts the Latin hypercube sampling method, which is a multi-dimensional space uniform sampling technique that can better cover the parameter space with a relatively small number of sampling points. In specific implementation, first, the safety boundary range of each key process parameter is equally divided into N intervals. Then, a value is randomly selected from each interval of each parameter, and these values are randomly combined to form N parameter combination schemes. This method avoids the over-concentration or sparsity of parameters in certain regions and ensures a full exploration of the parameter space. The generated parameter combination schemes are then input into the digital mapping body for verification and evaluation. Based on the previously constructed double-layer data structure, the digital mapping body simulates the forging process under different parameter combinations and calculates the performance index values such as deformation uniformity, tissue integrity, and production efficiency corresponding to each scheme, forming a performance dataset of parameter combination schemes.

[0092] Performing multi-objective comprehensive analysis on the performance data is a key step in screening the optimal process parameter combination. The multi-objective analysis uses the Pareto optimization method, which can handle multiple competing optimization objectives and find the solution set with the optimal comprehensive performance. In specific implementation, first, the scores of each parameter combination on each objective index are calculated based on the performance data. Then, the dominance relationship between parameter combinations is judged. If one combination is not worse than another combination in all indexes and is better than another combination in at least one index, then the first combination is said to dominate the second combination. The combinations that are not dominated by any other combination form the non-dominated solution set, also known as the Pareto front. This front represents the set of optimal balance points that cannot be further improved among various objectives under the current conditions. By extracting the non-dominated solution set, a set of process parameter combinations that simultaneously meet the requirements of multiple quality indexes such as deformation uniformity, tissue integrity, and production efficiency is screened out. These combinations are superior to other schemes in terms of comprehensive performance.

[0093] Perform process adaptability evaluation on the parameter combinations on the Pareto front, considering the existing equipment capabilities, operating habits, and production constraints, and screen out those schemes that are theoretically feasible but difficult to implement in practice. Then, conduct process stability analysis. By making small perturbations near the parameter values of each candidate scheme to simulate the parameter fluctuations during the production process, observe the change range of the performance indexes. The scheme with a small change range has better stability, is not sensitive to parameter fluctuations, and is more reliable in actual production. Finally, comprehensively consider the comprehensive performance scores and stability scores of each candidate scheme, and select the parameter combination with the highest comprehensive evaluation as the finally determined process parameter combination, which will guide the parameter setting in the actual forging production process.

[0094] In a specific embodiment, the process of performing the screening step from the set of process parameter combinations may specifically include the following steps:

[0095] Conduct production adaptability evaluation on each parameter combination in the set of process parameter combinations, and screen out a subset of implementable parameter combinations according to the existing equipment capabilities and production conditions;

[0096] Conduct process stability analysis on the subset of implementable parameter combinations, calculate the sensitivity of each parameter combination to disturbance factors, and obtain process stability indicators;

[0097] Based on the process stability indicators, conduct parameter combination ranking, and select the top three parameter combinations with the highest stability indicators as alternative solutions;

[0098] Input the alternative solutions into the digital mapping body for virtual production testing, simulate the multi-batch production process, and record the quality fluctuation range of each solution;

[0099] Conduct statistical analysis on the quality fluctuation range, calculate the quality consistency scores of each alternative solution, and form the final evaluation result;

[0100] According to the final evaluation result, select the parameter combination with the highest quality consistency score as the process parameter combination.

[0101] Specifically, for production adaptability evaluation, it is first necessary to collect the capacity parameters of existing forging equipment, including technical parameters such as the maximum forging pressure, maximum stroke, anvil surface size, heating temperature range, and deformation rate range. These parameters constitute the hard constraints for forging production. At the same time, collect production condition data, including factors such as the existing die specifications, the skill levels of operating workers, and the capabilities of auxiliary equipment. These constitute the soft constraints for forging production. Subsequently, compare each solution in the set of process parameter combinations with these constraints, design an adaptability evaluation matrix. The rows of the matrix represent each parameter combination, and the columns represent each constraint. The matrix element values represent the degree to which the parameter combination meets the corresponding constraint. For hard constraints, if the parameter value exceeds the equipment capacity range, it is directly determined as non-implementable; for soft constraints, different scores are assigned according to the degree of satisfaction. By calculating the comprehensive adaptability score of each parameter combination and setting an adaptability threshold, screen out the parameter combinations with scores higher than the threshold to form a subset of implementable parameter combinations. The solutions in this subset meet the requirements of the existing equipment capabilities and production conditions and have practical operability.

[0102] Process stability analysis uses the perturbation sensitivity analysis method. This method evaluates the sensitivity of parameter combinations to perturbations by applying small perturbations near the nominal values of parameter combinations and observing the degree of change in output performance indicators. In specific implementation, first, the main perturbation factors in the forging process are identified, including operation errors (such as heating temperature deviation, reduction amount deviation), environmental changes (such as environmental temperature fluctuations, cooling condition changes), and material fluctuations (such as composition deviation, initial state differences), etc. Then, for each implementable parameter combination, perturbations with the standard deviation amplitude are applied to each perturbation factor respectively, and the perturbed parameter values are input into the digital mapping body to calculate the change amount of key performance indicators (such as deformation uniformity, tissue integrity). The smaller the change amount, the lower the sensitivity of the parameter combination to the perturbation factor and the higher the stability. By comprehensively considering the influence of each perturbation factor, the process stability index of each parameter combination is calculated, and this index reflects the robustness of the parameter combination under actual production conditions. In the sorting process, first, the parameter combinations are arranged from high to low according to the process stability index. The scheme with a high process stability index is insensitive to perturbation factors and is more likely to obtain stable quality performance in actual production. Considering that the importance of different perturbation factors may be different, perturbation factor weights can be introduced in the sorting process to perform weighted processing on the sensitivity of each perturbation factor. For example, heating temperature deviation is more common and has a greater impact in actual production, so a higher weight is given; while material composition fluctuation is relatively small and difficult to control, so a lower weight is given. After weighted calculation, a more realistic process stability index is obtained, and the sorting is re-done accordingly. Finally, the top three parameter combinations with the highest stability index are selected from the sorting results as alternative solutions. These solutions have better anti-perturbation ability while meeting the functional requirements and are more suitable for the actual production environment.

[0103] Inputting the alternative solutions into the digital mapping body for virtual production testing is an important means to verify the actual effects of the solutions. Virtual production testing simulates the multi-batch production process in actual production, including within-batch variation and between-batch variation. Within-batch variation mainly comes from random fluctuations in the production of the same batch, and between-batch variation mainly comes from systematic differences between different batches. During the virtual testing process, first, based on historical production data, a probability distribution model of perturbation factors is established, including normal distribution, uniform distribution, or Weibull distribution, etc. Then, using the Monte Carlo random simulation method, random samples are taken from these distributions to generate multiple sets of parameter sets with perturbations, and these parameter sets simulate the actual process conditions in multi-batch production. These parameter sets are input into the digital mapping body in sequence to perform simulation calculations to obtain the product quality prediction results for each batch. By comparing the quality differences between different batches, the quality fluctuation ranges of each alternative solution under multi-batch production conditions are recorded, including statistical quantities such as the maximum value, minimum value, average value, and standard deviation. These data reflect the quality stability of each alternative solution in the actual production environment.

[0104] Statistical analysis of the quality fluctuation range data is a key step in evaluating the consistency of the solutions. Statistical analysis first calculates the distribution characteristics of quality indicators for each alternative solution in multiple batches of production, including statistics such as mean, standard deviation, skewness, and kurtosis. Among them, the mean reflects the average level of quality, the standard deviation reflects the degree of dispersion of quality, the skewness reflects the asymmetry of the distribution, and the kurtosis reflects the peakedness of the distribution. Then, based on these statistics, the quality consistency scores of each alternative solution are calculated. The quality consistency score comprehensively considers the quality level and quality stability. The higher the score, the better the solution can maintain a high quality level while ensuring the smallest quality difference between batches. The calculation method usually adopts a weighted combination of the mean and the standard deviation, that is, giving priority to the solution with a higher mean (high quality level) and a smaller standard deviation (small quality fluctuation). In addition, the shape characteristics of the quality distribution can also be considered, such as preferring a distribution with a skewness close to zero (symmetric distribution) and a moderate kurtosis (neither too concentrated nor too dispersed). Through these analyses, the final evaluation results of each alternative solution are formed. Selecting the parameter combination with the highest quality consistency score as the determined process parameter combination according to the final evaluation result is the last link of the decision-making process. In this link, first check the quality consistency scores of each alternative solution and directly select the parameter combination with the highest score. If the scores of several solutions are very close and it is difficult to directly distinguish the advantages and disadvantages, then further consider other factors, such as production cost, operation difficulty, and compatibility with the existing process. Based on the comprehensive evaluation of these factors, an optimal process parameter combination is finally determined. This combination not only has the best quality performance in theory but also has good stability and consistency under actual production conditions and can meet the requirements of mass production. The finally determined process parameter combination includes specific parameter values such as forging temperature, deformation amount, deformation rate, and pass distribution, and these parameters will guide the operation setting of the actual forging production process.

[0105] The above describes the digital twin-based forging part processing process control method in the embodiments of the present application. Next, the digital twin-based forging part processing process control system in the embodiments of the present application will be described. Please refer to Figure 2 In an embodiment, the digital twin-based forging part processing process control system in the embodiments of the present application includes:

[0106] An acquisition module 201, configured to collect the temperature field and stress field at multiple points in different regions of the large hydraulic press equipment and the forging during the forging process to obtain spatio-temporal process data streams;

[0107] An identification module 202, configured to perform organization-defect correlation pattern recognition on the spatio-temporal process data streams to generate a multi-dimensional forging process database;

[0108] A partitioning module 203, configured to construct a forging digital twin model by using a constitutive unit nesting method based on a forging process database, wherein a forging material is divided into multiple macro deformation units, and each macro deformation unit embeds a microstructural evolution unit to obtain a digital mapping body;

[0109] A simulation module 204, configured to perform non-linear progressive simulation on the digital mapping body to obtain critical deformation conditions and microstructural defect formation thresholds;

[0110] A solving module 205, configured to perform reverse solving of a process window on the critical deformation conditions and microstructural defect formation thresholds to obtain a process parameter combination.

[0111] Through the collaborative cooperation of the above-mentioned components, by collecting the multi-point temperature field and stress field of the large hydraulic press equipment and forgings during the forging process, a process data stream with spatio-temporal characteristics is obtained, realizing the comprehensive monitoring and digital expression of the key physical fields during the forging process, and providing a high-quality data basis for subsequent analysis; by identifying the organization-defect correlation pattern of the spatio-temporal process data stream, a multi-dimensional forging process database is generated, and the mapping relationship between process parameters, material microstructure evolution, and defect formation is established, breaking through the limitations of traditional experience-based process design; by using the constitutive unit nesting method based on the forging process database to construct a forging digital twin model, wherein the forging material is divided into multiple macro deformation units, and each macro deformation unit embeds a microstructural evolution unit, realizing the two-way coupling calculation of macro deformation and microstructural evolution, and solving the problem of macro-micro separation in traditional simulation methods; performing non-linear progressive simulation on the digital mapping body to obtain critical deformation conditions and microstructural defect formation thresholds, accurately identifying the safety boundaries of materials under different deformation conditions, and providing a scientific basis for process design; performing reverse solving of the process window on the critical deformation conditions and microstructural defect formation thresholds to obtain a process parameter combination that takes into account quality requirements and production stability, realizing the closed-loop control from defect prevention to process optimization. The hierarchical clustering algorithm adopted in the solution is used to identify the pattern of the spatio-temporal process data stream, the constitutive unit nesting method is used to realize macro-micro coupling modeling, the non-linear progressive simulation algorithm is used to predict deformation behavior and microstructure evolution, and the process window reverse solving algorithm is used to optimize process parameters. These artificial intelligence algorithms and models together constitute a complete data-driven and intelligent decision-making forging process control system. Especially in the field of extreme manufacturing of large alloy steel components, it significantly improves the consistency and stability of forging quality, reduces the process parameter deviation rate and the blank size out-of-tolerance rate, improves the first-pass inspection qualification rate, shortens the manufacturing cycle, and reduces energy consumption and manpower requirements.

[0112] Above Figure 2The embodiment of the present invention will be described in detail from the perspective of modular functional entities for the forging processing process control system based on digital twin. Next, the forging processing process control device based on digital twin in the embodiment of the present invention will be described in detail from the perspective of hardware processing.

[0113] Figure 3 FIG. 4 is a schematic structural diagram of a forging processing process control device based on digital twin provided by an embodiment of the present invention. The forging processing process control device 300 based on digital twin may vary greatly due to configuration or performance differences, and may include one or more processors (central processing units, CPU) 310 (for example, one or more processors) and a memory 320, and one or more storage media 330 for storing application programs 333 or data 332 (for example, one or more mass storage device terminals). Among them, the memory 320 and the storage media 330 may be transient storage or persistent storage. The program stored in the storage media 330 may include one or more modules (not shown in the figure), and each module may include a series of instruction operations on the forging processing process control device 300 based on digital twin. Further, the processor 310 may be configured to communicate with the storage media 330 and execute a series of instruction operations in the storage media 330 on the forging processing process control device 300 to implement the steps of the forging processing process control method based on digital twin described above.

[0114] The forging processing process control device 300 based on digital twin may further include one or more power supplies 340, one or more wired or wireless network interfaces 350, one or more input / output interfaces 360, and / or one or more operating systems 331, such as Windows Serve, Mac OS X, Unix, Linux, FreeBSD, and so on. Those skilled in the art can understand that Figure 3 the shown structural diagram of the forging processing process control device based on digital twin does not limit the forging processing process control device based on digital twin provided by the present invention, and may include more or fewer components than shown in the figure, or combine some components, or have different component arrangements.

[0115] The present invention also provides a computer-readable storage medium, which may be a non-volatile computer-readable storage medium or a volatile computer-readable storage medium. Instructions are stored in the computer-readable storage medium. When the instructions run on a computer, the computer is caused to execute the steps of the forging processing process control method based on digital twin.

[0116] Those skilled in the art can clearly understand that for the convenience and conciseness of description, the specific working processes of the above-described systems, systems, and units can refer to the corresponding processes in the foregoing method embodiments and will not be elaborated herein.

[0117] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for causing a digital-twin-based forging process control device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The foregoing storage medium includes: various media such as USB flash drives, mobile hard disks, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical discs that can store program codes.

[0118] The above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit them. Although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions recorded in the foregoing embodiments or perform equivalent replacements for some of the technical features. These modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims

1. A method for controlling the forging processing based on digital twin, characterized in that, The method includes: Collecting the temperature field and stress field at multiple points in different zones of large hydraulic press equipment and forgings during the forging process to obtain spatio-temporal process data streams; Performing organization-defect correlation pattern recognition on the spatio-temporal process data streams to generate a multi-dimensional forging process database; Constructing a forging digital twin model based on the forging process database using the constitutive unit nesting method, where the forging material is divided into multiple macroscopic deformation units, and each macroscopic deformation unit embeds a microscopic tissue evolution unit to obtain a digital mapping body, including: extracting the geometric data and material property data of the forging from the forging process database, performing mesh division processing on the forging to obtain a three-dimensional mesh data structure containing multiple macroscopic deformation units; screening the historical temperature field and stress field data from the forging process database, assigning corresponding thermodynamic parameters and mechanical parameters to each macroscopic deformation unit to generate a macroscopic deformation data set; extracting the tissue evolution data and defect formation data from the forging process database, constructing a microscopic tissue evolution data set, and embedding it into each macroscopic deformation unit to form a double-layer data structure; creating a data transfer channel between the double-layer data structures, setting data conversion rules, enabling the deformation data of the macroscopic unit to be converted into the input data of the microscopic unit, and at the same time, the tissue characteristic data of the microscopic unit to be converted into the material property correction data of the macroscopic unit; obtaining the equipment operation parameters from the forging process database, generating a virtual equipment data template, and establishing a data mapping relationship with the double-layer data structure to form a data interaction network for the forging process; designing a real-time data collection and update mechanism, inputting the collected process parameter data into the data interaction network, updating the state values of each data node, and obtaining a digital mapping body that can dynamically change with the process; Performing non-linear progressive simulation on the digital mapping body to obtain the critical deformation conditions and the threshold for the formation of microstructural defects; Performing reverse solution of the process window for the critical deformation conditions and the threshold for the formation of microstructural defects to obtain the process parameter combination.

2. The method for controlling the forging processing process based on digital twin according to claim 1, wherein The collecting the temperature field and stress field at multiple points in different zones of large hydraulic press equipment and forgings during the forging process to obtain spatio-temporal process data streams includes: Installing temperature sensors, thermal imagers, and stress-strain sensors on the hydraulic cylinder, slider, and workbench surface of the hydraulic press equipment to monitor the equipment parameters and generate a parameter matrix containing temperature values, pressure values, and displacement values; Dividing the mesh monitoring area on the surface of the forging, obtaining the temperature data of each grid point through thermal imaging scanning, and performing time-frequency analysis on the temperature field using Fourier transform to obtain the temperature field distribution feature vector; Arranging pressure sensors on the forging surface, collecting the pressure distribution data during the forging process, calculating the internal stress field distribution of the forging through the finite element inversion method, and constructing a three-dimensional stress tensor field; Synchronizing the data and aligning the timestamps of the parameter matrix, temperature field distribution feature vector, and three-dimensional stress tensor field to obtain multi-source data; Inputting the multi-source data into a tensor decomposition algorithm, and decomposing the temperature field and stress field into a time mode matrix and a space mode matrix through the Tucker decomposition method to construct a four-dimensional time-space data model; Perform singular value decomposition on the four-dimensional data model, calculate the eigenvectors corresponding to the singular values, and perform data compression through discrete cosine transform to finally form a spatio-temporal process data stream containing the spatio-temporal evolution characteristics of the temperature field and the spatio-temporal evolution characteristics of the stress field.

3. The method for controlling the forging processing process based on digital twin according to claim 1, wherein, Perform organization-defect correlation pattern recognition on the spatio-temporal process data stream to generate a multi-dimensional forging process database, including: Perform data segmentation processing on the spatio-temporal process data stream, divide the forging process into three process stages: heating, deformation, and cooling, and obtain the process characteristic data of each stage; Calculate the grain size, phase fraction, and segregation degree of the forging in different regions according to the process characteristic data of each stage to generate microstructure characteristic data; Classify and organize the forging defect data, establish a defect feature library, and spatially correspond the defect positions to the spatio-temporal process data stream to form a defect-process mapping relationship; Based on the microstructure characteristic data and the defect-process mapping relationship, analyze the correlation between process parameters and tissue evolution and defect formation, and establish an organization-defect correlation rule; Integrate the organization-defect correlation rule with the spatio-temporal process data stream, construct a data structure containing the relationships between materials, processes, tissues, and properties, and form a multi-dimensional data framework; Establish a logical connection and causal relationship network between entities for the multi-dimensional data framework, store and manage it through a database, and generate a multi-dimensional forging process database.

4. The method for controlling the forging processing process based on digital twin according to claim 1, characterized in that Perform non-linear progressive simulation on the digital mapping body to obtain the critical deformation conditions and the threshold for the formation of microstructural defects, including: Set the initial state and boundary conditions of the digital mapping body, apply incremental loads and thermal boundary conditions to the digital mapping body, perform the first simulation calculation, and obtain the deformation amount and temperature distribution of the macroscopic deformation unit; Transfer the deformation amount and temperature distribution data of the macroscopic deformation unit to the corresponding microstructural evolution unit, calculate the change amount of the microstructure, and obtain the grain size, phase fraction, and dislocation density data; Based on the grain size, phase fraction, and dislocation density data, update the local mechanical property parameters of the material, and transfer the updated parameters back to the macroscopic deformation unit for the next progressive simulation; Set deformation monitoring points for the macroscopic deformation unit, record the stress, strain, and temperature data at different deformation stages, construct a deformation path spectrogram, and analyze the correspondence between the deformation path and the microstructural evolution; According to the deformation path spectrogram and the microstructural evolution data, identify the critical points where the material undergoes plastic instability or the formation of tissue defects under different deformation conditions, and obtain the critical deformation condition data; Perform statistical analysis on the critical deformation condition data, establish an association mapping between process parameters and defect formation, and determine the threshold for the formation of microstructural defects such as cracks, folds, and tissue inhomogeneity in the material.

5. The method for controlling the forging processing process based on digital twin according to claim 1, wherein, Perform reverse solution of the process window for the critical deformation conditions and the threshold for the formation of microstructural defects to obtain a combination of process parameters, including: Organize the data of the critical deformation conditions and the threshold for the formation of microstructural defects, determine the critical process conditions for the generation of various defects, and establish a safe boundary range for process parameters; Set multiple process optimization objectives based on the safe boundary range of the process parameters, including deformation uniformity, tissue integrity, and production efficiency indicators, to form multi-objective process parameter screening conditions; Rank the importance of the multi-objective process parameter screening conditions, calculate the influence weights of each process parameter on the forging quality, and screen out the key process parameters of temperature, deformation rate, and deformation amount; Generate process parameter combination schemes within the safe boundary range according to the key process parameters, and input these schemes into the digital mapping body for verification and evaluation to obtain the performance data of each parameter combination; Conduct multi-objective comprehensive analysis on the performance data to screen out the set of process parameter combinations that simultaneously meet the requirements of multiple quality indicators; Screen from the set of process parameter combinations to determine the process parameter combination.

6. The method for controlling the forging processing process based on digital twin according to claim 5, characterized in that The screening from the set of process parameter combinations to determine the process parameter combination includes: Conduct production adaptability evaluation on each parameter combination in the set of process parameter combinations, and screen out the subset of implementable parameter combinations according to the existing equipment capabilities and production conditions; Conduct process stability analysis on the subset of implementable parameter combinations, calculate the sensitivity of each parameter combination to disturbance factors, and obtain the process stability index; Rank the parameter combinations based on the process stability index, and select the top three parameter combinations with the highest stability index as alternative solutions; Input the alternative solutions into the digital mapping body for virtual production testing, simulate the multi-batch production process, and record the quality fluctuation range of each solution; Conduct statistical analysis on the quality fluctuation range, calculate the quality consistency score of each alternative solution, and form the final evaluation result; Select the parameter combination with the highest quality consistency score as the process parameter combination according to the final evaluation result.

7. A forging processing control system based on digital twin, characterized in that For implementing the digital twin-based forging part processing process control method according to any one of claims 1-6, the digital twin-based forging part processing process control system includes: An acquisition module for acquiring the temperature field and stress field of the large hydraulic press equipment and forgings in the forging process at multiple points in different zones to obtain the spatio-temporal process data stream; An identification module for identifying the tissue-defect correlation pattern of the spatio-temporal process data stream to generate a multi-dimensional forging process database; A partitioning module, configured to construct a forging digital twin model by using a constitutive unit nesting method based on the forging process database, wherein the forging material is divided into multiple macro deformation units, and each macro deformation unit embeds a microstructure evolution unit to obtain a digital mapping body, including: extracting the forging geometric data and material property data from the forging process database, performing mesh partitioning processing on the forging to obtain a three-dimensional mesh data structure containing multiple macro deformation units; screening the temperature field and stress field historical data from the forging process database, allocating corresponding thermodynamic parameters and mechanical parameters to each macro deformation unit to generate a macro deformation data set; extracting the tissue evolution data and defect formation data from the forging process database, constructing a microstructure evolution data set, and embedding it into each macro deformation unit to form a double-layer data structure; creating a data transfer channel between the double-layer data structures, setting data conversion rules, enabling the deformation data of the macro unit to be converted into the input data of the micro unit, and at the same time, the tissue characteristic data of the micro unit to be converted into the material property correction data of the macro unit; obtaining the equipment operation parameters from the forging process database, generating a virtual equipment data template, and establishing a data mapping relationship with the double-layer data structure to form a data interaction network for the forging process; designing a real-time data acquisition and update mechanism, inputting the collected process parameter data into the data interaction network, updating the state values of each data node, and obtaining a digital mapping body that can dynamically change with the process; A simulation module, configured to perform a non-linear progressive simulation on the digital mapping body to obtain the critical deformation condition and the microstructure defect formation threshold; A solving module, configured to perform a reverse solution of the process window for the critical deformation condition and the microstructure defect formation threshold to obtain a process parameter combination.

8. A forging processing control device based on digital twin, characterized in that, It includes a memory and a processor, the memory stores a computer program that can run on the processor, and when the processor executes the computer program, it implements the digital twin-based forging part processing process control method according to any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is run by the processor, the processor is caused to execute the digital twin-based forging part processing process control method according to any one of claims 1 to 6.

Images