Hot Processing Control Method, System, Equipment and Storage Medium Based on Digital Twin

Abstract

The present application relates to the field of hot processing control technology, and discloses a hot processing management and control method, system, device and storage medium based on digital twin. The method includes: obtaining process parameters and quality data of smelting, casting, forging and heat treatment processes through a multi-sensor fusion Internet of Things data acquisition system, constructing a full-process dataset of hot processing and building a multi-dimensional process data middle platform, realizing parameter mapping and grid conversion through an adaptive interface algorithm, establishing a collaborative simulation platform, constructing a digital twin model and deploying it to an edge computing gateway, and generating a visual management and control system. Calculating process parameter combinations and resource priority allocation schemes. The present application can, on the basis of real-time obtaining a large amount of process data, realize cross-process collaborative optimization and intelligent decision-making, especially under the conditions of multi-physical field coupling and complex time-varying characteristics, how to effectively monitor and adjust process parameters in real time to ensure the optimization of product quality and production efficiency.

Description

Technical Field

[0001] This application relates to the technical field of hot processing control, and particularly to a hot processing management and control method, system, device, and storage medium based on digital twin. Background Art

[0002] With the continuous development of the manufacturing industry, hot processing technologies have been widely applied in multiple fields, such as casting, forging, heat treatment, etc. Especially when producing large components and high-strength products, hot processing technologies play a crucial role in product quality and production efficiency. In the existing hot processing process, the optimization of process parameters and quality control usually rely on experience and manual operations, and ensure product quality by monitoring the equipment status and adjusting process conditions. However, with the increase in product complexity and the improvement of requirements for product accuracy and consistency, traditional control methods and technical means are difficult to meet the needs of efficient and high-quality production. Most of the existing technologies lack real-time performance and adaptability, cannot process complex multi-physical field data in real time, and cannot perform automatic adjustment and optimization in a timely manner during the process.

[0003] Although many modern manufacturing systems have introduced data acquisition and monitoring systems, such as temperature sensors, pressure sensors, and on-line detection devices, etc., to collect process data, these systems are usually isolated and lack the ability to integrate and intelligently analyze data across processes and multiple process links. At the same time, the existing hot processing control systems cannot effectively handle the time-varying characteristics and non-linear relationships of process parameters, and it is also difficult to perform dynamic optimization and adjustment through real-time data feedback and prediction. Due to the complexity of the system, the existing control methods still rely on a large amount of manual intervention, and it is difficult to achieve autonomous intelligent optimization and decision-making, resulting in resource waste, low production efficiency, and unstable product quality. Summary of the Invention

[0004] This application provides a hot processing management and control method, system, device, and storage medium based on digital twin, which is used to achieve cross-process collaborative optimization and intelligent decision-making on the basis of real-time acquisition of a large amount of process data. Especially under the conditions of multi-physical field coupling and complex time-varying characteristics, how to effectively monitor and adjust process parameters in real time to ensure the optimization of product quality and production efficiency.

[0005] In a first aspect, the present application provides a digital-twin-based hot processing control method. The digital-twin-based hot processing control method includes: using a multi-sensor fusion Internet of Things data acquisition system to collect process parameters and quality data for smelting, casting, forging, and heat treatment processes, so as to obtain a hot processing full-process data set; constructing a multi-dimensional process data center for hot processing according to the hot processing full-process data set; based on the multi-dimensional process data center for hot processing, through an adaptive interface algorithm, performing parameter mapping and grid conversion on multi-stage simulation results to establish a hot processing full-process collaborative simulation platform; performing hierarchical parametric modeling based on the hot processing full-process collaborative simulation platform to construct a digital twin model; deploying the digital twin model to an edge computing gateway, and generating a visual control system through a virtual-real data interaction mechanism and multi-level situation awareness; inputting the real-time data of the manufacturing process collected by the visual control system into a multi-round distributed process optimization engine to calculate the target process parameter combination and the priority allocation scheme of limited resources.

[0006] In a first implementation manner of the first aspect, the step of using a multi-sensor fusion Internet of Things data acquisition system to collect process parameters and quality data for smelting, casting, forging, and heat treatment processes, so as to obtain a hot processing full-process data set includes: extracting smelting temperature, pressure, and composition data from configuration software and PLC controllers by using the OPC communication method, and collecting process parameters of charging, rough smelting, refining, vacuum, top casting, and bottom casting processes in the smelting process to obtain smelting process parameter data; accessing on-site instrumentation equipment of ingots through the Modbus protocol, collecting and digitally processing process requirement data, process execution data, and auxiliary material data to obtain casting process parameter data; performing multi-sensor access transformation on large hydraulic presses, presses, and supporting forging heating equipment, collecting temperature field, stress field, and deformation field data during the forging process to generate forging process parameter data; digitally processing quality inspection samples of the heat treatment process based on image recognition and machine learning technologies, converting organizational structure features and mechanical property features into numerical representations, and constructing a heat treatment quality data set; performing time series marking and correlation analysis on the smelting process parameter data, the casting process parameter data, the forging process parameter data, and the heat treatment quality data set to construct a process chain data structure with process inheritance relationships; performing outlier detection and data cleaning on the process chain data structure, and generating a hot processing full-process data set after standardization processing and feature extraction.

[0007] In the second implementation manner of the first aspect, constructing a multi-dimensional process data center for hot processing based on the hot processing full-process data set includes: performing data cleaning, trend item separation, and noise suppression processing on the hot processing full-process data set, extracting main process features, and obtaining preprocessed hot processing data; designing a hierarchical data structure including a metadata management layer, a task management layer, a data conversion layer, and a view management layer according to the preprocessed hot processing data, and constructing a multi-level data storage architecture; based on the multi-level data storage architecture, classifying and storing the preprocessed hot processing data according to smelting, casting, forging, and heat treatment process types, and generating a professional process database module; performing working condition segmentation and standardization processing on the data in the professional process database module, analyzing the mapping relationship between material-process-structure-property, and establishing a multi-dimensional data association model; using the multi-dimensional data association model, generating a process data service component through a standardized data service interface based on middleware;

[0008] Combining the functions of the process data service component through service orchestration and a rule engine to construct a multi-dimensional process data center for hot processing with data management, process parameter retrieval, and process sequence correlation analysis functions.

[0009] In the third implementation manner of the first aspect, based on the multi-dimensional process data center for hot processing, establishing a full-process collaborative simulation platform for hot processing by performing parameter mapping and mesh conversion on multi-stage simulation results through an adaptive interface algorithm, including: calling the material thermal physical property parameters and process condition parameters of each process of casting, forging, and heat treatment from the multi-dimensional process data center for hot processing to construct independent simulation models for each process; performing data extraction and structured processing on the calculation results of the independent simulation models for each process to generate a process simulation data set including temperature field, stress field, structure field, and defect distribution; based on the process simulation data set, using a spatial interpolation algorithm to convert the upstream process simulation grid point data into the grid form required for downstream process simulation to obtain a cross-process grid mapping model; performing error analysis and calibration processing on the cross-process grid mapping model to determine the error range of each parameter transfer and establish a data transfer error compensation mechanism; according to the data transfer error compensation mechanism, through simulation interface software, realizing the automatic conversion of the upstream process calculation results to the initial conditions of the downstream process, and generating a seamlessly coupled link system between processes; integrating the link system into the full-process simulation software of casting, forging, and heat treatment, and generating a full-process collaborative simulation platform for hot processing through hardware acceleration and computing resource scheduling integration.

[0010] In the fourth implementation manner of the first aspect, hierarchical parametric modeling is performed based on the full-process collaborative simulation platform for hot processing to construct a digital twin model, including: extracting the forging equipment structure and multi-physical field characteristic data from the full-process collaborative simulation platform for hot processing, using geometric-physical field coupling-driven modeling to analyze the characteristic relationships, and obtaining the basic data set for the large component manufacturing process; performing skin model processing on the large forging equipment structure in the basic data set, extracting the key feature points and associated parameters on the equipment surface, and constructing the digital model of the equipment; based on the digital model of the equipment, dividing the design parameters of the large hot processing component into three levels: overall contour parameters, key component parameters, and process control parameters, and establishing a hierarchical parametric structure system; performing three-dimensional scanning and measurement on the hot processing component in the hierarchical parametric structure system, comparing and calibrating the obtained geometric dimensions and temperature field data with the simulation calculation results, and generating a time-varying feature database; inputting the time-varying feature database into the upsetting and drawing process analysis modules of the forging process of the hot processing component, calculating the variation rules of temperature, strain, and damage distribution, and generating process model data; applying the model accuracy evaluation index system to the process model data for accuracy, lightweight, and real-time evaluation and optimization, and constructing a digital twin model.

[0011] In the fifth implementation manner of the first aspect, deploying the digital twin model to the edge computing gateway and generating a visual management and control system through virtual-real data interaction mechanism and multi-level situation awareness, including: deploying the digital twin model to the edge computing gateway between the device end and the digital system, constructing a data channel supporting 5G or wired transmission, and generating an edge computing service framework; performing distributed processing on the key equipment, production process, and workshop environment data collected by the edge computing service framework, establishing a real-time mapping matrix of virtual and physical data, and generating a virtual-real data interaction mechanism; based on the virtual-real data interaction mechanism, constructing a multi-level situation awareness model including equipment layer situation, process layer situation, and safety layer situation, and generating a real-time monitoring data stream of the production process; applying a fast scanning and multi-dimensional modeling analysis algorithm to the quality characteristics of the hot processing process in the real-time monitoring data stream of the production process, and generating a precise analysis data set for the forging allowance; based on the precise analysis data set for the forging allowance, constructing a fusion control system including a human-machine remote control module, an interactive operation control module, and a production safety management module, and generating a safety warning index for the manufacturing process; integrating the safety warning index for the manufacturing process and the real-time monitoring data stream of the production process into a unified interface framework through cloud configuration technology, and optimizing the parameters of the edge computing gateway according to the cloud strategy, and generating a visual management and control system.

[0012] In the sixth implementation manner of the first aspect, the input of the real-time manufacturing process data collected by the visualization control system into the multi-round distributed process optimization engine to calculate the target process parameter combination and the limited resource priority allocation scheme includes: extracting the process parameters and quality inspection data of each process running in real time from the visualization control system to construct a real-time process execution data set; applying a parameter evaluation algorithm based on knowledge reasoning and deep neural network to the real-time process execution data set to calculate the deviation matrix between the current process state and the target quality, and generating multi-dimensional process risk assessment indicators; according to the multi-dimensional process risk assessment indicators, extracting historical optimization cases from the process database, and applying the gradient boosting decision tree method to construct a non-linear mapping relationship between process parameters and quality characteristics to generate a multi-round distributed process optimization engine; setting multi-dimensional optimization goals such as the best quality, the minimum energy consumption, and the shortest cycle in the multi-round distributed process optimization engine, and generating a candidate process parameter combination scheme set through parallel iterative optimization calculation; virtually testing the candidate process parameter combination scheme set through a digital twin model, calculating the quality prediction scores and resource consumption indicators of each scheme, and determining the target process parameter combination; based on the target process parameter combination, applying a multi-attribute analysis method to comprehensively evaluate the importance of the workpiece, the process difficulty, and the production urgency, calculating the priority allocation matrix of equipment and resources, and generating a limited resource priority allocation scheme.

[0013] In a second aspect, the present application provides a hot processing control system based on digital twin. The hot processing control system based on digital twin includes:

[0014] An acquisition module, configured to collect process parameters and quality data of smelting, casting, forging, and heat treatment processes by using an Internet of Things data acquisition system with multi-sensor fusion to obtain a hot processing full-process data set;

[0015] A construction module, configured to construct a multi-dimensional process data center platform for hot processing according to the hot processing full-process data set;

[0016] A conversion module, configured to perform parameter mapping and grid conversion on the multi-stage simulation results through an adaptive interface algorithm based on the multi-dimensional process data center platform for hot processing to establish a full-process collaborative simulation platform for hot processing;

[0017] A modeling module, configured to perform hierarchical parametric modeling based on the full-process collaborative simulation platform for hot processing to construct a digital twin model;

[0018] A perception module, configured to deploy the digital twin model to an edge computing gateway, and generate a visualization control system through a virtual-real data interaction mechanism and multi-level situation awareness;

[0019] An input module for inputting the real-time data of the manufacturing process collected by the visualization control system into a multi-round distributed process optimization engine to calculate the target process parameter combination and the finite resource priority allocation scheme.

[0020] In a third aspect, a digital-twin-based hot processing control device is provided, including: a memory and at least one processor, wherein instructions are stored in the memory; the at least one processor calls the instructions in the memory to enable the digital-twin-based hot processing control device to execute the above-mentioned digital-twin-based hot processing control method.

[0021] In a fourth aspect, a computer-readable storage medium is provided, in which instructions are stored, and when they run on a computer, the computer is enabled to execute the above-mentioned digital-twin-based hot processing control method.

[0022] In the technical solution provided by this application, through the multi-sensor fusion Internet of Things data acquisition system, the process parameters and quality data of the smelting, casting, forging and heat treatment processes are collected in real time, and a full-process dataset of hot processing is constructed. It can realize the comprehensive monitoring of the entire production process. Through the adaptive interface algorithm, the system can perform accurate parameter mapping and grid conversion on the multi-stage simulation results to ensure seamless cross-process data connection, thereby realizing the collaborative simulation of the entire hot processing process. The cooperation efficiency and data consistency between processes are improved. The construction of the hierarchical parametric modeling and digital-twin model enables the process to reflect the actual production situation in real time. The deployment of the digital-twin model to the edge computing gateway, combined with the virtual-real data interaction mechanism and multi-level situation awareness, not only realizes efficient data processing and real-time monitoring, but also enables each link in the production process to obtain accurate feedback. The intelligent level and automation ability of the hot processing process are improved, manual intervention is reduced, and production efficiency and quality are optimized. Using the multi-round distributed process optimization engine, the process parameters can be optimized and calculated, and a finite resource priority allocation scheme can be generated, so as to ensure the optimal balance of production efficiency and product quality under limited resources. Description of the Drawings

[0023] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for the description of the embodiments will be briefly introduced below. Obviously, the drawings in the following description 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.

[0024] Figure 1 It is a schematic diagram of an embodiment of the digital-twin-based hot processing control method in the embodiments of this application;

[0025] Figure 2It is a schematic diagram of an embodiment of the hot processing control system based on digital twin in the embodiments of the present application;

[0026] Figure 3 It is a structural schematic block diagram of the hot processing control device based on digital twin in the embodiments of the present invention. Specific implementation manners

[0027] The embodiments of the present application provide a hot processing control method, system, device and storage medium based on digital twin. Terms such as "first", "second", "third", "fourth", etc. (if any) in the specification, claims and the 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 other than those illustrated or described here. In addition, the terms "including" or "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product or device that includes 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.

[0028] For the convenience of understanding, the specific process of the embodiments of the present application will be described below. Please refer to Figure 1 An embodiment of the hot processing control method based on digital twin in the embodiments of the present application includes:

[0029] Step S101: Use an Internet of Things data acquisition system with multi-sensor fusion to collect process parameters and quality data for the smelting, casting, forging and heat treatment processes, and obtain a full-process dataset of hot processing;

[0030] Step S102: Build a multi-dimensional process data center platform for hot processing according to the full-process dataset of hot processing;

[0031] Step S103: Based on the multi-dimensional process data center platform for hot processing, perform parameter mapping and grid conversion on the multi-stage simulation results through an adaptive interface algorithm, and establish a full-process collaborative simulation platform for hot processing;

[0032] Step S104: Perform hierarchical parametric modeling based on the full-process collaborative simulation platform for hot processing to build a digital twin model;

[0033] Step S105: Deploy the digital twin model to the edge computing gateway, and generate a visual control system through the virtual-real data interaction mechanism and multi-level situation awareness;

[0034] Step S106: Input the real-time data of the manufacturing process collected by the visualization control system into the multi-round distributed process optimization engine to calculate the target process parameter combination and the finite resource priority allocation plan.

[0035] It can be understood that the execution entity of this application can be a hot processing control system based on digital twin, or 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.

[0036] Specifically, data such as temperature, pressure, composition, and strain are collected through various sensors, and these data are transmitted to the system through OPC communication and Modbus protocol. In the smelting process, information such as the temperature, pressure, and composition of the smelting furnace is interacted with the PLC controller through OPC to obtain real-time data during raw material processing, refining, etc. In the casting process, the Modbus protocol is used to connect the on-site instruments of casting to obtain process requirement data, execution data, and auxiliary material data. These data are converted into digital process parameter data through standardized processing. In the forging process, multiple sensors are connected to large hydraulic presses and presses to collect data on the temperature field, stress field, and deformation field during forging. These data are integrated into a data set through a data acquisition system. At the same time, the quality data during the heat treatment process is transformed into numerical representations through image recognition and machine learning techniques, including organizational structure characteristics and mechanical property characteristics, so as to construct a heat treatment quality data set. These data are time-sequentially marked and subjected to correlation analysis after being collected, and a process chain data structure is established. Through outlier detection, data cleaning, and standardized processing, a complete hot processing full-process data set including data on processes such as smelting, casting, forging, and heat treatment is finally generated.

[0037] The hot processing full-process data set is used to construct a multi-dimensional process data middle platform for hot processing. Through data preprocessing techniques, such as removing noise, separating the trend components in the data, and standardizing various data, the consistency and comparability of the data are ensured. For example, for the temperature data in the smelting process, there may be small fluctuations due to equipment errors. These fluctuations are removed through the trend separation method, and the long-term change trend is retained. The construction of the data middle platform adopts a hierarchical data structure, including a metadata management layer, a task management layer, a data conversion layer, a view management layer, etc. In this process, the flexible invocation and distribution of data are realized through middleware to ensure the accurate transmission of data between various process modules. At this stage, the process data is classified and stored according to different process types of smelting, casting, forging, and heat treatment to generate professional process databases. Through multi-dimensional correlation analysis of the data, a mapping relationship between materials, processes, organizations, and properties is established to generate a multi-dimensional data correlation model.

[0038] Through the thermal processing multi-dimensional process data center, the multi-stage simulation results are parameter-mapped and grid-converted using the adaptive interface algorithm, thus establishing a full-process collaborative simulation platform for thermal processing. During the process, the simulation models of different processes are interconnected through the interface algorithm to ensure accurate data transfer between different processes. Taking the casting, forging, and heat treatment processes as examples, the simulation results of each process, such as the temperature field and stress field, will be extracted and structured. The grid data of the upstream process is mapped to the grid form required by the downstream process through the spatial interpolation algorithm. For example, the temperature field data in the casting process is converted into grid data suitable for the forging process through the interpolation algorithm. After error analysis and calibration, these data ensure the accurate transfer of simulation results between different processes.

[0039] Based on the full-process collaborative simulation platform for thermal processing, hierarchical parametric modeling is carried out to construct a digital twin model. The construction of the digital twin model first extracts the basic data set in the thermal processing process, such as the structural and multi-physical field characteristic data of large forging equipment. Through the modeling method of geometric and physical field coupling, the basic data set of the equipment is obtained. After being processed by the skin model, the key feature points and associated parameters on the equipment surface are extracted to construct the digital model of the forging equipment. For the hot processing process of forgings, the design parameters are divided into three levels: overall profile parameters, key component parameters, and process control parameters. This hierarchical parametric structure enables precise control of the digital twin model at each level. Through three-dimensional scanning and measurement, the geometric dimensions and temperature field data obtained are compared and calibrated with the simulation calculation results to generate a time-varying feature database.

[0040] The digital twin model is deployed to the edge computing gateway to achieve virtual-real data interaction and multi-level situation awareness, generating a visual management and control system. After the digital twin model is deployed, data is transmitted between the device side and the digital system through a data channel that supports 5G or wired transmission. Through the edge computing service framework, various data of key equipment, production process, and workshop environment are collected and processed in real time. After these data are processed, they are interacted with virtual data to generate a virtual-real data mapping matrix. With the help of this matrix, a multi-level situation awareness model of equipment layer situation, process layer situation, and safety layer situation is constructed to generate a real-time monitoring data stream of the production process. Through the rapid scanning and multi-dimensional modeling analysis of these data streams, the forging allowance can be accurately analyzed, and based on these analysis results, the process control and production safety warning can be optimized.

[0041] The real-time data collected by the input visualization control system is used to calculate the deviation between the process state and the target quality through a parameter evaluation algorithm based on knowledge reasoning and deep neural networks. For example, assuming that there is a certain deviation between the temperature field data of a forging process and the target value, a deviation matrix can be obtained through the algorithm, and an optimization model can be generated by the gradient boosting decision tree method. During the optimization process, historical optimization cases in the process database are used to construct the non-linear mapping relationship between process parameters and quality characteristics. Through parallel iterative optimization, the target process parameter combination is obtained to ensure the best process quality, the minimum energy consumption, and the shortest production cycle. Combining with the resource allocation priority matrix, the optimal allocation of limited resources is ensured. Through these calculation results, the intelligent optimization and precise control of the entire hot processing process are realized.

[0042] In the embodiment of the present application, through the Internet of Things data acquisition system with multi-sensor fusion, the process parameters and quality data of the smelting, casting, forging, and heat treatment processes are collected in real time, and a full-process dataset of hot processing is constructed. It can realize the comprehensive monitoring of the entire production process. Through the adaptive interface algorithm, the system can perform accurate parameter mapping and grid conversion on the multi-stage simulation results to ensure seamless connection of data across processes, thereby realizing the collaborative simulation of the entire hot processing process. The cooperation efficiency and data consistency between processes are improved. The construction of the hierarchical parametric modeling and digital twin model enables the process to reflect the actual production situation in real time. The deployment of the digital twin model to the edge computing gateway, combined with the virtual-real data interaction mechanism and multi-level situation awareness, not only realizes efficient data processing and real-time monitoring, but also enables each link in the production process to obtain accurate feedback. The intelligent level and automation ability of the hot processing process are improved, manual intervention is reduced, and production efficiency and quality are optimized. Using the multi-round distributed process optimization engine, the process parameters can be optimized and calculated, and a limited resource priority allocation plan can be generated, so as to ensure the optimal balance of production efficiency and product quality under limited resources.

[0043] In a specific embodiment, the process of executing step S101 may specifically include the following steps:

[0044] Use the OPC communication method to extract the smelting temperature, pressure, and composition data from the configuration software and the PLC controller, and collect the process parameters of the stock preparation, rough smelting, refining, vacuum, top casting, and bottom casting processes in the smelting process to obtain the smelting process parameter data;

[0045] Access the ingot casting on-site instrument equipment through the Modbus protocol, collect and digitally process the process requirement data, process execution data, and auxiliary material data to obtain the casting process parameter data;

[0046] Retrofit large hydraulic presses, presses, and supporting forging heating equipment with multi-sensors to collect data on the temperature field, stress field, and deformation field during forging, and generate forging process parameter data;

[0047] Digitally process quality inspection samples during the heat treatment process based on image recognition and machine learning technologies, convert the organizational structure characteristics and mechanical property characteristics into numerical representations, and construct a heat treatment quality data set;

[0048] Perform time series marking and correlation analysis on the smelting process parameter data, casting process parameter data, forging process parameter data, and heat treatment quality data set, and construct a process chain data structure with process inheritance relationships;

[0049] Perform outlier detection and data cleaning on the process chain data structure, and generate a full-process heat treatment data set after standardization processing and feature extraction.

[0050] Specifically, the smelting, casting, forging, and heat treatment processes use a variety of sensors and data protocols to accurately collect real-time data, which are processed, analyzed, and an optimization model is generated to drive various decisions in the hot working process. The data collection of the smelting process uses the combination of OPC communication method and PLC controller to collect key data during the smelting process, such as temperature, pressure, and composition. OPC (OLE for Process Control) is a widely used communication protocol in industrial automation that can read real-time data from field devices through a standardized interface and transmit it to the control system. During the smelting process, devices such as temperature sensors, pressure sensors, and chemical composition analyzers transmit data to the PLC controller, and the PLC sends this data to the configuration software for monitoring and storage through the OPC protocol. At this time, real-time data such as various parameters of the smelting process, such as furnace temperature, pressure, and liquid level, are collected and a data stream is generated. These data provide key process parameters for the entire smelting process. In addition, parameters of various links involved in the smelting process, such as stock preparation, rough smelting, refining, vacuum, top casting, and bottom casting, are also collected through corresponding devices. These data include the temperature, pressure, chemical composition, and flow rate of the materials, and are continuously collected and transmitted to the data processing platform through the cooperation of the OPC protocol and the PLC to generate smelting process parameter data.

[0051] In the steps of the casting process, Modbus protocol is used to access the ingot casting on-site instrument equipment. Modbus is a serial communication protocol that is widely used in industrial automation and is mainly used to read or write data from devices. Through the Modbus protocol, the system can interact with casting equipment (such as heating furnaces, mold temperature control systems, etc.) to collect process requirement data, process execution data, and auxiliary material data. Process requirement data may include parameters such as temperature, pressure, and time during the casting process, while process execution data reflects the operating status of the equipment in actual operation, and auxiliary material data refers to the auxiliary materials used in casting, such as the usage amount of alloy additives, etc. After being digitized through the Modbus protocol, the data enters the data acquisition system.

[0052] In the forging process, data acquisition relies on equipment modified with multi-sensor access. For example, the pressure, displacement, and temperature sensors of large hydraulic presses and presses can obtain the temperature field, stress field, and deformation field data during the forging process in real-time. Through these sensors, various parameters of the forging process can be comprehensively monitored, such as the temperature change, stress distribution, and deformation degree of the forgings. It is transmitted to the control system through the edge computing gateway and is fed back to the digital twin model in real-time. Through the multiple fusion of sensors, the forging process can be comprehensively monitored.

[0053] In the heat treatment process, data acquisition is completed through image recognition and machine learning techniques for the digital processing of quality inspection samples. Image recognition technology is used to scan the surface of heat-treated workpieces to identify defects such as cracks and pores on the surface, while machine learning algorithms are used to analyze the relationship between these defects and process parameters. Through the learning of a large number of data samples, the system can convert the characteristics of the microstructure (such as grain size, phase transformation type, etc.) and mechanical property characteristics (such as hardness, tensile strength, etc.) into numerical representations, and these numerical data form the heat treatment quality data set.

[0054] The multi-dimensional process data platform for hot processing performs time series marking and correlation analysis on the data of each process to generate the process chain data structure. For the time series data of different process links, it will be marked to ensure that the data can be correctly correlated according to the process sequence. For example, the temperature changes during the smelting process and the temperature changes during the casting process need to be marked according to the time point and process steps. Subsequently, based on these time series data, the system will perform correlation analysis to find out the inheritance relationship between each process. For example, the temperature change during the smelting process directly affects the cooling rate of the casting process, and the quality of the casting process will affect the forging accuracy. Through these correlation analyses, a complete process chain data structure is constructed.

[0055] After data association is completed, the adaptive interface algorithm will be used to perform parameter mapping and grid conversion on the multi-stage simulation results. In the simulation results of processes such as smelting, casting, and forging, the system will automatically map the data generated by different processes to a common parameter space and grid. For example, the temperature field and pressure field in the smelting process need to be mapped to the simulation models of the casting and forging processes through grid conversion to ensure seamless data docking between different processes.

[0056] Through data cleaning and standardization processing, the smelting process parameter data, casting process parameter data, forging process parameter data, and heat treatment quality data are sorted out, noise is removed, missing values are filled, features are extracted, and a full-process hot working dataset is generated. This dataset contains complete data from smelting to heat treatment and serves as the basis for subsequent process optimization.

[0057] For example, during the hot working process, the temperature of the smelting furnace gradually rises from 1500 °C to 1650 °C, and the temperature control during the casting process is very sensitive. The temperature data collected through the Modbus protocol may contain certain fluctuations, such as the temperature changing by about ±5 °C per minute. These data will be standardized to eliminate the influence of equipment errors and environmental factors and ensure the accuracy of subsequent simulations and optimizations. These temperature data are mapped to the temperature field simulation in the forging process through grid conversion, enabling the forging process to be optimized based on more accurate smelting and casting process data.

[0058] In a specific embodiment, the process of executing step S102 may specifically include the following steps:

[0059] Perform data cleaning, trend item separation, and noise suppression processing on the full-process hot working dataset, extract the main process features, and obtain the preprocessed hot working data;

[0060] According to the preprocessed hot working data, design a hierarchical data structure including a metadata management layer, a task management layer, a data conversion layer, and a view management layer, and construct a multi-level data storage architecture;

[0061] Based on the multi-level data storage architecture, classify and store the preprocessed hot working data according to the process types of smelting, casting, forging, and heat treatment to generate a professional process database module;

[0062] Perform working condition segmentation and standardization processing on the data in the professional process database module, analyze the mapping relationship between material-process-structure-property, and establish a multi-dimensional data association model;

[0063] Utilize the multi-dimensional data association model to generate process data service components through a standardized data service interface based on middleware;

[0064] Functionally combine the process data service components through service orchestration and a rule engine to build a multi-dimensional process data platform for hot processing with data management, process parameter retrieval, and process association analysis functions.

[0065] Specifically, for the data collected from the smelting, casting, forging, and heat treatment processes, it is processed through data cleaning, trend item separation, and noise suppression to ensure the accuracy and availability of the data. During the smelting process, the collection of temperature, pressure, and composition data is often affected by factors such as equipment failures and environmental fluctuations, resulting in abnormal fluctuations in the data. Abnormal fluctuations may be caused by sensor failures, sudden changes in environmental temperature, or operational errors. By setting thresholds, these data that do not conform to normal patterns can be quickly identified and eliminated. For example, when the temperature sensor reading of a certain smelting furnace jumps to an abnormal value at a certain moment, through the set abnormal value identification mechanism, the system will automatically identify and mark these data as invalid data and make corresponding corrections, such as filling them using interpolation methods of the front and back data or repairing them using historical data trends.

[0066] Trend item separation is to separate the long-term trend and short-term fluctuations in time series data. For example, during the smelting process, the temperature usually has a long-term upward trend over time, manifested as the gradual heating process of the smelting furnace. However, due to equipment adjustment or improper operation, the temperature will have short-term fluctuations. By using filtering algorithms (such as low-pass filtering or moving average method), these short-term fluctuations can be effectively removed, and only the long-term temperature change trend is retained. This helps to analyze the temperature change law during the smelting process, reduce the interference of equipment periodic fluctuations, and ensure the accuracy of smelting process data.

[0067] During the forging process, temperature and stress field data are often affected by the external environment, such as factors like workshop temperature fluctuations and equipment vibrations, which will generate noise in the data. For example, the vibration of forging equipment may cause fluctuations in the data read by the measurement sensor. Through technologies such as wavelet transform, the system can decompose the data into different frequency components, remove the high-frequency noise components, and only retain the low-frequency signals related to the process parameters. This helps to significantly improve the signal-to-noise ratio of the data and ensure that the temperature field and stress field data more accurately reflect the actual situation of the forging process.

[0068] After completing data cleaning, trend item separation, and noise suppression, it enters the stage of extracting the main process characteristics. In the smelting process, the core characteristics include temperature stability, pressure change range, composition change, etc.; in the casting process, they include casting speed, mold temperature change, etc.; in the forging process, the important characteristics may be the distribution of stress, temperature, and degree of deformation. These characteristics are the key control variables in each process step. When constructing a multi-level data storage architecture, a hierarchical structure including a metadata management layer, a task management layer, a data conversion layer, and a view management layer is designed. The metadata management layer is used to manage the basic information of the data, such as data source, data collection time, equipment number, etc.; the task management layer is responsible for scheduling and managing data processing tasks to ensure that each processing task proceeds in the correct order and dependency relationship; the data conversion layer undertakes the conversion tasks between different process data formats to ensure that data from different processes such as smelting, casting, forging, and heat treatment can be uniformly processed; the view management layer provides data query and visualization interfaces.

[0069] Classify and store the data according to different process types to generate professional process database modules such as smelting, casting, forging, and heat treatment. For example, the temperature, pressure, and composition data in the smelting process are stored separately in the smelting process database, while the process parameters and auxiliary material data in the casting process are stored in the casting process database. This classification storage method helps to improve the data access efficiency, enabling engineers to quickly locate the relevant data of a specific process and conduct analysis.

[0070] Perform working condition segmentation and standardization processing on the data in these process database modules. Working condition segmentation refers to differentiating the data according to different process conditions (such as different material types, different production batches, etc.). For example, different furnace charges in the smelting process will result in different temperature curves, so it is necessary to segment the data according to the type of furnace charge to ensure that the performance of each process under different working conditions can be analyzed separately. Standardization processing is used to eliminate the dimensional differences between different data sources to ensure that various data can be compared and analyzed on the same platform. In the smelting and casting processes, due to different equipment types, there may be inconsistent data units, and standardization processing can unify different process data to the same dimension.

[0071] Using a multi-dimensional data association model, the system quantitatively analyzes the relationships among materials, processes, microstructures, and properties. These association relationships can reveal the impact of each process on the properties of materials. For example, the changes in temperature and chemical composition during the smelting process on the mechanical properties such as strength and toughness of the final product. Through the model, a comprehensive analysis of the impact relationships between different process steps and different material characteristics can be achieved. Based on the data association model, process data service components are generated through a standardized data service interface. These service components can provide parameters, quality data, etc. of different processes such as smelting, casting, and forging to other modules for use. The process data service components combine functions through service orchestration and a rule engine, and finally build a multi-dimensional process data middle platform for hot processing that can provide data management, process parameter retrieval, and process correlation analysis functions. Through the middle platform, users can quickly query relevant data of different processes, monitor process parameters in real time, and make decisions based on the analysis results of the data to ensure the efficiency and accuracy of the production process.

[0072] For example, during the casting process, through the process data middle platform of this system, the temperature change trends of the casting temperature and the mold can be quickly obtained, and then the casting rate can be adjusted to ensure the quality of the castings. Through data association analysis, it is found that when the temperature during the smelting process is too high, the temperature control difficulty during the casting process increases. The system can automatically adjust the temperature range of the smelting process based on these findings to ensure the quality and efficiency of the entire production chain.

[0073] In a specific embodiment, the process of executing step S103 may specifically include the following steps:

[0074] Call the material thermophysical property parameters and process condition parameters of each process of casting, forging, and heat treatment from the multi-dimensional process data middle platform for hot processing to construct independent simulation models for each process;

[0075] Extract and structure the calculation results of the independent simulation models for each process to generate a process simulation data set including temperature field, stress field, microstructure field, and defect distribution;

[0076] Based on the process simulation data set, use a spatial interpolation algorithm to convert the upstream process simulation grid point data into the grid form required for downstream process simulation to obtain a cross-process grid mapping model;

[0077] Conduct error analysis and calibration processing on the cross-process grid mapping model to determine the error range of each parameter transfer and establish a data transfer error compensation mechanism;

[0078] According to the data transfer error compensation mechanism, through the simulation interface software, realize the automatic conversion of the upstream process calculation results to the initial conditions of the downstream process to generate a seamlessly coupled link system between processes;

[0079] Integrate the link system into the full - process simulation software for casting, forging, and heat treatment, and generate a collaborative simulation platform for the full process of hot processing through hardware acceleration and computing resource scheduling integration.

[0080] Specifically, call the material thermal property parameters and process condition parameters of each process of casting, forging, and heat treatment from the multi - dimensional process data platform for hot processing, and build independent simulation models for each process based on these data. Material thermal property parameters include properties such as specific heat capacity, thermal conductivity, melting point, etc., which describe the thermal response of materials at different temperatures, while process conditions include operating variables such as temperature, pressure, and time. These data are collected from different sensors and devices and transmitted to the data platform in real - time through the Internet of Things. Suppose in the casting process, the temperature of the molten steel and the thermal conductivity of the mold are the main influencing factors. By calling these data, the casting simulation model can simulate the temperature field and stress field during the solidification process based on the initial temperature of the molten steel and the mold characteristics. Similarly, in the forging process, by inputting data such as the yield strength and strain hardening index of the material, the forging model can simulate the stress and strain distribution of the material under different deformations.

[0081] After building the simulation models, extract and structure the calculation results of the independent simulation models for each process. The simulation model of each process will generate a series of process data sets, including temperature field, stress field, microstructure field, and defect distribution, etc. For example, during the casting process, the simulation model can generate a temperature field containing the temperature distribution of the molten steel, as well as distribution data of defects (such as pores, cracks) that may occur during solidification; in the forging process, the stress field distribution during the deformation process of the material and possible hot crack regions can be obtained; during the heat treatment process, the change of the microstructure field can be obtained by simulating the phase transformation process. These data need to be extracted and organized into structured simulation data sets to ensure that the simulation results of each process can be analyzed and used in a consistent format for subsequent steps.

[0082] Based on these process simulation data sets, use the spatial interpolation algorithm to convert the upstream process simulation grid point data into the grid form required for downstream process simulation, and build a cross - process grid mapping model. Convert the simulation results generated by each process through the spatial interpolation algorithm, so that the simulation data of the process can be compatible with the data format of the downstream process. For example, during the casting process, the temperature field may be calculated through a two - dimensional grid, while in the forging process, the temperature field may need to be converted into a three - dimensional grid for processing. Through the interpolation algorithm, the two - dimensional grid data in the casting process can be converted into the three - dimensional grid data required for the forging process. Interpolation algorithms (such as linear interpolation or bicubic interpolation) calculate the new position and value of each point according to the distance relationship between grid points, ensuring that the error is minimized during the spatial transformation of the data.

[0083] After the cross-process grid mapping is completed, error analysis and calibration processing need to be carried out on the mapping model. The goal of this stage is to ensure that the data transferred from one process to another is consistent in terms of space and physical meaning. The focus of error analysis is to identify any deviations that occur during the data transfer process and make adjustments through calibration algorithms. For example, during the grid mapping process from casting to forging, temperature data transfer errors may occur in some areas due to grid differences. At this time, it is necessary to analyze the differences in the temperature fields of the front and back processes, adjust the mapping relationship, and ensure the transfer accuracy of the data. The data transfer error compensation mechanism automatically corrects these errors by analyzing the differences between the simulation results and the actual production data.

[0084] The simulation interface software can automatically convert the calculation results of the upstream process into the initial conditions required for the downstream process, thus generating a seamless coupling link system between processes. Achieve precise data transfer between smelting, casting, forging, and heat treatment processes. For example, during the heat treatment process, it is necessary to control the heating and cooling rates based on the temperature and stress data obtained during the casting and forging processes. Through the simulation interface software, the system will automatically convert the temperature data during the casting process and the stress field data during the forging process into the initial temperature and stress conditions for the heat treatment process, so as to ensure that the heat treatment process can be accurately optimized according to the state of the previous process. Integrate the seamless coupling link system between processes into the full-process simulation software for casting, forging, and heat treatment to form a full-process collaborative simulation platform for hot processing. On the platform, through hardware acceleration and computing resource scheduling, the efficiency of simulation calculations is greatly improved. The simulation results can be real-time fed back to the production line, thus realizing the optimization of each link in the production process. Through parallel computing and efficient resource scheduling, the system can quickly process complex multi-physics problems under actual production conditions, ensuring that each process in the hot processing process can operate in the optimal state.

[0085] In a specific embodiment, the process of executing step S104 may specifically include the following steps:

[0086] Extract the structure and multi-physics field characteristic data of the forging equipment from the full-process collaborative simulation platform for hot processing, use geometric-physics field coupling-driven modeling to analyze the characteristic relationships, and obtain the basic data set for the large component manufacturing process;

[0087] Perform skin model processing on the large forging equipment structure in the basic data set, extract the key feature points and associated parameters on the equipment surface, and construct a digital model of the equipment;

[0088] Based on the digital model of the equipment, divide the design parameters of the large hot processing components into three levels: overall contour parameters, key component parameters, and process control parameters, and establish a hierarchical parameterized structure system;

[0089] Perform three-dimensional scanning and measurement on the hot processing components in the hierarchical parametric structure system, compare and calibrate the obtained geometric dimension and temperature field data with the simulation calculation results, and generate a time-varying feature database;

[0090] Input the time-varying feature database into the analysis modules for upsetting and drawing processes during the forging process of the hot processing components, calculate the variation laws of temperature, strain, and damage distribution, and generate process model data;

[0091] Apply the model accuracy evaluation index system to evaluate and optimize the accuracy, lightweight, and real-time performance of the process model data, and construct a digital twin model.

[0092] Specifically, extract the structural and multi-physical field feature data of the forging equipment from the full-process collaborative simulation platform. The structure of the forging equipment includes physical characteristics such as its geometric shape, material properties, and load distribution, while the multi-physical field features involve the temperature field, stress field, mechanical properties, etc. These data are collected by different sensors and simulation models and transmitted to the data processing platform. The geometric structure of the forging equipment may include the geometric information of components such as the main frame, oil cylinder, and connecting rod of the hydraulic press, while the physical field data such as the temperature field and stress field are sourced from real-time monitoring devices and simulation simulations. By analyzing these multi-physical field data, the thermal and mechanical response characteristics of each process stage can be obtained. Perform skin model processing on the structure of the large forging equipment. Skin model technology extracts the key surface feature points and associated parameters by analyzing the external characteristics of the equipment and the surface characteristics of key components. Taking a large hydraulic press as an example, the skin model can help identify key data points such as the positions of pressure sensors and heating areas on the outside of the hydraulic press. These data points can not only be used in the simulation model but also help monitor the working state of the forging equipment in real time. The key to the skin model is to convert the geometric data on the equipment surface into digital parameters for accurate modeling of the temperature and stress distribution on the equipment surface in the simulation.

[0093] Based on the above skin model, construct a digital model of the equipment, and divide the design parameters of the large hot processing components into three levels: overall contour parameters, key component parameters, and process control parameters. In the first stage, it is necessary to clarify the overall design of the forging equipment, such as the dimensions and shapes of the frame and oil cylinder (overall contour parameters). Secondly, focus on the design requirements of key components, such as the positions of pressure sensors and the temperature control range of the heating area (key component parameters). Finally, define the process control parameters, such as the temperature, speed, and pressure used during the forging process. Through these hierarchical structures, the equipment can be systematically modeled, and it can be ensured that the parameters at each level can be effectively utilized in subsequent process optimization. The parameters at each level not only reflect the geometric form of the equipment but also involve the functionality and working environment of the equipment.

[0094] For hot - working components in a hierarchical parametric structure system, three - dimensional scanning and measurement are carried out. The process obtains the geometric dimensions of the hot - working components through scanning equipment and calibrates the simulation calculation results in combination with the temperature field data. Three - dimensional scanning can accurately capture the geometric shape of the components and optimize the original model according to the scanned data. At the same time, the temperature field data can be used to monitor the temperature distribution on the surface of the components in real - time through temperature sensors and infrared measurement equipment. In this process, the geometric dimensions and temperature field data of the components are integrated and compared with the simulation calculation results to ensure that the simulation model can accurately reflect the real process conditions. By calibrating the three - dimensional scanning data and the simulation calculation results, a time - varying feature database is generated. The database records the geometric changes, temperature changes, and possible stress, deformation, etc. of the hot - working components at different time nodes. These time - varying feature data provide a dynamic basis for subsequent process control and optimization. For example, during the forging process, the changes in the geometric shape and temperature distribution of the components at different stages will affect the final forging result, and the time - varying feature database helps to track these changes and make real - time adjustments.

[0095] The time - varying feature database will be input into the forging process analysis module of the hot - working components. This module can calculate the change rules of temperature, strain, and damage distribution and generate process model data. In this process, the change rules of temperature, strain, and damage are not only obtained through numerical simulation calculations but also need to be optimized in combination with actual production data. For example, assume that during the forging process, the hot cracks caused by too high temperature in a certain part may affect the quality of the final product. The process model data can help identify potential quality problems and provide correction strategies by calculating these change rules.

[0096] Based on the process model data, a model accuracy evaluation index system is used to evaluate and optimize the accuracy, lightweight, and real - time performance to ensure that the generated digital twin model can operate efficiently and accurately in the actual production environment. In the process, the model accuracy evaluation index system helps to evaluate the deviation between the simulation results and the actual data to adjust the model parameters and ensure the accuracy of the model. At the same time, the lightweight and real - time performance evaluation helps to ensure that the model can operate efficiently during the production process on the premise of ensuring accuracy, avoiding waste of computing resources and excessive computing delays.

[0097] In a specific embodiment, the process of executing step S105 may specifically include the following steps:

[0098] Deploy the digital twin model to the edge - computing gateway between the device side and the digital system, construct a data channel supporting 5G or wired transmission, and generate an edge - computing service framework;

[0099] Distribute and process the data of key equipment, production processes, and workshop environments collected by the edge computing service framework, establish a real-time mapping matrix between virtual and physical data, and generate a mechanism for virtual-real data interaction;

[0100] Based on the virtual-real data interaction mechanism, construct a multi-level situation awareness model including equipment layer situation, process layer situation, and safety layer situation, and generate a real-time monitoring data stream for the production process;

[0101] Apply fast scanning and multi-dimensional modeling analysis algorithms to the quality characteristics of the hot processing process in the real-time monitoring data stream of the production process to generate a precise analysis dataset for forging allowances;

[0102] Based on the precise analysis dataset for forging allowances, construct an integrated control system including a human-machine remote control module, an interactive operation control module, and a production safety management module, and generate safety warning indicators for the manufacturing process;

[0103] Integrate the safety warning indicators of the manufacturing process and the real-time monitoring data stream of the production process into a unified interface framework through cloud configuration technology, optimize the parameters of the edge computing gateway according to cloud policies, and generate a visual management and control system.

[0104] Specifically, through the deployment of the edge computing gateway, the digital twin model enables data collection and processing to be carried out close to the data source, thereby significantly improving the response speed and computing efficiency. This process is to enable the system to have fast real-time data processing capabilities and enhance the intelligent decision-making capabilities during the production process. A reliable data transmission channel is established between the device side and the digital system to support the real-time transmission of a large amount of production data through 5G or wired networks. The introduction of 5G technology ensures low-latency and high-bandwidth data transmission, enabling quick response to the data provided by various sensors and devices during the production process. The device side, through the edge computing gateway, can obtain in real-time the data of key equipment, production processes, and workshop environments, such as equipment status, production parameters, and environmental conditions. These data are distributed and processed through the edge computing service framework to ensure the fast storage and analysis of data.

[0105] The edge computing service framework processes the collected key data and establishes a real-time mapping matrix between virtual and physical data. This mapping matrix docks the production data in the physical world with the data in the virtual model, realizing the interaction of virtual-real data. For example, the temperature, pressure, and other data collected in real-time by sensors will be compared with the physical fields (such as stress fields, temperature fields, etc.) in the digital twin model through the edge computing platform, enabling the virtual model to be updated in real-time to reflect the status of each link in the production process. The interaction mechanism ensures that all data during the production process can be accurately fed back into the virtual model.

[0106] Based on the virtual-real data interaction mechanism, a multi-level situation awareness model is further constructed. This model includes the device layer situation, the process layer situation, and the safety layer situation. The device layer situation focuses on the operating status of key devices, including temperature, pressure, load, etc.; the process layer situation focuses on various process parameters and actual performance during the production process, such as temperature field, stress field, deformation, etc.; the safety layer situation focuses on the safety of the workshop environment and operations, including personnel contact, equipment overload, environmental changes, etc. By integrating the situation awareness of these three layers, the system can monitor the status changes of each link in the production process in real time and feedback them to the process adjustment or safety management module in a timely manner. The real-time monitoring data stream during the production process is processed and analyzed through data stream processing and analysis algorithms to deeply analyze the quality characteristics of the hot processing process. During the process, fast scanning technology and multi-dimensional modeling analysis algorithms are used to extract and analyze the quality data during the forging process. These technologies can efficiently identify the key quality characteristics in the hot processing process, such as surface defects and dimensional deviations of forgings. By processing the data stream, the system can generate a precise analysis dataset of forging allowances in real time, analyze the actual allowances and forming errors of forgings, and provide data support for subsequent process optimization and adjustment.

[0107] Based on the precise analysis dataset of forging allowances, the system further constructs a fusion control system including a human-machine remote control module, an interactive operation control module, and a production safety management module. The core of the control system is to improve the operation flexibility and safety during the production process through intelligent algorithms and real-time data analysis. The human-machine remote control module allows operators to remotely monitor and control equipment for real-time adjustment and emergency handling; the interactive operation control module makes automatic adjustments according to the real-time collected data during the production process to optimize process parameters; the production safety management module prevents the occurrence of safety hazards, such as overheating, equipment failures, or operation errors, through real-time monitoring of the workshop environment and equipment status. Through the close cooperation of these modules, the smooth progress of the production process can be ensured and safety can be guaranteed.

[0108] To improve the intelligent level, safety warning indicators for the manufacturing process are generated and combined with the real-time monitoring data stream of the production process. Through cloud configuration technology, the system can integrate these warning indicators and real-time data streams into a unified interface framework. This enables operators to quickly obtain safety warning information during the production process and make timely decisions based on this information. The cloud strategy optimizes the parameters of the edge computing gateway, enabling the system to make dynamic adjustments according to different production requirements and production statuses, thereby realizing the intelligent scheduling of process parameters and resources during the production process.

[0109] In this way, the constructed visual control system can display all key data in real time and provide decision-making support. The data and warning information of each link are integrated and displayed through cloud configuration technology, enabling operators and managers to view, analyze, and make decisions on a unified platform. This integrated visual management not only improves production efficiency but also reduces risks during the production process, ensuring product quality and production safety.

[0110] For example, during the forging production process, the system collects real-time data such as the temperature and stress of forgings through an edge computing gateway and compares this data with the temperature field and stress field in the digital twin model in real time. Suppose the surface temperature of a forging shows an abnormality during the processing, exceeding the preset safety range. The system can immediately detect the abnormality through the virtual-real data interaction mechanism and issue an alarm in the visual control system. The operator can adjust the heating process of the equipment through the remote control module to ensure that the temperature of the forging does not rise further and avoid affecting product quality. At the same time, the system will automatically adjust the equipment parameters during the production process according to the data in the safety layer situation to ensure that the entire production process can proceed safely and efficiently.

[0111] In a specific embodiment, the process of executing step S106 may specifically include the following steps:

[0112] Extract the process parameters and quality inspection data of each process running in real time from the visual control system to construct a real-time process execution data set;

[0113] Apply a parameter evaluation algorithm based on knowledge reasoning and deep neural networks to the real-time process execution data set, calculate the deviation matrix between the current process state and the target quality, and generate multi-dimensional process risk assessment indicators;

[0114] According to the multi-dimensional process risk assessment indicators, extract historical optimization cases from the process database, and apply the gradient boosting decision tree method to construct a non-linear mapping relationship between process parameters and quality characteristics to generate a multi-round distributed process optimization engine;

[0115] Set multi-dimensional optimization goals such as the best quality, minimum energy consumption, and shortest cycle in the multi-round distributed process optimization engine, and generate a candidate process parameter combination scheme set through parallel iterative optimization calculation;

[0116] Conduct virtual tests on the candidate process parameter combination scheme set through the digital twin model, calculate the quality prediction scores and resource consumption indicators of each scheme, and determine the target process parameter combination;

[0117] Based on the target process parameter combination, apply the multi-attribute analysis method to comprehensively evaluate the importance of the workpiece, process difficulty, and production urgency, calculate the priority allocation matrix of equipment and resources, and generate a limited resource priority allocation scheme.

[0118] Specifically, process parameters and quality inspection data of each process running in real time are extracted from the visualization control system. The data includes process parameters such as real-time temperature, pressure, speed, displacement measured by various sensors, and quality inspection data such as dimensions, surface quality, hardness, etc. of forgings. These data are collected in real time and converted into a structured format to form a real-time process execution data set. The real-time process execution data set is a complete record of all key parameters and quality characteristics in the production process. Through the data collection and real-time monitoring system, the accuracy and timeliness of the data are ensured.

[0119] For the real-time process execution data set, a parameter evaluation algorithm based on knowledge reasoning and deep neural network is applied to calculate the deviation matrix between the current process state and the target quality. Evaluate the gap between the current process state (e.g., current temperature, pressure, etc.) and the target quality (such as expected forging dimensions, surface quality, etc.). The deep neural network (DNN) algorithm builds a relationship model between process parameters and quality characteristics by training a large amount of historical data, and helps to evaluate the deviation between the current state and the ideal target. For example, if the hardness requirement of the target forging is X, but the current production data shows that the actual hardness is Y, DNN will calculate the deviation between X and Y and generate a deviation matrix based on this. The deviation matrix not only reflects the gap between the current state and the target.

[0120] Based on the deviation matrix, multi-dimensional process risk assessment indicators are generated. The indicators comprehensively consider the impact of process parameters such as temperature, pressure, and time on quality, as well as risks brought by factors such as equipment and personnel. The multi-dimensional process risk assessment indicators provide a comprehensive risk assessment framework for process optimization through a comprehensive analysis of the process state, helping to identify links that may lead to unqualified product quality or low production efficiency. The process not only focuses on the current situation of the process, but also gives early warnings of potential risks in the production process, providing decision-making support for subsequent optimization. Extract historical optimization cases from the process database and apply the gradient boosting decision tree (GBDT) method to construct a non-linear mapping relationship between process parameters and quality characteristics. Historical optimization cases refer to successful cases recorded in previous production processes, including various process parameters and corresponding quality output results. By extracting these historical data, the GBDT model can establish a decision model reflecting the non-linear relationship between complex processes and quality. For example, in the casting process, GBDT can help determine the complex relationship between temperature and the final casting quality and provide decision-making support for new production tasks. The GBDT model will evaluate the impact of different process parameters in the current production process based on the experience of historical data and generate an optimization model.

[0121] The multi-round distributed process optimization engine performs optimization calculations based on multi-dimensional goals such as target quality, energy consumption, and production cycle through parallel computing and iterative optimization. The multi-round optimization algorithm can continuously approach the optimal solution through multiple iterative calculations according to predefined goals (such as optimal quality, minimum energy consumption, shortest cycle, etc.). For example, during the forging process, the process optimization engine can perform a series of simulation calculations based on historical optimization data to generate multiple candidate process parameter combinations, such as combinations of different temperatures, pressures, and deformation rates. Finally, the system selects the optimal process parameter combination to ensure the quality and efficiency of the production process.

[0122] Virtual tests are performed on the set of candidate process parameter combination plans. Through the digital twin model, the system simulates each candidate plan and calculates the quality prediction scores and resource consumption indicators of each plan. The digital twin model reproduces the actual production process through simulation technology, calculates indicators such as the quality, dimensions, and surface quality of the product under each candidate process plan, and at the same time evaluates the resource consumption such as energy and time required for each plan. The purpose of this test step is to compare the theoretical results of each process plan with the actual production results to ensure that the selected process parameter combination can not only achieve the quality goal but also minimize resource waste and production cycle.

[0123] Based on the target process parameter combination, the multi-attribute analysis method is applied to comprehensively evaluate the importance, process difficulty, and production urgency of the workpiece, calculate the priority allocation matrix of equipment and resources, and generate a priority allocation plan for limited resources. During the production process, resources (such as equipment, raw materials, time, etc.) are limited, so it is necessary to reasonably allocate resources according to the needs of different workpieces. Through multi-attribute analysis, the system can evaluate the priority of each workpiece. For example, for a certain production batch of forgings, perhaps its quality requirements are particularly high, or its production urgency is relatively large, and the system will allocate more production resources to these high-priority workpieces. The optimized allocation plan ensures the maximum utilization of limited resources and also guarantees that key workpieces can be completed on time and meet the quality standards.

[0124] For example, assume that in a certain production batch, through the deep neural network algorithm, it is found that there are large deviations in the hardness and dimensions of the current forgings (the deviation matrix shows high risk). Based on this, the process optimization engine proposes three different process parameter combinations, which adjust the temperature, pressure, and deformation rate respectively. Through the virtual test of the digital twin model, it is found that the forgings of the second process combination have the best quality but higher energy consumption. After multi-attribute analysis, considering the production urgency and equipment load, the system finally selects the third process combination, which can ensure the quality while having lower energy consumption and meeting the production cycle requirements. Through multi-dimensional optimization and decision support, the system realizes the reasonable allocation of resources and ensures the dual optimization of production efficiency and product quality.

[0125] The above describes the digital-twin-based hot processing control method in the embodiments of the present application. Next, the digital-twin-based hot processing control system in the embodiments of the present application will be described. Please refer to Figure 2 , an embodiment of the digital-twin-based hot processing control system in the embodiments of the present application includes:

[0126] An acquisition module 201, configured to collect process parameters and quality data for the smelting, casting, forging, and heat treatment processes using a multi-sensor fusion Internet of Things data acquisition system, and obtain a hot processing full-process data set;

[0127] A construction module 202, configured to construct a multi-dimensional process data middle platform for hot processing according to the hot processing full-process data set;

[0128] A conversion module 203, configured to, based on the multi-dimensional process data middle platform for hot processing, perform parameter mapping and grid conversion on the multi-stage simulation results through an adaptive interface algorithm, and establish a hot processing full-process collaborative simulation platform;

[0129] A modeling module 204, configured to perform hierarchical parametric modeling based on the hot processing full-process collaborative simulation platform to construct a digital twin model;

[0130] A perception module 205, configured to deploy the digital twin model to an edge computing gateway, and generate a visual control system through a virtual-real data interaction mechanism and multi-level situation awareness;

[0131] An input module 206, configured to input the real-time data of the manufacturing process collected by the visual control system into a multi-round distributed process optimization engine to calculate the target process parameter combination and the limited resource priority allocation scheme.

[0132] Through the collaborative cooperation of the above-mentioned various components, the Internet of Things data acquisition system based on multi-sensor fusion can collect the process parameters and quality data of the smelting, casting, forging, and heat treatment processes in real time, and construct a full-process dataset for hot processing. It can achieve comprehensive monitoring of the entire production process. Through the adaptive interface algorithm, the system can perform accurate parameter mapping and mesh conversion on the multi-stage simulation results to ensure seamless data connection across processes, thereby realizing the collaborative simulation of the full process of hot processing. This improves the collaboration efficiency and data consistency between processes. The construction of the hierarchical parametric modeling and digital twin model enables the process to reflect the actual production situation in real time. The deployment of the digital twin model to the edge computing gateway, combined with the virtual-real data interaction mechanism and multi-level situation awareness, not only realizes efficient data processing and real-time monitoring, but also enables each link in the production process to obtain accurate feedback. This improves the intelligent level and automation ability of the hot processing process, reduces manual intervention, and optimizes production efficiency and quality. Using the multi-round distributed process optimization engine, it is possible to optimize the calculation of process parameters and generate a priority allocation plan for limited resources, thereby ensuring the optimal balance between production efficiency and product quality under limited resources.

[0133] Above Figure 2 From the perspective of modular functional entities, the digital-twin-based hot processing control system in the embodiments of the present invention is described in detail. Next, the digital-twin-based hot processing control equipment in the embodiments of the present invention is described in detail from the perspective of hardware processing.

[0134] Figure 3 FIG. is a schematic structural diagram of a digital-twin-based hot processing control equipment provided by an embodiment of the present invention. The digital-twin-based hot processing control equipment 300 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 digital-twin-based hot processing control equipment 300. 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 digital-twin-based hot processing control equipment 300 to implement the steps of the above-mentioned digital-twin-based hot processing control method.

[0135] The digital twin-based hot processing control device 300 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, etc. Those skilled in the art can understand that Figure 3 The structure of the digital twin-based hot processing control device shown does not constitute a limitation on the digital twin-based hot processing control device provided by the present invention. It may include more or fewer components than shown, or combine certain components, or have different component arrangements.

[0136] The present invention also provides a computer-readable storage medium. The computer-readable storage medium 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 digital twin-based hot processing control method.

[0137] Those skilled in the art can clearly understand that for the convenience and brevity 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.

[0138] 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 this 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 to cause a digital twin-based hot processing control device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in 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.

[0139] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it; 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 described in the foregoing embodiments, or perform equivalent replacements on some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims

1. A hot processing control method based on digital twin, characterized in that The method includes: Using an Internet of Things data acquisition system with multi-sensor fusion to collect process parameters and quality data for the smelting, casting, forging, and heat treatment processes, obtaining a full-process dataset of hot processing; Constructing a multi-dimensional process data center for hot processing based on the full-process dataset of hot processing; Based on the multi-dimensional process data center for hot processing, through an adaptive interface algorithm, performing parameter mapping and grid conversion on the multi-stage simulation results to establish a full-process collaborative simulation platform for hot processing, including: calling the material thermal physical property parameters and process condition parameters of the casting, forging, and heat treatment processes from the multi-dimensional process data center for hot processing to construct independent simulation models for each process; performing data extraction and structured processing on the calculation results of the independent simulation models for each process to generate a process simulation dataset including temperature field, stress field, microstructure field, and defect distribution; based on the process simulation dataset, using a spatial interpolation algorithm to convert the upstream process simulation grid point data into the grid form required for downstream process simulation, obtaining a cross-process grid mapping model; performing error analysis and calibration processing on the cross-process grid mapping model to determine the error range of each parameter transfer and establishing a data transfer error compensation mechanism; according to the data transfer error compensation mechanism, through simulation interface software, realizing the automatic conversion of the upstream process calculation results to the initial conditions of the downstream process and generating a seamlessly coupled link system between processes; integrating the link system into the full-process simulation software of casting, forging, and heat treatment, and generating a full-process collaborative simulation platform for hot processing through hardware acceleration and computing resource scheduling integration; Performing hierarchical parametric modeling based on the full-process collaborative simulation platform for hot processing to construct a digital twin model; Deploying the digital twin model to an edge computing gateway, and generating a visual management and control system through a virtual-real data interaction mechanism and multi-level situation awareness; Inputting the real-time data of the manufacturing process collected by the visual management and control system into a multi-round distributed process optimization engine to calculate the target process parameter combination and the priority allocation scheme of limited resources.

2. The digital-twin-based hot processing control method according to claim 1, wherein, The step of using an Internet of Things data acquisition system with multi-sensor fusion to collect process parameters and quality data for the smelting, casting, forging, and heat treatment processes, obtaining a full-process dataset of hot processing, includes: Extracting smelting temperature, pressure, and composition data from the configuration software and PLC controller using the OPC communication method, and collecting the process parameters of the stock preparation, rough smelting, refining, vacuum, up-casting, and down-casting processes in the smelting process to obtain smelting process parameter data; Accessing the ingot casting on-site instrument equipment through the Modbus protocol, collecting and digitally processing the process requirement data, process execution data, and auxiliary material data to obtain casting process parameter data; Performing multi-sensor access transformation on large oil presses, presses, and supporting forging heating equipment, collecting the temperature field, stress field, and deformation field data during the forging process, and generating forging process parameter data; Digitally processing the quality inspection samples of the heat treatment process based on image recognition and machine learning technologies, converting the microstructure characteristics and mechanical property characteristics into numerical representations, and constructing a heat treatment quality dataset. Perform time series tagging and correlation analysis on the smelting process parameter data, the casting process parameter data, the forging process parameter data, and the heat treatment quality data set, and construct a process chain data structure with process inheritance relationships; Perform outlier detection and data cleaning on the process chain data structure, and generate a full-process hot processing data set through standardization processing and feature extraction.

3. The digital-twin-based hot processing control method according to claim 1, characterized in that, Construct a multi-dimensional process data middle platform for hot processing based on the full-process hot processing data set, including: Perform data cleaning, trend item separation, and noise suppression processing on the full-process hot processing data set, extract the main process features, and obtain the preprocessed hot processing data; According to the preprocessed hot processing data, design a hierarchical data structure including a metadata management layer, a task management layer, a data conversion layer, and a view management layer, and construct a multi-level data storage architecture; Based on the multi-level data storage architecture, classify and store the preprocessed hot processing data according to smelting, casting, forging, and heat treatment process types, and generate a professional process database module; Perform working condition segmentation and standardization processing on the data in the professional process database module, analyze the mapping relationship between material-process-structure-property, and establish a multi-dimensional data association model; Utilize the multi-dimensional data association model to generate process data service components through a standardized data service interface based on middleware; Combine the functions of the process data service components through service orchestration and a rules engine to construct a multi-dimensional process data middle platform for hot processing with data management, process parameter retrieval, and process association analysis functions.

4. The digital twin-based hot processing control method according to claim 1, wherein Perform hierarchical parametric modeling based on the full-process hot processing collaborative simulation platform to construct a digital twin model, including: Extract the forging equipment structure and multi-physical field characteristic data from the full-process hot processing collaborative simulation platform, and use geometric-physical field coupling-driven modeling to analyze the characteristic relationships to obtain the basic data set for the large component manufacturing process; Perform skin model processing on the large forging equipment structure in the basic data set, extract the key feature points and associated parameters on the equipment surface, and construct an equipment digital model. The skin model extracts the key feature points and associated parameters on the surface by analyzing the external features of the equipment and the surface characteristics of the key components. Among them, the skin model converts the geometric data on the equipment surface into digital parameters; Based on the equipment digital model, divide the design parameters of the large hot processing component into three levels: overall contour parameters, key component parameters, and process control parameters, and establish a hierarchical parametric structure system; Perform three-dimensional scanning and measurement on the hot processing components in the hierarchical parametric structure system, compare and calibrate the obtained geometric dimensions and temperature field data with the simulation calculation results, and generate a time-varying feature database; Input the time-varying feature database into the upsetting and drawing process analysis module of the hot processing component forging process, calculate the change laws of temperature, strain, and damage distribution, and generate process model data; Apply the model accuracy evaluation index system to the process model data for accuracy, lightweight, and real-time evaluation and optimization, and construct a digital twin model.

5. The digital twin-based hot processing control method according to claim 1, wherein Deploying the digital twin model to an edge computing gateway, and generating a visual control system through a virtual-real data interaction mechanism and multi-level situation awareness, including: Deploying the digital twin model to an edge computing gateway between the device side and the digital system, constructing a data channel supporting 5G or wired transmission, and generating an edge computing service framework; Distributively processing the data of key equipment, production process, and workshop environment collected by the edge computing service framework, establishing a real-time mapping matrix of virtual and physical data, and generating a virtual-real data interaction mechanism; Based on the virtual-real data interaction mechanism, constructing a multi-level situation awareness model including equipment layer situation, process layer situation, and security layer situation, and generating a real-time monitoring data stream of the production process; Applying a fast scanning and multi-dimensional modeling analysis algorithm to the quality characteristics of the hot working process in the real-time monitoring data stream of the production process, and generating a precise analysis dataset of forging allowance; Based on the precise analysis dataset of forging allowance, constructing a fusion control system including a human-machine remote control module, an interactive operation control module, and a production safety management module, and generating a safety warning index for the manufacturing process; Integrating the safety warning index of the manufacturing process and the real-time monitoring data stream of the production process into a unified interface framework through cloud configuration technology, and optimizing the parameters of the edge computing gateway according to the cloud strategy, and generating a visual control system.

6. The digital twin-based hot processing control method according to claim 1, wherein Inputting the real-time data of the manufacturing process collected by the visual control system into a multi-round distributed process optimization engine to calculate the target process parameter combination and the limited resource priority allocation scheme, including: Extracting the process parameters and quality inspection data of each process running in real time from the visual control system, and constructing a real-time process execution dataset; Applying a parameter evaluation algorithm based on knowledge reasoning and deep neural network to the real-time process execution dataset, calculating the deviation matrix between the current process state and the target quality, and generating a multi-dimensional process risk assessment index; According to the multi-dimensional process risk assessment index, extracting historical optimization cases from the process database, and applying the gradient boosting decision tree method to construct a non-linear mapping relationship between process parameters and quality characteristics, and generating a multi-round distributed process optimization engine; Setting multi-dimensional optimization goals such as optimal quality, minimum energy consumption, and shortest cycle in the multi-round distributed process optimization engine, and generating a candidate process parameter combination scheme set through parallel iterative optimization calculation; Virtually testing the candidate process parameter combination scheme set through the digital twin model, calculating the quality prediction score and resource consumption index of each scheme, and determining the target process parameter combination; Based on the target process parameter combination, applying a multi-attribute analysis method to comprehensively evaluate the importance of the workpiece, the process difficulty, and the production urgency, calculating the priority allocation matrix of equipment and resources, and generating a limited resource priority allocation scheme.

7. A hot processing control system based on digital twin, characterized in that For implementing the digital twin-based hot working control method according to any one of claims 1-6, the digital twin-based hot working control system includes: The acquisition module is used to collect process parameters and quality data for the smelting, casting, forging, and heat treatment processes using an Internet of Things data acquisition system with multi-sensor fusion, obtaining a full-process dataset of hot processing. The construction module is used to construct a multi-dimensional process data center for hot processing based on the full-process dataset of hot processing. The conversion module is used to, based on the multi-dimensional process data center for hot processing, perform parameter mapping and grid conversion on the multi-stage simulation results through an adaptive interface algorithm to establish a full-process collaborative simulation platform for hot processing, including: calling the material thermal property parameters and process condition parameters of the casting, forging, and heat treatment processes from the multi-dimensional process data center for hot processing to construct independent simulation models for each process; extracting and structuring the calculation results of the independent simulation models for each process to generate a process simulation dataset containing the temperature field, stress field, microstructure field, and defect distribution; based on the process simulation dataset, using a spatial interpolation algorithm to convert the upstream process simulation grid point data into the grid form required for downstream process simulation to obtain a cross-process grid mapping model; performing error analysis and calibration processing on the cross-process grid mapping model to determine the error range of each parameter transfer and establishing a data transfer error compensation mechanism; according to the data transfer error compensation mechanism, through simulation interface software, realizing the automatic conversion of the upstream process calculation results to the initial conditions of the downstream process to generate a seamlessly coupled link system between processes; integrating the link system into the full-process simulation software for casting, forging, and heat treatment, and generating a full-process collaborative simulation platform for hot processing through hardware acceleration and computing resource scheduling integration. The modeling module is used to perform hierarchical parametric modeling based on the full-process collaborative simulation platform for hot processing to construct a digital twin model. The perception module is used to deploy the digital twin model to an edge computing gateway and generate a visual management and control system through a virtual-real data interaction mechanism and multi-level situation awareness. The input module is used to input the real-time data of the manufacturing process collected by the visual management and control system into a multi-round distributed process optimization engine to calculate the target process parameter combination and the priority allocation scheme of limited resources.

8. A hot 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. When the processor executes the computer program, it implements the digital twin-based hot processing management and control method according to any one of claims 1 to 7.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is run by the processor, it causes the processor to execute the digital twin-based hot processing management and control method according to any one of claims 1 to 7.

Images