Monitoring method for food puffing and drying process based on digital twinning technology

Abstract

The application relates to a food puffing and drying process monitoring method based on a digital twin technology, wherein the method comprises the following steps: constructing a virtual twin model of a device and a product; establishing an initial simulation model through an experience model and a mechanism model to simulate and predict the device running state, the material state and the product quality index under different input parameters; collecting device process data and product quality data and importing the data into a server; integrating a three-dimensional model and the simulation model, predicting the possible output of the next step and visually displaying the output; meanwhile, setting a machine learning model in the server, correcting the model by comparing the simulation results of the initial model with the actual operation monitoring data of the physical world, so that the iteration and improvement of the twin are realized; on the basis, introducing a large language model as an automatic, intelligent and real-time interaction means between users, entity devices and twins, so that the monitoring efficiency can be further improved.

Description

Technical Field

[0001] This application relates to the field of extruder process monitoring technology, and in particular to a method for monitoring food extrusion and drying processes based on digital twin technology. Background Technology

[0002] Currently, most food extrusion process monitoring schemes are based on empirical models (EM), which rely heavily on the experience of technicians and operators for design and improvement. This process involves numerous approximations and assumptions, resulting in the following drawbacks:

[0003] 1. Low accuracy and unstable decision success rate: Because existing process monitoring methods based on mechanistic and empirical models rely on human operators' decisions based on experience and subjective judgment, the accuracy of parameter settings, quality problem warnings, and fault warnings during process monitoring is often low. Furthermore, human monitoring is easily affected by the operator's experience level and subjective factors. Different operators may sometimes make different judgments and reactions to the same process indicators and parameters. When the same operator faces new working conditions, new products, or fails to observe the changes in relevant parameters or indicators in a timely manner, biases in judgment or decision-making are likely to occur.

[0004] 2. Poor real-time early warning: During manual monitoring, when process parameters or product indicators deviate, on-site operators and process technicians still need to wait for a certain period of time to collect data, make judgments and decisions. The reaction speed is slow, and sometimes some problems that are not easy to detect are easily overlooked, which can cause a lot of material and product losses in the process.

[0005] 3. Human resource risks caused by reliance on specific employees in production management: In traditional methods, the effectiveness of process monitoring is highly dependent on whether technical personnel have sufficient experience and knowledge, whether they have enough skills to correctly design key components (such as export mold die plates) and to quickly and correctly operate (such as start-up, shutdown operations and correct troubleshooting when encountering faults or quality problems) and adjust parameters, etc. Moreover, due to the complexity and ambiguity of the model, it is difficult to accurately transfer knowledge and skills.

[0006] 4. Poor intuitiveness: Traditional control methods cannot simulate and display the material flow and changes, temperature distribution, and pressure distribution in the equipment in a real-time and intuitive manner. There are also few real-time analysis and simulation prediction results of equipment configuration. As a result, relevant control personnel can only understand and estimate the process status based on the combination of abstract process parameter data and the products collected on site. To a certain extent, this increases the difficulty of process monitoring and the difficulty of making correct judgments and actions.

[0007] 5. Inability to quickly collect and process large amounts of data: A large amount of data is generated during the extrusion process of pet food. Manually collecting, processing and analyzing data is time-consuming and costly, and important data is easily lost.

[0008] 6. Historical experience and data may not be applicable to process monitoring of specific equipment and products: The experience accumulated by technicians on other equipment over a long period of time may not be applicable to new equipment, because the wear and tear of each piece of equipment is different and the product design is different. Often, the process parameters and key equipment designs need to be recalculated and improved. When relying on past experience to monitor the process of new equipment and new products, deviations are likely to occur, resulting in waste or quality problems.

[0009] In summary, existing methods for monitoring food puffing processes have poor accuracy, real-time early warning capabilities, and intuitive monitoring. They also rely heavily on specific employees, making it difficult to quickly collect and process large amounts of data. These limitations in application scenarios urgently need to be addressed. Summary of the Invention

[0010] This application provides a monitoring method for food puffing and drying processes based on digital twin technology, which solves the problems of poor accuracy, real-time early warning and intuitive monitoring of existing food puffing process monitoring methods, as well as their reliance on specific employees, difficulty in quickly collecting and processing large amounts of data, and significant limitations in application scenarios.

[0011] The first aspect of this application provides a method for monitoring food puffing and drying processes based on digital twin technology, comprising the following steps: constructing a virtual twin model of a target device, and collecting equipment process data of the target device and product quality data of the target product, and preprocessing the equipment process data and the product quality data to obtain standard equipment process data and standard product quality data; constructing a mechanistic model and an empirical model corresponding to the target device, and establishing a simulation model of the target device through the mechanistic model and the empirical model, and collecting expert technical information corresponding to the target device through a pre-constructed large language model, and inputting the expert technical information, the standard equipment process data and the standard product quality data into a target server; integrating the virtual twin model and the simulation model, and combining the standard equipment process data and the standard product quality data to predict the equipment operating status information of the target device; constructing a target machine learning model of the virtual twin in the target server, and collecting actual operation monitoring data of the target device, and correcting the target machine learning model through the equipment operating status information and the actual operation monitoring data, so as to use the corrected target machine learning model to monitor the food puffing and drying processes of the target device and the target product.

[0012] Optionally, in one embodiment of this application, while monitoring the food puffing and drying process of the target equipment and the target product using the modified target machine learning model, the method further includes: generating a monitoring report corresponding to the target equipment and the target product; constructing multiple target intelligent agents through the large language model; and establishing a two-way information interaction system based on the multiple target intelligent agents; sending the monitoring report to the target user through the two-way information interaction system; and the target user providing visualized remote operation guidance for the puffing and drying process of the target equipment and the target product based on the monitoring report and through the two-way information interaction system and a preset AR strategy.

[0013] Optionally, in one embodiment of this application, the step of constructing a virtual twin model of the target device, collecting equipment process data of the target device and product quality data of the target product, and preprocessing the equipment process data and the product quality data to obtain standard equipment process data and standard product quality data includes: constructing a virtual twin model of the target device using a preset Unity3D, and establishing a communication connection between the target PLC and the target host computer, so that the virtual twin model can obtain the equipment process data and the product quality data through the target PLC; performing data cleaning and data normalization processing on the equipment process data and the product quality data to obtain the standard equipment process data and the standard product quality data; integrating the standard equipment process data and the standard product quality data through the large language model to obtain integrated natural language information, and displaying the integrated natural language information to the target user in the display interface of the virtual twin model, so that the target user can interact bidirectionally with multiple target intelligent agents constructed by the large language model based on the integrated natural language information.

[0014] Optionally, in one embodiment of this application, constructing the mechanism model and empirical model corresponding to the target device includes: performing physical and chemical process analysis on the target device to obtain analysis results, and constructing a mechanism model of the target device based on the analysis results; obtaining empirical information and actual production data of the target device, and constructing the empirical model based on the mechanism model, the empirical information and the actual production data.

[0015] Optionally, in one embodiment of this application, the step of integrating the virtual twin model and the simulation model, and combining the standard equipment process data and the standard product quality data, to predict the equipment operating status information of the target equipment includes: predicting the equipment operating status information based on the standard equipment process data, the standard product quality data, or simulation parameters customized by the target user through the virtual twin model, and combining the virtual twin model and the simulation model.

[0016] Optionally, in one embodiment of this application, the step of constructing the target machine learning model of the virtual twin in the target server and collecting actual operation monitoring data of the target device to correct the target machine learning model through the device operation status information and the actual operation monitoring data includes: training the target machine learning model through the actual operation monitoring data, and using a preset model evaluation index to evaluate the performance of the trained target machine learning model to obtain a performance evaluation result; and correcting the target machine learning model based on the performance evaluation result and a preset expert experience analysis model.

[0017] A second aspect of this application provides a monitoring device for food puffing and drying processes based on digital twin technology, comprising: a first construction module for constructing a virtual twin model of a target device, collecting equipment process data of the target device and product quality data of the target product, and preprocessing the equipment process data and the product quality data to obtain standard equipment process data and standard product quality data; and a second construction module for constructing a mechanism model and an empirical model corresponding to the target device, so as to establish a simulation model of the target device through the mechanism model and the empirical model, and collecting expert technical information corresponding to the target device through a pre-constructed large language model, so as to integrate the expert technical information with the target device. The technical information, the standard equipment process data, and the standard product quality data are input into the target server; the integration module is used to integrate the virtual twin model and the simulation model, and combine the standard equipment process data and the standard product quality data to predict the equipment operating status information of the target equipment; the correction module is used to construct the target machine learning model of the virtual twin in the target server, and collect the actual operation monitoring data of the target equipment, so as to correct the target machine learning model through the equipment operating status information and the actual operation monitoring data, so as to use the corrected target machine learning model to monitor the food puffing and drying process of the target equipment and the target product.

[0018] Optionally, in one embodiment of this application, it further includes: a generation module, configured to generate a monitoring report corresponding to the target equipment and the target product while monitoring the food puffing and drying process of the target equipment and the target product using the modified target machine learning model, and to construct multiple target intelligent agents through the large language model, and to establish a two-way information interaction system based on the multiple target intelligent agents; a large language model interaction module, configured to send the monitoring report to the target user through the two-way information interaction system, and the target user to perform visual remote operation guidance on the puffing and drying process of the target equipment and the target product based on the monitoring report, through the two-way information interaction system and a preset AR strategy.

[0019] Optionally, in one embodiment of this application, the first construction module includes: a communication unit, configured to construct a virtual twin model of the target device using a preset Unity3D, and establish a communication connection between the target PLC and the target host computer, so that the virtual twin model can obtain the equipment process data and the product quality data through the target PLC; a preprocessing unit, configured to perform data cleaning and data normalization processing on the equipment process data and the product quality data to obtain the standard equipment process data and the standard product quality data; and an interaction unit, configured to integrate the standard equipment process data and the standard product quality data through the large language model to obtain integrated natural language information, and display the integrated natural language information to the target user in the display interface of the virtual twin model, so that the target user can interact bidirectionally with multiple target intelligent agents constructed by the large language model based on the integrated natural language information.

[0020] Optionally, in one embodiment of this application, the second construction module includes: an analysis unit, configured to perform physical and chemical process analysis on the target device to obtain analysis results, and construct a mechanism model of the target device based on the analysis results; and an acquisition unit, configured to acquire empirical information and actual production data of the target device, and construct the empirical model based on the mechanism model, the empirical information, and the actual production data.

[0021] Optionally, in one embodiment of this application, the integration module includes: a prediction unit, used to predict the equipment operating status information based on the standard equipment process data, the standard product quality data, or simulation parameters customized by the target user through the virtual twin model, and in combination with the virtual twin model and the simulation model.

[0022] Optionally, in one embodiment of this application, the correction module includes: a training unit, configured to train the target machine learning model using the actual operation monitoring data, and to evaluate the performance of the trained target machine learning model using a preset model evaluation index to obtain a performance evaluation result; and an optimization unit, configured to correct the target machine learning model based on the performance evaluation result and a preset expert experience analysis model.

[0023] A third aspect of this application provides an electronic device, including: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the monitoring method for food puffing and drying processes based on digital twin technology as described in the above embodiments.

[0024] A fourth aspect of this application provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described monitoring method for food puffing and drying processes based on digital twin technology.

[0025] A fifth aspect of this application provides a computer program product, including a computer program that is executed to implement the above-described monitoring method for food puffing and drying processes based on digital twin technology.

[0026] Therefore, the embodiments of this application have the following beneficial effects:

[0027] The embodiments of this application can construct a virtual twin model of the target equipment, collect equipment process data of the target equipment and product quality data of the target product, and preprocess the equipment process data and product quality data to obtain standard equipment process data and standard product quality data; construct a mechanism model and an empirical model corresponding to the target equipment, and establish a simulation model of the target equipment through the mechanism model and empirical model; collect expert technical information corresponding to the target equipment through a pre-constructed large language model, and input the expert technical information, standard equipment process data and standard product quality data into the target server; integrate the virtual twin model and the simulation model, and combine them with the standard equipment process data and standard product quality data to predict the equipment operating status information of the target equipment; construct a target machine learning model of the virtual twin in the target server, and collect actual operation monitoring data of the target equipment, so as to correct the target machine learning model through the equipment operating status information and actual operation monitoring data, and use the corrected target machine learning model to monitor the food puffing and drying process of the target equipment and the target product, thereby achieving accurate, real-time, and visualized monitoring, and enabling precise and effective prediction, early warning and intervention of specific equipment operating conditions during process monitoring. This solves the problems of poor accuracy, real-time early warning, and intuitive monitoring of existing food puffing process monitoring methods, as well as their reliance on specific employees, difficulty in quickly collecting and processing large amounts of data, and significant limitations in application scenarios.

[0028] Additional aspects and advantages of this application will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of this application. Attached Figure Description

[0029] The above and / or additional aspects and advantages of this application will become apparent and readily understood from the following description of the embodiments taken in conjunction with the accompanying drawings, wherein:

[0030] Figure 1 This is a flowchart illustrating a monitoring method for food puffing and drying processes based on digital twin technology, according to an embodiment of this application.

[0031] Figure 2 A schematic diagram of conventional data collection and display provided for one embodiment of this application;

[0032] Figure 3 A schematic diagram of the overall framework of a virtual twin model provided for one embodiment of this application;

[0033] Figure 4 A schematic diagram of a decision support and simulation scheme framework based on digital twin technology is provided as an embodiment of this application;

[0034] Figure 5A schematic diagram of iterative optimization of a virtual twin based on machine learning is provided for one embodiment of this application;

[0035] Figure 6 A schematic diagram of a bidirectional interactive system architecture based on a large language model is provided for one embodiment of this application;

[0036] Figure 7 This is an example diagram of a monitoring device for a food puffing and drying process based on digital twin technology according to an embodiment of this application;

[0037] Figure 8 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application.

[0038] Among them, 10-a monitoring device for food puffing and drying process based on digital twin technology; 100-first building module, 200-second building module, 300-integration module, 400-correction module; 801-memory, 802-processor, 803-communication interface. Detailed Implementation

[0039] The embodiments of this application are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain this application, and should not be construed as limiting this application.

[0040] The following describes a monitoring method for food puffing and drying processes based on digital twin technology, according to embodiments of this application, with reference to the accompanying drawings. Addressing the issues mentioned in the background art, such as the heavy reliance on periodic complex inputs and outputs from ITOT experts, high iteration costs, slow iteration speed, and the inability of the outputs to specifically address the problems of managers and employees in different positions, this application provides a monitoring method for food puffing and drying processes based on digital twin technology. In this method, a virtual twin model of the target equipment is constructed, and equipment process data and product quality data of the target product are collected. The equipment process data and product quality data are preprocessed to obtain standard equipment process data and standard product quality data. A mechanistic model and an empirical model corresponding to the target equipment are constructed to establish a simulation model of the target equipment. The simulation model is then constructed using a pre-built large language model. This method gathers expert technical information corresponding to the target equipment, inputting this information, along with standard equipment process data and standard product quality data, into the target server. It integrates a virtual twin model and a simulation model, combining the standard equipment process data and standard product quality data to predict the target equipment's operating status. A target machine learning model of the virtual twin is constructed on the target server, and actual operational monitoring data of the target equipment is collected. This data is then used to correct the target machine learning model, which is then used to monitor the food puffing and drying processes of the target equipment and products. This achieves accurate, real-time, and visualized monitoring, and allows for precise and effective prediction, early warning, and intervention of specific equipment conditions during process monitoring. This solves the problems of existing food puffing process monitoring methods, such as poor accuracy, real-time early warning, and intuitive monitoring, reliance on specific personnel, difficulty in quickly collecting and processing large amounts of data, and significant limitations in application scenarios.

[0041] Specifically, Figure 1 This is a flowchart illustrating a monitoring method for food puffing and drying processes based on digital twin technology, provided in an embodiment of this application.

[0042] like Figure 1 As shown, the monitoring method for food puffing and drying processes based on digital twin technology includes the following steps:

[0043] In step S101, a virtual twin model of the target equipment is constructed, and the equipment process data of the target equipment and the product quality data of the target product are collected. The equipment process data and product quality data are preprocessed to obtain standard equipment process data and standard product quality data.

[0044] As one possible approach, this application uses the process monitoring of a single-screw extruder as an example to illustrate and introduce the monitoring method for food extrusion and drying processes based on digital twin technology.

[0045] The embodiments of this application first use modeling software such as Unity3D or Blender to construct a virtual twin model of the target device, namely a single-screw extruder, and establish a three-dimensional dynamic virtual model of the device and product as a carrier and interface for the visualization interface and the virtual world twin.

[0046] Subsequently, the embodiments of this application also require preprocessing operations such as data cleaning to obtain standard equipment process data and standard product quality data.

[0047] Optionally, in one embodiment of this application, a virtual twin model of the target device is constructed, and equipment process data of the target device and product quality data of the target product are collected. The equipment process data and product quality data are preprocessed to obtain standard equipment process data and standard product quality data. This includes: constructing a virtual twin model of the target device using a preset Unity3D, and establishing a communication connection between the target PLC and the target host computer, so that the virtual twin model can acquire equipment process data and product quality data through the target PLC; performing data cleaning and normalization on the equipment process data and product quality data to obtain standard equipment process data and standard product quality data; integrating the standard equipment process data and standard product quality data using a large language model to obtain integrated natural language information, and displaying the integrated natural language information to the target user in the display interface of the virtual twin model, so that the target user can interact bidirectionally with multiple target intelligent agents constructed by the large language model based on the integrated natural language information.

[0048] Specifically, embodiments of this application can first use engineering design software such as Solidworks / Solid Edge / ProE / UG or 3D Max / Maya to create a high-precision three-dimensional model, and then use modeling software such as Unity3D or Blender to construct a high-precision virtual model of the single-screw extruder, i.e., a virtual twin model; furthermore, embodiments of this application also need to integrate sensor data and use protocols such as OPC UA and MQTT to achieve real-time data synchronization between the virtual model and the actual equipment.

[0049] It should be noted that, during the modeling process, the embodiments of this application may use a real 3D design model of the equipment (such as a model from software like Solidworks / Solid Edge / UG) or utilize 3D scanning technology to obtain the actual size and details of the equipment to ensure the accuracy of the virtual model. In addition, the embodiments of this application may embed animation and interactive functions into the virtual model to simulate the actual operation of the extruder, and display the operating status, key process parameters, real-time data and other information of the extruder through a visual interface.

[0050] Subsequently, embodiments of this application monitor key parameters in the puffing process in real time by connecting temperature sensors, flow meters, pressure sensors, speed sensors, etc., and can use data acquisition cards or data acquisition systems (such as NI cDAQ, LabVIEW, etc.) to realize automatic recording and storage of data, or can be connected to a PLC.

[0051] In the specific implementation process, in order for the virtual twin model constructed using Unity3D to obtain parameters such as those from temperature sensors from the PLC, the embodiments of this application first need to establish communication between the PLC and the host computer (usually a computer running Unity3D) through various protocols such as Modbus, OPC UA, and Profibus, so that the host computer can read the data in the PLC, such as... Figure 2 As shown, the values ​​include those from temperature sensors, etc. Furthermore, embodiments of this application require the development of software or a script on a host computer to communicate with the PLC and read data. This software can be written using programming languages ​​such as C# or Python to periodically refresh the values ​​of each sensor and transmit these values ​​to the Unity3D model. Subsequently, embodiments of this application can write a script in Unity3D to receive sensor data from the host computer software. This script can be a simple data receiver that listens to data from the host computer and updates this data to the corresponding temperature and other display components in the virtual model.

[0052] Furthermore, embodiments of this application can clean, standardize, or normalize the collected raw data (i.e., equipment process data and product quality data) to remove outliers, noise, and duplicate data, and eliminate dimensional differences between different parameters to obtain standard equipment process data and standard product quality data. For example, embodiments of this application can simply remove singularities from data such as temperature, and use techniques such as sliding windows and Fourier transforms to extract data features of data such as motor vibration or noise to facilitate subsequent analysis.

[0053] In actual implementation, the cleaned data can be directly displayed on the twin interface or visualized through a Unity3D program. As one possible approach, embodiments of this application can utilize numerical or graphical methods, calling basic components such as Text, Image, Slider, and Progress Bar from Unity3D's built-in UI system to dynamically display the cleaned data. Secondly, embodiments of this application can also use third-party icon libraries such as UnityPlot and Plotly.Unity to display data in more diverse chart formats. In scenarios with higher requirements, data can also be displayed more intuitively and flexibly on the twin's virtual model through custom rendering, such as displaying fluid velocity in real-time as particle velocity, material density as particle density, and material temperature and pressure as different temperature values.

[0054] Therefore, the embodiments of this application construct a virtual twin model and collect equipment process data and product quality data, thereby providing reliable data guidance and basis for the monitoring of subsequent food puffing processes.

[0055] In step S102, a mechanism model and an empirical model corresponding to the target equipment are constructed to establish a simulation model of the target equipment. Expert technical information corresponding to the target equipment is collected through a pre-constructed large language model, and the expert technical information, standard equipment process data, and standard product quality data are input into the target server.

[0056] Furthermore, embodiments of this application also require the collection and parsing of oral information from experts and front-line operators, textual information on design and operation (such as equipment technical parameters and historical reports, information on relevant raw materials and products, design descriptions, maintenance guidelines, risk assessments, and FMEAs) and other expert technical information through a large language model, and the establishment of an initial simulation model through an experience model and a mechanism model.

[0057] In addition, such as Figure 3 As shown, embodiments of this application also require the equipment process data collected by relevant sensors at the physical device end (i.e., physical space) to be imported by the PLC / SCADA system into the server (local or cloud server) where the digital twin resides, such as... Figure 4 As shown, product quality data is imported into the same server via LIMS.

[0058] Optionally, in one embodiment of this application, constructing a mechanism model and an empirical model corresponding to the target device includes: performing physical and chemical process analysis on the target device to obtain analysis results, and constructing a mechanism model of the target device based on the analysis results; obtaining empirical information and actual production data of the target device, and constructing an empirical model based on the mechanism model, empirical information, and actual production data.

[0059] It should be noted that the embodiments of this application establish a mechanism model based on physical principles by analyzing the physical and chemical processes of a single-screw extruder, and solve the mechanism model using methods such as finite element analysis and computational fluid dynamics. Furthermore, the mechanism model is constructed based on existing experience and the specific equipment conditions. This mechanism model can be visualized and displayed in real time in Unity3D and used as an initial model.

[0060] Secondly, the embodiments of this application can also construct a preliminary model / approximate algorithm based on historical information of the same model and formula of the equipment, or similar models and formulas, to form a preliminary empirical calculation model, which is presented in a simple form such as Excel. After that, by collecting actual production data, including the operating data of the extruder and product quality data, the empirical model can be appropriately improved. The empirical model can be visualized and reflected in real time in Unity3D. As an empirical model, PLC control programs can also be designed accordingly based on the model.

[0061] In step S103, the virtual twin model and the simulation model are integrated, and standard equipment process data and standard product quality data are combined to predict the equipment operating status information of the target equipment.

[0062] Furthermore, embodiments of this application can also integrate virtual twin models and simulation models, enabling them to display relevant parameters in real time. Process parameters and material information, among other relevant parameters, can be input into the virtual twin model and simulation model in real time for calculation and simulation. This allows for the simulation and prediction of equipment operating status, material status, and product quality indicators under different input parameters, and their visualization.

[0063] Optionally, in one embodiment of this application, the virtual twin model and the simulation model are integrated, and standard equipment process data and standard product quality data are combined to predict the equipment operating status information of the target equipment. This includes: predicting the equipment operating status information based on standard equipment process data, standard product quality data, or simulation parameters customized by the target user through the virtual twin model, and combining the virtual twin model and the simulation model.

[0064] In actual implementation, after constructing the above-mentioned mechanism model and empirical model, the embodiments of this application further need to combine them with a digital twin constructed by Unity3D, etc., to realize the real-time display of model prediction results; at the same time, those skilled in the art can also input different parameters in the digital twin interface to simulate and view the simulation results in real time, which can be used to restore the causes of process or quality problems, pre-optimize parameters in the process of new product development and process improvement, and select and design key components, etc., thereby improving the efficiency and reliability of process monitoring.

[0065] In step S104, a virtual twin target machine learning model is constructed in the target server, and actual operation monitoring data of the target equipment is collected. The target machine learning model is corrected by the equipment operation status information and actual operation monitoring data, so as to monitor the food puffing and drying process of the target equipment and target products using the corrected target machine learning model.

[0066] Subsequently, in this embodiment of the application, a machine learning model needs to be set up in the server. The machine learning model is then corrected by comparing the simulation results of the initial model (i.e., device operating status information) with actual operating monitoring data from the physical world, in order to achieve continuous iterative improvement of the virtual twin. Figure 5 As shown, the food monitoring target equipment and target product's puffing and drying processes are utilized using an iteratively improved machine learning model.

[0067] Therefore, the embodiments of this application, based on digital twin technology and data-driven models, provide fault diagnosis, high visualization, and iterative improvement of virtual twins, greatly improving the efficiency of process monitoring.

[0068] Optionally, in one embodiment of this application, a target machine learning model of a virtual twin is constructed in the target server, and actual operation monitoring data of the target device is collected to correct the target machine learning model through device operation status information and actual operation monitoring data. This includes: training the target machine learning model through actual operation monitoring data, and using preset model evaluation indicators to evaluate the performance of the trained target machine learning model to obtain performance evaluation results; and correcting the target machine learning model based on the performance evaluation results and a preset expert experience analysis model.

[0069] Specifically, based on the mechanistic model and the empirical model, the embodiments of this application can also use machine learning algorithms such as multiple linear regression, support vector machine, neural network, random forest, gradient boosting tree, and deep learning to build a model and improve the system. In the process of testing and daily production, input and output parameters such as process parameters, formula design, actual working conditions and product quality are continuously collected to train the machine learning model, and the hyperparameters of the model are optimized through techniques such as cross-validation and grid search.

[0070] Based on this, the model is evaluated and optimized through iterations, and its predictive performance is assessed using metrics such as mean squared error and root mean square error to obtain performance evaluation results. Combining the experience of relevant experts with the performance evaluation results, the sources of error in the machine learning model are analyzed, and model optimization or data augmentation is performed on the parts with larger errors. After a period of time, an optimized machine learning model is formed, and a basic model is selected as the long-term basic model to guide long-term production and continuous improvement.

[0071] Subsequently, with the continuous collection of new data, changes in the production environment, and continuous improvement of equipment, processes, or operating procedures, the machine learning model is regularly updated and iterated, and technologies such as online learning and incremental learning are used to achieve continuous learning and adaptive adjustment of the machine learning model.

[0072] Therefore, the embodiments of this application can achieve visualized process monitoring, automatic iterative improvement of the pet food extrusion process model, intelligent early warning and decision suggestions with limited human intervention, thereby solving problems in actual production.

[0073] Optionally, in one embodiment of this application, while monitoring the food puffing and drying process of the target equipment and target product using the modified target machine learning model, the method further includes: generating monitoring reports corresponding to the target equipment and target product, constructing multiple target intelligent agents through a large language model, and establishing a two-way information interaction system based on the multiple target intelligent agents; sending the monitoring reports to the target user through the two-way information interaction system, and the target user, based on the monitoring reports, and through the two-way information interaction system and a preset AR strategy, providing visualized remote operation guidance for the puffing and drying process of the target equipment and target product.

[0074] It should be noted that the embodiments of this application may also be equipped with an intelligent visual feedback and control system, which enables the digital twin to simulate and confirm whether the equipment is ready to start, process flow or stop, or predict possible operation or quality problems during operation by using real-time acquired process and equipment data, and display the warning information on the large screen in the central control room and the remote terminal interface of relevant technical personnel (such as PC web APP, tablet computer, mobile APP, etc.).

[0075] At the same time, such as Figure 6As shown, this application embodiment can also construct a two-way information interaction system based on a large language model. It uses several different customized agents based on LLMs (Large Language Models) as the medium for real-time input and output information and the information input entry point for long-term iteration. It can conduct real-time two-way communication with relevant experts, managers, operators, etc., to further improve the digital twin model and generate customized reports. In the process of setting up several different customized agents, different permissions need to be assigned according to different departments and personnel.

[0076] Furthermore, embodiments of this application can also utilize AR tools to assist on-site personnel in observing the internal operation of a digital twin of the single-screw extruder in real time through AR glasses. This allows for precise and intuitive identification of potential problems or malfunctions, and real-time voice communication with remote experts and the previously set virtual expert agency, providing guidance throughout the debugging or maintenance process. When necessary, the digital twin model can be used for advance operational rehearsals and simulations to obtain corresponding operational optimization feedback. After the operation or debugging method is verified, it can be performed on the physical equipment, with the digital twin model reflecting the operational status in real time to generate further operational suggestions.

[0077] Therefore, the embodiments of this application realize real-time personalized interaction between the digital twin model and production management and technical personnel by using a large language model.

[0078] In summary, this application's embodiments are based on digital twin technology, using a large language model as an efficient and personalized real-time interactive means, VR / AR dynamic models as a visualization means, and a theoretical model as a foundation, which is then modified based on production practice experience. Furthermore, through long-term big data accumulation (such as information collected through MES systems and large language model interaction information collection), more key data is collected, and weakly supervised machine learning algorithms are continuously trained to achieve iterative optimization of the model. This effectively solves the problems of inaccurate modeling, high dependence on experience, and high reliance on manual monitoring by front-line technicians in complex processes such as food extrusion. In addition, this application's embodiments can significantly improve the advance warning and accuracy, and continuously provide suggestions for process improvement. This better addresses issues such as the rapidly accelerating pace of industry innovation, high personnel turnover, and the difficulty in recruiting, training, and retaining front-line technicians, thereby continuously improving quality and production efficiency.

[0079] The monitoring method for food puffing and drying processes based on digital twin technology proposed in this application involves: constructing a virtual twin model of the target equipment; collecting equipment process data and product quality data of the target product; preprocessing the equipment process data and product quality data to obtain standard equipment process data and standard product quality data; constructing a mechanistic model and an empirical model corresponding to the target equipment to establish a simulation model of the target equipment; collecting expert technical information corresponding to the target equipment through a pre-constructed large language model; inputting the expert technical information, standard equipment process data, and standard product quality data into the target server; integrating the virtual twin model and the simulation model, and combining the standard equipment process data and standard product quality data to predict the equipment operating status information of the target equipment; constructing a target machine learning model of the virtual twin in the target server, and collecting actual operation monitoring data of the target equipment; correcting the target machine learning model based on the equipment operating status information and actual operation monitoring data; and using the corrected target machine learning model to monitor the food puffing and drying processes of the target equipment and target products, thereby achieving accurate, real-time, and visualized monitoring, and enabling precise and effective prediction, early warning, and intervention of specific equipment operating conditions during process monitoring.

[0080] Secondly, with reference to the accompanying drawings, a monitoring device for food puffing and drying processes based on digital twin technology, according to an embodiment of this application, is described.

[0081] Figure 7 This is a block diagram of a monitoring device for food puffing and drying processes based on digital twin technology, according to an embodiment of this application.

[0082] like Figure 7 As shown, the monitoring device 10 for food puffing and drying processes based on digital twin technology includes: a first building module 100, a second building module 200, an integration module 300, and a correction module 400.

[0083] The first construction module 100 is used to construct a virtual twin model of the target equipment, collect the equipment process data of the target equipment and the product quality data of the target product, and preprocess the equipment process data and product quality data to obtain standard equipment process data and standard product quality data.

[0084] The second construction module 200 is used to construct the mechanism model and experience model corresponding to the target equipment, so as to establish the simulation model of the target equipment through the mechanism model and experience model, and to collect the expert technical information corresponding to the target equipment through the pre-constructed large language model, so as to input the expert technical information, standard equipment process data and standard product quality data into the target server.

[0085] The integration module 300 is used to integrate the virtual twin model and the simulation model, and combine them with standard equipment process data and standard product quality data to predict the equipment operating status information of the target equipment.

[0086] The correction module 400 is used to build a virtual twin of the target machine learning model in the target server and collect the actual operation monitoring data of the target equipment. The target machine learning model is corrected by the equipment operation status information and the actual operation monitoring data. The corrected target machine learning model is then used to monitor the food puffing and drying process of the target equipment and the target product.

[0087] Optionally, in one embodiment of this application, the monitoring device 10 for food puffing and drying processes based on digital twin technology further includes: a generation module and a large language model interaction module.

[0088] The generation module is used to generate monitoring reports for the target equipment and target products while monitoring the food puffing and drying processes of the target equipment and target products using the modified target machine learning model. It also constructs multiple target intelligent agents through a large language model and establishes a two-way information interaction system based on the multiple target intelligent agents.

[0089] The large language model interaction module is used to send monitoring reports to target users through a two-way information interaction system. Based on the monitoring reports, the target users can use the two-way information interaction system and preset AR strategies to provide visual remote operation guidance for the puffing and drying processes of target equipment and target products.

[0090] Optionally, in one embodiment of this application, the first building module 100 includes: a communication unit, a preprocessing unit, and an interaction unit.

[0091] The communication unit is used to construct a virtual twin model of the target device using a preset Unity3D, and to establish a communication connection between the target PLC and the target host computer, so that the virtual twin model can obtain equipment process data and product quality data through the target PLC.

[0092] The preprocessing unit is used to perform data cleaning and normalization on equipment process data and product quality data to obtain standard equipment process data and standard product quality data.

[0093] The interaction unit is used to integrate standard equipment process data and standard product quality data through a large language model to obtain integrated natural language information. The integrated natural language information is then displayed to the target user in the display interface of the virtual twin model, enabling the target user to interact bidirectionally with multiple target intelligent agents constructed by the large language model based on the integrated natural language information.

[0094] Optionally, in one embodiment of this application, the second construction module 200 includes an analysis unit and an acquisition unit.

[0095] The analysis unit is used to perform physical and chemical process analysis on the target device to obtain analysis results and construct a mechanism model of the target device based on the analysis results.

[0096] The acquisition unit is used to acquire experience information and actual production data of the target equipment, and to construct an experience model based on the mechanism model, experience information and actual production data.

[0097] Optionally, in one embodiment of this application, the integration module 300 includes: a prediction unit, used to predict equipment operating status information based on standard equipment process data, standard product quality data, or simulation parameters customized by the target user through a virtual twin model, and in combination with the virtual twin model and the simulation model.

[0098] Optionally, in one embodiment of this application, the correction module 400 includes a training unit and an optimization unit.

[0099] The training unit is used to train the target machine learning model using actual operation monitoring data, and to evaluate the performance of the trained target machine learning model using preset model evaluation metrics to obtain performance evaluation results.

[0100] The optimization unit is used to correct the target machine learning model based on the performance evaluation results and the preset expert experience analysis model.

[0101] It should be noted that the foregoing explanation of the embodiment of the monitoring method for food puffing and drying process based on digital twin technology also applies to the monitoring device for food puffing and drying process based on digital twin technology in this embodiment, and will not be repeated here.

[0102] The monitoring device for food puffing and drying processes based on digital twin technology proposed in this application includes a first construction module for constructing a virtual twin model of the target equipment, collecting equipment process data and product quality data of the target product, and preprocessing the equipment process data and product quality data to obtain standard equipment process data and standard product quality data; and a second construction module for constructing a mechanism model and an empirical model corresponding to the target equipment, establishing a simulation model of the target equipment through the mechanism model and empirical model, and collecting expert technical information corresponding to the target equipment through a pre-constructed large language model, so as to integrate the expert technical information, standard equipment process data, and standard product quality data. Data is input into the target server; the integration module integrates the virtual twin model and the simulation model, and combines standard equipment process data and standard product quality data to predict the equipment operating status information of the target equipment; the correction module builds the target machine learning model of the virtual twin in the target server, and collects the actual operation monitoring data of the target equipment to correct the target machine learning model through the equipment operating status information and actual operation monitoring data, so as to use the corrected target machine learning model to monitor the food extrusion and drying process of the target equipment and the target product, thereby achieving accurate, real-time, and visualized monitoring, and enabling precise and effective prediction, early warning and intervention of specific equipment operating conditions during process monitoring.

[0103] Figure 8 A schematic diagram of the structure of an electronic device provided in an embodiment of this application. The electronic device may include:

[0104] The memory 801, the processor 802, and the computer program stored on the memory 801 and capable of running on the processor 802.

[0105] When the processor 802 executes the program, it implements the monitoring method for food puffing and drying processes based on digital twin technology provided in the above embodiments.

[0106] Furthermore, electronic devices also include:

[0107] Communication interface 803 is used for communication between memory 801 and processor 802.

[0108] The memory 801 is used to store computer programs that can run on the processor 802.

[0109] The memory 801 may include high-speed RAM memory, and may also include non-volatile memory, such as at least one disk storage device.

[0110] If the memory 801, processor 802, and communication interface 803 are implemented independently, then the communication interface 803, memory 801, and processor 802 can be interconnected via a bus to complete communication between them. The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus, etc. Buses can be divided into address buses, data buses, control buses, etc. For ease of representation, Figure 8 The bus is represented by a single thick line, but this does not mean that there is only one bus or one type of bus.

[0111] Optionally, in a specific implementation, if the memory 801, processor 802, and communication interface 803 are integrated on a single chip, then the memory 801, processor 802, and communication interface 803 can communicate with each other through an internal interface.

[0112] The processor 802 may be a central processing unit (CPU), an application specific integrated circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of this application.

[0113] This application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described monitoring method for food puffing and drying processes based on digital twin technology.

[0114] This application also provides a computer program product, including a computer program, which, when executed, is used to implement the above-described monitoring method for food puffing and drying processes based on digital twin technology.

[0115] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of this application. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0116] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this application, "N" means at least two, such as two, three, etc., unless otherwise explicitly specified.

[0117] Any process or method described in the flowchart or otherwise herein can be understood as representing a module, segment, or portion of code comprising one or N executable instructions for implementing custom logic functions or processes, and the scope of the preferred embodiments of this application includes additional implementations in which functions may be performed not in the order shown or discussed, including substantially simultaneously or in reverse order depending on the functions involved, as should be understood by those skilled in the art to which embodiments of this application pertain.

[0118] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a processor-included system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transmit programs for use by, or in conjunction with, an instruction execution system, apparatus, or device. More specific examples (a non-exhaustive list) of computer-readable media include: an electrical connection having one or more wires (electronic device), a portable computer disk drive (magnetic device), random access memory (RAM), read-only memory (ROM), erasable and editable read-only memory (EPROM or flash memory), fiber optic devices, and portable optical disc read-only memory (CDROM). Alternatively, the computer-readable medium may be paper or other suitable media on which the program can be printed, since the program can be obtained electronically by optically scanning the paper or other medium, followed by editing, interpreting, or otherwise processing as necessary, and then stored in a computer memory.

[0119] It should be understood that the various parts of this application can be implemented using hardware, software, firmware, or a combination thereof. In the above embodiments, the N steps or methods can be implemented using software or firmware stored in memory and executed by a suitable instruction execution system. If implemented in hardware, as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0120] Those skilled in the art will understand that all or part of the steps of the methods in the above embodiments can be implemented by a program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, the program includes one or a combination of the steps of the method embodiments.

[0121] Furthermore, the functional units in the various embodiments of this application can be integrated into a processing module, or each unit can exist physically separately, or two or more units can be integrated into a module. The integrated module can be implemented in hardware or as a software functional module. If the integrated module is implemented as a software functional module and sold or used as an independent product, it can also be stored in a computer-readable storage medium.

[0122] The storage medium mentioned above can be a read-only memory, a disk, or an optical disk, etc. Although embodiments of this application have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting this application. Those skilled in the art can make changes, modifications, substitutions, and variations to the above embodiments within the scope of this application.

Claims

1. A method for monitoring food puffing and drying processes based on digital twin technology, characterized in that, Includes the following steps: A virtual twin model of the target equipment is constructed, and the equipment process data of the target equipment and the product quality data of the target product are collected. The equipment process data and the product quality data are preprocessed to obtain standard equipment process data and standard product quality data. A mechanism model and an empirical model corresponding to the target device are constructed to establish a simulation model of the target device. Expert technical information corresponding to the target device is collected through a pre-constructed large language model. The expert technical information, the standard equipment process data, and the standard product quality data are then input into the target server. By integrating the virtual twin model and the simulation model, and combining the standard equipment process data and the standard product quality data, the equipment operating status information of the target equipment is predicted; A target machine learning model of the virtual twin is constructed in the target server, and actual operation monitoring data of the target device is collected. The target machine learning model is corrected by the device operation status information and the actual operation monitoring data, so as to monitor the food puffing and drying process of the target device and the target product using the corrected target machine learning model.

2. The method according to claim 1, characterized in that, While monitoring the food puffing and drying processes of the target equipment and the target product using a modified target machine learning model, the method also includes: The system generates monitoring reports corresponding to the target device and the target product, constructs multiple target intelligent agents through the large language model, and establishes a two-way information interaction system based on the multiple target intelligent agents. The monitoring report is sent to the target user through the two-way information interaction system, and the target user, based on the monitoring report, provides visual remote operation guidance for the puffing and drying processes of the target equipment and the target product through the two-way information interaction system and the preset AR strategy.

3. The method according to claim 1, characterized in that, The process involves constructing a virtual twin model of the target equipment, collecting equipment process data and product quality data of the target product, and preprocessing the equipment process data and product quality data to obtain standard equipment process data and standard product quality data, including: A virtual twin model of the target device is constructed using a preset Unity3D, and a communication connection is established between the target PLC and the target host computer, so that the virtual twin model can obtain the device's process data and product quality data through the target PLC. The equipment process data and the product quality data are cleaned and normalized to obtain the standard equipment process data and the standard product quality data. The standard equipment process data and standard product quality data are integrated and processed by the large language model to obtain integrated natural language information. The integrated natural language information is then displayed to the target user in the display interface of the virtual twin model, so that the target user can interact bidirectionally with multiple target intelligent agents constructed by the large language model based on the integrated natural language information.

4. The method according to claim 3, characterized in that, The construction of the mechanistic model and empirical model corresponding to the target device includes: Physical and chemical process analyses are performed on the target device to obtain analytical results, and a mechanistic model of the target device is constructed based on the analytical results. The empirical information and actual production data of the target equipment are obtained, and the empirical model is constructed based on the mechanism model, the empirical information and the actual production data.

5. The method according to claim 4, characterized in that, The process of integrating the virtual twin model and the simulation model, and combining the standard equipment process data and the standard product quality data, to predict the equipment operating status information of the target equipment includes: Based on the standard equipment process data, the standard product quality data, or the simulation parameters customized by the target user through the virtual twin model, and in combination with the virtual twin model and the simulation model, the operating status information of the equipment is predicted.

6. The method according to claim 5, characterized in that, The step of constructing the target machine learning model of the virtual twin in the target server and collecting actual operation monitoring data of the target device to correct the target machine learning model using the device operation status information and the actual operation monitoring data includes: The target machine learning model is trained using the actual operation monitoring data, and the performance of the trained target machine learning model is evaluated using preset model evaluation metrics to obtain performance evaluation results. Based on the performance evaluation results and the preset expert experience analysis model, the target machine learning model is modified.

7. A monitoring device for food puffing and drying processes based on digital twin technology, characterized in that, include: The first construction module is used to construct a virtual twin model of the target equipment, collect the equipment process data of the target equipment and the product quality data of the target product, and preprocess the equipment process data and the product quality data to obtain standard equipment process data and standard product quality data. The second construction module is used to construct the mechanism model and experience model corresponding to the target equipment, so as to establish the simulation model of the target equipment through the mechanism model and the experience model, and to collect the expert technical information corresponding to the target equipment through the pre-constructed large language model, so as to input the expert technical information, the standard equipment process data and the standard product quality data into the target server; An integration module is used to integrate the virtual twin model and the simulation model, and combine the standard equipment process data and the standard product quality data to predict the equipment operating status information of the target equipment; The correction module is used to construct the target machine learning model of the virtual twin in the target server and collect the actual operation monitoring data of the target device, so as to correct the target machine learning model through the device operation status information and the actual operation monitoring data, so as to use the corrected target machine learning model to monitor the food puffing and drying process of the target device and the target product.

8. An electronic device, characterized in that, include: The memory, the processor, and the computer program stored in the memory and executable on the processor, the processor executing the program to implement the monitoring method for food puffing and drying processes based on digital twin technology as described in any one of claims 1-6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, The program is executed by the processor to implement the monitoring method for food puffing and drying processes based on digital twin technology as described in any one of claims 1-6.

10. A computer program product, comprising a computer program, characterized in that, The computer program is executed to implement the monitoring method for food puffing and drying processes based on digital twin technology as described in any one of claims 1-6.

Images