A large model digital twin system
By leveraging the collaborative efforts of a data integration and analysis platform, a real-time rendering engine, an instruction router, and a generative large model, the problem of limited interaction methods in digital twin technology has been solved. This enables deep integration and interaction between physical and digital spaces, enhancing user experience and system intelligence, and supporting multi-turn dialogues and business simulations.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- CHONGQING INST OF GREEN & INTELLIGENT TECH CHINESE ACAD OF SCI
- Filing Date
- 2024-12-30
- Publication Date
- 2026-05-12
AI Technical Summary
Existing digital twin technologies have limited interaction methods and are inefficient, making it difficult to integrate securely and reliably with generative large models, thus restricting their application scope and user experience.
Through the collaborative work of components such as the data integration and analysis platform, real-time controllable rendering engine, command router, human-computer interaction subsystem, and generative large model, the system achieves deep integration and interaction between physical and digital spaces, supports multi-turn voice dialogue and diverse interaction methods, and combines a content processing subsystem with virtual simulation and business simulation functions.
It achieves deep integration and interaction between physical and digital spaces, improves user experience and human-computer interaction efficiency, supports multi-turn dialogue and intelligent decision-making, reduces practical operational risks and costs, and enhances the system's flexibility and scalability.
Smart Images
Figure CN119885874B_ABST
Abstract
Description
Technical Field
[0001] This invention belongs to the field of artificial intelligence or digital twins, and relates to a large-scale digital twin system. Background Technology
[0002] With the accelerating pace of digital transformation in the economy and society, digital twin technology, as a significant force, is increasingly becoming a hot topic of exploration across various industries. Leveraging its unique advantages, digital twin technology has become a crucial tool and core driving force for promoting enterprise digital transformation and fostering the development of the digital economy. However, despite the enormous potential and value of digital twin technology, current interaction methods are still primarily limited to traditional and relatively simple means such as mouse operation and 3D glasses. This limited interaction and inefficiency restricts the application scope and user experience of digital twin systems. In recent years, generative artificial intelligence technology based on large-scale language models has emerged as a powerful force, bringing unprecedented changes with its rapid development and superior performance. This technology enables smooth and reliable multi-turn voice dialogue, greatly enriching the ways and scenarios of human-computer interaction. Users can seamlessly communicate with digital twin systems through natural language, thus obtaining the information and services they need more intuitively and conveniently. However, despite the significant advantages of large-scale language models in voice interaction, how to securely and reliably integrate generative large-scale models with digital twin systems remains an unsolved problem. Summary of the Invention
[0003] In view of this, the purpose of the present invention is to provide a large-scale digital twin system.
[0004] To achieve the above objectives, the present invention provides the following technical solution:
[0005] A large-scale digital twin system includes:
[0006] The data integration and analysis platform constructs a digital copy of the actual physical system in the digital space to reflect the status, performance and changes of the actual physical system in real time, thereby realizing the fusion and interaction between the physical space and the digital space.
[0007] The physical space is digitally mapped into a static three-dimensional model in the digital space;
[0008] Real-time dynamic data used to drive the virtual visualization of the static 3D model, which is provided by the data integration and analysis platform;
[0009] The real-time controllable rendering engine integrates and renders the static 3D model and dynamic data in real time to visualize the actual physical system operation; it receives rendering control commands parsed by the command router, adjusts the rendering area, rendering method and rendering parameters, sends the rendering results to the display terminal of the human-computer interaction subsystem, and sends the command execution results back to the command router.
[0010] The instruction router receives and parses instructions from the instruction issuer, forwards the parsed instructions to the corresponding instruction execution unit, collects the instruction execution results and sends them back to the instruction issuer, and simultaneously transmits them to the content generation processing subsystem for processing and returns the generative large model to realize the global context; the instruction issuer includes a human-computer interaction subsystem and a content generation processing subsystem.
[0011] The human-computer interaction subsystem receives user input, identifies interactive actions, processes data, parses intent, invokes relevant systems to implement intent, and presents results. When interaction involves control commands, the commands are transmitted to the command router and then forwarded to the appropriate execution unit. This subsystem is connected to a generative big data model. When a user enters natural dialogue text in a specific text box on the interface, or when a user uses voice activation to input natural dialogue text, the text is sent to the generative big data model. The model processes the text and returns a response to the subsystem.
[0012] The generative large model receives the natural dialogue text input by the human-computer interaction subsystem, generates two types of generated content based on the text: instructional and non-instructional, and sends them to the generated content processing subsystem; it receives the content returned by the generated content processing subsystem and sends it to the human-computer interaction subsystem, while recording the returned content in the context to support subsequent multi-round question-and-answer interaction dialogue and decision-making.
[0013] The content processing subsystem receives generated content from the generative large model; for non-instructional generated content, it directly returns it to the generative large model; for instructional generated content, it parses it; first, it sends the parsing result to the instruction router, and then it collects the instruction execution results returned by the instruction router, processes them, and returns them to the generative large model.
[0014] The physical control command distributor receives the physical control commands parsed by the command router and transmits them to the actuators in the physical system to control physical objects; it also transmits the execution results of the physical control commands back to the command router.
[0015] Furthermore, the instructions issued by the instruction issuer include a single instruction, an instruction queue, and a directed graph of instruction logic; the instruction router maintains the sequentiality, synchronization, asynchronicity, and context execution relationships of the instruction queue and the directed graph of instruction logic, and sends the instruction execution results back to the instruction issuer in batches or in whole batches according to the instruction logic and instruction requirements; the execution unit includes a real-time controllable rendering engine and a physical control instruction distributor.
[0016] Furthermore, the content generation processing subsystem includes:
[0017] The content generation router analyzes the generated content of the generative large model, transmits the instruction-type generated content to the corresponding instruction parser, and transmits the non-instruction-type generated content back to the generative large model.
[0018] An instruction parser converts the generated instruction content into an instruction set and forms the instruction queue or the instruction logic directed graph, which is then transmitted to the instruction router.
[0019] An execution result encapsulator is used to modify and encapsulate the execution results fed back by the instruction router and transmit them to the generative large model.
[0020] Furthermore, the human-computer interaction subsystem includes a speech synthesis module, which is used to play the human-computer interaction results.
[0021] Furthermore, it also includes a virtual simulation model, which performs virtual simulation calculations based on user input, generated instructions, or predefined business scenarios, and applies the simulation calculation results to the static three-dimensional model to visually display the results of the virtual simulation calculations in advance to verify the business plan.
[0022] Furthermore, the human-computer interaction subsystem supports a sub-window mode to display the results of the virtual simulation calculation.
[0023] Furthermore, it also includes: a business simulation component, which performs business simulation calculations based on user input, generated instructions, or predefined business scenarios, and applies the results of the business simulation calculations to the static three-dimensional model to visually simulate key business processes, important facilities and equipment, and operating logic or operation methods of the enterprise to support training and maintenance.
[0024] Furthermore, the human-computer interaction subsystem supports a sub-window mode to display the business simulation calculation results.
[0025] The beneficial effects of this invention are as follows:
[0026] First, the data integration and analysis platform, along with a real-time controllable rendering engine, reflects the state, performance, and changes of the physical system in real time, achieving deep integration and interaction between physical and digital spaces. Real-time data feedback and visualization provide users with intuitive and comprehensive system monitoring and management tools. Second, due to the collaborative work of the human-computer interaction subsystem and the generative large model, users can communicate efficiently with the system through an intuitive interface. Whether through text input or voice interaction, the system accurately identifies and processes user intent, providing immediate feedback and improving user experience and human-computer interaction efficiency. The introduction of the generative large model enables the system to intelligently process users' natural dialogue text, generating both instructional and non-instructional content, providing users with more flexible and diverse interaction methods. It supports multi-turn dialogues between users and the large model digital twin system, enhancing the system's intelligence level. The generated content processing subsystem intelligently parses and processes the output of the generative large model, ensuring accurate execution of instructional content and reasonable feedback of non-instructional content, improving the system's automation level, and guaranteeing the accuracy and consistency of information. Furthermore, the instruction router can maintain the sequentiality, synchronization, asynchronicity, and contextual execution relationships of the instruction queue and the directed graph of instruction logic, ensuring accurate instruction execution and timely feedback of results. In addition, the digital twin system proposed in this invention also supports virtual simulation and business simulation functions. Users can perform virtual simulation calculations and business simulation calculations by inputting instructions or predefined business scenarios, thereby pre-verifying business solutions or simulating key business processes of the enterprise. This not only helps reduce the risks and costs of actual operation but also improves the system's flexibility and scalability. Finally, the optimization and improvement of the human-computer interaction subsystem allows users to interact with the system more conveniently through the speech synthesis module, sub-window mode, and other methods, improving user experience and satisfaction.
[0027] First, the data integration and analysis platform, along with a real-time controllable rendering engine, reflects the state, performance, and changes of the physical system in real time, achieving deep integration and interaction between physical and digital spaces. Real-time data feedback and visualization provide users with intuitive and comprehensive system monitoring and management tools. Second, due to the collaborative work of the human-computer interaction subsystem and the generative large model, users can communicate efficiently with the system through context-supported multi-turn dialogues via an intuitive interface. Whether through text input or voice interaction, the system accurately identifies and processes user intent, providing immediate feedback and improving user experience and human-computer interaction efficiency. The introduction of the generative large model enables the system to intelligently process users' natural dialogue text, generating both instructional and non-instructional content, providing users with more flexible and diverse interaction methods. This supports multi-turn dialogues between users and the large model digital twin system, enhancing the system's intelligence level. The generated content processing subsystem intelligently parses and processes the output of the generative large model, ensuring accurate execution of instructional content and reasonable feedback of non-instructional content, improving the system's automation level, and guaranteeing the accuracy and consistency of information. Furthermore, the instruction router can maintain the sequentiality, synchronization, asynchronicity, and contextual execution relationships of the instruction queue and the directed graph of instruction logic, ensuring accurate instruction execution and timely feedback of results. In addition, the digital twin system proposed in this invention also supports virtual simulation and business simulation functions. Users can perform virtual simulation calculations and business simulation calculations by inputting instructions or predefined business scenarios, thereby pre-verifying business solutions or simulating key business processes of the enterprise. This not only helps reduce the risks and costs of actual operation but also improves the system's flexibility and scalability. Finally, the optimization and improvement of the human-computer interaction subsystem allows users to interact with the system more conveniently through the speech synthesis module, sub-window mode, and other methods, improving user experience and satisfaction.
[0028] Other advantages, objectives, and features of the invention will be set forth in part in the description which follows, and in part will be apparent to those skilled in the art from the following examination, or may be learned from practice of the invention. The objectives and other advantages of the invention can be realized and obtained through the following description. Attached Figure Description
[0029] To make the objectives, technical solutions, and advantages of the present invention clearer, the preferred embodiments of the present invention will be described in detail below with reference to the accompanying drawings, wherein:
[0030] Figure 1 This is a schematic diagram of a large-scale digital twin system according to an embodiment of the present invention;
[0031] Figure 2 This is a schematic diagram of a content generation processing subsystem according to an embodiment of the present invention;
[0032] Figure 3 This is a schematic diagram of a large-scale digital twin system including a virtual simulation model, according to an embodiment of the present invention. Detailed Implementation
[0033] The following specific examples illustrate the implementation of the present invention. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and various details in this specification can be modified or changed based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that the illustrations provided in the following embodiments are only schematic representations of the basic concept of the present invention. Unless otherwise specified, the following embodiments and features can be combined with each other.
[0034] The accompanying drawings are for illustrative purposes only and are schematic diagrams, not actual pictures. They should not be construed as limiting the invention. To better illustrate the embodiments of the invention, some parts in the drawings may be omitted, enlarged, or reduced, and do not represent the actual product dimensions. It is understandable to those skilled in the art that some well-known structures and their descriptions may be omitted in the drawings.
[0035] In the accompanying drawings of the embodiments of the present invention, the same or similar reference numerals correspond to the same or similar components. In the description of the present invention, it should be understood that if terms such as "upper," "lower," "left," "right," "front," and "rear" indicate the orientation or positional relationship based on the orientation or positional relationship shown in the drawings, they are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, the terms used to describe positional relationships in the drawings are only for illustrative purposes and should not be construed as limiting the present invention. For those skilled in the art, the specific meaning of the above terms can be understood according to the specific circumstances.
[0036] Please see Figure 1 This is a schematic diagram of a large-scale digital twin system according to an embodiment of the present invention. This embodiment details a large-scale digital twin system with comprehensive functions and a clear structure, in which its internal components work collaboratively to achieve a complete closed loop from user input to physical world feedback. This large-scale digital twin system mainly consists of the following core components, which cooperate with each other to realize the various functions of the digital twin:
[0037] The data integration and analysis platform receives data from sensors and edge computing devices in the actual physical system, and constructs a digital copy of the actual physical system in the digital space to reflect the status, performance and changes of the actual physical system in real time, thereby realizing the fusion and interaction between the physical space and the digital space.
[0038] A static 3D model is a digital mapping of physical space into digital space; this static 3D model is a digital model with three-dimensional spatial characteristics constructed based on computer graphics technology.
[0039] Real-time dynamic data is used to drive the virtual visualization of the static 3D model. This real-time dynamic data is provided by the data integration and analysis platform. The dynamic data includes Internet of Things (IoT) sensor data, access video, or real-time data from video recognition.
[0040] The real-time controllable rendering engine, as the core of the digital twin result presentation, is responsible for receiving interaction and rendering instructions from the instruction router, integrating and fusing static 3D models with dynamic data, rendering realistic digital twin effects, and feeding the rendering results back to the display terminal of the human-computer interaction subsystem.
[0041] The generative big model, as the core of intelligent decision-making and generation, is connected to the human-computer interaction subsystem. It intelligently processes input information such as user input, file reading, and network requests, and outputs both instructional and non-instructional generated content. This generated content not only provides direct feedback to the user but also serves as context for the next round of dialogue, supporting multi-turn conversations between the user and the big model digital twin system. The generative big model also receives execution results and uses them as context to support subsequent multi-turn conversations and decisions. The application of the generative big model also gives the digital twin system a stronger context-awareness capability. By intelligently inferring the user's true intent based on the user's historical dialogue and current context, it directly generates target instructions that match this intent. Compared to the traditional step-based instruction generation method, this approach is no longer limited to executing a series of intermediate steps to achieve a goal but directly focuses on meeting the user's actual needs. This shift greatly improves the coherence and fluency of the interaction process, making each conversation more closely centered on the user's core needs. Users no longer need to endure tedious steps and redundant information but can directly and efficiently obtain the services or information they need. Generative big data models also improve the interaction mode between voice and speakers, allowing users to communicate directly with the digital twin system using natural language. Specifically, users first issue commands or ask questions to the speaker via voice. The speaker, acting as a sound signal receiver, clearly captures and transmits the user's voice information to the generative big data model. During this process, the speaker's high-fidelity sound pickup technology and noise suppression algorithms ensure the clarity and accuracy of the voice signal, providing a solid foundation for subsequent processing. Subsequently, the generative big data model uses advanced natural language processing technology to parse and understand the user's voice. It can identify the user's intent, extract key information, and generate corresponding responses or execute corresponding tasks based on this information and the interaction history context. In this process, the big data model's deep learning capabilities enable it to handle complex language structures and semantic relationships, thereby accurately understanding the user's true needs. After understanding the user's intent, the generative big data model interacts with the big data digital twin system, transforming the user's commands or questions into a form that the digital twin system can parse and execute. The digital twin system, as a digital mapping of the physical world, can simulate and reflect the states and behaviors of the real world. Driven by generative large models, the digital twin system can adjust itself or execute tasks according to user instructions. Finally, the execution result or response of the digital twin system is fed back to the user via voice through a speaker. This process not only achieves intelligent and efficient human-computer dialogue but also provides users with an intuitive and convenient interactive experience. Users do not need complex operations or input; they can communicate instantly with the digital twin system simply through voice and enjoy the convenience of intelligent services.
[0042] The content processing subsystem receives generated content from the generative large model; for non-instructional generated content, it directly returns it to the generative large model; for instructional generated content, it parses it; first, the parsing result is sent to the instruction router, then the instruction execution results returned by the instruction router are collected, processed, and returned to the generative large model. Please refer to [link to relevant documentation]. Figure 2 This is a schematic diagram of a content generation processing subsystem according to an embodiment of the present invention; it illustrates the logical structure of a specific implementation of the content generation processing subsystem, which includes:
[0043] The content generation router intelligently selects the appropriate instruction parser based on the characteristics of the output content of the generative large model; it then transmits the non-instruction generated content back to the generative large model; the instruction router can maintain the order, synchronization, asynchronicity, and context execution relationships of the instruction queue and the directed graph of instruction logic, ensuring accurate execution of instructions and timely feedback of results.
[0044] The instruction parser converts generated instructions into an instruction set, forming an instruction queue or directed graph of instruction logic, which is then transmitted to the instruction router. The instruction router, as the core node for instruction processing, is responsible for parsing the instruction queue into control instructions or interaction and rendering instructions. Control instructions are transmitted to the physics control instruction distributor to control physical objects, while interaction and rendering instructions are transmitted to the real-time controllable rendering engine for rendering. Simultaneously, the instruction router is also responsible for receiving execution results and, through an execution result encapsulator, transmitting these results to the generative large model to support subsequent multi-turn dialogues and decision-making.
[0045] The execution result encapsulator is used to modify and encapsulate the execution results fed back by the instruction router and send them to the generative large model.
[0046] To enable interaction and feedback between the digital twin system and the physical world, this system also includes the following components:
[0047] Physical control command distributor: This distributor is responsible for transmitting control commands to the actuators in the actual physical system to control the physical objects. Simultaneously, it receives feedback results from the actuators and transmits these results to the command router to complete the closed loop of the entire command execution process.
[0048] In building a large-scale digital twin system, specific business or domain-specific private knowledge is effectively integrated and applied to the system. This is typically achieved by constructing a domain knowledge vector library, which contains digital representations of various data, rules, logic, or experiences related to a specific business scenario. The key to this integration approach is that it goes beyond simply storing or importing private domain knowledge into the system; it involves a deep and organic integration of this knowledge with the large-scale digital twin system. This means that the large-scale digital twin system can fully consider this private domain knowledge when generating instructions, making decisions, or conducting virtual simulations, ensuring its application in specific business scenarios is both secure and accurate. Furthermore, specific organizational or domain-specific private knowledge data, along with a series of task requirements or guidelines, are used to further train and optimize the generative large-scale model through feedback reinforcement learning methods to improve its ability to complete specific tasks. This process aims to better adapt the model to specific scenarios and needs, thereby improving its performance and usability.
[0049] The large-scale digital twin system described in this embodiment achieves a complete closed loop from user input to physical world feedback through highly integrated components and a logically clear instruction processing and transmission mechanism. This system not only improves the efficiency and accuracy of digital twins but also provides users with a more intuitive and intelligent interactive experience.
[0050] Please see Figure 3 This is a schematic diagram of a large-scale digital twin system including a virtual simulation model, according to another embodiment of the present invention; Figure 1Building upon the proposed large-scale digital twin system, its functionality has been further expanded with the addition of a crucial component: a simulation model. This simulation model is a highly integrated computational module designed to perform in-depth virtual simulation calculations based on specific user needs, instructions generated by the large-scale model, or a series of predefined typical business scenarios. This model not only enhances the practicality and flexibility of the digital twin system but also provides strong support for its application across multiple industries. First, the simulation model receives direct input from the user, which may include various parameter settings and operation instructions. Simultaneously, it can receive instructions generated by the large-scale digital twin system based on user input or predefined conditions; these instructions can involve complex computational tasks or specific simulation scenarios. Upon receiving input, the simulation model activates the corresponding computational engine to perform virtual simulation calculations based on the input information. These calculations can involve multiple disciplines such as physics, chemistry, and biology, aiming to simulate complex phenomena or processes in the real world. After the calculations are completed, the simulation model integrates the results data with the 3D model in the digital twin system. Using 3D visualization technology, these results data are rendered onto the output interface, presenting them to the user in an intuitive and vivid way. This large-scale digital twin system not only possesses powerful data processing and intelligent decision-making capabilities, but also enables virtual simulation of complex phenomena and visualization of results, thereby allowing for pre-simulation and verification of events or business scenarios that have not yet occurred.
[0051] exist Figure 1Building upon the proposed large-scale digital twin system, its functionality has been further expanded with the addition of a crucial component: a business simulation model. This model simulates key business processes and critical facilities and equipment to meet the needs of various application scenarios, including training, maintenance, and support. Through this model, users can gain a deeper understanding of each stage of the business process, improving operational skills and troubleshooting capabilities. The business simulation model accurately simulates key business processes, including production processes, supply chain management, and customer service. Users can simulate business processes under different conditions in a virtual environment, observe their operation, and identify potential problems and areas for improvement. For critical facilities and equipment, such as production line equipment, power facilities, oil depot storage and distribution, and facility maintenance, the business simulation model can simulate their operating status and procedures. Users can practice equipment operation, maintenance, and troubleshooting in a virtual environment, improving their practical skills. This business simulation model can also serve as a training tool, providing employees with a vivid practical environment. By simulating real business scenarios, employees can learn business knowledge in a risk-free environment, improving their work skills and teamwork abilities. In terms of facility and equipment maintenance, the business simulation model can simulate equipment failure scenarios, providing maintenance personnel with training in fault diagnosis and repair solutions. Simultaneously, the model can also provide historical data and predictive information on equipment maintenance, helping maintenance personnel develop more reasonable maintenance plans. This business simulation model can be invoked through instructions generated by the large-scale digital twin system. Therefore, users can customize personalized simulation scenarios and tasks according to specific needs through the large-scale digital twin system, thereby further enhancing the simulation effect and application value.
[0052] exist Figure 1Based on the proposed large-scale digital twin system, a traditional digital twin interactive interface is also included. This interface serves as a crucial means for information exchange between the large-scale digital twin system and the user. Based on traditional graphical user interface (GUI) design principles and incorporating the characteristics of digital twin technology, this interface provides users with an intuitive, user-friendly, and feature-rich operating environment. Through this interface, users can easily browse models, query data, and input commands, achieving comprehensive control over the digital twin system. In the large-scale digital twin system, virtual simulation results are displayed through the traditional digital twin interactive interface. To meet different user needs, the system offers two display methods: users can choose to directly display the virtual simulation results in the current window. This method is suitable for users who want to quickly view and compare different simulation results within the same interface. By overlaying the virtual simulation results in the original window, users can intuitively see how the model changes under different conditions, thus making more accurate judgments. In addition to displaying in the original window, users can also choose to open a new window to display the virtual simulation results. This method is suitable for users who need to perform detailed analysis of the simulation results or share them with others. The newly opened window can be independently resized and positioned, allowing users to observe and operate it in detail. Users can also export simulation results as images, videos, and other formats for subsequent analysis and sharing.
[0053] The large-scale digital twin system proposed in this invention firstly reflects the state, performance, and changes of the physical system in real time through a data integration and analysis platform and a real-time controllable rendering engine, achieving deep integration and interaction between physical and digital spaces. Real-time data feedback and visualization provide users with intuitive and comprehensive system monitoring and management tools. Secondly, due to the collaborative work of the human-computer interaction subsystem and the generative large-scale model, users can communicate efficiently with the system through an intuitive interface. Whether through text input or voice interaction, the system can accurately identify and process user intentions, achieving instant feedback and improving user experience and human-computer interaction efficiency. The introduction of the generative large-scale model enables the system to intelligently process users' natural dialogue text, generating both instructional and non-instructional content, providing users with more flexible and diverse interaction methods. It supports multi-turn dialogues between users and the large-scale digital twin system, enhancing the system's intelligence level. The generated content processing subsystem intelligently parses and processes the output of the generative large-scale model, ensuring the accurate execution of instructional content and the reasonable feedback of non-instructional content, improving the system's automation level, and guaranteeing the accuracy and consistency of information. Furthermore, the instruction router can maintain the sequentiality, synchronization, asynchronicity, and contextual execution relationships of the instruction queue and the directed graph of instruction logic, ensuring accurate instruction execution and timely feedback of results. In addition, the digital twin system proposed in this invention also supports virtual simulation and business simulation functions. Users can perform virtual simulation calculations and business simulation calculations by inputting instructions or predefined business scenarios, thereby pre-verifying business solutions or simulating key business processes of the enterprise. This not only helps reduce the risks and costs of actual operation but also improves the system's flexibility and scalability. Finally, the optimization and improvement of the human-computer interaction subsystem allows users to interact with the system more conveniently through the speech synthesis module, sub-window mode, and other methods, improving user experience and satisfaction.
[0054] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.
Claims
1. A large-scale digital twin system, characterized in that, include: The data integration and analysis platform constructs a digital copy of the actual physical system in the digital space to reflect the state, performance, and changes of the actual physical system in real time, realizing the fusion and interaction between the physical space and the digital space; the physical space is digitally mapped into a static three-dimensional model in the digital space; real-time dynamic data is used to drive the visualization and virtual operation of the static three-dimensional model, and this real-time dynamic data is provided by the data integration and analysis platform. The real-time controllable rendering engine integrates and renders the static 3D model and dynamic data in real time to visualize the actual physical system operation; it receives rendering control commands parsed by the command router, adjusts the rendering area, rendering method and rendering parameters, and sends the rendering results to the display terminal of the human-computer interaction subsystem. The execution result of the instruction is sent back to the instruction router, and the rendering result and the real-time dynamic data are sent back to the instruction router as the execution result reflecting the actual physical system operating status. The instruction router receives and parses instructions from the instruction issuer, forwards the parsed instructions to the corresponding instruction execution unit, collects the instruction execution results and sends them back to the instruction issuer, and simultaneously transmits them to the content generation processing subsystem for processing and returns the generative large model to realize the global context; the instruction issuer includes a human-computer interaction subsystem and a content generation processing subsystem. The human-computer interaction subsystem receives user input, identifies interactive actions, processes data, parses intent, invokes relevant systems to implement intent, and presents results. When interaction involves control commands, the commands are transmitted to the command router and then forwarded to the appropriate execution unit. This subsystem is connected to a generative big data model. When a user enters natural dialogue text in a specific text box on the interface, or when a user uses voice activation to input natural dialogue text, the text is sent to the generative big data model. The model processes the text and returns a response to the subsystem. The generative large model receives the natural dialogue text input by the human-computer interaction subsystem, generates two types of generated content based on the text: instructional and non-instructional, and sends them to the generated content processing subsystem; it receives the content returned by the generated content processing subsystem and sends it to the human-computer interaction subsystem, while recording the instruction execution result in the context, which is used to support subsequent multi-turn question-and-answer interactive dialogue and decision-making. The real-time dynamic data is used to characterize the system's operating status; A content processing subsystem receives generated content from the generative large model; For the non-instructional generated content, it is directly returned to the generative large model; for the instruction-type generated content, it is parsed; first, the parsing result is sent to the instruction router, then the instruction execution results returned by the instruction router are collected, processed, and returned to the generative large model. The content generation processing subsystem includes: The content generation router analyzes the generated content of the generative large model, transmits the instruction-type generated content to the corresponding instruction parser, and transmits the non-instruction-type generated content back to the generative large model. The instruction parser converts the generated instruction content into an instruction set and forms an instruction queue or instruction logic directed graph, which is then transmitted to the instruction router. An execution result encapsulator is used to modify and encapsulate the execution results fed back by the instruction router and transmit them to the generative large model. The physical control command distributor receives the physical control commands parsed by the command router and transmits them to the actuators in the physical system to control physical objects; it also transmits the execution results of the physical control commands back to the command router.
2. The large-scale digital twin system according to claim 1, characterized in that: The instructions issued by the instruction issuer include a single instruction, an instruction queue, and a directed graph of instruction logic. The instruction router maintains the sequentiality, synchronization, asynchronicity, and context execution relationships of the instruction queue and the directed graph of instruction logic, and sends the instruction execution results back to the instruction issuer in batches or in whole batches according to the instruction logic and instruction requirements. The execution unit includes a real-time controllable rendering engine and a physical control instruction distributor.
3. The large-scale digital twin system according to any one of claims 1-2, characterized in that: The human-computer interaction subsystem includes a speech synthesis module, which is used to play the results of the human-computer interaction.
4. The large-scale digital twin system according to any one of claims 1-2, characterized in that: It also includes a virtual simulation model, which performs virtual simulation calculations based on user input, generated instructions, or predefined business scenarios, and applies the simulation calculation results to the static three-dimensional model to visually display the results of the virtual simulation calculations in order to pre-verify the business plan.
5. The large-scale digital twin system according to claim 4, characterized in that: The human-computer interaction subsystem supports a sub-window mode to display the results of the virtual simulation calculation.
6. The large-scale digital twin system according to any one of claims 1-2, characterized in that, Also includes: The business simulation component performs business simulation calculations based on user input, generated instructions, or predefined business scenarios. The results of the business simulation calculations are then applied to the static 3D model to visually simulate key business processes, important facilities and equipment, and operational logic or methods to support training and maintenance.
7. The large-scale digital twin system according to claim 6, characterized in that: The human-computer interaction subsystem supports a sub-window mode to display the results of the business simulation calculation.