Large-model digital twin system

Abstract

The invention relates to a large-model digital twinning system, and belongs to the field of artificial intelligence or digital twinning. The system comprises a man-machine interaction subsystem, a generative large model, a generated content processing subsystem, an instruction router and a real-time controllable rendering engine, the man-machine interaction subsystem receives user input and analyzes an intention to realize man-machine interaction; the large generative model processes user voice or text input to generate instruction and non-instruction generation content and transmits the instruction and non-instruction generation content to the generation content processing subsystem to analyze the instruction content and send the analyzed instruction content to the instruction router, and the non-instruction content and an execution result are transmitted back to the large generative model to support multiple rounds of question and answer interaction; and the instruction router maintains an execution relationship between the instruction queue and the logic directed graph and returns an execution result. According to the method, deep fusion and interaction of physical and digital spaces are realized, a user can intuitively communicate with a system in a context-supported multi-round dialogue form and obtain instant, accurate and efficient feedback, and the experience feeling and the interaction efficiency are improved.

Description

Technical Field

[0001] The present invention belongs to the field of artificial intelligence or digital twins and relates to a large-model digital twin system. Background Art

[0002] As the pace of economic and social digital transformation continues to accelerate, digital twin technology, as an important force, is increasingly becoming a hot technology that various industries are competing to explore. With its unique advantages, digital twin technology has become an important tool and core driving force to promote the digital transformation of enterprises and promote the development of the digital economy. However, although digital twin technology has shown great potential and value, the current way of interacting with it is still mainly limited to traditional and relatively single means of interaction such as mouse operation and 3D glasses. The interaction method is relatively single and inefficient, which limits the application scope and user experience of the digital twin system. In recent years, generative artificial intelligence technology based on large language models has emerged, bringing unprecedented changes with its rapid development speed and excellent performance. The emergence of this technology makes smooth and reliable multi-round voice dialogues possible, greatly enriching the ways and scenarios of human-computer interaction. Users can achieve seamless communication with the digital twin system through natural language, so as to obtain the required information and services more intuitively and conveniently. However, although large language models have shown great advantages in voice interaction, how to organically integrate generative large models and digital twin systems safely and reliably remains an unresolved problem. Summary of the invention

[0003] In view of this, an object of the present invention is to provide a large-model digital twin system.

[0004] In order to achieve the above object, the present invention provides the following technical solutions:

[0005] A large-model digital twin system, comprising:

[0006] A data integration and analysis platform, which constructs a digital copy of the actual physical system of the physical space in the digital space to reflect the status, performance and changes of the actual physical system in real time, and realizes the integration and interaction of the physical space and the digital space;

[0007] A static three-dimensional model of a digital mapping of the physical space in the digital space;

[0008] Real-time dynamic data used to drive the virtual operation of the static three-dimensional model visualization, the real-time dynamic data is provided by the data integration and analysis platform;

[0009] A real-time controllable rendering engine integrates and renders the static 3D model with the dynamic data in real time to realize the visualization of the operation of the actual physical system; receives the rendering control instruction parsed by the instruction router, adjusts the rendering area, rendering mode and rendering parameters, and sends the rendering result to the display terminal of the human-computer interaction subsystem; and transmits the instruction execution result back to the instruction router;

[0010] The instruction router receives and parses the instruction from the instruction issuer, and forwards the parsed instruction to the corresponding instruction execution unit; collects the instruction execution results and returns them to the instruction issuer, and at the same time transmits them to the generated content processing subsystem for processing and returns them to the generated large model to realize the global context; the instruction issuer includes a human-computer interaction subsystem and a generated content processing subsystem;

[0011] The human-computer interaction subsystem receives the user's human-computer interaction input, identifies the human-computer interaction actions, processes the human-computer interaction data, analyzes the interaction intention, calls the relevant system to realize the interaction intention, and presents the human-computer interaction results; when the human-computer interaction involves control instructions, the control instructions are transmitted to the instruction router and then forwarded to the corresponding instruction execution unit; the human-computer interaction subsystem is connected to the generative big model, when the user enters the natural dialogue text in the specific text box of the interaction interface, or when the user uses voice wake-up to enter the natural dialogue text, the natural dialogue text will be transmitted to the generative big model, and the generative big model will return the response result to the human-computer interaction subsystem after processing;

[0012] The generative large model receives the natural dialogue text input by the human-computer interaction subsystem, generates two types of generated content, command-type and non-command-type, according to the text, and transmits them to the generated content processing subsystem; receives the content returned by the generated content processing subsystem and transmits it to the human-computer interaction subsystem, and records the returned content into the context to support subsequent multiple rounds of question-answering interactive dialogue and decision-making;

[0013] The generated content processing subsystem receives the generated content from the generative big model; directly returns the non-instruction generated content to the generative big model; parses the instruction generated content; first sends the parsing result to the instruction router, then collects the instruction execution results returned by the instruction router, processes them, and then returns them to the generative big model;

[0014] The physical control instruction distributor receives the physical control instruction parsed by the instruction router and transmits it to the actuator in the physical system to control the physical object; and transmits the execution result of the physical control instruction back to the instruction router.

[0015] Furthermore, the instructions of the instruction issuer include a single instruction, an instruction queue and an instruction logic directed graph; the instruction router maintains the sequentiality, synchronization, asynchrony and context execution relationship of the instruction queue and the instruction logic directed graph, and packages the instruction execution results in batches or in batches according to the instruction logic and instruction requirements and transmits them back to the instruction issuer; the execution unit includes a real-time controllable rendering engine and a physical control instruction distributor.

[0016] Furthermore, the generated content processing subsystem includes:

[0017] A generated content router analyzes the generated content of the generative big model, transmits the instruction-type generated content to a corresponding instruction parser, and transmits the non-instruction-type generated content back to the generative big model;

[0018] An instruction parser converts the instruction-type generated content into an instruction set and forms the instruction queue or the instruction logic directed graph and transmits it to the instruction router;

[0019] The execution result encapsulator is used to transform and encapsulate the execution result fed back by the instruction router and transmit it to the generative large model.

[0020] Furthermore, the human-computer interaction subsystem is provided with a speech synthesis module, and the speech synthesis module is used to play the human-computer interaction result.

[0021] Furthermore, it also includes a virtual simulation model, which performs virtual simulation calculations according to 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 to pre-verify the business plan.

[0022] Furthermore, the human-computer interaction subsystem supports a sub-window mode to display the result 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, applies the business simulation calculation results to the static three-dimensional model, and visually simulates the enterprise's key business processes, important facilities and equipment, and operating logic or operating methods 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 the present invention are:

[0026] First, the data integration and analysis platform and the real-time controllable rendering engine reflect the state, performance and changes of the physical system in real time, realizing the deep integration and interaction of physical space and digital space; real-time data feedback and visualization display provide users with intuitive and comprehensive system monitoring and management methods. Secondly, since the human-computer interaction subsystem works with the generative big model, users can communicate with the system efficiently through an intuitive interactive interface. Whether through text input or voice interaction, the system can accurately identify and process the user's intention, achieve instant feedback, and improve the user experience and human-computer interaction efficiency; the introduction of the generative big model enables the system to intelligently process the user's natural dialogue text, generate command-type and non-command-type content, provide users with more flexible and diverse interaction methods, support multiple rounds of dialogue between users and the big model digital twin system, and improve the intelligence level of the system; the generated content processing subsystem intelligently analyzes and processes the output of the generative big model, ensuring the accurate execution of command-type content and the reasonable feedback of non-command-type content, improving the automation of the system and ensuring the accuracy and consistency of information. Furthermore, the instruction router can maintain the order, synchronization, asynchrony and contextual execution relationship of the instruction queue and the instruction logic directed graph, ensuring the accurate execution of the instructions and the timely feedback of the results. In addition, the digital twin system proposed in the present invention also supports virtual simulation and business simulation functions. Users can perform virtual simulation calculations and business simulation calculations by inputting instructions or pre-defining business scenarios, thereby pre-verifying business plans or simulating key business processes of enterprises, which not only helps to reduce the risks and costs of actual operations, but also improves the flexibility and scalability of the system. Finally, the optimization and improvement of the human-computer interaction subsystem enables users to interact with the system more conveniently through speech synthesis modules, sub-window modes, etc., improving the user experience and satisfaction.

[0027] First, the data integration and analysis platform and the real-time controllable rendering engine reflect the state, performance and changes of the physical system in real time, realizing the deep integration and interaction of physical space and digital space; real-time data feedback and visualization display provide users with intuitive and comprehensive system monitoring and management methods. Secondly, because the human-computer interaction subsystem works together with the generative big model, users can communicate with the system efficiently in the form of multi-round dialogues supported by context through an intuitive interactive interface. Whether through text input or voice interaction, the system can accurately identify and process the user's intentions, achieve instant feedback, and improve the user experience and human-computer interaction efficiency; the introduction of the generative big model enables the system to intelligently process the user's natural dialogue text, generate command-type and non-command-type content, provide users with more flexible and diverse interaction methods, support multi-round dialogues between users and the big model digital twin system, and improve the intelligence level of the system; the generated content processing subsystem intelligently analyzes and processes the output of the generative big model, ensuring the accurate execution of command-type content and the reasonable feedback of non-command-type content, improving the automation of the system and ensuring the accuracy and consistency of information. Furthermore, the instruction router can maintain the order, synchronization, asynchrony and contextual execution relationship of the instruction queue and the instruction logic directed graph, ensuring the accurate execution of the instructions and the timely feedback of the results. In addition, the digital twin system proposed in the present invention also supports virtual simulation and business simulation functions. Users can perform virtual simulation calculations and business simulation calculations by inputting instructions or pre-defining business scenarios, thereby pre-verifying business plans or simulating key business processes of enterprises, which not only helps to reduce the risks and costs of actual operations, but also improves the flexibility and scalability of the system. Finally, the optimization and improvement of the human-computer interaction subsystem enables users to interact with the system more conveniently through speech synthesis modules, sub-window modes, etc., improving the user experience and satisfaction.

[0028] Other advantages, objectives and features of the present invention will be described in the following description to some extent, and to some extent, will be obvious to those skilled in the art based on the following examination and study, or can be taught from the practice of the present invention. The objectives and other advantages of the present invention can be realized and obtained through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] In order to make the purpose, technical solutions and advantages of the present invention more clear, the present invention will be described in detail below in conjunction with the accompanying drawings, wherein:

[0030] Figure 1 A schematic diagram of a large-model digital twin system according to an embodiment of the present invention;

[0031] Figure 2 A schematic diagram of a generated content processing subsystem according to an embodiment of the present invention;

[0032] Figure 3 A schematic diagram of a large-model digital twin system including a virtual simulation model according to an embodiment of the present invention. DETAILED DESCRIPTION

[0033] The following describes the embodiments of the present invention by specific examples, and those skilled in the art can easily understand other advantages and effects of the present invention from the contents disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed in various ways 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 only illustrate the basic concept of the present invention in a schematic manner, and the following embodiments and features in the embodiments can be combined with each other without conflict.

[0034] Among them, the drawings are only used for illustrative explanations, and they only represent schematic diagrams rather than actual pictures, and should not be understood as limitations on the present invention. In order to better illustrate the embodiments of the present invention, some parts of the drawings may be omitted, enlarged or reduced, and do not represent the size of actual products. For those skilled in the art, it is understandable that some well-known structures and their descriptions in the drawings may be omitted.

[0035] The same or similar numbers in the drawings of the embodiments of the present invention correspond to the same or similar parts; in the description of the present invention, it should be understood that if the terms "upper", "lower", "left", "right", "front", "rear", etc. indicate the orientation or position relationship, they are based on the orientation or position relationship shown in the drawings, which is only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation. Therefore, the terms describing the position relationship in the drawings are only used for illustrative purposes and cannot be understood as limiting the present invention. For ordinary technicians in this field, the specific meanings of the above terms can be understood according to specific circumstances.

[0036] See also Figure 1 , is a schematic diagram of a large-scale digital twin system of an embodiment of the present invention; this embodiment describes in detail a large-scale digital twin system with comprehensive functions and clear structure, whose internal components work together to realize a complete closed loop from user input to physical world feedback. This large-scale digital twin system is mainly composed of the following core components, which work together to realize the various functions of digital twins:

[0037] The data integration and analysis platform receives data from sensors and edge computing devices in the actual physical system, builds a digital copy of the actual physical system in the physical space in the digital space to reflect the status, performance and changes of the actual physical system in real time, and realizes the integration and interaction of the physical space and the digital space.

[0038] Static three-dimensional model, the digital mapping of physical space in digital space; this static three-dimensional model is a digital model with three-dimensional space characteristics built based on computer graphics technology.

[0039] Real-time dynamic data is used to drive the visualization virtual operation of the static three-dimensional model. The real-time dynamic data is provided by a data integration and analysis platform; the dynamic data includes Internet of Things (IoT) sensor data, access video or real-time data of video recognition.

[0040] The real-time controllable rendering engine, as the presentation core of the digital twin results, is responsible for receiving interaction and rendering instructions from the command router, integrating the static three-dimensional model with the dynamic data, rendering a realistic digital twin effect, and feeding back the rendering results 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 can intelligently process input information such as user input, file reading, network requests, and output two types of generated content: command-type and non-command-type. These generated contents not only provide direct feedback to users, but also serve as the context for the next round of dialogue, supporting multiple rounds of dialogue between users and the big model digital twin system. The generative big model is also used to receive execution results and use the execution results as context to support subsequent multiple rounds of dialogue and decision-making. The application of the generative big model also enables the digital twin system to have stronger context-awareness. By intelligently inferring the user's true intention based on the user's historical dialogue and current context, and directly generating target instructions that match this intention. Compared with the traditional step-based instruction generation method, this is no longer limited to executing a series of intermediate steps step by step to achieve the goal, but directly focusing on meeting the user's actual needs. This transformation not only greatly improves the coherence and fluency of the interaction process, but also makes each dialogue more closely revolving around the user's central needs. Users no longer need to endure cumbersome steps and redundant information, but can directly and efficiently obtain the required services or information. The generative big model also improves the interaction mode of voice and speaker coordination. Users can communicate with the digital twin system directly in natural language. Specifically, the user first issues instructions or asks questions to the speaker through voice. The speaker, as a receiving device for sound signals, can clearly capture and transmit the user's voice information to the generative big model. In this process, the speaker's high-fidelity pickup technology and noise suppression algorithm ensure the clarity and accuracy of the voice signal, providing a solid foundation for subsequent processing. Subsequently, the generative big model uses advanced natural language processing technology to parse and understand the user's voice. It can identify the user's intention, extract key information, and generate corresponding responses or perform corresponding tasks based on this information combined with the interaction history context. In this process, the deep learning ability of the big model enables it to process complex language structures and semantic relationships, so as to accurately understand the user's real needs. After understanding the user's intention, the generative big model will interact with the big model digital twin system and convert the user's instructions or questions into a form that the digital twin system can parse and execute. As a digital mapping of the physical world, the digital twin system can simulate and reflect the state and behavior of the real world. Driven by the generative big model, the digital twin system can make corresponding adjustments or perform corresponding tasks according to the user's instructions. Finally, the execution results or responses of the digital twin system will be fed back to the user in the form of voice through the speaker again. This process not only realizes intelligent and efficient human-computer dialogue, but also provides users with an intuitive and convenient interactive experience. Users do not need complicated operations or inputs, they can communicate with the digital twin system instantly through voice and enjoy the convenience brought by intelligent services.

[0042] The generated content processing subsystem receives the generated content from the generative big model; directly returns the non-instruction generated content to the generative big model; parses the instruction generated content; first sends the parsing result to the instruction router, then collects the instruction execution results returned by the instruction router and returns them to the generative big model after processing. Figure 2 , is a schematic diagram of a generated content processing subsystem according to an embodiment of the present invention; it shows the logical structure of a specific implementation form of the generated content processing subsystem, and the generated content processing subsystem includes:

[0043] The generated content router intelligently selects the corresponding instruction parser according to the characteristics of the output content of the generative big model; the non-instruction generated content is returned to the generative big model; the instruction router can maintain the sequentiality, synchronization, asynchrony and contextual execution relationship of the instruction queue and the instruction logic directed graph, ensuring the accurate execution of the instructions and the timely feedback of the results.

[0044] The instruction parser converts the instruction-type generated content into an instruction set and forms an instruction queue or an instruction logic directed graph to be transmitted to the instruction router; the instruction router, as the core node of instruction processing, is responsible for parsing the instruction queue into control instructions or interaction and rendering instructions. The control instructions are transmitted to the physical control instruction distributor to control the physical objects, while the interaction and rendering instructions are transmitted to the real-time controllable rendering engine for rendering. At the same time, the instruction router is also responsible for receiving the execution results and transmitting these results to the generative large model through the execution result encapsulator to support subsequent multi-round dialogues and decisions.

[0045] The execution result encapsulator is used to transform and encapsulate the execution results fed back by the instruction router and transmit them to the generative large model.

[0046] In order to realize the interaction and feedback between the digital twin system and the physical world, this system also includes the following components:

[0047] Physical control instruction distributor: This distributor is responsible for transmitting control instructions to the actuators in the actual physical system to achieve control of physical objects. At the same time, it also receives feedback results from the actuators and transmits these results to the instruction router to complete the closed loop of the entire instruction execution process.

[0048] In the process of building a large-scale digital twin system, private knowledge of a specific business or domain is effectively integrated and applied to the system. This is usually achieved by building a domain knowledge vector library, which contains digital representations of various data, rules, logic or experience information related to specific business scenarios. The key to this integration method is that it is not just a simple storage or introduction of private domain knowledge into the system, but a deep and organic combination 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 performing virtual simulations, thereby ensuring that its application in specific business scenarios is both safe and accurate. The private knowledge data of a specific organization or domain and a series of task requirements or guidance are further trained and optimized through feedback reinforcement learning methods to improve its ability to complete specific tasks. This process aims to make the model better adaptable to specific scenarios and needs, thereby improving its performance and practicality.

[0049] The large-scale digital twin system described in this embodiment realizes a complete closed loop from user input to physical world feedback through highly integrated components and logically clear instruction processing and transmission mechanisms. 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] See also Figure 3 , is a schematic diagram of a large-model digital twin system including a virtual simulation model according to another embodiment of the present invention; Figure 1Based on the proposed large-scale digital twin system, its functions have been further expanded, and an important component has been added, which is the simulation model. The simulation model is a highly integrated computing module designed to perform in-depth virtual simulation calculations based on the specific needs of users, the instructions generated by the large model, or a predefined series of typical business scenarios. This model not only improves the practicality and flexibility of the digital twin system, but also provides strong support for its application in multiple industries. First, the simulation model receives direct input from the user, which may include various parameter settings, operation instructions, etc. At the same time, it can also receive instructions generated by the large-scale digital twin system based on user input or predefined conditions, which may involve complex computing tasks or specific simulation scenarios. After receiving the input, the simulation model will start the corresponding computing engine and perform virtual simulation calculations based on the input information. These calculations may involve multiple disciplines such as physics, chemistry, and biology, and are designed to simulate complex phenomena or processes in the real world. After the calculation is completed, the simulation model will integrate the result data with the three-dimensional model in the digital twin system. Through three-dimensional visualization technology, these result data are rendered on the output interface and presented to the user in an intuitive and vivid way. This large-model digital twin system not only has powerful data processing and intelligent decision-making capabilities, but also realizes virtual simulation of complex phenomena and visual display of results, thereby enabling pre-simulation and verification of events or business scenarios that have not occurred.

[0051] exist Figure 1Based on the proposed large-scale digital twin system, its functions have been further expanded, and an important component has been added, which is the business simulation model. The business simulation model is used to simulate key business processes or important facilities and equipment to meet the needs of various application scenarios such as training and teaching, maintenance support, etc. Through this model, users can have a deep understanding of each link of the business process and improve their operation skills and fault response capabilities. The business simulation model can accurately simulate the key business processes of the enterprise, including production processes, supply chain management, customer service, etc. Users can simulate business processes under different conditions in a virtual environment and observe their operation to find potential problems and improvement points. For important facilities and equipment, such as production line equipment, power facilities, oil depot storage and distribution, and facility equipment inspection and maintenance, the business simulation model can simulate their operating status and operation process. Users can practice equipment operation, maintenance and troubleshooting in a virtual environment to improve their actual operation capabilities. The business simulation model can be used as a training tool to provide employees with a vivid practice environment. By simulating real business scenarios, employees can learn business knowledge in a risk-free environment and improve their work skills and teamwork capabilities. In terms of facility and equipment maintenance, the business simulation model can simulate equipment failures and provide maintenance personnel with training on fault diagnosis and maintenance solutions. At the same time, the model can also provide historical data and forecast information on equipment maintenance to help maintenance personnel develop more reasonable maintenance plans. The business simulation model can be called through instructions generated by the large-model digital twin system. Therefore, users can customize personalized simulation scenarios and tasks through the large-model digital twin system according to specific needs, thereby further improving the simulation effect and application value.

[0052] exist Figure 1The proposed large-scale digital twin system also includes a traditional digital twin interactive interface, which is an important means for the large-scale digital twin system to exchange information with users. This interface is based on the traditional graphical user interface (GUI) design principles and combines the characteristics of digital twin technology to provide users with an intuitive, easy-to-use, and feature-rich operating environment. Through this interface, users can easily perform operations such as model browsing, data query, and command input to achieve comprehensive control of the digital twin system. In the large-scale digital twin system, the virtual simulation results are displayed through the traditional digital twin interactive interface. In order to meet the different needs of users, the system provides two display methods: users can choose to display the virtual simulation results directly in the current window. This method is suitable for users who want to quickly view and compare different simulation results in the same interface. By superimposing the virtual simulation results in the original window, users can intuitively see the changes of the model under different conditions, so as to make 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 analyze the simulation results in detail or share them with others. The newly opened window can be resized and positioned independently, which is convenient for users to observe and operate in detail. At the same time, users can also export simulation results into formats such as pictures and videos for subsequent analysis and sharing.

[0053] The large-scale digital twin system proposed in the present invention, firstly, reflects the state, performance and change of the physical system in real time through the data integration and analysis platform and the real-time controllable rendering engine, realizing the deep integration and interaction between the physical space and the digital space; the real-time data feedback and visual display provide users with intuitive and comprehensive system monitoring and management means. Secondly, since the human-computer interaction subsystem works with the generative large model, the user can communicate with the system efficiently through an intuitive interactive interface. Whether through text input or voice interaction, the system can accurately identify and process the user's intention, realize instant feedback, and improve the user experience and human-computer interaction efficiency; the introduction of the generative large model enables the system to intelligently process the user's natural dialogue text, generate command-type and non-command-type content, provide users with more flexible and diverse interaction methods, support multiple rounds of dialogue between users and the large-scale digital twin system, and improve the intelligence level of the system; the generated content processing subsystem intelligently analyzes and processes the output of the generative large model, ensuring the accurate execution of command-type content and the reasonable feedback of non-command-type content, improving the automation of the system, and ensuring the accuracy and consistency of information. Furthermore, the instruction router can maintain the order, synchronization, asynchrony and contextual execution relationship of the instruction queue and the instruction logic directed graph, ensuring the accurate execution of the instructions and the timely feedback of the results. In addition, the digital twin system proposed in the present invention also supports virtual simulation and business simulation functions. Users can perform virtual simulation calculations and business simulation calculations by inputting instructions or pre-defining business scenarios, thereby pre-verifying business plans or simulating key business processes of enterprises, which not only helps to reduce the risks and costs of actual operations, but also improves the flexibility and scalability of the system. Finally, the optimization and improvement of the human-computer interaction subsystem enables users to interact with the system more conveniently through speech synthesis modules, sub-window modes, etc., improving the user experience and satisfaction.

[0054] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solution of the present invention can be modified or replaced by equivalents without departing from the purpose and scope of the technical solution, which should be included in the scope of the claims of the present invention.

Claims

1. A large-scale digital twin system, characterized in that: include: A data integration and analysis platform, which constructs a digital copy of the actual physical system of the physical space in the digital space to reflect the status, performance and changes of the actual physical system in real time, and realizes the integration and interaction of the physical space and the digital space; A static three-dimensional model of a digital mapping of the physical space in the digital space; Real-time dynamic data for driving the virtual operation of the static three-dimensional model visualization, the real-time dynamic data being provided by the data integration and analysis platform; A real-time controllable rendering engine integrates and fuses the static 3D model with the dynamic data in real time and renders the model to visualize the operation of the actual physical system; receives the rendering control instruction parsed by the instruction router, adjusts the rendering area, rendering mode and rendering parameters, and sends the rendering result to the display terminal of the human-computer interaction subsystem; Return the command execution result to the command router; The instruction router receives and parses the instruction from the instruction issuer, and forwards the parsed instruction to the corresponding instruction execution unit; collects the instruction execution results and returns them to the instruction issuer, and at the same time transmits them to the generated content processing subsystem for processing and returns them to the generated large model to realize the global context; the instruction issuer includes a human-computer interaction subsystem and a generated content processing subsystem; The human-computer interaction subsystem receives the user's human-computer interaction input, identifies the human-computer interaction actions, processes the human-computer interaction data, analyzes the interaction intention, calls the relevant system to realize the interaction intention, and presents the human-computer interaction results; when the human-computer interaction involves control instructions, the control instructions are transmitted to the instruction router and then forwarded to the corresponding instruction execution unit; the human-computer interaction subsystem is connected to the generative big model, when the user enters the natural dialogue text in the specific text box of the interaction interface, or when the user uses voice wake-up to enter the natural dialogue text, the natural dialogue text will be transmitted to the generative big model, and the generative big model will return the response result to the human-computer interaction subsystem after processing; The generative large model receives the natural dialogue text input by the human-computer interaction subsystem, generates two types of generated content, command-type and non-command-type, according to the text, and transmits them to the generated content processing subsystem; receives the content returned by the generated content processing subsystem and transmits it to the human-computer interaction subsystem, and records the returned content into the context to support subsequent multiple rounds of question-answering interactive dialogue and decision-making; A generated content processing subsystem receives generated content from the generative macro model; For the non-instruction type generated content, directly return it to the generative big model; for the instruction type generated content, parse it; first send the parsing result to the instruction router, then collect the instruction execution results returned by the instruction router, process them and return them to the generative big model; The physical control instruction distributor receives the physical control instruction parsed by the instruction router and transmits it to the actuator in the physical system to control the physical object; and transmits the execution result of the physical control instruction back to the instruction router.

2. The large-scale digital twin system according to claim 1, characterized in that: The instructions of the instruction issuer include a single instruction, an instruction queue and an instruction logic directed graph; the instruction router maintains the sequentiality, synchronization, asynchrony and context execution relationship of the instruction queue and the instruction logic directed graph, and packages the instruction execution results in batches or in batches according to the instruction logic and instruction requirements and transmits them back to the instruction issuer; the execution unit includes a real-time controllable rendering engine and a physical control instruction distributor.

3. The large-model digital twin system according to claim 2, characterized in that: The generated content processing subsystem includes: A generated content router analyzes the generated content of the generative big model, transmits the instruction-type generated content to a corresponding instruction parser, and transmits the non-instruction-type generated content back to the generative big model; An instruction parser converts the instruction-type generated content into an instruction set and forms the instruction queue or the instruction logic directed graph and transmits it to the instruction router; The execution result encapsulator is used to transform and encapsulate the execution result fed back by the instruction router and transmit it to the generative large model.

4. The large-model digital twin system according to any one of claims 1 to 3, characterized in that: The human-computer interaction subsystem is provided with a speech synthesis module, and the speech synthesis module is used for playing the human-computer interaction result.

5. The large-model digital twin system according to any one of claims 1 to 3, characterized in that: It also includes a virtual simulation model, which performs virtual simulation calculations according to 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 to pre-verify the business plan.

6. The large-scale digital twin system according to claim 5, characterized in that: The human-computer interaction subsystem supports a sub-window mode to display the result of the virtual simulation calculation.

7. The large-model digital twin system according to any one of claims 1 to 3, characterized in that: Also includes: The business simulation component performs business simulation calculations based on user input, generated instructions or predefined business scenarios, applies the business simulation calculation results to the static three-dimensional model, and visually simulates the enterprise's key business processes, important facilities and equipment, and operating logic or operating methods to support training and maintenance.

8. The large-scale digital twin system according to claim 7, characterized in that: The human-computer interaction subsystem supports a sub-window mode to display the business simulation calculation results.

Images