An intelligent three-dimensional engine system for industrial internet of things

Abstract

This invention discloses an intelligent 3D engine system for the Industrial Internet of Things (IIoT), comprising the following modules: a modeling module for creating 3D models of industrial scenes and establishing corresponding data linked lists; a rendering module for rendering 3D scenes in real time from the user's perspective; a picking module for responding to user interactions and selecting 3D model data; a data association module for dynamically associating 3D models with external data sources; and a data processing module for performing algorithmic analysis and processing on the associated data. All modules collaborate efficiently as a whole through unified kernel management, data structures, and thread mechanisms. All modules access model data and real-time data through pointer interfaces, enabling data sharing and synchronization among modules. This system provides a platform for building industrial 3D virtual scenes and associates relevant production data, maintenance data, attribute data, etc., with 3D models, and introduces mathematical models to construct intelligent digital twins.

Description

TECHNICAL FIELD

[0001] The present application relates to computer graphics and three-dimensional engine creation, and belongs to the technical field of computers. The present application designs an intelligent three-dimensional engine system applied to industrial Internet of Things, which can provide a platform for building an industrial three-dimensional virtual scene, associate production data, operation and maintenance data, attribute data and the like with a three-dimensional model, introduce a mathematical model, and construct an intelligent digital twin. BACKGROUND

[0002] The rapid development and wide application of computer graphics technology have promoted the development of digital construction and digital twinning in the industrial field. For an industrial enterprise, by creating a three-dimensional model and associating the three-dimensional model with attribute data, production data, operation and maintenance data and the like, the industrial enterprise can help to connect various data sources, and based on a three-dimensional scene and data flow, carry out comprehensive simulation, early warning, training and the like, so as to improve production efficiency, reduce operating costs, and thus achieve the goal of maximizing benefits. In this process, the selection and use of a three-dimensional engine system is crucial, and if not properly selected, various application problems will be caused, and even the effect of digital twinning construction will deviate from the original intention, wasting manpower and resources.

[0003] At present, there are two major categories of three-dimensional engine systems commonly used in the industrial field, the first category being a three-dimensional design engine of Europe and the United States, and the second category being a game development engine of Europe and the United States (in actual application, there is a third category of engine, i.e., a secondary development product based on the above two categories of engines, and the kernel of this engine is still the above two categories of engines, so it is not listed separately here). The application of these two categories of engines (hereinafter referred to as "European and American engines") in the digital construction of industrial enterprises has the following problems.

[0004] First, both of these two categories of engines are developed by foreign enterprises and have been developed for many years, and when applied in the domestic industrial field (especially when important industries and sensitive data are involved), their security must be highly concerned. With the development of the information technology application innovation industry, this is particularly important.

[0005] Second, these two categories of engines are not developed for industrial Internet of Things construction. The three-dimensional design engine of Europe and the United States mainly serves industrial design and production, and pursues modeling accuracy, with a model structure that is extremely detailed. The game development engine of Europe and the United States mainly serves game development, and pursues appearance and dynamic effects (such as light and shadow effects, particle effects, etc.). Therefore, the three-dimensional model created by the former has a huge size and occupies too much memory space, while the three-dimensional scene created by the latter has high requirements for computer hardware, so that ordinary users cannot use the development results smoothly on a computer with general configuration. As a result, the three-dimensional digital system built at great expense often becomes a decoration in the process of enterprise operation and maintenance, causing great waste.

[0006] Thirdly, the data association and data processing functions of the two types of engines are imperfect. The three-dimensional design engine in Europe and the United States can only associate static data and cannot associate dynamic data. The data association of the game development engine in Europe and the United States requires a large amount of work and can cause the memory occupied by the three-dimensional scene to increase exponentially, thereby requiring higher computer hardware. In addition, the two types of engines cannot process data according to the actual needs of enterprises and cannot mine data value. The basis of digital twinning is a three-dimensional model, and the core is data and algorithms. Only by processing data association can digital twinning be separated from the misunderstanding of "vase", thereby promoting the application and development of industrial Internet of Things.

[0007] Fourthly, users can only develop based on the two types of engines and cannot obtain the highest authority of the drawing chain table pointer created by the engine, which restricts the convenience of the engine in real-time data association and drawing speed optimization, greatly increases the artificial workload, improves the requirements of the engine on hardware, and ultimately restricts the development of the engine in the field of industrial Internet of Things.

[0008] As can be seen from the above, the current popular three-dimensional engine system is neither a domestic software nor can it well meet the application needs of industrial Internet of Things. In view of the above problems, the present application designs an intelligent three-dimensional engine system for industrial Internet of Things. SUMMARY

[0009] The technical purpose of the present application is to design an intelligent three-dimensional engine system that is completely domestic and covers the whole life cycle of industrial enterprise Internet of Things data. The system can create / use lightweight three-dimensional models, associate static and dynamic data, and complete data processing. The system kernel is completely independently developed and does not depend on foreign three-dimensional engines. The memory occupied by the industrial enterprise scene model created is small, and the system can run smoothly on a general computer with an independent graphics card. The system obtains real-time data from various data sources through a communication thread and displays it in real time in the scene. The three-dimensional topological structure and three-dimensional orientation data attached to the scene can help data processing.

[0010] The technical scheme adopted by the present application is an intelligent three-dimensional engine system for industrial Internet of Things, which comprises the following modules: A modeling module for creating an industrial scene three-dimensional model and establishing a corresponding data linked list; A drawing module for real-time drawing of a three-dimensional scene according to a user's perspective; A picking module for responding to user interaction and selecting a three-dimensional model; A data association module for dynamically associating a three-dimensional model with an external data source; A data processing module for algorithm analysis and processing of associated data.

[0011] The modules are connected through a unified kernel management, data structure and thread mechanism to form a whole through efficient cooperation. All the modules access model data and real-time data through a pointer interface to realize data sharing and synchronization among the modules.

[0012] Further, the modeling module is connected with the rendering module. After the modeling module loads various three-dimensional models (devices, pipelines, instruments, etc.) through the CLoad3DS class, the models are stored in a rendering linked list managed by CMy3DICDoc. The rendering module obtains model data from the rendering linked list in each frame rendering period to call the Draw function of each device class for rendering. The model properties (such as material, color, and primary key code) defined in the modeling process of the modeling module become the basis for subsequent rendering, picking, and data association.

[0013] The picking module is interactively connected with the rendering module. The picking module captures user operations through a mouse event function (such as OnLButtonDown) in the CMy3DICView class to generate a picking ray. The picking ray collides with the scene model generated by the rendering module to return the pointer of the selected model through a picking buffer. Further, after selection, the picking module feeds back to the rendering module and highlights the corresponding model to realize visual feedback.

[0014] The rendering module is interactively connected with the data association module. In the rendering process, the rendering module determines whether to trigger data update according to the device state tag in CMy3DICDoc. The data association module obtains real-time data from an external data source through an independent thread (such as m_ReadDataThread) and updates a data linked list. The rendering module reads the latest data from the data linked list in the next rendering period to dynamically update model color, state icon, and other visual elements.

[0015] The data association module is connected with the data processing module. The data association module provides real-time data (such as temperature, pressure, and flow) bound to the three-dimensional model to the data processing module. The data processing module calls built-in algorithms or external mathematical models according to business requirements to calculate and analyze the data (such as health assessment and fault warning). The processing result of the data can be fed back to the rendering module for visual display or output to other industrial systems through a three-dimensional topological structure.

[0016] The analysis result of the data processing of the data processing module (such as an abnormal state of a device) can reversely drive the rendering module to adjust model display (such as alarm flashing). The processing result can also be used to optimize the model property setting of the modeling module or guide the interactive logic of the picking module. Through three-dimensional topological relationship data, the data processing module can enhance the understanding and expression ability of the modules for an industrial scene.

[0017] Furthermore, the thread mechanism is responsible for the drawing module, the picking module, and user interface responses. The communication thread ensures that the data association module operates independently, asynchronously acquiring data from systems such as SIS / MIS through a multi-threaded mechanism without affecting the smoothness of the main thread.

[0018] Furthermore, the modeling module includes the CLoad3DS class, CBMPLoad class, and CAxis class, which are used to load 3D models, textures, and coordinate axis information.

[0019] Furthermore, the drawing module includes the Draw function in each device class, which optimizes the drawing range through region definition and distance judgment, and only draws the model within the user's field of view.

[0020] Furthermore, the picking module includes the OnKeyDown, OnLButtonDown, OnMouseMove, and OnMouseWheel functions in the CMy3DICView class, which are used to enable user interaction with the 3D model.

[0021] Furthermore, the data association module includes thread functions in the CMy3DICDoc class and functions in each device class that access data through the primary key of the 3D model, enabling real-time data acquisition and updating.

[0022] Furthermore, the data processing module includes calculation functions for each device class, and supports the introduction of external algorithms for data analysis and processing.

[0023] Furthermore, the system also includes a CMy3DICDoc class, which manages drawing lists and data lists, and grants access permissions to various device classes and view classes via pointers.

[0024] Furthermore, the drawing linked lists include OBJECT linked lists for equipment, PIPE3D linked lists for pipeline nodes, PIPE3Dinfo linked lists for pipes, CONDUCT3D linked lists for cable tray nodes, and INSTRUMENT3D linked lists for instruments. Each linked list node records the relationships between models through a brotherNode array. In the kernel management, the CMy3DICDoc class serves as the core data management unit of the system. The CMy3DICDoc class is responsible for creating and maintaining the drawing linked lists and data linked lists. The drawing linked lists record the geometric, topological, and state information of all 3D models; the data linked lists store real-time data obtained from external systems.

[0025] Furthermore, the system acquires data from an external real-time database through an independent data communication thread and updates the model state in real time in the main thread, reducing the data access load. Complex algorithms in the data processing module of the computation thread are executed in an independent thread, avoiding blocking scene rendering and interaction.

[0026] Compared with existing technologies, firstly, the 3D engine system described in this invention is completely independently developed, without relying on the engines of related products or their secondary development platforms, thus avoiding the risk of "backdoors." Secondly, while creating the drawing linked list, this 3D engine system records the topological relationships and orientation data between 3D models, which can assist in subsequent data association and data processing for industrial enterprises. Thirdly, the 3D engine system of this invention optimizes the created industrial enterprise scene model, significantly reducing the memory requirements and reducing the workload of the GPU by using the CPU's field of view determination, enabling the 3D engine system to run smoothly on ordinary computers equipped with dedicated graphics cards. Finally, the 3D engine system of this invention combines the drawing linked list and the data linked list. The data linked list is obtained from various data sources through a communication thread independent of the main process, which allows the drawing linked list to display real-time data, while effectively controlling the real-time access volume of the data linked list. Attached Figure Description

[0027] Figure 1 This is a schematic diagram demonstrating the system of the present invention. Detailed Implementation

[0028] The present invention will now be described in detail with reference to the accompanying drawings and embodiments.

[0029] This invention employs the C++ language and, based on a basic graphics library, independently develops a 3D engine kernel. The CMainFrame class is responsible for window frame creation, the CMy3DICView class for 3D graphics rendering and user keyboard and mouse event handling, the CCamera class for user navigator calculation, and the CMy3DICDoc class for creating drawing and data lists. Classes for various industrial IoT devices are responsible for model loading, rendering, data association, and data processing for their respective devices. Various dialog classes receive user interactions with the scene. Data can be shared between these classes. During each rendering cycle of CMy3DICView, the pointers to the drawing lists created by CMy3DICDoc can be directly programmed. Device classes can directly modify the data lists of CMy3DICDoc based on user interactions. This allows 3D application development based on this engine to be unrestricted by kernel functionality, enabling intelligent development according to the needs of 3D display, data association, or data processing within each class. This significantly improves rendering and data access efficiency while reducing hardware requirements.

[0030] In the Industrial Internet of Things (IIoT) ecosystem, the CLoad3DS class is responsible for loading and rendering the 3D models and animations for all devices. Therefore, textures and 3D models within the same project can be shared, reducing memory requirements. Furthermore, during the scene rendering process in the RenderScene function of the CMy3DICView class, the 3D engine can adjust the colors and display blocks of textures and models based on real-time data from each device class and the CMy3DICDoc data list. This reduces the amount of data occupied by the 3D model while enriching the final model's expressive effect.

[0031] When using European and American engines for secondary development, users cannot obtain the highest level of access to the drawing linked list. To edit the data of a specific node in the drawn linked list, access must first be obtained through polling, which is inefficient. However, the CMy3DICDoc class in the 3D engine system described in this invention manages and maintains the drawing linked list through pointers and is open to various device classes and CMy3DICView classes. Users can directly obtain the data of the corresponding device through pointers, greatly improving both drawing efficiency and data access efficiency.

[0032] When the main thread of the 3D engine system described in this invention uses the RenderScene function in the CMy3DICView class to draw a 3D model, the CMy3DICDoc class can start several data communication threads (such as m_ReadDataThread, m_GetSISDataThread, etc.) according to the number of data sources to maintain communication with the external real-time database, obtain the real-time data corresponding to the 3D model within the field of view, and ensure the real-time display of the 3D model status with minimal communication, which reduces the data communication volume of the 3D engine and also reduces the load on the data server.

[0033] When drawing a 3D model, the 3D engine system of this invention not only retains the drawing linked lists (including the OBJECT linked list for equipment, the PIPE3D linked list for pipeline nodes, the PIPE3Dinfo linked list for pipes, the CONDUCT3D linked list for cable tray nodes, and the INSTRUMENT3D linked list for instruments, etc.), but also uses the brotherNode[] array in each linked list node to record the relationship between the current node and other linked list nodes, ultimately forming a network structure between models. This relationship can assist in data processing.

[0034] The 3D engine system described in this invention reduces the workload of 3D rendering by defining regions (CDistrict class) and determining distances (CDistance class). Specifically, it uses the current viewing position (the 3D engine system camera, ex, ey, ez) as the center, rendering only the 3D scene within that center and a portion of the scene in adjacent areas along the line of sight. For relatively isolated instruments and model details in industrial scenes, rendering is determined based on distance. In this way, by utilizing the CPU's decision-making capabilities to optimize the GPU's rendering range, the rendering workload of the 3D engine system can be reduced.

[0035] The technical architecture and workflow are as follows: The 3D engine system described in this invention comprises five modules: modeling module, drawing module, picking module, data association module, and data processing module.

[0036] The modeling module provides users with graphical modeling tools for creating industrial enterprise scenes. It allows users to create 3D models of these scenes and includes classes such as CLoad3DS, CBMPLoad, and CAxis. The rendering module is responsible for rendering the industrial enterprise scene and ensuring smooth dynamics, facilitating user browsing of the 3D scene. The rendered scene changes according to the user's perspective and dynamically displays the real-time status of corresponding equipment using the 3D models within the scene. This module includes the Draw function for each equipment class. The picking module responds to user mouse and keyboard inputs, determines the selected 3D model in the current scene, and highlights it. This module includes the OnKeyDown, OnLButtonDown, OnMouseMove, and OnMouseWheel functions in the CMy3DICView class. The data association module associates the industrial enterprise data source with the user-selected 3D model. This module includes corresponding thread functions in the CMy3DICDoc class and functions in each equipment class that access built-in data via the 3D model's primary key encoding. The data processing module introduces various algorithms to display the corresponding calculation results to the user and includes calculation functions in each equipment class.

[0037] This invention describes the workflow of the entire system and the connections between modules through the following steps.

[0038] 1. Model Building Model building is primarily accomplished through the modeling module within the 3D engine system. This process utilizes classes such as CCreateHouseObject, CNewPipeNodeDlg, and CInstrument. The model building process is also a data input process. The 3D engine system establishes various data lists (including the OBJECT list for equipment, the PIPE3D list for pipeline nodes, the PIPE3Dinfo list for pipes, the CONDUCT3D list for cable tray nodes, and the INSTRUMENT3D list for instruments), recording model feature data (such as the PIPE3Dinfo list containing color r, g, b, name sPipeName, pipe material sMaterial, pipe medium sFluid, pipe start point startC and end point endC, pipe radius, and pipe attribute data).

[0039] 2. Model Interaction Model interaction is the process of enabling users to interact with 3D models, primarily responding to user mouse operations. This is mainly accomplished by the `ProcessSelection` function in the `CMy3DICView` class. Each interaction corresponds to a linear algebra calculation. The 3D engine system treats the user's "clicking the mouse button" as emitting a "ray." The `ProcessObject` function in the `CMy3DICView` class calculates in real time which models this ray (originating point `basePt`, originating direction `dirPt`) intersects with, and draws them one by one to the pick buffer using the `Draw` function in each device class. The pick buffer returns a pointer `ptr` to the layer name number corresponding to the nearest 3D model. The 3D engine system compares the layer name number with the drawing list to find the corresponding code for the model, which is then passed to the subsequent data association module for data querying. Simultaneously, the 3D engine system defines the model as "selected" in the drawing list, prompting the drawing module to highlight it.

[0040] 3. Scene drawing Scene rendering is the process of displaying a 3D scene in real time from the user's perspective. Because users constantly change their perspective during software operation, the 3D engine system needs to redraw the model for the user from the new perspective in a timely manner. If the rendering process is sluggish, then "model interaction" and "data association" cannot be realized smoothly.

[0041] Therefore, to ensure smooth visuals, the 3D engine system continuously calls the RenderScene function in the CMy3DICView class to render the scene. Simultaneously, the Draw function in each device class optimizes each frame using region definition and distance checks, ensuring a refresh rate of at least 15 frames per second. The 3D engine system first uses the ComputeAllDistNum function in the CDistrict class to determine the region number of each device, then uses the IsInKen function in the CCamera class to determine if each region is within the current field of view. It then performs additional IsInKen checks on each device within the camera's area, ultimately determining which models are outside the current field of view and which, although theoretically within the field of view, are too far away to be meaningful for display. For these two types of models, the 3D engine system chooses to ignore them during refresh, i.e., it does not render them. This working mode saves computational resources and conforms to the physiological mechanisms of the human eye.

[0042] 4. Data Association After rendering the scene, the 3D engine system associates the models displayed in the scene with the data sources, primarily through the data association module. After years of digital transformation, industrial enterprises typically have primary keys in their data sources. For example, the primary key in a thermal power plant's SIS / MIS system is kks (power plant identification system code). Using kks as the theoretical primary key for the 3D model, the 3D engine system can obtain the corresponding kks by selecting a model, and then use the access functions provided by each data source to retrieve data from each source. To obtain data faster and better, the 3D engine system first uses ComputeAllDistNum to determine if the device model is in the same area as the 3D engine camera; secondly, it uses IsInKen to determine if the device model is in front of the 3D engine camera; then, it uses the CDistance class to determine if the distance between them meets the display requirements; and finally, it tags each device in the rendering list with whether it needs real-time data updates, thus reducing data access volume and the load on the data server. Furthermore, because the data association model uses a threaded access mode, it does not consume the main process's processing time, ensuring smooth operation of the main process.

[0043] 5. Data Processing Data processing encompasses both the data processing within the 3D engine system itself and its auxiliary data processing for other enterprise systems. Regarding the former, the 3D engine system can incorporate various algorithms and mathematical models to process data associated with 3D models, enabling functions such as operation and maintenance optimization. Regarding the latter, the 3D engine system can output its processing results and unique 3D topology structures to assist other systems in achieving more comprehensive data processing. Taking a thermal power plant as an example, the 3D engine system can employ different algorithms to analyze equipment inspection data, enabling equipment health monitoring. Utilizing the 3D topology structures provided by the 3D engine system, other systems can quickly clarify the relationships between equipment, achieving data filtering and dimensionality reduction.

Claims

1. A smart 3D engine system for the Industrial Internet of Things, characterized in that, Includes the following modules: The modeling module is used to create 3D models of industrial scenes and establish corresponding data linked lists. The rendering module is used to render 3D scenes in real time according to the user's perspective; The picking module is used to respond to user interaction and select the 3D model. The data association module is used to dynamically associate 3D models with external data sources; The data processing module is used for algorithmic analysis and processing of related data; The various modules collaborate efficiently as a whole through a unified kernel management, data structure, and thread mechanism; all modules access model data and real-time data through pointer interfaces, enabling data sharing and synchronization among the modules.

2. The intelligent 3D engine system for the Industrial Internet of Things according to claim 1, characterized in that, The modeling module is connected to the drawing module. The modeling module loads various 3D models it creates through the CLoad3DS class and stores them in the drawing linked list managed by CMy3DICDoc. In each frame rendering cycle, the rendering module retrieves model data from the aforementioned rendering chain list to call the Draw function of each device class for rendering; during the modeling process, the modeling module defines model attributes that form the basis for subsequent rendering, picking, and data association. The picking module and the drawing module are interconnected. The picking module captures user operations through mouse event functions in the CMy3DICView class and generates a picking ray. The picking ray performs collision detection with the scene model generated by the drawing module and returns a pointer to the selected model through the picking buffer. Furthermore, after selection, the picking module feeds back to the drawing module and highlights the corresponding model to achieve visual feedback. The rendering module and the data association module interact with each other. During the rendering process, the rendering module determines whether to trigger a data update based on the device status label in CMy3DICDoc. The data association module obtains real-time data from an external data source through an independent thread and updates the data list. In the next rendering cycle, the rendering module reads the latest data from the data list and dynamically updates visual elements such as model color and status icons. The data association module is connected to the data processing module, and the data association module provides the data processing module with real-time data bound to the 3D model; The data processing module calls built-in algorithms or external mathematical models to perform calculations and analyses on the data according to business needs; the data processing results can be fed back to the drawing module for visualization, or output to other industrial systems through a three-dimensional topology structure; The data processing module's analysis results drive the drawing module to adjust the model display; the processing results can also be used to optimize the model attribute settings of the modeling module, or guide the interaction logic of the picking module; through 3D topological relationship data, the data processing module can enhance each module's understanding and expression of industrial scenarios.

3. The intelligent 3D engine system for the Industrial Internet of Things according to claim 1, characterized in that, The thread mechanism is responsible for the drawing module, the picking module, and the user interface response; the communication thread is used to ensure that the data association module runs independently and obtains data asynchronously from the SIS / MIS system through a multi-threaded mechanism.

4. The intelligent 3D engine system for the Industrial Internet of Things according to claim 1, characterized in that, The modeling module includes the CLoad3DS class, CBMPLoad class, and CAxis class, which are used to load 3D models, textures, and coordinate axis information.

5. The intelligent 3D engine system for the Industrial Internet of Things according to claim 1, characterized in that, The drawing module includes the Draw function in each device class, which optimizes the drawing range by defining the region and judging the distance, and only draws the model within the user's field of view.

6. The intelligent 3D engine system for the Industrial Internet of Things according to claim 1, characterized in that, The picking module includes the OnKeyDown, OnLButtonDown, OnMouseMove, and OnMouseWheel functions in the CMy3DICView class, which are used to enable user interaction with the 3D model.

7. The intelligent 3D engine system for the Industrial Internet of Things according to claim 1, characterized in that, The data association module includes thread functions in the CMy3DICDoc class and functions in each device class that access data through the primary key of the 3D model, enabling real-time data acquisition and updating.

8. The intelligent 3D engine system for the Industrial Internet of Things according to claim 1, characterized in that, The data processing module includes calculation functions for each device class and can introduce external algorithms for data analysis and processing.

9. The intelligent 3D engine system for the Industrial Internet of Things according to claim 1, characterized in that, The system also includes the CMy3DICDoc class, which manages drawing lists and data lists, and grants access permissions to various device classes and view classes through pointers.

10. The intelligent 3D engine system for the Industrial Internet of Things according to claim 1, characterized in that, The drawing linked lists include OBJECT (equipment class), PIPE3D (pipeline network node class), PIPE3Dinfo (pipeline class), CONDUCT3D (cable tray node class), and INSTRUMENT3D (instrument class). Each linked list node records the relationships between models through a brotherNode array. In the kernel management, the CMy3DICDoc class serves as the core data management unit of the system. The CMy3DICDoc class is responsible for creating and maintaining the drawing linked lists and data linked lists. The drawing linked lists record the geometric, topological, and state information of all 3D models; the data linked lists store real-time data obtained from external systems. The system obtains data from an external real-time database through an independent data communication thread and updates the model state in real time in the main thread, reducing the data access load; the complex algorithms in the data processing module of the calculation thread are executed in an independent thread to avoid blocking scene rendering and interaction.

Images