Operation and maintenance digital twin rendering method based on dynamic and static entity separation and twin differential synchronization

Abstract

The present application relates to the field of digital twinning, and more particularly to a kind of operation and maintenance digital twin rendering method based on dynamic and static entity separation and twin differential synchronization, obtain the digital twin model data of target operation and maintenance scene, according to the dynamic change attribute of each entity, entity is divided into static entity and dynamic entity;Based on the model data of static entity, static scene base is constructed and pretreated, and static rendering resource is generated;The state data of dynamic entity is acquired in real time and compared with historical state data, and the state change difference is extracted;According to state change difference, determine the target dynamic entity that occurs state change, only to target dynamic entity is incrementally rendered and updated, and dynamic rendering data is generated;Static rendering resource and dynamic rendering data are rendered and fused, and real-time rendering picture of operation and maintenance digital twin is generated and output. Through dynamic and static separation and differential synchronization combination, the computing overhead of digital twin rendering is reduced, and the real-time performance of operation and maintenance scene rendering is improved.

Description

Technical Field

[0001] This application relates to the field of digital twin technology, and in particular to an operation and maintenance digital twin rendering method based on the separation of dynamic and static entities and the synchronization of twin differentials. Background Technology

[0002] In existing digital twin rendering solutions, the rendering management of static and dynamic entities is typically done relatively independently. Static entities are rendered based on pre-built model data, while dynamic entities are updated based on real-time acquired state data, with the two being merged and output at the rendering level. However, in existing solutions, the management of static and dynamic entities is relatively independent. Once the static scene has completed preprocessing, its visual content no longer adjusts with the presence of dynamic entities, resulting in a lack of effective collaboration mechanisms between the static base and dynamic entities. Furthermore, state changes of dynamic entities are usually handled through direct rendering, lacking effective filtering and refinement of change data, and computational overhead increases significantly with the number of dynamic entities.

[0003] Furthermore, in existing technologies, state changes among multiple dynamic entities often do not occur in isolation, but are interconnected through physical connections or logical relationships. A state change in one entity may trigger a chain reaction of changes in multiple related entities. Existing solutions lack mechanisms for identifying and synchronizing these interconnected changes, typically treating each entity's change as an independent event and processing them one by one. This leads to misalignments in the timeline of state updates of related entities in the virtual twin view. Simultaneously, the relationships between entities themselves may change during operation and maintenance, making existing static identification methods based on fixed relationships ill-suited to this dynamic nature. On the other hand, existing solutions set relatively rigid boundaries in the division between static and dynamic entities, assuming entities are either completely static or completely dynamic. However, in operation and maintenance scenarios, there are entities whose geometric models and spatial locations remain unchanged while their operational status information continuously changes, making it difficult for existing classification methods to reasonably categorize them. Summary of the Invention

[0004] To address one or more problems in the existing technology, the main objective of this application is to provide a digital twin rendering method for operation and maintenance based on the separation of dynamic and static entities and the synchronization of twin differentials.

[0005] To achieve the aforementioned objectives, this application proposes a digital twin rendering method for operation and maintenance based on the separation of dynamic and static entities and the synchronization of twin differentials. The method includes: Obtain digital twin model data for the target operation and maintenance scenario; The dynamic change attributes of each entity in the digital twin simulation data under the running state are analyzed, and the entities are divided into static entities and dynamic entities based on the analysis results. Based on the model data of the static entities, a static scene base is constructed, and the static scene base is preprocessed to generate static rendering resources. The state data and corresponding historical state data of the dynamic entity are acquired in real time, and the state change difference of the dynamic entity is extracted based on the difference between the current state data and the historical state data. Based on the state change difference, the target dynamic entity that has undergone a state change is determined from the dynamic entities, and incremental rendering updates are performed only on the target dynamic entity to generate dynamic rendering data. The static rendering resources and the dynamic rendering data are rendered and fused to generate and output a digital twin real-time rendering screen of the target operation and maintenance scene.

[0006] This application also provides an operation and maintenance digital twin rendering system based on dynamic and static entity separation and twin differential synchronization, including: The acquisition module is used to acquire digital twin model data for the target operation and maintenance scenario; The analysis module is used to analyze the dynamic change attributes of each entity in the digital twin simulation data under the running state, and divide the entities into static entities and dynamic entities based on the analysis results. The construction module is used to construct a static scene base based on the model data of the static entity, and to preprocess the static scene base to generate static rendering resources. The extraction module is used to acquire the state data and corresponding historical state data of the dynamic entity in real time, and extract the state change difference of the dynamic entity based on the difference between the current state data and the historical state data. The update module is used to determine the target dynamic entity that has undergone a state change from the dynamic entities based on the state change difference, and to perform incremental rendering updates only on the target dynamic entity to generate dynamic rendering data. The output module is used to render and fuse the static rendering resources with the dynamic rendering data to generate and output a digital twin real-time rendering screen of the target operation and maintenance scene.

[0007] This application also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps of any of the methods described above.

[0008] This application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of any of the methods described above.

[0009] The operation and maintenance digital twin rendering method based on static and dynamic entity separation and differential synchronization in this application pre-divides entities in the operation and maintenance scenario into static and dynamic entities, allowing for separate processing of the static and dynamic parts without interference. Static entities are pre-processed to generate reusable rendering resources, eliminating the need for repeated rendering calculations in every frame and fundamentally reducing the computational burden of real-time rendering. Dynamic entities are incrementally updated only for changed parts through differential extraction, avoiding the resource waste caused by re-rendering the entire scene. More importantly, static rendering resources and dynamic incremental update data maintain visual consistency during rendering fusion, ensuring high-quality presentation of the static parts in the twin image while ensuring that the dynamic parts can change in real time with the actual state. The two paths of static pre-rendering and dynamic differential synchronization each perform their respective functions while cooperating with each other, jointly achieving smooth, real-time, and accurate presentation of digital twin images in large-scale operation and maintenance scenarios. Compared with the method of directly rendering the entire scene in full, or the method of only distinguishing between static and dynamic at the data level but still processing at a uniform granularity in the rendering layer, rendering efficiency and real-time response capability are significantly improved. Attached Figure Description

[0010] Figure 1 This is a flowchart illustrating an embodiment of the digital twin rendering method for operation and maintenance based on the separation of dynamic and static entities and the synchronization of twin differentials according to this application. Figure 2 This is a flowchart illustrating another embodiment of the digital twin rendering method for operation and maintenance based on the separation of dynamic and static entities and the synchronization of twin differentials. Figure 3 This is a schematic block diagram of an operation and maintenance digital twin rendering system based on dynamic and static entity separation and twin differential synchronization according to an embodiment of this application; Figure 4 This is a schematic block diagram of the structure of a computer device according to an embodiment of this application.

[0011] The realization of the purpose, functional features and advantages of this application will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0012] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0013] Reference Figure 1 This application provides a digital twin rendering method for operation and maintenance based on the separation of dynamic and static entities and the synchronization of twin differentials. The method includes: S1. Obtain digital twin model data for the target operation and maintenance scenario; S2. Analyze the dynamic change attributes of each entity in the digital twin simulation data under the running state, and divide the entities into static entities and dynamic entities according to the analysis results; S3. Based on the model data of the static entity, construct a static scene base and preprocess the static scene base to generate static rendering resources; S4. Real-time acquisition of the state data and corresponding historical state data of the dynamic entity, and extraction of the state change difference of the dynamic entity based on the difference between the current state data and the historical state data. S5. Based on the state change difference, determine the target dynamic entity that has undergone a state change from the dynamic entities, and perform incremental rendering updates only on the target dynamic entity to generate dynamic rendering data. S6. The static rendering resources and the dynamic rendering data are rendered and fused to generate and output a digital twin real-time rendering screen of the target operation and maintenance scene.

[0014] As described in steps S1-S3 above, this method first acquires the digital twin model data of the target operation and maintenance scenario. This data typically includes the geometric models, spatial coordinates, equipment attribute parameters, and interconnections of all devices, facilities, and environments within the scenario. Data sources can include engineering design drawings, 3D scan data, building information models, or existing equipment ledgers for the operation and maintenance scenario. After format conversion and standardization, the data can be used in subsequent processes. After acquiring the model data, all entities in the scenario need to be classified. The classification is based on whether each entity undergoes continuous changes in spatial location, geometric shape, or key operating parameters during operation. In the digital twin platform, static and dynamic entities in the scenario can be categorized based on the nature of the graphical interface elements. Layer elements with fixed locations and unchanging shapes, such as map base maps, building outlines, and road networks, can be classified as static entities, while information elements with continuously changing locations or presentation states, such as floating data labels, dynamic trajectory lines, and real-time updated status icons, are classified as dynamic entities. After this classification, the entities in the scenario are divided into two sets, each entering different processing paths. For static entities, a static scene base is constructed based on their model data and spatial location information. This involves arranging all static entities according to their actual spatial distribution to form a complete 3D spatial structure. Based on this, the static scene base is preprocessed to generate static rendering resources. Preprocessing can be implemented in various ways, such as generating rendering cache data after performing rendering calculations on the static scene, or generating rendering resources such as texture maps, light maps, and shadow maps for each static entity in the scene. These pre-rendered resources are stored and can be directly reused in each subsequent frame of rendering without needing to be recalculated.

[0015] As described in steps S4-S6 above, for dynamic entities, the real-time monitoring and processing flow begins. The current state data of each dynamic entity is acquired in real time, including spatial location, attitude angle, and key operating parameters such as current, voltage, temperature, and pressure. Simultaneously, the historical state data of each dynamic entity is maintained in the system. When new state data arrives, the current state data is compared item by item with the corresponding historical state data to identify the differences. These differences constitute the state change difference of the entity. The specific content of the difference depends on which items in the state data have changed; it could be a change in location, a change in a physical parameter, or a change in the connection relationships between entities. After extracting the difference, entities that have truly undergone state changes are selected from all dynamic entities and identified as target dynamic entities. For the target dynamic entity, incremental rendering updates are performed based on the changes recorded in the change difference. This involves extracting the specific parameters that have changed from the difference, such as new position coordinates, new angle values, or new operating parameter values. Then, on top of the static scene base, only the rendering area corresponding to the entity is locally refreshed, without re-rendering the entire scene. For dynamic entities that remain unchanged, their appearance remains the same, and no update operations are performed. The rendering workload for each frame is proportional to the number of entities that have actually changed in the scene, not to the total number of dynamic entities. Finally, the static rendering resources are rendered and fused with the incrementally updated dynamic rendering data to generate a complete real-time rendering of the operation and maintenance digital twin and output it to the display device. Fusion can employ conventional techniques in graphics rendering, such as using layer overlay or depth testing to composite the rendering results of dynamic entities with the static scene base, ensuring that the dynamic entities are presented on the static scene with the correct spatial position and posture. Once fusion is complete, the rendering process for one frame is finished. By repeatedly executing the above steps of dynamic entity state acquisition, differential extraction, and incremental update, continuous real-time updates of the operation and maintenance scene digital twin can be achieved.

[0016] As described above, by pre-dividing entities in the operation and maintenance scenario into static and dynamic entities, the static and dynamic parts can be processed separately without interference. Static entities are pre-processed to generate reusable rendering resources, eliminating the need for repeated rendering calculations in every frame and fundamentally reducing the computational burden of real-time rendering. Dynamic entities, on the other hand, are incrementally updated only for the changed parts through differential extraction, avoiding the resource waste caused by re-rendering the entire scene. More importantly, static rendering resources and dynamic incremental update data maintain visual consistency during rendering fusion, ensuring both high-quality presentation of the static parts in the digital twin and real-time changes in the dynamic parts that follow the actual state. The two paths of static pre-rendering and dynamic differential synchronization each perform their respective functions while cooperating with each other, jointly achieving a smooth, real-time, and accurate presentation of the digital twin in large-scale operation and maintenance scenarios. Compared with directly rendering the entire scene in full, or with only distinguishing between static and dynamic at the data level but still processing at a uniform granularity in the rendering layer, rendering efficiency and real-time response capabilities are significantly improved.

[0017] In one embodiment, extracting the state change difference of the dynamic entity includes: Obtain the state vector of each dynamic entity at the current moment and the corresponding state vector at the historical moment; wherein, the state vector is used to represent the complete running state information of the dynamic entity in the operation and maintenance scenario; The current state vector and the historical state vector are deconstructed in multiple dimensions according to semantic type. Based on the result of the dimensional deconstruction, a differentiated subset of states including spatial pose dimension, physical parameter dimension and topological connection dimension is generated. For each dimension, a difference measurement strategy that matches the characteristics of the dimension is adopted to calculate the difference measurement value between the current state subset and the historical state subset under each dimension; Obtain a preset set of dimension weight coefficients, which includes weight values ​​corresponding to each dimension. Each weight value is preset based on the degree of influence of the state change of the corresponding dimension on the continuity of the operation and maintenance scene display and the accuracy of the state expression. Based on the set of dimensional weight coefficients, multiple difference metrics are weighted and fused to generate a comprehensive change feature value that characterizes the overall change intensity of the dynamic entity. The comprehensive change feature value is matched with a preset multi-level change threshold range. Based on the matching result, the state change difference of the dynamic entity is assigned a corresponding change level identifier, and the difference analysis result carrying the change level identifier is output.

[0018] As mentioned above, after acquiring the state data of each dynamic entity, the process doesn't involve directly comparing the current complete state vector with the historical complete state vectors. Instead, the state vectors are first deconstructed at the dimensional level. Specifically, the state vector of each dynamic entity contains complete operational information about that entity in the maintenance scenario, such as spatial location, attitude angle, current, voltage, temperature, and pressure. These information belong to different semantic types: some represent the entity's spatial occupancy, some represent the entity's physical operating conditions, and some represent the entity's connection relationships with other entities. Deconstructing the state vector according to these semantic types yields three differentiated state subsets: the spatial pose dimension state subset, the physical parameter dimension state subset, and the topological connection dimension state subset. Through this deconstruction, the originally mixed overall state vector becomes three components with clear semantic boundaries, each representing a specific aspect of the entity's operational state. After the state vector undergoes dimensional deconstruction... For each dimension, a difference measurement strategy matching the characteristics of that dimension is used to calculate the difference between the current moment and historical moments. The three dimensions have different physical meanings, and their difference measurement methods also differ accordingly. Changes in the spatial pose dimension are manifested as offsets in position coordinates and attitude angles, suitable for a spatial distance measurement strategy to measure the magnitude of the change; changes in the physical parameter dimension are manifested as fluctuations in continuous values ​​such as current, voltage, and temperature, suitable for a numerical change rate measurement strategy to measure the severity of the change; changes in the topological connectivity dimension are manifested as the establishment or dissolution of connections, suitable for a structural similarity measurement strategy to measure the degree of change in connection status. Through this dimension-matching measurement method, each dimension obtains its own difference measurement value, reflecting the degree of change of the entity in that dimension. The difference measurement values ​​calculated for each of the three dimensions have different dimensions and physical meanings, and cannot be directly added or compared. Therefore, a set of preset dimension weight coefficients needs to be introduced to map the three difference measurement values ​​to a unified measurement scale. These weighting coefficients reflect the degree of influence of state changes in different dimensions on the final twin image presentation effect. Specifically, changes in spatial pose directly affect the position and posture of entities in the image, having the most direct impact on the continuity of the image. Changes in physical parameters are usually presented through data labels or color mapping, which have a greater impact on the accuracy of state expression but a smaller impact on the continuity of the image. Changes in topological connections reflect changes in the connection relationships between entities, and their impact is somewhere in between.

[0019] Through weighted fusion calculation, the three difference metrics are combined into a single comprehensive change feature value, which characterizes the overall change intensity of the dynamic entity at the current moment on a unified scale. The comprehensive change feature value is a continuous value, but continuous values ​​are difficult to directly drive subsequent rendering strategy selection. Therefore, it needs to be mapped to discrete change levels. The system presets multi-level change threshold ranges, dividing the range of the comprehensive change feature value into several level intervals, each interval corresponding to a change level identifier. Matching the comprehensive change feature value with these threshold intervals assigns the corresponding change level identifier to the entity's state change difference. The difference analysis result carrying the change level identifier is used as the output of this step for subsequent incremental rendering update steps. Incremental rendering updates can select different rendering strategies based on different change level identifiers; for example, slight changes only refresh data labels, moderate changes refresh a local rendering area of ​​the entity, and drastic changes trigger a larger-scale rendering update.

[0020] As an optional implementation, the difference measurement of the spatial pose dimension can employ a spatial distance metric strategy. This strategy represents the subset of states in the spatial pose dimension as a combination of position vectors and attitude vectors. For position changes, the Euclidean distance between the current position vector and the position vector at a historical time is calculated as the position change; for attitude changes, the angle between the current attitude quaternion and the attitude quaternion at a historical time is calculated as the attitude change. The position change and attitude change are normalized and then weighted and summed to obtain the difference measurement value of the spatial pose dimension. Taking a dynamic floating tag in a digital twin platform as an example, when the data source bound to the tag is updated, causing the anchor point position of the tag in the 3D scene to migrate, its spatial position coordinates change, the Euclidean distance of the position vector increases accordingly, and the difference measurement value of this dimension also increases accordingly.

[0021] For the physical parameter dimension, a numerical change rate measurement strategy can be adopted. This strategy calculates the relative change rate between the current value and the historical value for each parameter in the state subset of the physical parameter dimension; that is, the absolute value of the difference divided by the absolute value of the historical value. After normalization of the relative change rates of each parameter, the maximum value or weighted average is taken as the difference measure of the physical parameter dimension. Taking the real-time indicator card on the data panel as an example, when the data bound to this indicator card is updated from 60 to 75, the relative change rate is 25%, which can accurately reflect the drastic change of this data item.

[0022] For the topological connectivity dimension, a structural similarity measurement strategy can be employed. This strategy represents the state subset of the topological connectivity dimension as the set of connecting edges between that entity and other entities. The Jaccard distance between the current set of connecting edges and the set of connecting edges from previous times is calculated; that is, the ratio of the intersection size to the union size of the two sets is used as the structural similarity. The difference between 1 and this similarity is used as the dissimilarity measure of the topological connectivity dimension. When connections between entities are established or broken, this dissimilarity measure will change accordingly, effectively reflecting changes in the topological structure.

[0023] The difference metrics calculated using the three metrics strategies described above are all dimensionless values, ranging from 0 to 1, and can be directly used for subsequent weighted fusion calculations.

[0024] In one embodiment, the step of determining the target dynamic entity that has undergone a state change from the dynamic entities based on the state change difference, and performing incremental rendering updates only on the target dynamic entity to generate dynamic rendering data, includes: Analyze the state change difference, and select the target dynamic entity from the dynamic entities according to the change level identifier; Based on the dimension corresponding to the state change difference, the change type of the target dynamic entity is determined, wherein the change type includes spatial pose change type, physical parameter change type and topological connection change type, which correspond to the spatial pose dimension, physical parameter dimension and topological connection dimension, respectively. Based on the change level identifier and the change type, a corresponding incremental rendering strategy is determined. The incremental rendering strategy is control information selected from a preset strategy set based on the change level identifier and the change type, which is used to limit the rendering update range and rendering update precision of the target dynamic entity. According to the incremental rendering strategy, the local change parameters of the target dynamic entity in the corresponding dimension are extracted from the state change difference. By combining the local change parameters and the spatial positioning information of the static scene base, the local rendering area corresponding to the target dynamic entity is refreshed on the static rendering resource to generate the dynamic rendering data.

[0025] As described above, the state change difference is first analyzed to extract entities carrying change level identifiers, and then filtered based on these identifiers. Entities with a change level identifier of 0 indicate that their change level has not yet reached the minimum threshold required to trigger a rendering update, and these entities are filtered out. Entities with a change level identifier of non-zero indicate that their change level has reached the condition for triggering a rendering update and are identified as target dynamic entities. Through this filtering step, only entities whose change level reaches the preset threshold will enter the subsequent rendering update process. After identifying the target dynamic entity, it is necessary to further analyze what type of change the entity has undergone. The state change difference is derived from a subset of states after multi-dimensional deconstruction. The difference itself carries information about the source of the dimension—whether it originates from the spatial pose dimension, the physical parameter dimension, or the topological connection dimension. Based on the dimension from which the difference originates, the change type of the target dynamic entity can be determined, specifically including three types: spatial pose change, physical parameter change, and topological connection change, corresponding to spatial position movement or attitude angle deflection, numerical fluctuation of operating parameters, and the establishment or dissolution of connections between entities, respectively.

[0026] In some cases, the state change difference of an entity may originate from multiple dimensions simultaneously, thus the entity exhibits multiple change types. The change level identifier represents the severity or importance of the change, while the change type represents the classification of the change in nature; both together constitute the basis for selecting the incremental rendering strategy. The system pre-defines a strategy set, storing various incremental rendering strategies, each with a mapping relationship to a specific combination of change level identifier and change type. Based on the change level identifier and change type of the current target dynamic entity, the corresponding incremental rendering strategy is matched from the strategy set. This strategy limits the update scope (whether only the entity itself is updated, or its surrounding associated regions need to be updated) and update precision (whether high-precision full rendering or simplified low-precision rendering is used). After strategy matching is complete, according to the update scope and update precision defined by the strategy, the specific parameter values ​​that have changed are extracted from the state change difference. For example, for spatial pose changes, new position coordinates and new attitude angles are extracted; for physical parameter changes, new current or temperature values ​​are extracted; for topological connection changes, new connection state information is extracted. Then, based on these local change parameters and the spatial positioning information of the static scene base, the local rendering area corresponding to the entity is refreshed on the static rendering resource.

[0027] The core of incremental rendering updates lies in the word "incremental." It doesn't re-render the entire dynamic entity from start to finish; instead, it updates only the changed local content. Specifically, if an entity only changes position, only its spatial transformation matrix needs updating, without reloading its geometry or regenerating its textures. If an entity only changes physical parameters, only the displayed content of its data labels needs updating, without refreshing its 3D geometry. If an entity only changes topological connections, only the visual representation of the connections needs updating. By employing different granularity update methods for different types and levels of change, a reasonable match is achieved between rendering overhead and the degree of change.

[0028] Reference Figure 2 In one embodiment, the step of constructing a static scene base based on the model data of the static entity and preprocessing the static scene base to generate static rendering resources includes: S31. Obtain the model data and spatial position information of each static entity in the static entity set, and arrange the model data of each static entity in space according to the spatial position information to generate three-dimensional spatial structure data. S32. Organize the three-dimensional spatial structure data into a scene graph to generate a static scene graph, wherein the static scene graph includes multiple nodes, each node corresponds to one or more static entities, and carries the model data and spatial transformation information of the corresponding static entities. S33. Traverse each node in the static scene graph, and perform spatial transformation and pre-rendering processing on the model data carried by the node according to the spatial transformation information carried by the node, so as to generate pre-rendered resources corresponding to each node. S34. Combine the pre-rendered resources corresponding to each node according to the node hierarchy structure of the static scene graph to generate the static rendering resources.

[0029] As described above, the first step is to acquire the model data of each static entity in the static entity set and its spatial location information in the operation and maintenance scenario. The model data includes visual information such as the entity's 3D geometry, texture, and material properties, while the spatial location information records the entity's coordinates and orientation in the actual scene. These static entities are arranged according to their respective spatial location information, so that they are restored in digital space to a spatial distribution consistent with the actual physical scene, thus forming the 3D spatial structure data of the entire static scene base. The 3D spatial structure data is essentially a complete scene description containing all static entities, but at this point, the organizational relationships between the entities are not yet structured enough.

[0030] To facilitate subsequent preprocessing and management, the 3D spatial structure data needs to be organized into a scene graph to generate a static scene graph. A scene graph is a hierarchical data structure that organizes entities in the scene into a tree-like structure according to their spatial location and logical relationships. In the static scene graph, each node corresponds to one or more static entities, carrying the complete model data and spatial transformation information of that entity. The spatial transformation information specifically records the entity's translation, rotation, and scaling relative to its parent node or relative to the world coordinate system; these are essential parameters for transforming the model data from the local coordinate system to the world coordinate system. After the static scene graph is generated, each node in the graph is traversed, and the model data carried by each node undergoes spatial transformation and pre-rendering processing. Since the geometric coordinates in the model data are usually defined in the local coordinate system, they need to be transformed to the world coordinate system based on the spatial transformation information carried by the node, ensuring it appears in the correct position and orientation.

[0031] After spatial transformation, pre-rendering processing is performed on the entity corresponding to the node, generating pre-rendered resources for that node. Pre-rendering processing may include rendering cache calculations, texture map generation, and lighting information baking. The results of these processes are stored as pre-rendered resources for later direct use. After all pre-rendered resources for all nodes are generated, they are combined according to the node hierarchy of the static scene graph to generate a complete static rendering resource. During the combination process, the spatial transformation relationships between parent and child nodes are preserved, ensuring that the relative positional relationships between entities after combination are consistent with the original scene. Through this combination process, the pre-rendered resources originally scattered across various nodes are integrated into a complete static rendering resource that can be directly used for rendering output. Through these steps, the static scene transforms from raw model data into a structured, pre-rendered, and directly reusable rendering resource. Scene graphs provide a structured way to organize static scenes, making them no longer a bunch of scattered model data, but a hierarchical and structured whole; node-level pre-rendering processing ensures that each static entity is optimized individually, and the pre-rendering results can be directly called; the combined static rendering resources can be directly used in each subsequent frame rendering without repeating spatial transformation, texture generation, lighting calculation and other operations, thus eliminating the computational load of the static part on the real-time rendering pipeline from the root.

[0032] It's worth noting that in existing technologies, static scenes are typically processed by directly loading all static model data, performing a single rendering calculation, and then caching the results. The problem with this approach is that the static scene is treated as a single unit, lacking a structured organization. If any part needs updating, the entire scene must be reprocessed. For example, if the position of a static device in a substation is adjusted, or the texture of a section of pipe needs to be changed, the entire static scene needs to be re-pre-rendered according to current practices. Although static parts change infrequently, they are not entirely static. This "one change, all changes" approach still results in considerable resource waste when the static scene is large. This embodiment, however, introduces a hierarchical data structure called a scene graph. This breaks down the static scene from a single unit into multiple nodes with spatial transformation information. Each node corresponds to one or a group of static entities, possesses independent model data and transformation parameters, and can independently perform pre-rendering to generate its own pre-rendered resources. The direct benefit of this node-based organization is that when a static entity changes, only the pre-rendered resources corresponding to that node need to be updated, and then the elements can be recombined; the entire scene does not need to be reprocessed. Meanwhile, the spatial transformation information of each node is preserved in the scene graph, providing a spatial positioning reference for the incremental rendering of subsequent dynamic entities. That is, when dynamic entities are updated through incremental rendering, they need to know the spatial position of each part in the static scene base, and the scene graph provides this information.

[0033] In one specific embodiment, in existing solutions, incremental rendering updates of dynamic entities are performed by processing each entity independently. Each dynamic entity detects whether it has changed and performs its own rendering update. This approach is effective in most scenarios, but a special case exists in operational scenarios: multiple dynamic entities are interconnected through physical connections or logical relationships, and a change in the state of one entity often triggers changes in other related entities in a short period of time, forming a chain reaction. Taking the data linkage scenario of a digital twin platform as an example, when the data of a master data source changes, multiple chart components, data labels, and status indicators bound to that data source will be updated synchronously. However, according to the method of processing each entity independently, the state changes of these related entities are detected separately, the differences are extracted separately, and the updates are performed separately. The state changes of each entity in the twin screen appear sequentially, and the presented change process does not match the actual physical process. In other words, when dealing with related changes, the update unit, the individual entity, the actual unit of change in the physical world, and the overall related change composed of multiple entities are misaligned. To address this issue, this embodiment further introduces an association change identification and synchronization processing mechanism based on the aforementioned steps. Specifically, the step of determining the target dynamic entity that has undergone a state change from the dynamic entities based on the state change difference, and performing incremental rendering updates only on the target dynamic entity to generate dynamic rendering data, also includes: Within a preset time window, detect entities in the dynamic entity that undergo state changes and generate a candidate set of changed entities; Obtain the entity relationship graph of the target operation and maintenance scenario, wherein the entity relationship graph includes multiple nodes and connecting edges, the nodes correspond one-to-one with the dynamic entities, and the connecting edges are used to characterize the mutual influence relationship of the running states between two connected dynamic entities; Map each entity in the candidate variable entity set to the entity association graph, perform association matching based on the connection edges of the entities in the entity association graph, and divide the entities in the candidate variable entity set that are directly or indirectly connected by connection edges into the same association variable cluster. For all dynamic entities within the associated change cluster, incremental rendering updates are performed synchronously within the same rendering cycle. For isolated entities in the candidate changed entity set that do not belong to any associated changed cluster, incremental rendering updates are performed in an entity-by-entity independent manner.

[0034] As mentioned above, under the established static / dynamic separation rendering framework, incremental rendering updates of dynamic entities are typically processed independently for each entity. However, in operational scenarios where a large number of dynamic entities change simultaneously, this independent processing approach can lead to temporal misalignment in the state updates of related entities on the screen. To address this issue, this embodiment introduces a pre-processing step. First, within a preset time window, all dynamic entities that have undergone state changes are detected, and these entities are aggregated to generate a candidate set of changed entities. The time window is introduced to avoid dispersing changes that are close in time but belong to the same event chain into different processing cycles. Taking a data cascading update scenario as an example, when a certain underlying data source changes, the statistical values ​​of multiple upstream data aggregation nodes change successively within a short period of time. After obtaining the candidate set of changed entities, it is necessary to use an entity relationship graph to determine which entities have mutual influence relationships.

[0035] An entity relationship graph is a structured description of the physical connections or logical relationships between entities in an operations and maintenance scenario. Each node in the graph corresponds to a dynamic entity, and the connecting edges between nodes indicate a mutual influence relationship in the operational state between the two connected entities. For example, in a digital twin platform, a data source node has a data binding relationship with multiple visualization components subscribed to that data source. This relationship is recorded in the relationship graph. When the data of the data source node changes, the presentation state of all visualization components subscribed to that data source will be affected. After obtaining the relationship graph, each entity in the candidate change entity set is mapped to the corresponding node in the graph, and then association matching is performed based on the connecting edges between nodes. Specifically, the graph searches for connection paths between candidate change entities. If two or more entities are directly connected by connecting edges in the graph, or indirectly connected by one or more intermediate nodes, it indicates that these entities have a relationship and should be classified into the same associated change cluster. If an entity cannot form a connection path with any other candidate change entity through connecting edges in the graph, the entity is considered an isolated entity.

[0036] After identifying clusters of related changes, incremental rendering updates are performed synchronously on all dynamic entities within a cluster within the same rendering cycle. That is, when a cluster contains multiple entities, the system does not process them individually, but rather processes the state change differences of these entities together within the same frame and completes the rendering update uniformly, ensuring that the state changes of these related entities in the twin view are presented at the same moment. For isolated entities, since their changes are not related to other entities, incremental rendering updates are still performed according to the existing entity-by-entity independent method. Through the above processing, when a large number of related entities change simultaneously, the division of related change clusters ensures that all entities on the same event chain are identified and bundled together for processing, avoiding timing misalignment caused by individual processing. For isolated entities without any relationship, the original independent processing method is maintained, without incurring unnecessary additional overhead due to the addition of correlation analysis. The coexistence of correlation analysis and independent processing modes ensures both consistency of image timing and high processing efficiency.

[0037] In this embodiment, a pre-processing step based on a relational graph is added before the original entity-by-entity independent processing flow, upgrading the update unit from "single entity" to "related change clusters". Specifically, all changing entities are first collected within a time window, and then the relational graph is used to identify which entities are related. Related entities are packaged into clusters, and all entities within a cluster are updated within the same rendering cycle, ensuring they are presented synchronously on the twin screen. Unrelated entities revert to the original independent processing method. This "related packaging, unrelated independence" diversion processing mechanism solves the problem of synchronous updating of related changing entities without introducing additional complexity. For scenarios with large-scale chain changes, the related change cluster mechanism ensures that the twin screen can accurately reflect the synchronicity of changes in the physical world; for scenarios with isolated changes, the original independent processing mechanism ensures that processing efficiency is not affected. The reasonable coexistence of the two modes enables this method to effectively solve the problem of screen temporal consistency in large-scale related change scenarios without adding unnecessary overhead.

[0038] In one embodiment, obtaining the entity relationship graph of the target operation and maintenance scenario includes: Obtain the initial entity relationship graph of the target operation and maintenance scenario; Real-time monitoring of topology change events occurring in the target operation and maintenance scenario, wherein the topology change events include connection relationship establishment events, connection relationship termination events, and connection relationship attribute change events between entities; Based on the topology change event, determine the affected nodes and / or connecting edges in the initial entity association graph, and generate change operation instructions; The initial entity association graph is updated according to the change operation instruction. Based on the update result, a dynamic association graph is generated at the current time, and the dynamic association graph is used as the entity association graph for subsequent association matching and association change cluster division.

[0039] As mentioned above, the aforementioned embodiments introduced an association graph as the basis for determining whether an association exists between entities. However, once the association graph is constructed, if it is not updated, it cannot reflect the actual changes in the connection relationships between entities in the operation and maintenance scenario. Taking a factory production line as an example, when product models are switched, some equipment may be bypassed or connected, and the process flow path changes accordingly. After these topology changes occur, if the association graph is not updated accordingly, the association change cluster division based on the graph may be biased, that is, entities that have been decoupled are still treated as associated entities, or newly established associations are not identified. To address this issue, this method adopts a dynamic maintenance mechanism when acquiring the association graph. First, the initial entity association graph of the target operation and maintenance scenario is acquired. This initial graph can come from the static design data of the operation and maintenance scenario, such as the wiring diagram of the power grid, the process flow diagram of the factory, the connectivity diagram of the pipeline network, etc., or it can come from the actual topology detection results during the first run. The initial graph constructs the baseline state of the association relationships between entities. Based on the initial graph, topology change events occurring in the operation and maintenance scenario are monitored in real time. A topology change event refers to a substantial alteration in the connectivity relationships between entities, specifically including the establishment, dissolution, and modification of connectivity attributes. For example, in a power grid, a switch changing from open to closed signifies the establishment of a new connection between two previously isolated nodes—this is a connectivity establishment event. Conversely, a line being disconnected for maintenance signifies the dissolution of the connection between two nodes—this is a connectivity dissolution event. A modification of connectivity attributes refers to a change in the nature of the connection, even if the connection itself is not established or dissoluted; for instance, the impedance parameters of a line may change due to a switching operation. After each topology change event is detected, it is necessary to determine the specific impact of the event on the initial correlation graph.

[0040] Specifically, this involves identifying which node and edge in the graph corresponds to the location of an event, and determining the necessary modifications to that node or edge after the event. Based on these analysis results, change operation instructions are generated. The specific type of these instructions depends on the nature of the topology change event: adding a new entity node requires a node addition instruction, removing a node requires a node deletion instruction, establishing a new connection between two entities requires an edge addition instruction, breaking an existing connection requires an edge deletion instruction, and a change in the weight or attribute of an edge requires an edge weight modification instruction. The generated change operation instructions are then used to update the current relationship graph, reflecting the changes described in the instructions in the graph. After each update, the latest dynamic relationship graph is obtained, serving as the basis for subsequent relationship matching and cluster partitioning. Through this mechanism, the relationship graph is no longer static but evolves synchronously with actual changes in the topology during the operation and maintenance scenario. Each time the clusters of association changes are divided, the graph used reflects the latest topological state at the current moment. The association matching results based on this graph can accurately identify entities that actually have association relationships, avoiding misidentification or omission due to outdated graph information.

[0041] In one embodiment, analyzing the dynamic change attributes of each entity in the digital twin model data under its operating state, and dividing the entities into static entities and dynamic entities based on the analysis results, includes: Based on the dynamic change attributes of each entity, the abnormal entities in the digital twin model data are identified as having a dynamic-static separation. The abnormal entities are entities whose geometric model and spatial location do not change, but whose operating status information changes. Based on the identification results, if the abnormal entity of dynamic-static separation exists, the geometric model data and spatial location information of the abnormal entity of dynamic-static separation will be classified into the static entity and participate in the construction of the static scene base. The dynamic and static separation abnormal entity operation status information is divided into dynamic data, which participates in the extraction of the state change difference and incremental rendering update.

[0042] As mentioned above, in the static / dynamic classification method, entities are divided into static and dynamic entities based on their dynamic change attributes. This classification method is based on an implicit assumption: an entity is either static or dynamic, with a clear boundary between the two. However, in actual operation and maintenance scenarios, there exists a special type of entity that does not fully conform to this binary classification. Taking a pipeline system as an example, the geometry and spatial location of the pipeline itself do not change during operation, but the operating status information of the medium inside the pipeline, such as temperature, pressure, and flow rate, is constantly changing. Taking the data flow configuration of a digital twin platform as an example, during operation, users reconfigure the input source or output destination of the data pipeline, which changes the data flow topology in the platform. This type of entity does not fully conform to the characteristics of a static entity, nor does it fully conform to the characteristics of a dynamic entity.

[0043] For these types of entities, this method adds an identification step to the static / dynamic classification process. Specifically, when classifying entities in a scene into static and dynamic categories, the method first judges based on the dynamic change attributes of each entity, identifying those entities whose geometric models and spatial positions do not change but whose operational status information continuously changes, and marking them as static / dynamic separation anomalous entities. These entities are called anomalous not because they have any inherent problems, but because existing static / dynamic binary classification frameworks cannot directly accommodate them; their geometric features conform to static entities, but their information features conform to dynamic entities. After identifying anomalous entities, instead of classifying the entity as a whole into either static or dynamic categories in the traditional way, it is decomposed into two levels for separate processing. At the geometric level, the entity's geometric model data and spatial position information are classified as static entities, allowing them to participate in the construction and pre-rendering process of the static scene base. Since the geometric model and spatial position of these entities do not change during operation, treating them as static entities avoids unnecessary repeated rendering and fully leverages the resource reuse advantages of static pre-rendering. At the information level, the entity's operational status information is categorized into dynamic data, allowing it to participate in the extraction of state change differences and incremental rendering update processes. This ensures that every change in operational status information is detected and reflected in the twin view in a timely manner, guaranteeing real-time information accuracy. After this two-layer division, the geometric representation of this type of entity is carried by the static scene base, while its operational status information is presented by the dynamic data layer. The two are merged in the final rendering fusion stage, allowing the twin view to see both the entity's correct geometric shape and spatial location, as well as its real-time updated operational status information.

[0044] In one feasible embodiment, dynamic data does not exist in isolation. The state change carried by dynamic data often has an informational impact beyond the entity itself. Taking the increase in the temperature of the medium inside a pipeline as an example, the temperature change is not just a numerical update; it signifies a change in the thermal state within the pipeline. This change can affect the temperature sensor readings along the pipeline through heat conduction or radiation, alter the thermal expansion state of adjacent pipe sections, and even affect the maintenance personnel's assessment of the overall pipeline's condition. In other words, every change in dynamic data can potentially propagate along physical or logical connections to other entities within its influence range. However, the problem is that most of these affected entities are static entities. In existing solutions, once a static entity is incorporated into a static scene base and pre-rendered, its visual presentation is fixed and no longer receives any updates from dynamic data. Temperature sensors are static (position unchanged), but the monitored values ​​they display are dynamic; pipe sections are static (geometrically unchanged), but their thermal state labels are dynamic. When dynamic data changes, the informational impact has already propagated to the doorsteps of these static entities, but the system, lacking a corresponding propagation mechanism, can only allow this information to be interrupted at the doorstep. The final result is that the data of the dynamic entity is updated, but the surrounding static entities affected by it do not respond synchronously. The information in the twin screen that should have changed accordingly remains outdated, forming a gap in information dissemination.

[0045] Therefore, to solve this problem, this embodiment proposes that the method further includes: Obtain the information influence relationship configuration information between static entities and dynamic entities in the target operation and maintenance scenario. The information influence relationship configuration information is used to characterize which static entities' data presentation status needs to be synchronously propagated to when the dynamic entity's running status data changes. When an update to the running status data of a dynamic entity is detected, the static entities affected by the dynamic entity are determined based on the information influence relationship configuration information. Based on the updated running status data of the dynamic entity, a data layer refresh instruction for the static entity is generated; the data layer refresh instruction is used to indicate the update of the data presentation status of the static entity, and the refresh range of the data layer refresh instruction is limited to the data layer of the static entity and does not involve the geometric layer of the static entity. The data layer refresh instruction is executed to incrementally refresh the data presentation state of the static entity, and based on the incrementally refreshed data presentation state, the data label area corresponding to the static entity in the static scene base is updated. The updated static scene base participates in subsequent rendering fusion.

[0046] In this embodiment, configuration information regarding the information influence relationship between static and dynamic entities in an operation and maintenance scenario is obtained. This configuration information records which static entities' data presentation status needs to be synchronously propagated when the operating status data of a dynamic entity changes. Taking a factory production line as an example, when the operating power of a key machine changes, the material level indicator in the upstream material buffer, the judgment threshold of the downstream quality inspection equipment, and the energy consumption display panel of the area where the machine is located may all be affected. These relationships are determined by the business logic of the operation and maintenance scenario and can be pre-configured in the system. When an update to the operating status data of a dynamic entity is detected, the system queries the information influence relationship configuration information based on the identifier of the dynamic entity, finds all static entities that have an information influence relationship with the dynamic entity, and generates a list of static entities affected by the dynamic entity. Each static entity in this list needs to receive the change information of the dynamic data and update its own data presentation status. After determining the list of affected static entities, a corresponding data layer refresh instruction is generated for each affected static entity based on the updated operating status data of the dynamic entity. The data layer refresh command indicates what kind of update is needed for the data presentation status of the static entity, such as updating a monitoring value, switching a status icon, or changing the color label of a certain area. The refresh scope of the refresh command is explicitly limited to the data layer of the static entity and does not involve the geometry layer.

[0047] In other words, the 3D model and spatial position of static entities are unaffected by these refreshes; only the data labels, status indicators, or information display areas attached to the entity change. After generating data layer refresh instructions, these instructions are executed to incrementally refresh the data presentation state of the static entity. Incremental refresh means updating only the display content that has actually changed, rather than regenerating all display information related to the entire static entity. After the refresh is complete, based on the updated data presentation state, the data label area corresponding to the static entity in the static scene base is updated. The geometry layer in the static scene base remains unchanged, while the content of the data label area is replaced with the latest display content. The updated static scene base participates in subsequent rendering and fusion, presenting the changes in the data layer on the final twin screen. Through this mechanism, changes in dynamic data can propagate along information influence relationships to related static entities, allowing the information displayed on the static entities to be updated synchronously with the changes in dynamic data, ensuring that the information presented on the screen remains consistent.

[0048] Reference Figure 3 This application also provides an operation and maintenance digital twin rendering system based on the separation of dynamic and static entities and the synchronization of twin differentials, including: Module 1 is used to acquire digital twin model data of the target operation and maintenance scenario; Analysis module 2 is used to analyze the dynamic change attributes of each entity in the digital twin simulation data under the running state, and divide the entities into static entities and dynamic entities based on the analysis results; Module 3 is used to construct a static scene base based on the model data of the static entity, and to preprocess the static scene base to generate static rendering resources. Extraction module 4 is used to acquire the state data and corresponding historical state data of the dynamic entity in real time, and extract the state change difference of the dynamic entity based on the difference between the current state data and the historical state data. Update module 5 is used to determine the target dynamic entity that has undergone a state change from the dynamic entities based on the state change difference, and to perform incremental rendering updates only on the target dynamic entity to generate dynamic rendering data. Output module 6 is used to render and fuse the static rendering resources with the dynamic rendering data to generate and output a digital twin real-time rendering screen of the target operation and maintenance scene.

[0049] As described above, it is understood that each component of the operation and maintenance digital twin rendering system based on dynamic and static entity separation and twin differential synchronization proposed in this application can realize the function of any one of the operation and maintenance digital twin rendering methods based on dynamic and static entity separation and twin differential synchronization as described above, and the specific structure will not be repeated.

[0050] Reference Figure 4 This application also provides a computer device, which may be a server, and its internal structure may be as follows: Figure 4 As shown, the computer device includes a processor, memory, network interface, and database connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and database. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The database stores monitoring data and other data. The network interface communicates with external terminals via a network connection. When the computer program is executed by the processor, it implements a digital twin rendering method for operation and maintenance based on the separation of dynamic and static entities and the synchronization of twin differentials.

[0051] The processor described above executes the operation and maintenance digital twin rendering method based on the separation of static and dynamic entities and the synchronization of twin differentials, including: acquiring digital twin model data of the target operation and maintenance scenario; analyzing the dynamic change attributes of each entity in the digital twin simulation data under the running state, and dividing the entities into static entities and dynamic entities according to the analysis results; constructing a static scene base based on the model data of the static entities, and preprocessing the static scene base to generate static rendering resources; acquiring the state data and corresponding historical state data of the dynamic entities in real time, and extracting the state change difference of the dynamic entities according to the difference between the current state data and the historical state data; determining the target dynamic entity that has undergone a state change from the dynamic entities according to the state change difference, and performing incremental rendering updates only on the target dynamic entity to generate dynamic rendering data; and rendering and fusing the static rendering resources and the dynamic rendering data to generate and output the real-time rendering screen of the digital twin of the target operation and maintenance scenario.

[0052] One embodiment of this application also provides a computer-readable storage medium storing a computer program. When executed by a processor, the computer program implements a digital twin rendering method for operation and maintenance based on the separation of dynamic and static entities and the synchronization of twin differentials. The method includes the following steps: acquiring digital twin model data of a target operation and maintenance scenario; analyzing the dynamic change attributes of each entity in the digital twin simulation data under its operating state, and dividing the entities into static entities and dynamic entities based on the analysis results; constructing a static scene base based on the model data of the static entities, and preprocessing the static scene base to generate static rendering resources; acquiring the state data and corresponding historical state data of the dynamic entities in real time, and extracting the state change differential of the dynamic entities based on the difference between the current state data and the historical state data; determining the target dynamic entity whose state has changed from the dynamic entities based on the state change differential, and performing incremental rendering updates only on the target dynamic entity to generate dynamic rendering data; and rendering and fusing the static rendering resources and the dynamic rendering data to generate and output a real-time digital twin rendering screen of the target operation and maintenance scenario.

[0053] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in this application and in the embodiments can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDR SDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), RAMbus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0054] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, apparatus, article, or method that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, apparatus, article, or method. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, apparatus, article, or method that includes that element.

[0055] The above description is only a preferred embodiment of this application and does not limit the patent scope of this application. Any equivalent structural or procedural changes made based on the content of this application's specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of this application.

Claims

1. A digital twin rendering method for operation and maintenance based on separation of dynamic and static entities and synchronization of twin differentials, characterized in that, The method includes: Obtain digital twin model data for the target operation and maintenance scenario; The dynamic change attributes of each entity in the digital twin simulation data under the running state are analyzed, and the entities are divided into static entities and dynamic entities based on the analysis results. Based on the model data of the static entities, a static scene base is constructed, and the static scene base is preprocessed to generate static rendering resources. The state data and corresponding historical state data of the dynamic entity are acquired in real time, and the state change difference of the dynamic entity is extracted based on the difference between the current state data and the historical state data. Based on the state change difference, the target dynamic entity that has undergone a state change is determined from the dynamic entities, and incremental rendering updates are performed only on the target dynamic entity to generate dynamic rendering data. The static rendering resources and the dynamic rendering data are rendered and fused to generate and output a digital twin real-time rendering screen of the target operation and maintenance scene.

2. The operation and maintenance digital twin rendering method based on dynamic and static entity separation and twin differential synchronization according to claim 1, characterized in that, The extraction of the state change difference of the dynamic entity includes: Obtain the state vector of each dynamic entity at the current moment and the corresponding state vector at the historical moment; wherein, the state vector is used to represent the complete running state information of the dynamic entity in the operation and maintenance scenario; The current state vector and the historical state vector are deconstructed in multiple dimensions according to semantic type. Based on the result of the dimensional deconstruction, a differentiated subset of states including spatial pose dimension, physical parameter dimension and topological connection dimension is generated. For each dimension, a difference measurement strategy that matches the characteristics of the dimension is adopted to calculate the difference measurement value between the current state subset and the historical state subset under each dimension; Obtain a preset set of dimension weight coefficients, which includes weight values ​​corresponding to each dimension. Each weight value is preset based on the degree of influence of the state change of the corresponding dimension on the continuity of the operation and maintenance scene display and the accuracy of the state expression. Based on the set of dimensional weight coefficients, multiple difference metrics are weighted and fused to generate a comprehensive change feature value that characterizes the overall change intensity of the dynamic entity. The comprehensive change feature value is matched with a preset multi-level change threshold range. Based on the matching result, the state change difference of the dynamic entity is assigned a corresponding change level identifier, and the difference analysis result carrying the change level identifier is output.

3. The operation and maintenance digital twin rendering method based on dynamic and static entity separation and twin differential synchronization according to claim 2, characterized in that, The step of determining the target dynamic entity that has undergone a state change from the dynamic entities based on the state change difference, and performing incremental rendering updates only on the target dynamic entity to generate dynamic rendering data, includes: Analyze the state change difference, and select the target dynamic entity from the dynamic entities according to the change level identifier; Based on the dimension corresponding to the state change difference, the change type of the target dynamic entity is determined, wherein the change type includes spatial pose change type, physical parameter change type and topological connection change type, which correspond to the spatial pose dimension, physical parameter dimension and topological connection dimension, respectively. Based on the change level identifier and the change type, a corresponding incremental rendering strategy is determined. The incremental rendering strategy is control information selected from a preset strategy set based on the change level identifier and the change type, which is used to limit the rendering update range and rendering update precision of the target dynamic entity. According to the incremental rendering strategy, the local change parameters of the target dynamic entity in the corresponding dimension are extracted from the state change difference. By combining the local change parameters and the spatial positioning information of the static scene base, the local rendering area corresponding to the target dynamic entity is refreshed on the static rendering resource to generate the dynamic rendering data.

4. The operation and maintenance digital twin rendering method based on dynamic and static entity separation and twin differential synchronization according to claim 1, characterized in that, The process of constructing a static scene base based on the model data of the static entities, and preprocessing the static scene base to generate static rendering resources, includes: Obtain the model data and spatial location information of each static entity in the static entity set, and arrange the model data of each static entity in space according to the spatial location information to generate three-dimensional spatial structure data; The three-dimensional spatial structure data is organized into a scene graph to generate a static scene graph, wherein the static scene graph includes multiple nodes, each node corresponds to one or more static entities, and carries the model data and spatial transformation information of the corresponding static entities; Traverse each node in the static scene graph, and perform spatial transformation and pre-rendering processing on the model data carried by the node according to the spatial transformation information carried by the node, to generate pre-rendered resources corresponding to each node. The pre-rendered resources corresponding to each node are combined according to the node hierarchy structure of the static scene graph to generate the static rendering resources.

5. The operation and maintenance digital twin rendering method based on dynamic and static entity separation and twin differential synchronization according to claim 1, characterized in that, The step of determining the target dynamic entity that has undergone a state change from the dynamic entities based on the state change difference, and performing incremental rendering updates only on the target dynamic entity to generate dynamic rendering data, further includes: Within a preset time window, detect entities in the dynamic entity that undergo state changes and generate a candidate set of changed entities; Obtain the entity relationship graph of the target operation and maintenance scenario, wherein the entity relationship graph includes multiple nodes and connecting edges, the nodes correspond one-to-one with the dynamic entities, and the connecting edges are used to characterize the mutual influence relationship of the running states between two connected dynamic entities; Map each entity in the candidate variable entity set to the entity association graph, perform association matching based on the connection edges of the entities in the entity association graph, and divide the entities in the candidate variable entity set that are directly or indirectly connected by connection edges into the same association variable cluster. For all dynamic entities within the associated change cluster, incremental rendering updates are performed synchronously within the same rendering cycle. For isolated entities in the candidate changed entity set that do not belong to any associated changed cluster, incremental rendering updates are performed in an entity-by-entity independent manner.

6. The operation and maintenance digital twin rendering method based on dynamic and static entity separation and twin differential synchronization according to claim 5, characterized in that, The step of obtaining the entity relationship graph of the target operation and maintenance scenario includes: Obtain the initial entity relationship graph of the target operation and maintenance scenario; Real-time monitoring of topology change events occurring in the target operation and maintenance scenario, wherein the topology change events include connection relationship establishment events, connection relationship termination events, and connection relationship attribute change events between entities; Based on the topology change event, determine the affected nodes and / or connecting edges in the initial entity association graph, and generate change operation instructions; The initial entity association graph is updated according to the change operation instruction. Based on the update result, a dynamic association graph is generated at the current time, and the dynamic association graph is used as the entity association graph for subsequent association matching and association change cluster division.

7. The operation and maintenance digital twin rendering method based on dynamic and static entity separation and twin differential synchronization according to claim 1, characterized in that, The analysis examines the dynamic changes of each entity in the digital twin model data under its operational state. Based on the analysis results, the entities are divided into static entities and dynamic entities, including: Based on the dynamic change attributes of each entity, the abnormal entities in the digital twin model data are identified as having a dynamic-static separation. The abnormal entities are entities whose geometric model and spatial location do not change, but whose operating status information changes. Based on the identification results, if the abnormal entity of dynamic-static separation exists, the geometric model data and spatial location information of the abnormal entity of dynamic-static separation will be classified into the static entity and participate in the construction of the static scene base. The dynamic and static separation abnormal entity operation status information is divided into dynamic data, which participates in the extraction of the state change difference and incremental rendering update.

8. A digital twin rendering system for operation and maintenance based on separation of dynamic and static entities and synchronization of twin differentials, characterized in that, include: The acquisition module is used to acquire digital twin model data for the target operation and maintenance scenario; The analysis module is used to analyze the dynamic change attributes of each entity in the digital twin simulation data under the running state, and divide the entities into static entities and dynamic entities based on the analysis results. The construction module is used to construct a static scene base based on the model data of the static entity, and to preprocess the static scene base to generate static rendering resources. The extraction module is used to acquire the state data and corresponding historical state data of the dynamic entity in real time, and extract the state change difference of the dynamic entity based on the difference between the current state data and the historical state data. The update module is used to determine the target dynamic entity that has undergone a state change from the dynamic entities based on the state change difference, and to perform incremental rendering updates only on the target dynamic entity to generate dynamic rendering data. The output module is used to render and fuse the static rendering resources with the dynamic rendering data to generate and output a digital twin real-time rendering screen of the target operation and maintenance scene.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 7.

Images