Data-driven updating method for digital twin during simulation operation

Abstract

The invention discloses a data-driven updating method for a digital twin during simulation operation. The data-driven updating method comprises the following steps: step S1, a data access and preprocessing step; s2, a state updating and driving step; step S3, a multi-dimensional visual rendering step; s4, a simulation behavior verification and feedback step; s5, a dynamic evaluation and report generation step; step S6, a multi-person collaborative synchronization step; and S7, a simulation process backtracking and replaying step. According to the method, real-time access and fusion of multi-source heterogeneous data are realized, state updating of digital twin driven by multi-dimensional data such as a sensor, a service system and a simulation model is supported, and the real-time performance and data integrity of a simulation system are improved; through high-precision collision detection, a physical engine and a state synchronization mechanism, the behavior authenticity and consistency of the digital twin in scenes of interactive operation, fault simulation, multi-person cooperation and the like are guaranteed.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of computer systems, in particular to a data-driven updating method of digital twin in simulation runtime. BACKGROUND

[0002] Digital twin technology realizes the simulation and prediction of device state, behavior and performance by constructing a virtual mapping of physical entities. In mechanical simulation, equipment operation and maintenance, training and other scenarios, the real-time performance and accuracy of digital twin are crucial.

[0003] However, the existing digital twin simulation system relies on a single data source for runtime updates, and the data fusion capability is insufficient, resulting in a lag in twin state updates, a disconnection between behavior simulation and physical entities, and difficulty in achieving high-precision, high-real-time interactive simulation and fault prediction. In addition, the existing system lacks a verification and feedback mechanism for data-driven behaviors during simulation, limiting its application effect in complex operation training, collaborative work and other scenarios.

[0004] Therefore, there is an urgent need for a data-driven updating method that can fuse multi-source heterogeneous data, drive digital twin state updates in real time, and support simulation behavior verification and dynamic evaluation. SUMMARY

[0005] To this end, the present application provides a data-driven updating method of digital twin in simulation runtime to solve the problem of lag in twin state updates, disconnection between behavior simulation and physical entities, and difficulty in achieving high-precision, high-real-time interactive simulation and fault prediction in the prior art.

[0006] To achieve the above-mentioned purpose, the present application provides the following technical solutions:

[0007] The data-driven updating method of digital twin in simulation runtime comprises the following steps:

[0008] Step S1, data access and preprocessing step: real-time access to sensor data from physical entities, state data from business systems and parameter data from simulation models, analyze, clean and format the multi-source heterogeneous data, and extract state information and event information that can be used to drive digital twin updates;

[0009] Step S2, state update and driving step: update the corresponding component state, system parameters and operating mode in the digital twin according to the preprocessed data, calculate collision detection, motion trajectory and mechanical response through the physics engine, and drive the digital twin to perform corresponding actions, state changes or fault performance in the virtual scene;

[0010] Step S3, multi-dimensional visualization rendering step: based on the updated digital twin state, a high-fidelity three-dimensional visualization picture is generated using a real-time rendering engine, supporting multi-dimensional display of structure explosion view, cross-sectional view, semi-transparent view, and principle animation, and synchronously updating state prompts, operation guidelines, and data panels in the UI interface;

[0011] Step S4, simulation behavior verification and feedback step: during the interaction operation between the user and the digital twin, the correctness of the operation logic, the compliance of the tool use, and the accuracy of the step sequence are detected in real time, and immediate guidance is provided to the user through prompt information, error records, and state feedback, and operation data is recorded to the behavior database;

[0012] Step S5, dynamic evaluation and report generation step: based on the user operation record and the digital twin state change data, the simulation training process is scored and comprehensively evaluated in real time according to the preset evaluation rules, a training report containing operation right and wrong, time efficiency, and fault elimination capability indicators is generated, and report export and historical query are supported.

[0013] Preferably, in step S1, the multi-source heterogeneous data includes:

[0014] Real-time monitoring data from equipment sensors, including temperature, pressure, vibration, and position signals;

[0015] Maintenance records, operation procedures, fault codes, and material information from business systems;

[0016] Kinematics parameters, dynamics parameters, material properties, and fault tree logic data from simulation models;

[0017] Data preprocessing includes data filtering, outlier removal, timestamp alignment, and unit unification, and different sources of data are mapped to a unified state model of the digital twin through a data fusion algorithm.

[0018] Preferably, in step S2, the physics engine supports high-precision collision detection with collision accuracy reaching millimeter level, can calculate contact, separation, constraints, and mechanical response between parts in real time, and supports fault injection and state propagation based on the fault tree model, simulating the process of fault occurrence, diffusion, and influence.

[0019] Preferably, in step S3, the visualization rendering supports physics-based global lighting model, dynamic softening shadow, and real-time scene clipping, can present a three-dimensional model scene containing 16 million facets at a frame rate of no less than 60 frames per second, and supports immersive display and interactive operation of VR, AR, and MR head-mounted devices.

[0020] Preferably, in step S4, the simulation behavior verification includes:

[0021] Verification of correctness of tool selection for disassembly, assembly, inspection, maintenance operation;

[0022] Verification of logical compliance of operation step sequence;

[0023] Verification of role division and instruction coordination in multi-person collaborative operation;

[0024] The verification result is fed back to the user in real time through text prompts, highlight display, sound feedback or vibration prompt in the UI interface.

[0025] Preferably, it further comprises step S6: multi-person collaborative synchronization step, in the multi-person collaborative training scene, the state, operation action and visual angle information of the digital twin of each terminal user are synchronized in real time through the collaborative service system.

[0026] Preferably, the multi-person collaborative synchronization step comprises:

[0027] The collaborative creator creates a training course and sets the participating roles and responsibilities;

[0028] The collaborative joiner joins the collaborative scene through the network and obtains the initial state and role allocation;

[0029] During the training process, the operation action, tool state and part position of each terminal user are uploaded to the collaborative server in real time, and after being processed by the state synchronization algorithm, they are distributed to all terminals;

[0030] Support for event-triggered collaborative instruction delivery.

[0031] Preferably, it further comprises step S7: simulation process backtracking and replaying step, based on the user operation sequence and digital twin state change history recorded in the behavior database, the full-process backtracking, key node jumping and slow replaying of the simulation training process are supported.

[0032] The present application has the following advantages:

[0033] Real-time access and fusion of multi-source heterogeneous data are realized, digital twin state updates are driven by multi-dimensional data such as sensors, business systems and simulation models, and the real-time performance and data integrity of the simulation system are improved;

[0034] Through high-precision collision detection, physical engine and state synchronization mechanism, the behavior authenticity and consistency of the digital twin in interactive operation, fault simulation, multi-person collaboration and other scenes are guaranteed;

[0035] Combined with visual rendering and behavior verification feedback, visual monitoring and real-time evaluation of operation correctness of the simulation process are realized, and the immersion and teaching effect of the training are enhanced;

[0036] Supporting fault prediction and state tracking based on fault tree model, realizing intelligent driving of fault simulation and discharge process, and improving the practicality and efficiency of maintenance training;

[0037] With good scalability and compatibility, supporting multi-platform release, multi-terminal collaboration and secondary development, suitable for virtual operation, maintenance training and digital twin application of complex equipment. BRIEF DESCRIPTION OF DRAWINGS

[0038] In order to more intuitively illustrate the prior art and the present application, the following exemplary drawings are given. It should be understood that the specific shape, structure shown in the drawings should not be regarded as a limitation condition for realizing the present application; for example, based on the technical concept and exemplary drawings disclosed in the present application, those skilled in the art can easily make routine adjustments or further optimization to some units (components) in terms of increase / decrease / attribute division, specific shape, positional relationship, connection mode, size ratio relationship, etc.

[0039] Figure 1 The flowchart of the data-driven updating method of the digital twin during simulation running provided by the embodiment of the present application. DETAILED DESCRIPTION

[0040] The embodiments of the present application are described below by specific specific embodiments, and those skilled in the art can easily understand other advantages and effects of the present application from the content disclosed in the present specification. Obviously, the described embodiments are part of the embodiments of the present application, not all. It should be understood that these embodiments are only for further illustration of the present application, and cannot be understood as a limitation on the protection scope of the present application. The technical engineers in this field can make some non-essential improvements and adjustments to the present application according to the content of the above-mentioned application; based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0041] Please refer to Figure 1 , the data-driven updating method of the digital twin during simulation running, comprising the following steps:

[0042] Step S1, data access and preprocessing step: real-time access to sensor data from physical entities, state data of business systems and parameter data of simulation models, analyze, clean and format unify multi-source heterogeneous data, extract state information and event information that can be used to drive digital twin update;

[0043] Step S2, state update and driving step: according to the preprocessed data, update the corresponding part state, system parameter and operation mode in the digital twin, calculate the collision detection, motion trajectory and mechanical response through the physical engine, drive the digital twin to perform corresponding actions, state changes or fault performances in the virtual scene;

[0044] Step S3, multi-dimensional visualization rendering step: based on the updated digital twin state, generate high-fidelity three-dimensional visualization pictures using real-time rendering engine, support multi-dimensional display of structure explosion view, cross-section view, semi-transparent view, principle animation, and synchronously update the state prompt, operation guide and data panel in the UI interface;

[0045] Step S4, simulation behavior verification and feedback step: during the interaction between the user and the digital twin, real-time detect the correctness of the operation logic, the compliance of the tool usage and the accuracy of the step sequence, provide immediate guidance to the user through prompt information, error record and state feedback, and record the operation data to the behavior database;

[0046] Step S5, dynamic evaluation and report generation step: based on the user operation record and the digital twin state change data, according to the preset evaluation rules, real-time score and comprehensively evaluate the simulation training process, generate the training report containing operation right or wrong, time efficiency, fault elimination ability index, and support report export and historical query.

[0047] In the implementation of the scheme, step S1 solves the problem of single data source and difficult fusion of traditional simulation system by real-time accessing and preprocessing multi-source heterogeneous data, can unify the analysis and mapping of sensor monitoring data, business logic and simulation model parameters, and provide comprehensive and real-time state input for the digital twin. Step S2 drives the digital twin state and behavior change based on the updated data, and realizes high-precision collision detection and fault simulation combined with the physical engine, ensures that the action response, mechanical performance of the virtual model is consistent with the real physical world, and improves the reality and credibility of the simulation. Step S3 realizes multi-dimensional and interactive visualization display by using high-fidelity rendering technology, supports users to understand the device state from multiple perspectives such as structure, principle and operation, and enhances the immersion and intuitiveness of the simulation process. Step S4 verifies the correctness of the user operation in real time during the interaction and gives immediate feedback, helps the user to correct errors and standardize operation processes in time, and improves the pertinence and efficiency of the training. Step S5 generates quantitative evaluation report based on the whole process data, provides multi-dimensional evaluation such as operation right or wrong, time, compliance, supports traceability and evaluation of training effect, and provides basis for teaching review and ability certification. This method realizes the closed loop from data access, state driving, visual interaction to behavior evaluation, significantly improves the real-time, interactive reality and teaching practicality of the digital twin simulation system.

[0048] In step S1, the multi-source heterogeneous data includes:

[0049] Real-time monitoring data from equipment sensors, including temperature, pressure, vibration, and position signals;

[0050] Maintenance records, operating procedures, fault codes, and material information from business systems;

[0051] Kinematics parameters, dynamics parameters, material properties, and fault tree logic data from simulation models;

[0052] Data preprocessing includes data filtering, outlier removal, timestamp alignment, and unit unification, and through data fusion algorithms, different sources of data are mapped to a unified state model of the digital twin.

[0053] By structuring the data sources and implementing data cleaning, alignment, and fusion, different dimensions and formats of information can be effectively integrated into a unified state model, avoiding information silos and data conflicts, improving the completeness and accuracy of the digital twin state description, and making it more realistic to reflect the comprehensive condition of the physical entity, providing a reliable data foundation for subsequent simulation driving and decision analysis.

[0054] In step S2, the physics engine supports high-precision collision detection with millimeter-level collision accuracy, which can calculate the contact, separation, constraints, and mechanical response between parts in real time, and supports fault injection and state propagation based on the fault tree model, simulating the process of fault occurrence, spread, and impact.

[0055] Millimeter-level collision accuracy can realistically simulate the assembly relationship, contact mechanics, and interference between parts, ensuring the physical credibility of disassembly, maintenance, and other operations. Combined with the fault tree model, it can achieve automatic injection of fault phenomena, simulation of propagation path, and impact analysis, enabling the digital twin to have fault prediction and discharge deduction capabilities, greatly enhancing the practical value of the simulation system in maintenance training, safety assessment, and other scenarios.

[0056] In step S3, the visualization rendering supports physics-based global lighting models, dynamic soft shadows, and real-time scene clipping, which can present a three-dimensional model scene containing 16 million polygons at a frame rate of not less than 60 frames per second, and support immersive display and interactive operation of VR, AR, and MR head-mounted devices.

[0057] By using the physical-based rendering technology and high frame rate output, the scene of complex model can be kept smooth and the details can be kept real, and the visual experience close to the real object can be provided. Meanwhile, the VR / AR / MR and other immersive terminals are supported, so that the user can operate and train in a natural interactive way, and the limitations of the traditional desktop simulation are broken, and the immersion, reality and practicality of the training are significantly improved.

[0058] In step S4, the simulation behavior verification includes:

[0059] verification of correctness of tool selection for disassembly, assembly, inspection and maintenance operations;

[0060] verification of logical compliance of operation step sequence;

[0061] verification of role division and instruction coordination in multi-person collaborative operation;

[0062] The verification result is fed back to the user in real time through the text prompt, highlight display, sound feedback or vibration prompt in the UI interface.

[0063] The above scheme can identify common errors and violations in user operation in real time through real-time logical verification of tool selection, step sequence and collaborative division, and give prompt through multi-channel feedback mechanism, which not only reduces the learning cost caused by misoperation, but also standardizes the operation process and strengthens the standard operation consciousness, which has positive significance for cultivating rigorous and standardized operation habits.

[0064] Further comprising a multi-person collaborative synchronization step S6: In the multi-person collaborative training scene, the digital twin state, operation action and perspective information of each terminal user are synchronized in real time through the collaborative service system, the delay of each terminal scene display is ensured to be not more than 20 ms, and the role division, instruction transmission and collaborative process editing and driving are supported, and through the low-delay state synchronization and role division management, the consistent perspective and coordinated action of each terminal are ensured, which can simulate the communication, cooperation and process connection in real team cooperation, expand the application scene of the simulation system, and is suitable for assembly, maintenance, emergency drill and other comprehensive training requiring multi-person cooperation, and effectively improves the training effect of team cooperation ability.

[0065] The multi-person collaborative synchronization step includes:

[0066] The collaborative creator creates a training course and sets the participating roles and responsibilities;

[0067] The collaborative joiner joins the collaborative scene through the network, obtains the initial state and role allocation;

[0068] During the training process, the operation action, tool state and part position of each terminal user are uploaded to the collaborative server in real time, and after being processed by the state synchronization algorithm, they are distributed to all terminals.

[0069] Support event-triggered cooperative instruction transmission, such as part transmission, tool handover, and joint operation, to ensure logical consistency and timing correctness of cooperative actions.

[0070] The scheme specifies the implementation process of multi-person cooperation, including creation, joining, state synchronization, and instruction transmission. Through the event-triggered instruction mechanism, the interactive behaviors such as tool transmission and joint operation in real collaboration can be simulated, ensuring the logical consistency and timing correctness of cooperative actions. This makes the multi-person training process controllable and editable, and can flexibly adapt to different cooperative courses and role settings, improving the organization flexibility and training pertinence of the cooperative simulation system.

[0071] It also includes step S7: simulation process backtracking and replaying step, based on the user operation sequence and digital twin state change history recorded in the behavior database, supporting full-process backtracking, key node jumping, and slow replaying of the simulation training process, for operation review, error analysis, and teaching demonstration. By recording the operation sequence and state snapshot, it supports playback at any time, key node jumping, and slow analysis of the training process, facilitating post-mortem review, error tracing, and teaching demonstration. This design enhances the teaching analysis capability of the simulation system, helping instructors and students understand operation details and summarize lessons learned, enabling review, analysis, and optimization of the training process, and improving the systematicness and scientificity of simulation training.

[0072] The above only describes the preferred embodiments of the present application and is not intended to limit the present application. Any modifications, equivalent replacements, and improvements made within the spirit and principles of the present application shall be included in the protection scope of the present application.

Claims

1. A data-driven update method for digital twins during simulation runtime, characterized in that, Includes the following steps: Step S1, Data Access and Preprocessing Steps: Real-time access to sensor data from physical entities, status data from business systems, and parameter data from simulation models; parsing, cleaning, and format unification of multi-source heterogeneous data; extracting status and event information that can be used to drive the update of the digital twin. Step S2, State Update and Driving Step: Based on the preprocessed data, update the corresponding component states, system parameters and operating modes in the digital twin, calculate collision detection, motion trajectory and mechanical response through the physics engine, and drive the digital twin to perform corresponding actions, state changes or fault manifestations in the virtual scene. Step S3, Multi-dimensional Visualization Rendering Step: Based on the updated digital twin state, a real-time rendering engine is used to generate a highly realistic 3D visualization screen, supporting multi-dimensional display of exploded structural views, cross-sectional views, semi-transparent views, and principle animations, while simultaneously updating the status prompts, operation guides, and data panels in the UI interface. Step S4, Simulation Behavior Verification and Feedback Step: During the interaction between the user and the digital twin, the correctness of the operation logic, the compliance of the tool use, and the accuracy of the step sequence are detected in real time. The user is given immediate guidance through prompts, error logs, and status feedback, and the operation data is recorded in the behavior database. Step S5, Dynamic Evaluation and Report Generation: Based on user operation records and digital twin state change data, the simulation training process is scored and comprehensively evaluated in real time according to preset evaluation rules, generating a training report that includes indicators such as operation correctness, time efficiency, and troubleshooting ability, and supports report export and historical query.

2. The data-driven update method for a digital twin during simulation runtime according to claim 1, characterized in that, In step S1, the multi-source heterogeneous data includes: Real-time monitoring data from equipment sensors, including temperature, pressure, vibration, and position signals; Maintenance records, operating procedures, fault codes, and material information from the business system; Kinematic parameters, dynamic parameters, material properties, and fault tree logic data from the simulation model; Data preprocessing includes data filtering, outlier removal, timestamp alignment and unit unification, and data fusion algorithms map data from different sources to a unified state model of the digital twin.

3. The data-driven update method for a digital twin during simulation runtime according to claim 1, characterized in that, In step S2, the physics engine supports high-precision collision detection with millimeter-level accuracy. It can calculate the contact, separation, constraint and mechanical response between components in real time, and supports fault injection and state propagation based on fault tree model to simulate the occurrence, spread and impact of faults.

4. The data-driven update method for a digital twin during simulation runtime according to claim 1, characterized in that, In step S3, the visualization rendering supports physically based global illumination models, dynamic soft shadows, and real-time scene clipping. It can present a 3D model scene containing 16 million polygons at a frame rate of no less than 60 frames per second and supports immersive display and interactive operation of VR, AR, and MR head-mounted displays.

5. The data-driven update method for a digital twin during simulation runtime according to claim 1, characterized in that, In step S4, the simulation behavior verification includes: Verify the correctness of tool selection for disassembly, assembly, inspection, and maintenance operations; Logical compliance verification of the sequence of operation steps; Verification of role division and command coordination in multi-person collaborative operations; The verification results are fed back to the user in real time through text prompts, highlighting, sound feedback, or vibration alerts in the UI interface.

6. The data-driven update method for a digital twin during simulation runtime according to claim 1, characterized in that, It also includes step S6: multi-person collaborative synchronization step, in which the digital twin status, operation actions and perspective information of each terminal user are synchronized in real time through the collaborative service system in a multi-person collaborative training scenario.

7. The data-driven update method for a digital twin during simulation runtime according to claim 6, characterized in that, The multi-person collaborative synchronization steps include: Collaborators create training courses and define participating roles and responsibilities; Collaborators join the collaborative scenario via the network and obtain their initial state and role assignments; During the training process, the operation actions, tool status, and part positions of each terminal user are uploaded to the collaboration server in real time, and then distributed to all terminals after being processed by the status synchronization algorithm. Supports event-triggered collaborative command delivery.

8. The data-driven update method for a digital twin during simulation runtime according to claim 1, characterized in that, It also includes step S7: simulation process backtracking and replay step, which supports full-process backtracking of the simulation training process, key node jumps and slow replay based on the user operation sequence and digital twin state change history recorded in the behavior database.

Images