Dynamic scene simulation system and method based on digital twinning

Abstract

The application relates to the technical field of dynamic simulation, and discloses a dynamic scene simulation system and method based on digital twinning. The method comprises the following steps: constructing a digital twinning model according to a physical object on a local computing node; uploading data of the model to a cloud server; the cloud server integrates the behavior relationship of multiple physical objects, constructs a system model, and supports multiple users to access the system model through permission management; a simulation task is assigned to the local computing node to perform local simulation, and the result is generated and synchronized to the cloud server in real time; the cloud server performs global scene simulation, generates an optimization strategy, and delivers the optimization strategy to the local model with permission, and the latter executes optimization simulation based on the optimization strategy and feeds back the result to control the operating parameters or state of the physical object. The method supports multiple users to share, cooperate and dynamically manage the simulation process, realizes efficient and safe distributed digital twinning simulation, and significantly improves the cooperation efficiency, real-time performance, data reliability and cross-organization cooperation flexibility.

Description

Technical Field

[0001] This application relates to the field of dynamic simulation technology, and in particular to a dynamic scene simulation system and method based on digital twins. Background Technology

[0002] With the continuous development of intelligent manufacturing and the Industrial Internet of Things (IIoT), digital twin technology is gradually becoming an important tool for optimizing production processes and improving equipment operating efficiency. Digital twin technology constructs virtual models corresponding to physical entities, reflecting the state, behavior, and interaction relationships of physical objects in real time, thus providing powerful support for the control, monitoring, and optimization of production processes. Especially in industrial production processes, digital twins can achieve efficient monitoring and management of complex systems such as production equipment, workshops, and production lines.

[0003] However, existing digital twin applications are mostly focused on modeling and simulating single physical objects or simple systems. For example, CN116861652A proposes a real-time simulation method based on digital twins. This method mainly involves constructing a twin model corresponding to the actual physical scene and collecting data related to simulation control in real time to generate virtual simulation results that are consistent with the behavior of the physical object. This method can achieve precise control of the target object and ensure dynamic synchronization between the digital twin model and the physical world.

[0004] While this technology can effectively achieve real-time simulation, its applicability remains somewhat limited. Current technologies primarily focus on the dynamic control of single target objects or small-scale simulation scenarios. However, for modeling and optimizing complex interactions between multiple physical objects, especially in collaborative simulation scenarios involving multiple nodes and multiple users, existing technologies often fail to meet practical needs. Summary of the Invention

[0005] This application provides a dynamic scene simulation system and method based on digital twins. Through a cloud-edge collaborative architecture and a multi-level permission management mechanism, a distributed digital twin simulation system is constructed to solve problems such as low efficiency of collaborative simulation, difficulty in secure data sharing, and insufficient real-time synchronization in complex dynamic scenes. It achieves efficient and secure distributed digital twin simulation and closed-loop optimization control, improving the real-time performance, accuracy, and scalability of simulation tasks.

[0006] Firstly, this application provides a dynamic scene simulation method based on digital twins, the method comprising: On local computing nodes, corresponding local digital twin models are built based on physical objects in the real world; The structured description information and interface data of multiple local digital twin models are uploaded to the cloud server, and a system model is built on the cloud server. The system model is used to integrate the behavioral relationships and interaction logic of multiple physical objects. The cloud server supports multiple users to access and share the system model, and sets the permissions of each user to the system model through a multi-level permission management mechanism; The simulation task is assigned to each local computing node. The local computing node performs local simulation based on the local digital twin model and real-time data, generates the first simulation result, and synchronizes the first simulation result to the cloud server in real time through the communication interface. The cloud server integrates and applies the received first simulation results to the corresponding regional modules in the system model, performs global scene simulation, generates optimization strategies, and then distributes the optimization strategies to local digital twin models with the corresponding permissions. The local digital twin model performs optimization simulation based on the optimization strategy, generates a second simulation result, and feeds the second simulation result back to the corresponding physical object to control its operating parameters or execution status.

[0007] In conjunction with the first aspect, in the first implementation of the first aspect of this application, the cloud server includes a participant management module for fine-grained control of access permissions to multiple local digital twin models and system models based on preset permission policies. The permission policies include: Each local digital twin model is bound to a corresponding user account, and access permissions to the local digital twin model are set according to the user's identity. The access permissions include the scope of access and operation permissions. Operation permissions include viewing, editing, deploying simulation, setting parameters, executing simulation and / or issuing control commands. Based on the simulation task, access to the corresponding modules is temporarily granted to collaborative users. The corresponding modules are the functional modules in the system model that correspond to the simulation task. Permission settings support cross-organizational configuration, allowing users in different organizations to obtain different levels of access permissions according to the needs of collaborative tasks, and allowing the local digital twin model owner to dynamically adjust or revoke access permissions according to task requirements or collaboration status; The system authenticates each access request, confirms access permissions, and records access logs.

[0008] In conjunction with the first aspect, in the second implementation of the first aspect of this application, the cloud server manages multiple local digital twin models and dynamically updates the system model by receiving data from each local computing node, specifically including: Bind each local digital twin model to its corresponding physical object and local computing node; Any local computing node receives first state data from the first physical object in real time and updates the first local digital twin model based on the first state data to ensure that the first local digital twin model reflects the current state and behavior of the first physical object. The first physical object and the first local digital twin model are the physical object and local digital twin model corresponding to any local computing node. The cloud server receives second status data uploaded by any local computing node in real time. When the status information and / or model data of any local computing node are detected to have changed, the system model is dynamically managed and updated based on the first status data to ensure that the behavior of a specific area module in the system model is synchronized with the specific local digital twin model. The specific local digital twin model is the local digital twin model corresponding to any local computing node, and the specific area module is the area module corresponding to the specific local digital twin model.

[0009] In conjunction with the first aspect, in the third implementation of the first aspect of this application, the local digital twin model is divided into multiple sub-modules, the basic sub-modules are encapsulated as executable units, and the remaining sub-modules are encrypted as model resource files and stored in a file storage module. A data adaptation module for storing the calling parameters of the model resource files is set on the local computing node. Before performing local simulation, the local computing node includes: When a local computing node receives a partial simulation task, it extracts all corresponding file data based on the calling parameters. Extract any file data, and the authorization proxy module generates an authentication request based on the identifier of the file data and the user identifier, and sends it to the authentication server; The authentication server verifies user permissions. If the permission verification is successful, the authorization proxy module decrypts any file data to obtain the corresponding model resource file. After traversing all file data, a local digital twin model is generated based on the executable unit and all model resource files.

[0010] In conjunction with the first aspect, in the fourth implementation of the first aspect of this application, the local digital twin model is formatted after the local simulation task is completed.

[0011] In conjunction with the first aspect, in the fifth implementation of the first aspect of this application, the local computing node includes the following steps before performing local simulation: Extract any real-time data, obtain the data credibility of any real-time data, determine whether the data credibility meets the corresponding preset standard, if not, obtain the data identifier of any real-time data, extract the first historical data within a first preset time based on the data identifier, perform statistical analysis on the first historical data, obtain statistical values, and set the statistical values ​​as any real-time data. After traversing all real-time data, new real-time data is generated.

[0012] In conjunction with the first aspect, in the sixth implementation of the first aspect of this application, when a user's model setting request is received, the local digital twin model is updated, including: The local digital twin model is divided into multiple sub-models based on functions or physical objects, where each sub-model corresponds to a specific function or a specific physical object of the local digital twin model; Extract the application target from the setting request, obtain the corresponding functional configuration model based on the application target, determine whether the input parameters of the functional configuration model can be obtained from the physical object, if not, send a prompt message, if yes, select a candidate model from multiple sub-models based on model function, data interface and / or logical relationship; Extract any candidate model, identify the common substructure between any candidate model and the functional configuration model through topology consistency analysis, and generate multiple dynamic alignment models based on any candidate model, the functional configuration model and the common substructure. Extract any dynamic alignment model, and generate candidate digital twin models based on any dynamic alignment model and other sub-models besides any candidate model; Extract the second historical data within the second preset time period, simulate the candidate digital twin model based on the second historical data, obtain performance index parameters, compare the performance index parameters with the historical performance index parameters, and quantify the offset of different performance indicators. After traversing all candidate models, all candidate digital twin models are sorted based on offsets and sent to the user. A new local digital twin model is then generated based on user feedback.

[0013] In conjunction with the first aspect, in the seventh implementation of the first aspect of this application, the constituent elements of the sub-model, the functional configuration model, and the dynamic alignment model include: entity topology descriptor, functional semantic label, entity derived signal set, logical execution chain, analysis model library, and parameter constraints and definitions.

[0014] Secondly, this application provides a dynamic scene simulation system based on digital twins. The system includes: a local computing node and a cloud server. The local computing node includes a local model building module, a simulation synchronization module and a simulation feedback module. The cloud server includes a system model building module, a permission setting module and a policy distribution module. The local model building module is used to build corresponding local digital twin models based on physical objects in the real world; The system model building module is used to build a system model after the structured description information and interface data of multiple local digital twin models are uploaded to the cloud server. The system model is used to integrate the behavioral relationships and interaction logic of multiple physical objects. The permission settings module is used to support multiple users to access and share the system model, and sets the permissions of each user to the system model through a multi-level permission management mechanism; The simulation synchronization module is used to perform local simulation based on the local digital twin model and real-time data after the cloud server distributes the simulation tasks to each local computing node, generate the first simulation result, and synchronize the first simulation result to the cloud server in real time through the communication interface. The strategy distribution module is used to integrate multiple received first simulation results and apply them to the corresponding regional modules in the system model, perform global scene simulation, generate optimization strategies, and then distribute the optimization strategies to the local digital twin models with the corresponding permissions. The simulation feedback module is used to enable the local digital twin model to perform optimization simulation according to the optimization strategy, generate a second simulation result, and feed the second simulation result back to the corresponding physical object to control its operating parameters or execution status.

[0015] Compared with the prior art, the beneficial effects of the technical solution of this application are at least as follows: 1. By collaborating with multiple local computing nodes and cloud servers, simulation task allocation and global optimization are achieved across multiple physical objects and computing nodes. Each node performs local simulation, while the cloud server integrates all simulation results, performs global scene simulation, and generates optimization strategies, greatly improving the response speed, processing power, and overall system efficiency of simulation tasks.

[0016] 2. Upload the structured description information and interface data of multiple local digital twin models to a cloud server, build a system model on the cloud server, and support multiple users to access and share the system model through a multi-level permission management mechanism. This allows for granular access and operation permissions for each user in different simulation tasks and system models, enabling multiple users to use the same digital twin model. This not only supports cross-organizational collaboration but also dynamically adjusts permissions according to different task requirements, ensuring system security while fully utilizing resources from multiple parties for optimized simulation.

[0017] 3. By integrating multiple local simulation results into the global system model through a cloud server, optimization strategies are generated and promptly fed back to the local digital twin model to control its behavior and execution status. This closed-loop feedback mechanism allows optimization decisions to directly impact the operation of physical objects, improving the system's dynamic response capabilities.

[0018] 4. Through data credibility assessment and historical data analysis mechanisms, the system ensures that the real-time data collected by each node meets the preset credibility standards. When the data credibility does not meet the standards, the system supplements and analyzes historical data, enhancing the reliability and accuracy of the data, thereby improving the quality of the simulation results.

[0019] 5. By encapsulating the core algorithms of the local digital twin model into an executable unit and utilizing dynamic permission verification and file data extraction mechanisms, the system can ensure the flexibility of model invocation while achieving secure management and permission control for the reuse of simulation model files. Attached Figure Description

[0020] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0021] Figure 1 This is a schematic diagram of one embodiment of the dynamic scene simulation method based on digital twins in this application. Figure 2 This is a schematic diagram of one embodiment of the local digital twin model file storage structure in this application. Figure 3 This is a schematic diagram of one embodiment of the dynamic alignment model processing flow in this application. Figure 4 This is a schematic diagram of one embodiment of the dynamic alignment model processing flow in this application. Figure 5 This is a schematic diagram of one embodiment of the dynamic scene simulation system based on digital twins in this application. Detailed Implementation

[0022] This application provides a dynamic scene simulation system and method based on digital twins. The terms "first," "second," "third," "fourth," etc. (if present) in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments described herein can be implemented in a sequence other than that illustrated or described herein. Furthermore, the terms "comprising" or "having" and any variations thereof are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or device that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or devices.

[0023] For ease of understanding, the specific process of the embodiments of this application is described below. Please refer to [link / reference]. Figure 1 One embodiment of the dynamic scene simulation method based on digital twins in this application includes: Step 1: On the local computing node, construct the corresponding local digital twin model based on the physical objects in the real world.

[0024] Specifically, a local computing node refers to a device or computing resource located in the physical environment of a distributed computing system, capable of performing computing tasks and interacting with other computing nodes and cloud servers. It can be a single computer, server, or even an embedded or edge computing device. Its function is to process, generate, and maintain digital twin models locally, while collaborating with the cloud or other computing nodes. By building digital twin models of physical objects on local computing nodes, a mapping from the physical world to the virtual world is achieved. The local digital twin model is a digital representation of the physical object, capable of simulating its behavior and state.

[0025] Physical objects refer to actual entities that need to be monitored, managed, or optimized in smart manufacturing and industrial IoT environments. These physical objects typically include various production equipment, machines, processes, production lines, and other hardware facilities involved in the production process. For example, production equipment such as robots, robotic arms, conveyor belts, and compressors are all physical objects; at a more macroscopic level, an entire production workshop or production line can also be considered a physical object. The main characteristic of these physical objects is that they exist in the physical world and perform specific functions or operations, such as processing, handling, assembly, and testing.

[0026] Step 2: Upload the structured description information and interface data of multiple local digital twin models to the cloud server, and build a system model on the cloud server. The system model is used to integrate the behavioral relationships and interaction logic of multiple physical objects.

[0027] Specifically, structured descriptive information refers to the basic information of the local digital twin model. This information typically includes the model's physical properties (such as size, weight, and materials), functional modules (such as sensors and actuators), state parameters (such as temperature and pressure), and behavioral logic (such as control flow and event response). This descriptive information is standardized and formatted to ensure it can be effectively parsed and processed by the cloud server. Interface data refers to the specific data exchanged between the local model and the cloud or other systems. It typically includes communication protocols, API interfaces, data formats, and operation instructions, enabling real-time data synchronization and sharing among the various local models.

[0028] By uploading information from multiple local digital twin models to a cloud server, a higher-level system model is constructed. When this structured descriptive information and interface data are uploaded to the cloud server, the server integrates the data from different local digital twin models to establish a global digital twin model (i.e., the system model). This system model integrates the behavioral relationships and interaction logic of multiple physical objects, helping the cloud achieve comprehensive control over the entire production system. The system model not only merges data from different local models but also models the dependencies and interaction processes between physical objects, ensuring that the collaborative work between various physical objects is accurately reflected. Ultimately, the cloud-based system model can support various functions such as multi-object collaborative simulation, global optimization, and decision-making, helping managers perform comprehensive production scheduling and optimization management.

[0029] Step 3: The cloud server supports multiple users to access and share the system model. Each user's permissions to the system model are set through a multi-level permission management mechanism.

[0030] Specifically, as a centralized management platform, the cloud server provides shared access to the system model, allowing users with different roles to view, analyze, and operate the data within the model as needed. A multi-level permission management mechanism assigns different access permissions to each user, ensuring system security and controllability while enabling multiple users to collaborate using the same model. Specifically, these permissions typically include: access permissions (different users can obtain different access scopes to the system model based on their roles and tasks), operation permissions (including viewing, editing, deleting, deploying simulations, setting parameters, etc.), cross-organizational collaboration support, and dynamic permission adjustment (the cloud server supports dynamically adjusting user permissions as tasks or collaboration needs change).

[0031] Step 4: Assign simulation tasks to each local computing node. The local computing node performs local simulation based on the local digital twin model and real-time data, generates the first simulation result, and synchronizes the first simulation result to the cloud server in real time through the communication interface.

[0032] Specifically, direct simulation in the system model requires centralized processing of massive amounts of data, which is limited by network bandwidth and cloud computing latency, making it difficult to meet real-time requirements. Edge servers, located close to the data source (such as sensors and physical objects), can process data and execute simulations locally quickly. They only upload high-value data that affects global decisions (such as abnormal equipment status (abnormal vibration, abnormal temperature), predictive conclusions (current equipment failure rate, equipment failure in one hour), etc.), rather than raw data streams, thus avoiding network latency and the explosive bandwidth costs of uploading massive amounts of equipment data.

[0033] Step 5: The cloud server integrates the received first simulation results into the corresponding regional modules in the system model, performs global scene simulation, generates optimization strategies, and then distributes the optimization strategies to local digital twin models with the corresponding permissions.

[0034] Specifically, after receiving the first simulation results uploaded by multiple local computing nodes, the cloud server performs data fusion and, combined with the behavioral patterns of corresponding regional modules in the system model, performs a global scenario simulation on all local simulation results. Global simulation is not merely a simple aggregation of results; it involves complex coordination and optimization based on the interrelationships between different modules. Through global simulation, the cloud server generates an optimization strategy and distributes this strategy to local digital twin models with corresponding permissions. The optimization strategy involves adjusting the operating parameters of physical objects and optimizing their behavior, aiming to improve overall system performance through optimized simulation. For example, the system model consists of a local digital twin model B1 for the production line of factory B, and local digital twin models A1 and A2 for the upstream and downstream production lines of factory A. When the production parameters of the upstream production line of factory A change, local digital twin model A1 simulates the changed production parameters and uploads the simulation results to the cloud server. Since the production progress of factory A is only related to its upstream and downstream production lines, the system model performs a global simulation based on the production parameters of the downstream production line of factory A, generating optimization strategies for adjusting the production progress, etc., and then distributing the corresponding optimization strategies for the upstream and downstream production lines to A1 and A2 respectively. For example, the system model consists of a local digital twin model C1 of the production line of factory C and a local digital twin model D1 of the production line of factory D. Both factories C and D have production equipment E (i.e., physical objects). Factory C optimizes the operating parameters of production equipment E and uploads the optimized operating parameters and simulation results to the cloud server. After the system model performs simulation, it determines that the operating parameters help improve production efficiency, and then sends the operating parameters to the local digital twin model D1 of the production line of factory D, which has the same production equipment E.

[0035] Step 6: The local digital twin model performs optimization simulation based on the optimization strategy, generates a second simulation result, and feeds the second simulation result back to the corresponding physical object to control its operating parameters or execution status.

[0036] Specifically, the local model adjusts its simulation parameters based on the optimization suggestions from the cloud server, performing more accurate local simulations to generate a second simulation result. This second simulation result reflects the optimized state of the physical object and, through a feedback mechanism, transmits these results back to the physical object itself. By applying the optimization results from the virtual world to the physical world, closed-loop control of the digital twin is achieved. The physical object can automatically adjust its operating state based on the simulation results, thus enabling continuous optimization and control.

[0037] In one specific embodiment, the cloud server includes a participant management module, used to perform fine-grained control over access permissions to multiple local digital twin models and system models based on preset permission policies. The permission policies include: (1) Bind each local digital twin model to the corresponding user account, and set the user's access permissions to the local digital twin model according to the user's identity. The access permissions include the scope of access and operation permissions. The operation permissions include viewing, editing, deploying simulation, setting parameters, executing simulation and / or issuing control commands.

[0038] (2) Based on the simulation task, temporarily open access permissions to the corresponding modules for collaborative users, where the corresponding modules are the functional modules corresponding to the simulation task in the system model.

[0039] (3) Permission settings support cross-organizational configuration, allowing users of different organizations to obtain different levels of access permissions according to the needs of collaborative tasks, and allowing the owner of the local digital twin model to dynamically adjust or revoke access permissions according to task requirements or collaboration status; (4) The system authenticates each access behavior, confirms access permissions, and records access logs.

[0040] Specifically, each local digital twin model is associated with a specific user account (such as an equipment administrator or operator), and different access permissions are set according to the user's identity (such as an administrator, engineer, or regular user). Access scope includes viewing only the model of a specific device or the model of the entire production line, etc. Operation permissions include only being able to view data, being allowed to edit the model, and issuing control commands, etc. For example, a factory administrator can view and adjust the parameters of the digital twin models of all devices, a production line operator can only view the simulation data of their own production line, and the administrator of factory A can view the parameters of the digital twin model of factory B.

[0041] When multiple users need to collaborate on a simulation task, the cloud server can temporarily grant access permissions to relevant modules. Once the task is completed, the permissions are automatically revoked or manually adjusted by the owner. For example, a factory collaborates with a supplier to optimize a piece of equipment. The supplier's engineer is temporarily granted "parameter setting" permissions for the equipment model, which are automatically revoked after the task ends.

[0042] It supports users from different enterprises or departments to obtain permissions on demand (such as supply chain collaboration and joint R&D). The model owner can adjust or revoke permissions at any time to adapt to changes in tasks or security requirements. For example, an automaker (Company G) and a battery supplier (Company H) share a battery digital twin model. Company H can only access the battery module, and its permissions can be revoked by Company G at any time.

[0043] By employing a refined permission management mechanism (such as user-model binding, hierarchical operation permissions, temporary permission opening, and cross-organizational collaborative configuration) and strict access control (authentication + log auditing), the security and collaborative flexibility of the digital twin system are ensured. This breaks down the barrier that the simulation results of a digital twin model cannot be shared with its related users, allowing multiple users to share the digital twin model. This not only prevents unauthorized operations but also supports dynamic task collaboration among multiple users and across organizations (such as joint simulation and supply chain optimization). Furthermore, through dynamic permission adjustment and operation traceability, the availability and reliability of the system in complex environments are improved.

[0044] The technical solution proposed in this application integrates and manages multiple local digital twin models from different owners (such as factories and suppliers) through a unified digital twin platform. It achieves secure sharing and collaboration through an access control mechanism, breaks down data silos, and enables the secure sharing of digital twin models from different enterprises. At the same time, it dynamically adjusts the sharing scope according to business needs (such as temporarily authorizing suppliers for debugging), achieving flexible collaboration and reducing trial and error costs.

[0045] In one specific embodiment, the cloud server manages multiple local digital twin models and dynamically updates the system model by receiving data from each local computing node, specifically including: (1) Bind each local digital twin model to its corresponding physical object and local computing node.

[0046] (2) Any local computing node receives the first state data from the first physical object in real time and updates the first local digital twin model based on the first state data to ensure that the first local digital twin model reflects the current state and behavior of the first physical object. The first physical object and the first local digital twin model are the physical object and local digital twin model corresponding to any local computing node.

[0047] (3) The cloud server receives the second status data uploaded by any local computing node in real time. When the status information and / or model data of any local computing node are detected to change, the system model is dynamically managed and updated according to the first status data to ensure that the behavior of a specific area module in the system model is synchronized with the specific local digital twin model. The specific local digital twin model is the local digital twin model corresponding to any local computing node, and the specific area module is the area module corresponding to the specific local digital twin model.

[0048] Specifically, the cloud server maintains a unique mapping relationship between physical objects, local computing nodes, and digital twin models. For example: machine tool K (physical object) ←→ computing node X ←→ digital twin model α.

[0049] The first state data is directly generated from the physical object. It consists of raw state data collected in real time by sensors or control systems, used to update the local digital twin model. This data includes operating parameters, operational status, environmental data, and location information. The second state data is structured state information uploaded to the cloud server after preprocessing by the local computing node. This data is used to update the system model. Examples include feature indicators (statistics extracted from the raw data, such as mean, peak value, and trend slope), model parameters (key parameters adjusted from the local digital twin model, such as kinetic coefficients and thermal resistance), event markers (abnormal alarms, such as overheat warnings, and state transitions, such as "standby → running"), and related data (interactions with other objects, such as material handover timestamps).

[0050] Take planned equipment removal as an example. Robot R (the physical object) on the intelligent production line is responsible for the welding process. The operator initiates a maintenance command for robot R through the MES system, stops data collection, and uploads second-state data (equipment removal declaration) to the local digital twin model. The cloud server marks robot R as offline for maintenance in the system model, dynamically reorganizes the production line digital thread, bypasses the physical object to generate a new process path, and updates the system model.

[0051] By enabling cloud servers to receive and dynamically update data from multiple local digital twin models, the system model can reflect the latest status and behavior of each physical object in real time. This not only improves the real-time performance and accuracy of the production process but also ensures that the system can quickly adjust and optimize the production process in the face of equipment failures, maintenance, or other unforeseen events. For example, when a piece of equipment or a process changes, the cloud server automatically reconfigures the system model to avoid impacting the production line, thereby improving overall production efficiency and flexibility. It also enables intelligent scheduling and seamless collaboration, ensuring the synchronization of physical object behavior with the system model and enhancing reliability and controllability in industrial production environments.

[0052] In one specific embodiment, the local digital twin model is split into multiple sub-modules, the basic sub-modules are encapsulated as executable units, and the remaining sub-modules are encrypted as model resource files and stored in a file storage module. A data adaptation module for storing the calling parameters of the model resource files is set on the local computing node. Before performing local simulation, the local computing node includes: (1) When the local computing node receives a local simulation task, it extracts all file data based on the calling parameters.

[0053] (2) Extract any file data. The authorization agent module generates an authentication request based on the identifier of any file data and the user identifier, and sends it to the authentication server.

[0054] (3) The authentication server verifies the user's permissions. If the permission verification is successful, the authorization agent module decrypts any file data and obtains the corresponding model resource file.

[0055] (4) After traversing all file data, generate a local digital twin model based on the executable unit and all model resource files.

[0056] In one specific embodiment, the local digital twin model is formatted after the local simulation task is completed.

[0057] Specifically, the aforementioned basic sub-modules are the system's original built-in or basic version models, belonging to embedded models. They are the parts of the local digital twin model that directly define the system's behavior, structure, or function, including all detailed information used for simulation calculations, such as mathematical equations, state machines, logical relationships, and physical parameters. The remaining sub-modules are models developed, optimized, and upgraded by third parties or users themselves based on the embedded models, possessing independent intellectual property rights. To protect the development results, encryption encapsulation and authorization mechanisms are required. Model resource files are independent data files storing the specific implementations of the remaining sub-modules. Through parameter calls from the data adaptation module, model resource files can be fully or partially encrypted. The file storage module can be located on a local computing node or a cloud server.

[0058] A schematic diagram of the structure of local digital twin model file storage is shown below. Figure 2 As shown. When a local simulation needs to be performed, the local computing node receives the local simulation task and extracts all corresponding model resource files 12 from the file storage module 300 based on the calling parameters in the data adaptation module 13. Then, it performs permission verification through the authentication server 400. After successful verification, the authorization proxy module 14 decrypts the model resource files.

[0059] For example, suppose a car manufacturer uses a simulation device to optimize its battery management system (i.e., a local digital twin model). This local digital twin model includes a data adaptation module IM_Battery, whose parameters point to an encrypted file named EF_Batt. In the file storage module, EF_Batt is encrypted using AES-256 and accompanied by a decryption program P_Batt with a SHA-256 signature. After the simulation device retrieves EF_Batt, the authorization agent module checks whether the authentication server authorizes the manufacturer to use this battery model. If the verification is successful, the authentication server returns a dynamically generated key k_Batt. The simulation device decrypts P_Batt to obtain the model resource file DF_Batt, which contains executable code containing the battery thermodynamic equations. After the simulation is complete, DF_Batt is immediately cleared from memory, while EF_Batt remains securely stored in the file storage module.

[0060] The system employs a hierarchical management system based on the importance and sensitivity of submodules. Basic or general components are encapsulated in plaintext as embedded models, while submodules with intellectual property value or security sensitivity are encrypted and stored as model resource files. The system dynamically references these model resource files through a data adaptation module and, after verification and authorization, loads and executes them as simulation models, thereby achieving a unified approach to model reuse, protection, and access control.

[0061] In one specific embodiment, the local computing node includes the following components before performing local simulation: (1) Extract any real-time data, obtain the data credibility of any real-time data, determine whether the data credibility meets the corresponding preset standard, if not, obtain the data identifier of any real-time data, extract the first historical data within the first preset time based on the data identifier, perform statistical analysis on the first historical data, obtain the statistical value, and set the statistical value as any real-time data.

[0062] (2) After traversing all real-time data, generate new real-time data.

[0063] Specifically, before performing local simulation, the data reliability of real-time data is obtained based on a preset calculation method. This preset calculation method includes, but is not limited to, generation interval, data range, maximum difference, and data source evaluation value. For example, the time interval between reception and generation is 5 seconds; the data reliability is 5 for direct sources and 3 for indirect sources; the data reliability is 0.1 for data acquisition equipment with a failure rate of 90% and 0.9 for a failure rate of 10%. Different data reliability calculation methods and preset standards can be set for each real-time data point.

[0064] If the reliability of the detected data does not meet the preset standard, the local computing node will mark this unreliable data and take remedial measures. Specific measures include: first, using historical data as a substitute; and second, performing statistical analysis on the historical data according to a preset time window to generate "reliable" statistical values, which will then be used as the values ​​of the real-time data.

[0065] Based on the above data screening method, data with acceptable credibility are selected for local simulation to ensure the accuracy and stability of the local simulation results.

[0066] In one specific embodiment, upon receiving a user's model setting request, updating the local digital twin model includes: (1) Divide the local digital twin model into multiple sub-models based on function or physical object, wherein each sub-model corresponds to a specific function or specific physical object of the local digital twin model.

[0067] (2) Extract the application target in the setting request, obtain the corresponding functional configuration model based on the application target, determine whether the input parameters of the functional configuration model can be obtained from the physical object, if not, send a prompt message, if yes, select a candidate model from multiple sub-models based on the model function, data interface and / or logical relationship.

[0068] (3) Extract any candidate model, identify the common substructure between any candidate model and the functional configuration model through topology consistency analysis, and generate multiple dynamic alignment models based on any candidate model, the functional configuration model and the common substructure.

[0069] (4) Extract any dynamic alignment model and generate candidate digital twin models based on any dynamic alignment model and other sub-models other than any candidate model.

[0070] (5) Extract the second historical data within the second preset time period, simulate the candidate digital twin model based on the second historical data, obtain the performance index parameters, compare the performance index parameters with the historical performance index parameters, and quantify the offset of different performance indicators.

[0071] (6) After traversing all candidate models, sort all candidate digital twin models based on offset and send them to the user. Generate a new local digital twin model based on user feedback.

[0072] Specifically, by extracting multiple common components (such as identical data preprocessing, sensor inputs, and logic modules) between candidate models and functional configuration models, multiple dynamic alignment models are constructed based on these common components. This involves combining "shared parts + newly added parts" in different ways. Each candidate digital twin model is evaluated for performance, resource consumption, and impact, and the evaluation results are fed back to the user, who then decides whether to adopt a particular candidate digital twin model. Preferably, only the evaluation results and the dynamic alignment model can be sent to the user. Using "common substructures" as a medium, a structurally sound and resource-efficient fusion solution is constructed, efficiently, cost-effectively, and securely integrating new functions into existing digital twin models, thus achieving rapid evolution and stable operation of the intelligent manufacturing system.

[0073] For example, the digital twin system of a smart factory includes the following sub-models: a spindle life prediction sub-model M1, which predicts the remaining life of the motor based on current, speed, and torque data; a temperature monitoring sub-model M2, which monitors the temperature of key parts of the equipment; and an energy consumption analysis sub-model M3, which analyzes the energy consumption patterns of the equipment. A user requests the addition of a "drill bit wear prediction" function, and the corresponding functional configuration model M4 is obtained based on the application objective. Through comparison, it is found that the spindle life prediction sub-model M1 and the functional configuration model M4 have the same input parameters and the same analysis model; therefore, the spindle life prediction sub-model M1 is selected as the candidate model.

[0074] by Figure 3 and Figure 4 The technical solution of this invention will be described using an example. The processing flow of the spindle life prediction sub-model M1 is S→N1→(N2, N5, N6)→N7→R2, where S is the input parameter, N1 (data preprocessing), N2 (torque calculation), N5 (shaft voltage calculation), N6 (shaft speed calculation), N7 (life prediction model) are the analysis models, and R2 is the output result. The processing flow of the functional configuration model M4 is S→N1→(N2, N3)→N4→R1, where S is the input parameter, N1 (data preprocessing), N2 (torque calculation), N3 (speed calculation), N4 (wear prediction model) are the analysis models, and R1 is the output result. Through topology consistency analysis, the common substructure between the spindle life prediction sub-model M1 and the functional configuration model M4 is identified as the input parameter S, and the analysis models N1 and N2. Figure 3 For dynamic alignment model 1, Figure 4 For Dynamic Alignment Model 2, the candidate digital twin models generated based on Dynamic Alignment Model 1 and Dynamic Alignment Model 2 are candidate 1 and candidate 2, respectively.

[0075] Using production data from the past week, compare the performance of the candidate models. Candidate 1 has a 2-second delay in tick adjustment and a 5% increase in CPU utilization; Candidate 2 has real-time tick response but a 15% increase in CPU utilization. The results are then sorted and sent to the user.

[0076] In one specific embodiment, the components of the sub-model, the functional configuration model, and the dynamic alignment model include: entity topology descriptor, functional semantic label, entity derived signal set, logical execution chain, analysis model library, and parameter constraints and definitions.

[0077] Specifically, the entity topology descriptor represents the hardware composition and connection relationships of the target device, including core component identifiers (such as drive units and execution units) and sensing unit configurations (such as current sensors and vibration sensors). Functional semantic tags identify the functional objectives of the digital model, including but not limited to lifespan prediction (such as remaining service life estimation), condition diagnosis (such as anomaly detection and wear assessment), and performance optimization (such as energy efficiency analysis). The entity-derived signal set is a collection of data directly or indirectly obtained from the target device, such as real-time sensor data streams (such as current and temperature), static parameters (such as equipment material and historical maintenance records), and virtual derived data (such as torque estimates generated by computational models). The logical execution chain comprises standardized processing steps from data input to output, including preprocessing nodes (such as data filtering and normalization), feature extraction nodes (such as frequency domain analysis and time domain statistics), and decision nodes (such as fault classification and lifespan prediction). The analysis model library consists of configurable algorithm modules within the logical execution chain, including machine learning models and physical equations. The parameter constraints and definitions specify the data interface specifications for each model in the analysis model library. For example, each module includes input parameter constraints (such as data type and value range) and output parameter definitions (such as failure probability and remaining life hours).

[0078] The above describes the dynamic scene simulation method based on digital twins in the embodiments of this application. The following describes the dynamic scene simulation system based on digital twins in the embodiments of this application. Please refer to [link / reference]. Figure 5 One embodiment of the dynamic scene simulation system based on digital twins in this application includes: a local computing node 100 and a cloud server 200. The local computing node 100 includes a local model building module 101, a simulation synchronization module 102 and a simulation feedback module 103. The cloud server 200 includes a system model building module 201, a permission setting module 202 and a policy distribution module 203.

[0079] The local model building module 101 is used to build corresponding local digital twin models based on physical objects in the real world.

[0080] The system model building module 201 is used to build a system model after the structured description information and interface data of multiple local digital twin models are uploaded to the cloud server. The system model is used to integrate the behavioral relationships and interaction logic of multiple physical objects.

[0081] The permission setting module 202 is used to support multiple users to access and share the system model, and sets the permissions of each user to the system model through a multi-level permission management mechanism.

[0082] The simulation synchronization module 102 is used to perform local simulation based on the local digital twin model and real-time data after the cloud server distributes the simulation task to each local computing node, generate the first simulation result, and synchronize the first simulation result to the cloud server in real time through the communication interface.

[0083] The strategy distribution module 203 is used to integrate and apply the received multiple first simulation results to the corresponding regional modules in the system model, perform global scene simulation, generate optimization strategies, and then distribute the optimization strategies to the local digital twin models with corresponding permissions.

[0084] The simulation feedback module 103 is used to enable the local digital twin model to perform optimization simulation according to the optimization strategy, generate a second simulation result, and feed the second simulation result back to the corresponding physical object to control its operating parameters or execution status.

[0085] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0086] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0087] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims

1. A dynamic scenario simulation method based on digital twinning, characterized in that, The method includes: On local computing nodes, corresponding local digital twin models are built based on physical objects in the real world; The structured description information and interface data of multiple local digital twin models are uploaded to a cloud server, and a system model is built on the cloud server. The system model is used to integrate the behavioral relationships and interaction logic of multiple physical objects. The cloud server supports multiple users to access and share the system model, and sets the permissions of each user to the system model through a multi-level permission management mechanism; The simulation task is assigned to each local computing node. The local computing node performs local simulation based on the local digital twin model and real-time data, generates a first simulation result, and synchronizes the first simulation result to the cloud server in real time through a communication interface. The cloud server integrates and applies the received first simulation results to the corresponding regional modules in the system model, performs global scene simulation, generates optimization strategies, and then distributes the optimization strategies to local digital twin models with corresponding permissions. The local digital twin model performs optimization simulation according to the optimization strategy, generates a second simulation result, and feeds the second simulation result back to the corresponding physical object to control its operating parameters or execution status. Upon receiving a user's model setting request, the local digital twin model is updated, including: The local digital twin model is divided into multiple sub-models based on the function or the physical object, wherein each sub-model corresponds to a specific function or a specific physical object of the local digital twin model; Extract the application target from the setting request, obtain the corresponding functional configuration model based on the application target, determine whether the input parameters of the functional configuration model can be obtained from the physical object, if not, send a prompt message, if yes, select a candidate model from multiple sub-models based on model function, data interface and / or logical relationship; Extract any candidate model, identify the common substructure between any candidate model and the functional configuration model through topology consistency analysis, and generate multiple dynamic alignment models based on any candidate model, the functional configuration model and the common substructure; Extract any dynamic alignment model, and generate a candidate digital twin model based on any dynamic alignment model and other sub-models besides any candidate model; Extract second historical data within a second preset time period, simulate the candidate digital twin model based on the second historical data, obtain performance index parameters, compare the performance index parameters with historical performance index parameters, and quantify the offset of different performance indicators. After traversing all candidate models, all candidate digital twin models are sorted based on the offset and sent to the user. A new local digital twin model is then generated based on the user's feedback.

2. The digital-twin-based dynamic scenario simulation method according to claim 1, characterized in that, The cloud server includes a participant management module, used to perform fine-grained control over access permissions to multiple local digital twin models and the system model based on preset permission policies. The permission policies include: Each local digital twin model is bound to a corresponding user account, and the user's access permissions to the local digital twin model are set according to the user's identity. The access permissions include the scope of access and operation permissions. The operation permissions include viewing, editing, deploying simulation, setting parameters, executing simulation and / or issuing control commands. Based on the simulation task, access permissions for the corresponding modules are temporarily granted to collaborating users, wherein the corresponding modules are the functional modules in the system model that correspond to the simulation task; The permission settings support cross-organizational configuration, allowing users from different organizations to obtain different levels of access permissions according to the needs of collaborative tasks, and allowing the owner of the local digital twin model to dynamically adjust or revoke the access permissions according to task requirements or collaboration status; The system authenticates each access request, confirms access permissions, and records access logs.

3. The dynamic scene simulation method based on digital twins according to claim 1, characterized in that, The cloud server manages multiple local digital twin models and dynamically updates the system model by receiving data from each local computing node, specifically including: Each local digital twin model is bound to its corresponding physical object and local computing node; Any local computing node receives first state data from the first physical object in real time and updates the first local digital twin model based on the first state data to ensure that the first local digital twin model reflects the current state and behavior of the first physical object. The first physical object and the first local digital twin model are the physical object and local digital twin model corresponding to any of the local computing nodes. The cloud server receives second status data uploaded by any of the local computing nodes in real time. When it detects a change in the status information and / or model data of any of the local computing nodes, it dynamically manages and updates the system model based on the first status data to ensure that the behavior of a specific region module in the system model is synchronized with a specific local digital twin model. The specific local digital twin model is the local digital twin model corresponding to any of the local computing nodes, and the specific region module is the region module corresponding to the specific local digital twin model.

4. The dynamic scene simulation method based on digital twins according to claim 1, characterized in that, The local digital twin model is divided into multiple sub-modules. The basic sub-modules are encapsulated as executable units, and the remaining sub-modules are encrypted as model resource files and stored in a file storage module. A data adaptation module is set up on the local computing node to store the calling parameters of the model resource files. Before executing the local simulation, the local computing node includes: When the local computing node receives a local simulation task, it extracts all corresponding file data based on the calling parameters. Extract any file data, and the authorization agent module generates an authentication request based on the identifier of any file data and the user identifier, and sends it to the authentication server; The authentication server verifies user permissions. If the permission verification is successful, the authorization proxy module decrypts any of the file data to obtain the corresponding model resource file. After traversing all the file data, the local digital twin model is generated based on the executable unit and all model resource files.

5. The dynamic scene simulation method based on digital twins according to claim 4, characterized in that, The local digital twin model is formatted after the local simulation task is completed.

6. The dynamic scene simulation method based on digital twins according to claim 1, characterized in that, Before executing the local simulation, the local computing node includes: Extract any real-time data, obtain the data credibility of any real-time data, determine whether the data credibility meets the corresponding preset standard, if not, obtain the data identifier of any real-time data, extract the first historical data within a first preset time based on the data identifier, perform statistical analysis on the first historical data, obtain statistical values, and set the statistical values ​​as any real-time data. After traversing all real-time data, new real-time data is generated.

7. The dynamic scene simulation method based on digital twins according to claim 1, characterized in that, The components of the sub-model, the functional configuration model, and the dynamic alignment model include: entity topology descriptor, functional semantic label, entity derived signal set, logical execution chain, analysis model library, and parameter constraints and definitions.

8. A dynamic scene simulation system based on digital twins, used to implement the method as described in any one of claims 1 to 7, characterized in that, The system includes: a local computing node and a cloud server. The local computing node includes a local model building module, a simulation synchronization module, and a simulation feedback module. The cloud server includes a system model building module, a permission setting module, and a policy distribution module. The local model building module is used to build corresponding local digital twin models based on physical objects in the real world; The system model building module is used to build a system model after the structured description information and interface data of multiple local digital twin models are uploaded to the cloud server. The system model is used to integrate the behavioral relationships and interaction logic of multiple physical objects. The permission setting module is used to support multiple users to access and share the system model, and sets the permissions of each user to the system model through a multi-level permission management mechanism; The simulation synchronization module is used to perform local simulation based on the local digital twin model and real-time data after the cloud server allocates the simulation task to each local computing node, generate a first simulation result, and synchronize the first simulation result to the cloud server in real time through the communication interface. The strategy distribution module is used to integrate and apply the received multiple first simulation results to the corresponding regional modules in the system model, perform global scene simulation, generate optimization strategies, and then distribute the optimization strategies to local digital twin models with corresponding permissions. The simulation feedback module is used to enable the local digital twin model to perform optimization simulation according to the optimization strategy, generate a second simulation result, and feed the second simulation result back to the corresponding physical object to control its operating parameters or execution status.

Images