A method for constructing digital twins of production equipment in an IoT production line
By constructing multi-entity pairs for digital twins and performing incremental updates, the problem of reconstruction efficiency of digital twins during dynamic changes on the production line is solved, realizing efficient digital twin reconstruction and lightweight control, and adapting to the needs of flexible manufacturing.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- WUXI RENJU INFORMATION TECH CO LTD
- Filing Date
- 2025-06-18
- Publication Date
- 2026-07-31
AI Technical Summary
In existing technologies, digital twins lack efficient reconfiguration capabilities when dealing with dynamic changes in production lines, resulting in system response delays and resource waste, making it difficult to support the needs of flexible manufacturing.
By collecting entity data from the physical space and assigning it a unique ID, and using a data platform to construct multi-entity pairs, the dynamic reconstruction of the digital twin is achieved by incrementally updating only the changed parts and triggering the rule engine locally.
It significantly reduces system response latency and computing power overhead, enables real-time adaptation of the digital twin to the production line topology, and ensures decision continuity and lightweight virtual control.
Smart Images
Figure CN120779806B_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of the Internet of Things (IoT), and more particularly to a method for constructing a digital twin of IoT production line equipment. Background Technology
[0002] With the deep integration of IoT technology into smart manufacturing, industrial internet systems use sensor networks to collect diverse data in real time, including physical equipment, environmental parameters, and personnel operations, constructing a production ecosystem where people, machines, and things are interconnected. Digital twins, as an advanced application of IoT, virtually map the operating status of physical entities and have gradually become a core tool for optimizing production processes and achieving predictive maintenance. For example, in production line management, digital twins can simulate equipment behavior, analyze energy consumption data, and provide decision-making support for remote control.
[0003] However, existing technologies suffer from a key bottleneck: digital twins lack efficient reconfiguration capabilities when dealing with dynamic changes in production lines. Current mainstream solutions rely on static modeling and full-update mechanisms—when production line structures are adjusted or equipment is added or removed, all entity data must be re-collected, topological relationships reconstructed, and the rule engine fully deployed, resulting in system lag. This rigid architecture not only creates computational redundancy (such as repeatedly processing data of unchanged entities) but also disrupts the continuity of production decisions due to the global restart of the rule engine. Especially in flexible production lines with high-frequency adjustments, the full-update mode significantly hinders the real-time performance of digital twins, making it difficult to support the demands of agile manufacturing. Summary of the Invention
[0004] To address the problem that existing digital twins lack efficient reconfiguration capabilities when dealing with dynamic changes in production lines, this application provides a method for constructing a digital twin of IoT production line equipment, comprising: collecting data from a physical space, wherein the data from the physical space includes physical entities, personnel entities, environmental entities, behavioral entities, and rule entities, and each entity contains attribute data and a unique ID code; The entity and its attributes are associated with features in the digital space through the ID encoding; The entities are stored in the data platform in the form of multi-entity pairs, and each multi-entity pair corresponds to an independent rule engine. Based on the characteristics of the multi-entity pairs, the rule engines, and the associated digital space, an initial digital twin is constructed. The topology of the physical space is updated according to changes in the factory's production lines, and new multi-entity pairs are constructed through the data platform. A new rule engine is loaded based on the new multi-entity pairs, and a new digital twin is generated based on the characteristics of the new multi-entity pairs, the new rule engine, and the associated digital space.
[0005] Specifically, the multi-entity pair includes at least one master entity and one slave entity. The master entity is a physical entity, and the slave entity is one or more of the personnel entity, the environment entity, the behavior entity, and the rule entity. The attributes of the slave entity include its dependency identifier on the master entity. The digital twin construction method also includes: When a production line change causes the physical entity to be updated, the data platform associates the dependent entity with the dependency identifier, generates new multi-entity pairs only for entities with changed dependency identifiers, and triggers incremental updates of the corresponding rule engine.
[0006] Specifically, the ID encoding includes physical layer ID, logical layer ID, and digital layer ID; The digital twin construction method also includes: When a new multi-entity pair is generated, the data platform automatically attaches the new entity pair to a preset digital space sub-module based on the mapping relationship between logical layer ID and digital layer ID, and activates the rule engine container of that sub-module.
[0007] Specifically, the digital twin construction method further includes: The dependency identifier is dynamically generated through real-time topology connection relationships. When there is direct physical interaction or data flow interaction between the master entity and the slave entity, the dependency identifier is automatically established. The data platform constructs an entity dependency network graph based on the dependency identifier, and generates new multi-entity pairs only for entities within connected subgraphs in the network graph.
[0008] Specifically, the rule engine's container is equipped with a cross-container communication interface; The digital twin construction method also includes: When multiple rule engines need to perform collaborative calculations, they exchange intermediate calculation results and perform consistency verification through the interface. If the verification fails, the attribute data of the associated entity pair will be re-collected.
[0009] Specifically, the logical layer ID is the functional role code of the equipment in the production line, including: production equipment, quality inspection equipment, logistics equipment, and energy equipment; The digital space submodule is pre-built according to functional roles, and newly added entity pairs are automatically attached to the corresponding submodule based on the functional role code.
[0010] Specifically, the digital twin construction method further includes: When the key attributes of the main entity continue to be abnormal, the main entity and its subordinate entities are automatically cloned through the data platform to generate a maintenance entity pair; The maintenance entity is loaded with a historical fault rule base, and maintenance decisions are output to the digital twin, with abnormal areas marked.
[0011] Specifically, the digital layer ID is a dynamic memory pointer. When the digital space submodule is mounted, a shared memory area is automatically allocated and a pointer is generated. Multiple rule engine containers can directly read and write data to the same memory area through this pointer.
[0012] Specifically, when the physical entity is marked as faulty, its logical layer ID functional role code is temporarily changed to an isolation device; The data platform will migrate the associated entity pairs to the isolated digital submodule and load the fault analysis rule engine container.
[0013] Specifically, the digital twin construction method further includes: When the environmental entity data undergoes a sudden change, an event cascading response is initiated, which includes: Locate the list of affected main entities; Freeze the state of the association rule engine container; Restart the rule engine container group in batches according to the device function role code; And record container state snapshots for anomaly analysis. This application has the following technical effects: This invention effectively solves the bottleneck of reconstruction efficiency caused by dynamic changes in production lines for digital twins. Through dynamic decoupling mapping between physical entities and digital features, it triggers local incremental updates only for the changed parts, significantly reducing system response latency and computing power overhead. In the high-frequency adjustment scenario of flexible production lines, this invention achieves for the first time precise linkage between the rule engine and entity changes, enabling the digital twin to adapt to topology adjustments in real time. This ensures decision continuity while avoiding global resource reset, providing lightweight and adaptive virtual control support for agile manufacturing. Attached Figure Description
[0014] The above and other objects, features, and advantages of exemplary embodiments of this application will become readily understood by reading the following detailed description with reference to the accompanying drawings. Several embodiments of this application are illustrated in the drawings by way of example and not limitation, and the same or corresponding reference numerals denote the same or corresponding parts.
[0015] Figure 1 This is a flowchart illustrating the steps of a method for constructing a digital twin of an IoT production line equipment in an embodiment of this application. Detailed Implementation
[0016] The technical solutions in the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings.
[0017] like Figure 1 As shown in this embodiment, a method for constructing a digital twin of production equipment in an IoT production line includes: collecting data from the physical space, whereby the data includes physical entities, personnel entities, environmental entities, behavioral entities, and rule entities, and each entity contains attribute data and a unique ID code; Entities and their attributes are associated with features in the digital space through ID encoding; Entities are stored in the data platform in the form of multi-entity pairs, and each multi-entity pair corresponds to an independent rule engine. Based on the characteristics of the multi-entity pairs, rule engines, and associated digital spaces, an initial digital twin is constructed. Update the physical space topology based on changes in the factory's production lines, and build new multi-entity pairs through the data platform; A new rule engine is loaded based on the new multi-entity pairs, and a new digital twin is generated based on the characteristics of the new multi-entity pairs, the new rule engine, and the associated digital space.
[0018] In this embodiment, physical entity data is first collected, covering physical entities (machine tools, robotic arms), personnel entities (operator identities and behaviors), environmental entities (temperature, humidity, lighting), behavioral entities (equipment start / stop, material handling), and rule entities (process standards, safety thresholds). Each entity contains attribute data and a unique three-layer ID code (physical layer MAC address, logical layer functional role code, and digital layer dynamic pointer). The ID code dynamically associates entity attributes with digital space characteristics. For example, the physical layer ID is bound to the equipment location coordinates, the logical layer ID is mapped to the production line functional areas, such as production / quality inspection / logistics areas, and the digital layer ID points to the data model updated in real time in memory.
[0019] In this embodiment, entities are stored as multi-entity pairs within a data platform built on Apache Kafka. Each multi-entity pair consists of a master entity, such as a CNC machine tool, and at least one slave entity, such as an associated temperature sensor or operator. The master entity is a physical entity, and the slave entities are one or more of the following: personnel entities, environment entities, behavior entities, and rule entities. The slave entity attributes include a dependency identifier on the master entity, such as "temperature monitoring belonging to machine tool A". Each entity pair independently deploys a containerized rule engine using Docker technology, driving the initial digital twin construction based on entity attributes and preset logic, such as "trigger shutdown if temperature exceeds limit".
[0020] When production line changes lead to updates to physical entities, the data platform associates entities based on dependency identifiers, generating new multi-entity pairs only for entities with changed dependency identifiers and triggering incremental updates to the corresponding rule engine. In this embodiment, when adding quality inspection equipment, the system updates the physical space topology using a laser scanner, and the data platform only generates new multi-entity pairs for the changed areas. For example, the newly added visual inspection equipment acts as the main entity, while its accompanying light source and image analysis algorithm act as secondary entities, with dependency identifiers automatically generated through real-time topology connections. A lightweight rule engine container is loaded based on the new multi-entity pairs, and combined with pre-associated digital spatial features (such as a 3D model of the quality inspection area), a locally updated digital twin is generated. The key technical detail lies in the incremental update mechanism: when physical entities are updated, the data platform associates entities based on dependency identifiers, generating new entity pairs only for entities with changed dependency identifiers. For example, when replacing machine tool cutting tools, only the cutting parameter rule engine associated with that tool is reconstructed, not the entire production line.
[0021] In this embodiment, to improve efficiency, dependency identifiers are dynamically generated based on real-time topology connections. When there is direct physical interaction between master and slave entities, such as a robotic arm grasping a workpiece, or data flow interaction, such as a sensor transmitting data to a PLC, the system automatically establishes dependencies and constructs an entity dependency network graph. The data platform constructs the entity dependency network graph based on the dependency identifiers, generating new multi-entity pairs only for entities within connected subgraphs of the network graph. That is, updates only process entities within connected subgraphs of the network graph; for example, adjustments to the logistics AGV path only affect the associated shelves and handling rule engine.
[0022] In terms of ID encoding design, ID encoding includes physical layer ID, logical layer ID, and digital layer ID. The logical layer ID is defined as the device's functional role code, such as "FAC-001" representing a stamping device. The digital space pre-constructs sub-modules according to function, such as a virtual model of the stamping area. When a new multi-entity pair is generated, the data platform automatically mounts the new entity pair to a preset digital space sub-module based on the mapping relationship between logical layer IDs and digital layer IDs, and activates the rule engine container of that sub-module. New entity pairs are automatically mounted to the corresponding sub-module according to their functional codes, and the rule engine container of that module is activated. The digital layer ID uses dynamic memory pointers, automatically allocating a shared memory area, such as a Redis cache, during mounting. Multiple rule engine containers directly read and write to the same memory area through pointers to exchange intermediate calculation results. The logical layer ID is the functional role code of the device in the production line, including: production equipment, quality inspection equipment, logistics equipment, and energy equipment. Digital space sub-modules are pre-constructed according to functional roles, and new entity pairs are automatically mounted to the corresponding sub-module according to their functional role codes. The digital layer ID is a dynamic memory pointer. When the digital space submodule is mounted, a shared memory area is automatically allocated and a pointer is generated. Multiple rule engine containers exchange data by directly reading and writing to the same memory area through this pointer.
[0023] In this embodiment, the rule engine container sets up a gRPC cross-container communication interface. When multiple engines need to perform collaborative calculations, such as energy efficiency optimization of an entire production line, they exchange data through the interface and perform consistency verification, such as hash value comparison. If the verification fails, such as due to energy efficiency calculation conflicts, a re-collection process for associated entity pairs is triggered to ensure data reliability.
[0024] In response to fault scenarios, when a key attribute of the main entity, such as the spindle vibration value, remains abnormal, the data platform automatically clones the main entity and its subordinate entities to generate a maintenance entity pair. It loads a historical fault rule library containing fault cases from the past three years and outputs maintenance decisions, such as "Bearing wear probability 92%, replacement recommended," to the digital twin, marking abnormal areas in the 3D model. If a physical entity is marked as faulty, such as a temperature sensor failure, its logical layer ID function role code is temporarily changed to an isolated device. The associated entity pair is migrated to an isolated digital submodule and loaded with a fault analysis rule engine, such as a signal compensation algorithm, to avoid polluting the main system's data flow.
[0025] In terms of environmental change response, when environmental entity data changes abruptly, such as a sudden drop in air pressure, the event cascading response process is initiated: first, the list of affected main entities, such as pneumatic equipment, is located; the status of the association rule engine container is frozen; a snapshot of the current parameters is saved; the container group is restarted in batches according to the equipment function role code; key production equipment is restarted first; and the container status snapshot is recorded for anomaly backtracking analysis.
[0026] Obviously, the embodiments described above are only some, not all, of the embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.
[0027] It should be understood that when the terms "first," "second," etc., are used in the claims, description, and drawings of this application, they are only used to distinguish different objects and not to describe a specific order. The terms "comprising" and "including" used in the description and claims of this application indicate the presence of the described features, integrals, steps, operations, elements, and / or components, but do not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components, and / or collections thereof.
Claims
1. A method for constructing a digital twin of production equipment in an Internet of Things (IoT) production line, characterized in that, include: Collect data from the physical space, which includes physical entities, personnel entities, environmental entities, behavioral entities, and rule entities. Each entity contains attribute data and a unique ID code. The entity and its attributes are associated with the features of the digital space through the ID encoding. The ID encoding is a three-layer ID encoding, which includes a physical layer MAC address, a logical layer functional role encoding, and a digital layer dynamic pointer. The physical layer MAC address is bound to the device location coordinates, the logical layer functional role encoding is mapped to the production line functional partition, the production line functional partition includes the production area, the quality inspection area, and the logistics area, and the digital layer dynamic pointer points to the data model that is updated in real time in memory. The entities are stored in the data platform in the form of multi-entity pairs, and each multi-entity pair corresponds to an independently deployed containerized rule engine. Based on the characteristics of the multi-entity pairs, the containerized rule engine and the associated digital space, an initial digital twin is constructed. The topology of the physical space is updated according to changes in the factory's production lines, and new multi-entity pairs are constructed through the data platform. A new rule engine is loaded based on the new multi-entity pair, and a new digital twin is generated based on the characteristics of the new multi-entity pair, the new rule engine, and the associated digital space. The multi-entity pair includes at least one master entity and one slave entity. The master entity is a physical entity, and the slave entity is one or more of the personnel entity, the environment entity, the behavior entity, and the rule entity. The attributes of the slave entity include its dependency identifier on the master entity. The digital twin construction method further includes: when a production line change causes the physical entity to be updated, the data platform associates the dependent entity based on the dependency identifier, generates new multi-entity pairs only for entities with changed dependency identifiers, and triggers incremental updates of the corresponding rule engine; the digital twin construction method further includes: The dependency identifier is dynamically generated through real-time topology connections. When there is direct physical interaction or data flow interaction between the master entity and the slave entity, the dependency identifier is automatically established. The data platform constructs an entity dependency network graph based on the dependency identifier, and generates new multi-entity pairs only for entities within connected subgraphs in the network graph. When the production line is changed to add quality inspection equipment, the topology of the physical space is updated by using a laser scanner. The newly added visual inspection equipment is taken as the main entity, and the matching light source and image analysis algorithm of the visual inspection equipment are taken as the secondary entities. Dependency identifiers are automatically generated based on the direct physical interaction or data flow interaction between the visual inspection equipment and its secondary entities. A lightweight rule engine container is loaded, and a locally updated digital twin is generated by combining the pre-associated 3D model of the quality inspection area. When the production line is changed to replace machine tool tools, only the cutting parameter rule engine associated with the machine tool tools is reconstructed, without a full reconstruction of the entire production line.
2. The method according to claim 1, characterized in that, The digital twin construction method also includes: When a new multi-entity pair is generated, the data platform automatically attaches the new entity pair to a preset digital space sub-module based on the mapping relationship between the logical layer functional role code and the digital layer dynamic pointer, and activates the rule engine container of that sub-module.
3. The method according to claim 1, characterized in that, The digital twin construction method also includes: When the key attributes of the main entity continue to be abnormal, the main entity and its subordinate entities are automatically cloned through the data platform to generate a maintenance entity pair; The maintenance entity is loaded with a historical fault rule base, and maintenance decisions are output to the digital twin, with abnormal areas marked.
4. The method according to claim 2, characterized in that, The logical layer functional role codes include: production equipment, quality inspection equipment, logistics equipment, and energy equipment; The digital space submodule is pre-built according to functional roles, and newly added entity pairs are automatically attached to the corresponding submodule based on the functional role code.
5. The method according to claim 2, characterized in that: The digital layer dynamic pointer is automatically allocated to a shared memory area and a pointer is generated when the digital space submodule is mounted; Multiple rule engine containers can directly read and write data to the same memory area through this pointer.
6. The method according to claim 4, characterized in that: When the physical entity is marked as faulty, its logical layer functional role code is temporarily changed to that of an isolation device. The data platform will migrate the associated entity pairs to the isolated digital submodule and load the fault analysis rule engine container.
7. The method according to claim 1, characterized in that, The rule engine's container is equipped with a cross-container communication interface; the digital twin construction method further includes: When multiple rule engines need to perform collaborative calculations, they exchange intermediate calculation results and perform consistency verification through the interface. If the verification fails, the attribute data of the associated entity pair will be re-collected.
8. The method according to claim 1, characterized in that, The digital twin construction method also includes: When the environmental entity data undergoes a sudden change, an event cascading response is initiated, which includes: Locate the list of affected main entities; Freeze the state of the association rule engine container; Restart the rule engine container group in batches according to the device function role code; It also records container state snapshots for anomaly analysis.