An industrial production line topology adaptive reconfiguration method, device, equipment and storage medium

Abstract

This invention relates to the field of industrial production technology, and particularly to a method, apparatus, equipment, and storage medium for adaptive reconfiguration of industrial production line topology. The method includes: initiating disturbance identification on the industrial production line; if a preset production line disturbance is identified, determining the disturbance information; scheduling edge computing nodes to remove the disturbance based on the disturbance information, and searching for feasible alternative paths using pre-stored process correlation data to generate a temporary reconfiguration strategy; controlling cloud supercomputing nodes to receive the temporary reconfiguration strategy, and calling a full-element digital twin model to perform global simulation verification of the temporary reconfiguration strategy; if the verification fails, initiating a preset optimization engine to adjust and optimize the temporary reconfiguration strategy, generating and outputting a final reconfiguration strategy that has passed verification; and distributing the final reconfiguration strategy to the production line execution control system. This method, through the two-level collaboration of edge lightweighting and cloud supercomputing, can balance real-time response, global decision-making, and operational security.

Description

Technical Field

[0001] The embodiments of the present invention relate to the field of industrial production technology, and in particular to an industrial production line topology adaptive reconfiguration method, apparatus, equipment and storage medium. Background Technology

[0002] Modern manufacturing is evolving towards higher flexibility, agility, and intelligence. Industrial production lines need to frequently cope with dynamic disturbances such as sudden equipment failures and the introduction of new processes, placing stringent demands on the real-time reconstruction and global optimization of production line topology. Against this backdrop, the computing architecture and topology modeling technology of industrial intelligent systems have become core supports, and their development and evolution directly determine the production line's ability to cope with dynamic scenarios.

[0003] In the field of computing architecture, traditional industrial intelligent systems generally adopt a centralized cloud computing model, where all production data must be uploaded to the cloud for centralized processing and decision generation. While this model possesses strong global computing capabilities, it suffers from inherent drawbacks such as high response latency (typically greater than 1 second), high network bandwidth consumption, and significant single-point failure risks, making it difficult to adapt to scenarios with extremely high real-time requirements, such as emergency handling of equipment failures. To compensate for the shortcomings of centralized cloud computing, industrial edge computing technology has emerged. By offloading some computing tasks to edge nodes closer to the equipment terminals, it achieves millisecond-level response and local closed-loop control, effectively solving the problems of real-time performance and bandwidth pressure. However, the pure edge computing model is limited by the scale of node computing power and data coverage, lacks a global optimization perspective, and cannot generate reorganization plans covering the entire production line. It still relies on manual rescheduling and cannot support global resource scheduling and optimal decision-making for complex production lines.

[0004] Therefore, how to overcome the limitations of traditional topology modeling, which relies on manual labor and has slow updates, and achieve automated adaptation and safe and controllable reconstruction of production line topology is an urgent problem to be solved by those skilled in the art. Summary of the Invention

[0005] The purpose of this invention is to provide at least one evaluation method, apparatus, equipment, and storage medium for industrial production lines, which can realize second-level intelligent regeneration and hour-level model reconstruction of production lines, taking into account real-time response, global decision-making, and operational safety, and providing reliable technical support for highly dynamic manufacturing environments.

[0006] To address the aforementioned technical problems, at least one embodiment of this application provides an industrial production line topology adaptive reconfiguration method, comprising:

[0007] Initiate disturbance identification for industrial production lines; if a preset disturbance is detected, determine the disturbance information. The edge computing nodes are scheduled to remove disturbances based on the disturbance information, and feasible alternative paths are searched through pre-stored process association data to generate a temporary reconfiguration strategy. The cloud supercomputing node receives the temporary reconstruction strategy and calls the full-element digital twin model to perform global simulation verification of the temporary reconstruction strategy. If the verification fails, the preset optimization engine is started to adjust and optimize the temporary reconstruction strategy, and the final reconstruction strategy that passes the verification is generated and output. The final reconfiguration strategy is sent to the production line execution control system to execute the final reconfiguration strategy.

[0008] In one embodiment, the scheduling edge computing node performs disturbance removal based on the disturbance information, including: The edge computing node is controlled to call the current production line topology model to match associated nodes based on the disturbance information; the production line topology model is constructed with production line equipment as nodes and process dependencies as edges; the disturbance information includes: disturbance type and associated impact data; Perform a topology adjustment operation on the associated node that is adapted to the disturbance type.

[0009] In one embodiment, performing a topology adjustment operation adapted to the disturbance type on the associated node includes: If the disturbance type is a sudden equipment failure, then the associated node is the failed equipment node corresponding to the failure, and a removal operation is performed on the failed equipment node. If the disturbance type is a process iteration change, then the associated node is the target node involved in the process change. An attribute update operation is performed on the target node, or an add / delete operation is performed on the associated edges of the target node.

[0010] In one embodiment, the industrial production line topology adaptive reconstruction method further includes: Real-time monitoring of equipment additions / reductions and process adjustments in the industrial production line, generating monitoring data that includes equipment node additions / reductions and process dependency adjustments; The monitoring data is input into a graph neural network, and the graph neural network is invoked to identify the changes in the dependency relationships between equipment nodes in the production line topology model based on the monitoring data. Based on the changes in the dependencies between the device nodes, the node information and edge relationships of the production line topology model are deduced and updated.

[0011] In one embodiment, the step of invoking a full-element digital twin model to perform global simulation verification of the temporary reconstruction strategy includes: The simulated production line operation status is verified based on preset verification indicators; the preset verification indicators include: the overall efficiency of the equipment meets the standards, the energy consumption cost control status, and the safety constraint compliance status. If the temporary reconstruction strategy meets all the preset verification indicators, it is determined that the verification is successful; If any of the preset verification indicators are not met, the verification is deemed to have failed.

[0012] In one embodiment, the identification of disturbances to the industrial production line includes: Real-time acquisition of multi-source heterogeneous data from industrial production lines; the multi-source heterogeneous data includes equipment layer data, control layer data, resource layer data, and business layer data; the equipment layer data includes: equipment health data and equipment operating status data; the control layer data includes: production line topology data and process logic relationship data; the resource layer data includes: work-in-process inventory data and energy supply capacity data; the business layer data includes: order urgency data and delivery window data. The system calls a preset rule engine and a lightweight machine learning model to process the multi-source heterogeneous data and identify whether there are preset production line disturbances.

[0013] In one embodiment, after the final reconfiguration strategy is issued to the production line execution control system, the method further includes: Continuously collect preset indicators of production line operation and compare them with preset thresholds; When the preset indicator deviates from the corresponding preset threshold, the historically stored stable strategy version is invoked, and a rollback command is sent to the production line execution control system to perform a strategy rollback operation.

[0014] At least one embodiment of this application also provides an industrial production line topology adaptive reconfiguration device, comprising: The disturbance identification unit is used to initiate disturbance identification on the industrial production line. If a preset disturbance is detected, the disturbance information is determined. The edge response unit is used to schedule edge computing nodes, remove disturbances based on the disturbance information, search for feasible alternative paths through pre-stored process association data, and generate a temporary reconfiguration strategy. The remote adjustment unit is used to control the cloud supercomputing node to receive the temporary reconstruction strategy and call the full-element digital twin model to perform global simulation verification of the temporary reconstruction strategy. If the verification fails, the preset optimization engine is started to adjust and optimize the temporary reconstruction strategy, and generate and output the final reconstruction strategy that passes the verification. The strategy execution unit is used to send the final reconfiguration strategy to the production line execution control system to execute the final reconfiguration strategy.

[0015] At least one embodiment of this application also provides an electronic device, including: at least one processor; and a memory communicatively connected to said at least one processor; The memory stores instructions that can be executed by the at least one processor, which are then executed by the at least one processor to enable the at least one processor to perform the industrial production line topology adaptive reconfiguration method.

[0016] At least one embodiment of this application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the industrial production line topology adaptive reconfiguration method.

[0017] The industrial production line topology adaptive reconfiguration method provided in this application initiates a continuous monitoring mechanism, relies on preset rules and a lightweight analysis model to determine disturbances, and forms structured disturbance information. This allows for real-time perception of disturbance events without human intervention, avoiding human error and response delays. Then, disturbance removal is performed locally on edge computing nodes deployed on the production line, significantly shortening the removal time and achieving rapid elimination of disturbance impacts. Simultaneously, feasible alternative paths are automatically searched based on pre-stored process correlation data, ensuring that the paths adapt to the production line equipment capabilities and process logic. The generated temporary strategies have direct emergency execution capabilities, guaranteeing the continuity of core production line processes and providing a foundational solution for efficient cloud-based global optimization. Furthermore, it leverages cloud supercomputing... The powerful computing capabilities of the nodes and the high-fidelity characteristics of the full-element digital twin model enable full-scenario, multi-dimensional global verification of temporary strategies. This effectively identifies potential problems such as resource conflicts and indicator imbalances that may exist in local emergency strategies. If the verification fails, the unqualified strategies are targeted and iteratively adjusted through a preset optimization engine to ensure that the final strategy takes into account multiple objectives such as overall equipment efficiency, energy consumption cost, and safety compliance. Finally, the verified final strategy is directly sent to the production line execution control system, avoiding instruction distortion and transmission delay in intermediate links. The execution control system and production line equipment are directly linked, which can accurately parse and execute instructions such as process adjustment and equipment allocation in the strategy, ensuring that the operation status of the reconstructed production line is consistent with the strategy expectation, and guaranteeing the stability and accuracy of the reconstruction implementation.

[0018] This method innovatively balances the timeliness of disturbance response with the globality of strategy optimization, resolving the contradiction that traditional single computing power architectures suffer from short-sighted decision-making due to fast response and high latency due to comprehensive optimization. By combining topology adaptive algorithms and digital twin models, it achieves a leapfrog improvement in production line reconfiguration from manual dependence and long cycles to automatic adaptation and short timeliness. Attached Figure Description

[0019] One or more embodiments are illustrated by way of example with reference to the accompanying drawings, and these illustrative descriptions do not constitute a limitation on the embodiments.

[0020] Figure 1This is a flowchart of an industrial production line topology adaptive reconfiguration method provided in one embodiment of this application; Figure 2 This is a schematic diagram of an industrial production line topology adaptive reconfiguration device provided in one embodiment of this application; Figure 3 This is a schematic diagram of the structure of an electronic device provided in one embodiment of this application. Detailed Implementation

[0021] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the various embodiments of this application will be described in detail below with reference to the accompanying drawings. However, those skilled in the art will understand that many technical details have been provided in the various embodiments of this application to help readers better understand this application. However, the technical solutions claimed in this application can be implemented even without these technical details and various changes and modifications based on the following embodiments. The division of the various embodiments below is for the convenience of description and should not constitute any limitation on the specific implementation of this application. The various embodiments can be combined with and referenced by each other without contradiction.

[0022] This invention proposes an adaptive topology reconstruction method for industrial production lines. The implementation details of the adaptive topology reconstruction method for industrial production lines in this embodiment are described below. The following content is only for the convenience of understanding and is not necessary for implementing this solution.

[0023] Example 1: The specific process of the industrial production line topology adaptive reconstruction method in this embodiment can be described as follows: Figure 1 As shown, it includes: Step 101: Start disturbance identification for the industrial production line. If a preset production line disturbance is identified, determine the disturbance information.

[0024] By connecting to relevant data acquisition channels on the production line through a pre-defined identification mechanism, the system compares and matches real-time production line operation information with pre-defined disturbance types and judgment criteria. It proactively monitors for any abnormal events or changes that meet pre-defined definitions during production line operation. If such disturbances are detected, the system accurately collects and clarifies key information related to the disturbance, forming complete disturbance information. This disturbance information specifically includes two core aspects: first, the disturbance type (clarifying whether it is a sudden equipment failure or a process iteration change); and second, related impact data (such as the faulty equipment identifier, specific parameter requirements of the process change, etc.), ensuring the completeness and usability of the disturbance information and providing a preliminary decision-making basis for subsequent production line topology reconfiguration.

[0025] Among them, disturbance refers to various dynamic events that affect the normal operation of the production line or require adjustment of the production line topology. Preset production line disturbance refers to a specific event type that is predefined and needs to trigger the topology reconstruction process. In this embodiment, the preset production line disturbance type is not limited. For example, it can be core equipment shutdown, performance abnormality, new product introduction, process adjustment, etc.

[0026] This method adopts an active and continuous identification mechanism to replace the traditional passive manual monitoring. It can complete the identification and information collection in the early stage of the disturbance event, which greatly shortens the time difference from the occurrence of the disturbance to the start of reconstruction, and provides a pre-emptive guarantee for second-level intelligent rebirth.

[0027] Step 102: Schedule edge computing nodes, remove disturbances based on disturbance information, and search for feasible alternative paths through pre-stored process association data to generate a temporary reconfiguration strategy.

[0028] Based on the disturbance identification results, a start command is automatically sent to the edge computing nodes deployed locally on the production line. The edge computing nodes are deployed locally on the production line, and the topology adjustment and strategy generation functions of the nodes are activated to ensure that computing tasks are executed locally. The scheduling and computing process does not rely on long-distance data transmission, avoiding data transmission delays and significantly shortening the time for topology adjustment and strategy generation, which meets the rapid response requirements in disturbance scenarios.

[0029] Edge computing nodes receive and parse disturbance information, determine the topology impact range corresponding to the disturbance, and make targeted adjustments to the current production line topology to eliminate the disruption to topology integrity and availability caused by the disturbance, ensuring that the remaining topology has basic operational capabilities.

[0030] Edge computing nodes call pre-stored process-related data and, combined with the remaining topology after disturbance removal, automatically retrieve alternative paths that conform to process rules, equipment capabilities, and material flow logic. These paths must be able to bypass the disturbance-affected area, ensure the continuous execution of core processes, and meet the basic operating requirements of the production line.

[0031] After determining a feasible alternative path, the edge computing nodes integrate key information such as process adjustment sequence, equipment allocation scheme, and material scheduling logic based on the path to form a structured temporary reconfiguration strategy. This ensures that the strategy has the basic conditions for direct execution and that the generation process does not rely on manual intervention.

[0032] Among them, disturbance removal refers to the targeted adjustment of the production line topology according to the disturbance type to eliminate topology conflicts caused by faults or changes; process association data refers to the pre-stored structured data related to the production line equipment capabilities and process logic, and the specific data type is not limited in this embodiment.

[0033] Step 103: Control the cloud supercomputing node to receive the temporary reconstruction strategy and call the full-element digital twin model to perform global simulation verification of the temporary reconstruction strategy. If the verification fails, start the preset optimization engine to adjust and optimize the temporary reconstruction strategy, generate and output the final reconstruction strategy that has passed the verification.

[0034] The cloud-based supercomputing node receives temporary reconfiguration strategies uploaded from the edge terminal via a pre-defined communication channel. These edge-generated strategies prioritize rapid emergency response and basic operational assurance, potentially lacking a global perspective and exhibiting issues such as resource conflicts and indicator imbalances. To avoid this, the cloud-based supercomputing node is activated and a pre-built full-element digital twin model is invoked. This model fully maps the core elements of the production line, including equipment configuration, process logic, and constraint rules, simulating the actual operating state of the production line, such as projecting the production line status for the next 1-2 hours. The temporary reconfiguration strategy is embedded into the full-element digital twin model, initiating a simulation process to simulate the strategy's operation on the actual production line, covering all scenarios including equipment collaboration, material flow, and energy consumption. Simultaneously, based on pre-defined verification standards, key operational indicators during the simulation are checked in real time to determine whether the temporary strategy meets global operational requirements. Through global simulation verification, deficiencies in the temporary strategy can be accurately identified, preventing global inefficiency or security risks caused by local optima.

[0035] If the global simulation verification results show that the temporary strategy meets all preset standards, it is directly determined as the final reconstruction strategy. If the verification fails, that is, if the temporary strategy has problems such as unmet indicators or operational conflicts, the preset optimization engine is immediately started. The process configuration, equipment allocation, scheduling logic, etc. of the temporary strategy are iteratively adjusted for the unmet links. After adjustment, it is re-submitted into the full-element digital twin model for simulation verification until the strategy passes all verifications and finally outputs the final reconstruction strategy that meets the requirements.

[0036] This step relies on the powerful computing power of cloud supercomputing and the high-fidelity simulation capabilities of full-element digital twin models to optimize strategies from the perspective of the entire production line, ensuring that the final solution achieves optimal adaptation under multiple objectives.

[0037] Step 104: The final reconfiguration strategy is sent to the production line execution control system to execute the final reconfiguration strategy.

[0038] The final reconstruction strategy, verified by cloud supercomputing nodes, is structured and encapsulated, clearly defining key information such as instruction types, execution priorities, and equipment association identifiers. This strategy is then precisely distributed to the production line's PLCs, DCSs, and other execution control systems. These systems, following the process sequences, equipment allocations, and scheduling logic outlined in the strategy, drive the production line to complete the actual topology reconstruction and operational adjustments, ultimately achieving effective adaptation to disturbances. The production line execution control system refers to the hardware control unit that directly drives the production line equipment. The strategy is directly distributed to this system, eliminating the need for manual intervention or secondary parsing, significantly shortening the strategy implementation time and meeting the needs of rapid production line reconstruction.

[0039] Based on the above introduction, the industrial production line topology adaptive reconfiguration method provided in this embodiment, by activating a continuous monitoring mechanism, relies on preset rules and a lightweight analysis model to perform disturbance judgment, forming structured disturbance information, and can achieve real-time perception of disturbance events without human intervention, avoiding human misjudgment and response lag; then, relying on the edge computing nodes deployed locally on the production line to perform disturbance removal processing nearby, significantly shortening the disturbance removal time and achieving rapid elimination of disturbance impact, while automatically searching for feasible alternative paths based on pre-stored process association data, ensuring that the path adapts to the production line equipment capabilities and process logic, and the generated temporary strategy has direct emergency execution capability, which not only ensures the continuity of the core processes of the production line, but also provides a basic solution for efficient connection to cloud-based global optimization; further leveraging the cloud The powerful computing capabilities of supercomputing nodes and the high-fidelity characteristics of the full-element digital twin model enable full-scenario, multi-dimensional global verification of temporary strategies. This effectively identifies potential problems such as resource conflicts and indicator imbalances in local emergency strategies. If the verification fails, the strategy that fails to meet the standards is targeted and iteratively adjusted through a preset optimization engine to ensure that the final strategy takes into account multiple objectives such as overall equipment efficiency, energy consumption cost, and safety compliance. Finally, the verified final strategy is directly sent to the production line execution control system, avoiding instruction distortion and transmission delay in intermediate links. The execution control system and production line equipment are directly linked, which can accurately parse and execute instructions such as process adjustment and equipment allocation in the strategy, ensuring that the operation status of the reconstructed production line is consistent with the strategy expectation, and guaranteeing the stability and accuracy of the reconstruction implementation.

[0040] This method innovatively balances the timeliness of disturbance response with the globality of strategy optimization, resolving the contradiction that traditional single computing power architectures suffer from short-sighted decision-making due to fast response and high latency due to comprehensive optimization. By combining topology adaptive algorithms and digital twin models, it achieves a leapfrog improvement in production line reconfiguration from manual dependence and long cycles to automatic adaptation and short timeliness.

[0041] Example 2: In the adaptive reconfiguration of industrial production line topology, disturbance removal is a core element in ensuring the timeliness and accuracy of the reconfiguration response. The specific methods for implementing disturbance removal are not limited in the above embodiments. Traditional production line topology adjustments often rely on manual experience and lack standardized model support and clear node operation logic. This leads to inaccurate positioning of the disturbance impact range and mismatch between topology adjustment actions and disturbance types, which can easily cause problems such as obstructed alternative path search and insufficient feasibility of temporary reconfiguration strategies.

[0042] To improve the accuracy of disturbance removal and the adaptability of topology adjustment, and to ensure the smoothness of subsequent alternative path search and the reliability of temporary reconstruction strategies, this embodiment proposes a disturbance removal implementation method. Through standardized topology model construction and targeted node adjustment operations, the disturbance removal process has quantifiable and reproducible execution basis, avoiding low reconstruction efficiency or strategy failure caused by ambiguous operations.

[0043] Specifically, in step 102, scheduling edge computing nodes and removing disturbances based on disturbance information can be performed according to the following steps: Step 21: Control the edge computing node to call the current production line topology model to match associated nodes based on disturbance information.

[0044] The edge computing nodes respond to the system's backend scheduling instructions by first calling the pre-stored current production line topology model. This model maps the actual operating logic of the production line in a structured way, using equipment as core nodes and dependencies between processes as connecting edges, making the relationships between production line equipment and processes visible and computable. Subsequently, the edge computing nodes parse the identified disturbance information, extracting the disturbance type (such as sudden equipment failure or process iteration changes) and related impact data (such as faulty equipment identification and process change parameters). This information is then compared and matched with the node and edge attributes in the topology model, ultimately identifying the related nodes directly affected by the disturbance.

[0045] Step 22: Perform topology adjustment operations on the associated nodes to match the disturbance type.

[0046] The edge computing nodes perform adaptation operations based on the associated nodes locked in step 21 and the specific types in the disturbance information. Since the above embodiments do not limit the specific disturbance types, this embodiment also does not limit the topology adjustment operations for various disturbance types, and can be set according to the actual use scenario.

[0047] To enhance understanding, this embodiment proposes a specific form of topology adjustment operation. If the disturbance type is a sudden equipment failure, the associated node is the failed equipment node corresponding to the failure, and the topology adjustment operation specifically involves removing the failed equipment node. If the disturbance type is a process iteration change, the associated node is the target node involved in the process change, and the topology adjustment operation specifically involves updating the attributes of the target node or adding / deleting the associated edges of the target node.

[0048] If the disturbance is a sudden equipment failure, the associated node is the model node corresponding to the failed equipment. In this case, a node removal operation is performed to remove the failed equipment from the topology model, severing its process dependencies with other nodes and preventing the failed equipment from interfering with subsequent path planning. If the disturbance is a process iteration change, the associated node is the target equipment node involved in the process change. In this case, according to the change requirements, either the attribute parameters of the target node are updated (such as equipment operating threshold, adaptable process type), or the associated edges of the target node are added or deleted (such as adding process dependencies, deleting redundant connections) to make the topology model adapt to the operating logic of the new process.

[0049] It should be noted that this embodiment only uses the above disturbance types and corresponding adjustment operations as examples, but is not limited to them. For example, when the disturbance type is an emergency order insertion, the associated node is the target process node corresponding to the order, and the node priority update operation or the addition of a process branch associated edge is performed; when the disturbance type is energy supply fluctuation, the associated node is a high-energy-consuming equipment node, and the node operating parameter dynamic adaptation operation or the temporary start-stop adjustment of non-core nodes is performed; when the disturbance type is abnormal work-in-process inventory, the associated node is a material flow-related equipment node, and the node associated edge path adjustment or material allocation logic adaptation operation is performed. Other examples can refer to the description in this embodiment, and will not be repeated here.

[0050] This method clarifies the structured construction of the production line topology model with equipment as nodes and process dependencies as edges, providing a clear logical framework for edge computing nodes to quickly match disturbance-related nodes, and ensuring accurate positioning of the disturbance's impact range.

[0051] Furthermore, disturbance removal operations rely on a topology model that reflects the actual state of the production line. However, traditional industrial production line topology models are mostly statically constructed. Even if targeted node adjustments can be made when a pre-set disturbance occurs, frequent non-pre-set dynamic changes during production line operation (such as daily additions and removals of equipment, fine-tuning of process flows, and equipment function upgrades) require manual remodeling and parameter configuration by professional engineers. The entire process can take several weeks, severely restricting the rapid iteration capability of the production line. This leads to the production line topology model becoming easily disconnected from the actual physical state and logical relationships. Consequently, each time a disturbance is removed, the topology model called by the edge computing nodes has information lag, resulting in problems such as mismatched related nodes and incompatibility between topology adjustments and the actual production line. This seriously affects the generation efficiency and feasibility of temporary reconstruction strategies.

[0052] To address this technical challenge, this embodiment further proposes a dynamic self-updating mechanism for the production line topology model. This mechanism ensures that the topology model can track the changing status of production line equipment and processes in real time, providing consistently effective model support for the accurate execution of disturbance removal operations and avoiding disruptions in the reconstruction process or failure of strategies due to model lag.

[0053] Specifically, in addition to the steps mentioned above, the following steps may be performed: Step 23: Monitor the equipment addition / reduction status and process adjustment of the industrial production line in real time, and generate monitoring data containing equipment node addition / reduction information and process dependency adjustment information.

[0054] The system monitors two core change dimensions of industrial production lines in real time: first, equipment addition and removal status, including changes in the number and attributes of equipment nodes such as new equipment being added, old equipment being taken offline, and standby equipment being activated; second, process adjustment status, including changes in process dependencies such as adding, deleting, changing the order of processes, and modifying constraints. For each monitored change, key information is automatically extracted and structured to generate standardized monitoring data containing details of equipment node additions and removals (such as equipment identification, model, and compatible processes) and details of process dependency adjustments (such as adjustment type, associated process identification, and changes in logical constraints). This ensures that the data can directly support subsequent model update processes and avoids model update delays caused by missing change information or inconsistent formats.

[0055] Step 24: Input the monitoring data into the graph neural network, and call the graph neural network to identify the changes in the dependency relationships between equipment nodes in the production line topology model based on the monitoring data.

[0056] The standardized monitoring data generated in step 23 is input into a graph neural network (GNN) in a preset format. This network has been pre-trained based on historical production line operation data and topological logic, and has the ability to capture the dependency relationship features between equipment nodes and processes. Then, the feature extraction and relationship recognition functions of the graph neural network are called. By analyzing the relationship logic of equipment addition and removal and process adjustment in the monitoring data, the network automatically identifies the change information of the dependency relationship between equipment nodes in the production line topology model, including the addition of new dependency relationships (such as the process connection between new equipment and existing equipment), the removal of dependency relationships (such as the disconnection of the connection between offline equipment and other equipment), and the adjustment of dependency strength (such as the change of dependency weight caused by the change of process priority). This ensures the accuracy and efficiency of dependency relationship change identification and provides a clear logical basis for model updates.

[0057] Step 25: Based on the information on changes in dependencies between equipment nodes, deduce and update the node information and edge relationships in the production line topology model.

[0058] Based on the changes in dependencies between equipment nodes identified in step 24, and combined with the production line process logic and equipment capacity constraints, topology relationship deduction is performed: For changes in equipment nodes, the node information of the topology model is updated synchronously, including the attribute entry of newly added equipment nodes, the removal of offline equipment nodes, and the parameter update for equipment attribute changes; for changes in process dependencies, the edge relationships of the topology model are adjusted synchronously, including the construction of edges corresponding to newly added process dependencies, the deletion of edges corresponding to failed dependencies, and the updating of edge attributes (such as dependency priority and constraint conditions) corresponding to logical adjustments. Through this deduction and update operation, it can be ensured that the production line topology model can map the actual physical structure and logical relationships of the production line in real time, avoiding the disconnect between the model and the actual situation on site, and providing accurate and reliable model support for subsequent reconstruction stages such as disturbance removal and alternative path search.

[0059] This method comprehensively captures dynamic changes in the production line by real-time monitoring of equipment additions and reductions and process adjustments, generating structured monitoring data to provide complete data input for model updates. Simultaneously, leveraging the self-learning capabilities of graph neural networks (GNNs), it automatically identifies changes in dependencies between equipment nodes, replacing traditional manual judgment. This improves the accuracy of relationship identification and significantly shortens the identification cycle. Based on the identified dependency changes, it infers and updates model node information and edge relationships, ensuring real-time synchronization between the topology model and the actual production line status. This ensures that operations such as node matching and topology adjustments are always based on an accurate model foundation, thereby improving the stability and reliability of the entire reconstruction process while reducing the operational costs of manual model maintenance.

[0060] Example 3: The above embodiments do not limit the specific verification method for calling the full-element digital twin model to perform global simulation verification of the temporary reconstruction strategy. The simulation verification of traditional production line reconstruction often relies on single-dimensional indicators or human experience judgment, lacking a standardized multi-dimensional verification system. This results in strong subjectivity and insufficient coverage in the verification process. For example, some solutions only focus on equipment operating efficiency while ignoring energy consumption cost control, while others do not fully consider safety constraints and compliance. Consequently, even verified strategies may still encounter problems such as inefficiency, excessive energy consumption, or even safety risks in actual implementation, failing to meet the multi-dimensional demands of modern industry for efficient, economical, and safe reconstruction strategies.

[0061] To construct a standardized, multi-dimensional global simulation verification system, avoid strategic risks caused by ambiguity in verification logic, and ensure that verified temporary reconstruction strategies can meet multiple objectives, this embodiment further proposes a global simulation verification method. In step 103, a full-element digital twin model is invoked to perform global simulation verification on the temporary reconstruction strategy. This can be performed as follows: the simulated production line operation status is verified based on preset verification indicators; if the temporary reconstruction strategy meets all preset verification indicators, it is determined to have passed verification; if any preset verification indicator is not met, it is determined to have failed verification. The preset verification indicators include, but are not limited to: equipment overall efficiency compliance, energy consumption cost control, and safety constraint compliance.

[0062] The verification method proposed in this embodiment systematically verifies the simulation results based on pre-set multi-dimensional verification indicators. These pre-set verification indicators are guided by the core operational requirements of the production line, focusing on the core objectives of efficient, economical, and safe industrial production. These include, but are not limited to, the overall equipment efficiency (OEE) meeting targets (e.g., whether the OEE reaches the preset target value), energy cost control (e.g., whether energy consumption per unit capacity does not exceed the preset budget), and safety constraint compliance (e.g., whether there are risks such as equipment overload, process parameter violations, or safety rule violations). Furthermore, the types of indicators can be flexibly expanded according to the industry characteristics and production needs of different production lines. During the verification process, a clear judgment rule is adopted: full compliance means pass, any non-compliance means fail. If the temporary reconstruction strategy meets the requirements of all pre-set verification indicators without any non-compliance, the strategy is deemed to have passed the global simulation verification and can be used as the basis for subsequent optimization or direct execution. If any pre-set verification indicator is not met, it indicates that the strategy has operational risks, resource conflicts, or optimization space in the corresponding dimension, and the verification is deemed to have failed, requiring the triggering of a pre-set optimization engine (e.g., PPO+NSGA-III) for targeted adjustments.

[0063] This method clarifies the verification indicators and judgment rules, including preset verification indicators covering the core dimensions of production line operation, such as the overall efficiency of equipment, energy consumption cost control, and safety constraint compliance. This ensures both the production efficiency of the strategy and the bottom line of cost control and safety compliance, effectively avoiding the decision-making bias caused by single-dimensional verification and ensuring the global adaptability of the strategy. At the same time, relying on the high-fidelity simulation of the full-element digital twin model, this method can identify potential hidden dangers that may exist in the actual operation of temporary strategies in advance, avoiding the production capacity loss, safety risks, or economic waste caused by the direct implementation of strategies, and greatly improving the global optimality and execution security of the final reconstructed strategy.

[0064] Example 4: Disturbance identification, as a core pre-processing step in adaptive topology reconfiguration of industrial production lines, directly determines the efficiency and effectiveness of subsequent reconfiguration processes due to its comprehensiveness and accuracy. Traditional industrial production line disturbance identification often relies on single-dimensional data (such as collecting only equipment operating status data), resulting in limited data coverage. Furthermore, it commonly employs simple threshold judgments or manual inspection methods, lacking a systematic data processing and intelligent identification mechanism. This leads to significant shortcomings in disturbance identification: on the one hand, it struggles to capture the multi-factor correlation characteristics behind disturbances such as equipment failures and process changes, easily resulting in missed or false alarms; on the other hand, it cannot quickly distinguish between critical disturbances and non-critical fluctuations, leading to erroneous triggering or delayed responses in the reconfiguration process, thereby affecting the stability of production line operation.

[0065] The above embodiments do not limit the specific source and processing method of the identification data. To ensure the reliability of the disturbance identification process, this embodiment proposes a disturbance identification method. In step 101, disturbance identification is initiated for the industrial production line, which can be performed according to the following steps: Step 11: Collect multi-source heterogeneous data from the industrial production line in real time.

[0066] The specific multi-source heterogeneous data collected includes, but is not limited to: device layer data, control layer data, resource layer data, and business layer data.

[0067] Equipment-level data directly reflects the status of the equipment itself, including but not limited to: equipment health data (such as vibration, temperature trends, and other indicators that reflect equipment aging or failure risk) and equipment operating status data (such as real-time operating parameters such as shutdown, offline, and load rate).

[0068] The control layer data focuses on the logical connections of the production line, including but not limited to: production line topology data (connection relationship between equipment and process) and process logical relationship data (process sequence, constraints, etc.).

[0069] Resource layer data focuses on production assurance capabilities, including but not limited to: work-in-process inventory data (material reserves and circulation status) and energy supply capacity data (the stability and surplus of energy supply such as electricity and steam).

[0070] Business layer data is linked to production target requirements, including but not limited to: order urgency data (order priority, delivery time limit requirements) and delivery window data (allowed production cycle range).

[0071] By collecting these multi-source heterogeneous data in a layered manner, we can comprehensively cover the entire chain of production line equipment, logic, resources, and business operations, avoiding the loss of disturbance features due to incomplete data, and laying a data foundation for subsequent accurate identification.

[0072] Step 12: Call the preset rule engine and lightweight machine learning model to process multi-source heterogeneous data and identify whether there is a preset production line disturbance.

[0073] The system invokes a pre-defined rule engine, which contains explicit rules for common production line disturbances (such as equipment health falling below a threshold or order urgency exceeding normal limits). This engine can quickly perform preliminary screening and matching of collected multi-source heterogeneous data, efficiently identifying potential disturbances that conform to explicit rules. Simultaneously, it invokes a lightweight machine learning model (such as a lightweight LSTM model). This model is pre-trained based on historical production line operation data and disturbance cases, and has the ability to mine implicit correlation features in the data. It can perform in-depth analysis of complex disturbance features hidden in multi-source data that the rule engine has not captured (such as the potential correlation between minor fluctuations in equipment operating parameters and changes in process logic).

[0074] By combining the rapid filtering of the rule engine with the deep mining of the lightweight machine learning model, it is possible to comprehensively determine whether the collected data meets the characteristics of the preset production line disturbance. If it does, the subsequent reconstruction process is triggered; if it does not, continuous monitoring is performed to ensure that the disturbance identification is both efficient and accurate, avoiding missed judgments, misjudgments, or delayed responses.

[0075] The disturbance identification method provided in this embodiment comprehensively covers the core dimensions of production line operation through multi-source heterogeneous data acquisition at four levels, avoiding the omission of disturbance features caused by traditional single data acquisition. Furthermore, it utilizes a rule engine for rapid screening of explicit disturbances and a lightweight machine learning model to mine implicit disturbance features. These two complementary approaches replace traditional manual judgment or single-threshold screening methods, significantly improving the efficiency and accuracy of disturbance identification and reducing the probability of missed or false positives. This method provides a unified execution standard for the disturbance identification process, eliminating the need for manual intervention throughout. This reduces manual operation costs and human error while ensuring the real-time performance of disturbance identification, providing a reliable prerequisite for the rapid triggering of subsequent production line topology reconfiguration.

[0076] Example 5: The operating environment of industrial production lines is subject to unpredictable dynamic fluctuations. Even if the final reconstruction strategy is verified by the cloud-based full-element digital twin model, the actual implementation process may still cause the production line operation indicators to deviate from expectations due to unforeseen variables such as sudden equipment aging, energy supply fluctuations, and material anomalies.

[0077] Traditional production line restructuring processes rely on manual discovery, analysis, and strategy switching, which is slow to respond and prone to secondary capacity losses. It may even lead to safety risks due to the continuous failure of strategies.

[0078] To ensure the continuous and stable operation of the production line, this embodiment proposes a closed-loop guarantee mechanism after the reconfiguration strategy is implemented. This mechanism can fill the technical gap in traditional processes where there is no emergency backup after the strategy is implemented. It ensures that even if there are unexpected situations such as deviations in indicators, the system can quickly switch to a stable operating state, minimize unplanned losses, and achieve self-healing guarantee for production line reconfiguration.

[0079] Specifically, after the final reconfiguration strategy is issued to the production line execution control system in step 104, the following steps can be further performed: Step 105: Continuously collect preset indicators of production line operation and compare them with preset thresholds.

[0080] Continuously and dynamically collect preset indicators that reflect the production line's operating status. These indicators focus on the core dimensions of strategy execution effectiveness and are consistent with the core focus dimensions of the cloud-based global simulation verification phase, including but not limited to: Overall Equipment Effectiveness (OEE), energy consumption cost control level, and safety constraint compliance, to ensure the comprehensiveness and relevance of monitoring dimensions.

[0081] At the same time, the data collected in real time will be compared with the corresponding pre-set thresholds (such as OEE compliance threshold, energy consumption budget limit, and safety rule compliance standards) to dynamically track whether the production line operation status meets expectations, avoid abnormal situations not being detected in time due to lack of monitoring, and lay a data foundation for subsequent emergency response.

[0082] Step 106: When the preset indicator deviates from the corresponding preset threshold, the historically stored stable strategy version is called and a rollback command is sent to the production line execution control system to perform the strategy rollback operation.

[0083] The system continuously receives the threshold comparison results from step 105. If any preset indicator deviates from the corresponding preset threshold, such as an OEE decrease exceeding 5%, energy consumption exceeding the budget, or signs of safety rule violations, the automatic emergency response mechanism is immediately activated. First, the system calls the historically stored stable strategy version. This version is a production line operation strategy that has been verified through long-term actual operation, with all indicators consistently meeting standards and no safety risks, possessing reliable adaptability and stability. Subsequently, the system precisely issues a rollback command to the production line execution control system, explicitly requiring the termination of the currently executing refactoring strategy and switching to the called stable strategy version. After receiving the command, the production line execution control system quickly drives equipment start-up and shutdown, process adjustment, material scheduling, and other aspects to resume operation according to the stable strategy, rapidly pulling the production line back to a steady state and avoiding capacity losses, safety risks, or economic waste caused by continuous abnormalities.

[0084] This method continuously collects core preset indicators and compares them with thresholds in real time, replacing the traditional manual inspection mode. This ensures that abnormal situations are captured immediately, curbing the expansion of risks at the source. When indicators deviate, no manual decision-making, analysis, or operation is required. The method automatically completes the invocation of stabilization strategies and the issuance of rollback instructions, significantly shortening emergency response time and avoiding the aggravation of losses caused by the lag in manual processing. The historical stabilization strategies have been verified through actual operation, and their adaptability, safety, and stability have been fully confirmed. After rollback, the production line can be quickly restored to normal operation. This method improves the stability, fault tolerance, and industrial applicability of the entire industrial production line topology adaptive reconstruction method, providing dual protection for the continuous safe and efficient operation of the production line.

[0085] Example 6: This embodiment relates to an industrial production line topology adaptive reconfiguration device. A schematic diagram of the industrial production line topology adaptive reconfiguration device in this embodiment is shown below. Figure 2 As shown, it includes: a disturbance identification unit 201, an edge response unit 202, a remote adjustment unit 203, and a policy execution unit 204.

[0086] The disturbance identification unit 201 is used to initiate disturbance identification on the industrial production line. If a preset production line disturbance is detected, the disturbance information is determined. The edge response unit 202 is used to schedule edge computing nodes, remove disturbances based on disturbance information, search for feasible alternative paths through pre-stored process correlation data, and generate temporary reconfiguration strategies. The remote adjustment unit 203 is used to control the cloud supercomputing node to receive the temporary reconstruction strategy and call the full-element digital twin model to perform global simulation verification of the temporary reconstruction strategy. If the verification fails, the preset optimization engine is started to adjust and optimize the temporary reconstruction strategy, and the final reconstruction strategy that passes the verification is generated and output. The strategy execution unit 204 is used to send the final reconfiguration strategy to the production line execution control system to execute the final reconfiguration strategy.

[0087] In the industrial production line topology adaptive reconfiguration device provided in this embodiment, the edge response unit relies on local computing power to achieve millisecond-level disturbance removal and temporary strategy generation, while the cloud adjustment unit utilizes supercomputing capabilities and a full-element digital twin model to complete global simulation verification and targeted optimization. These two complementary approaches overcome the bottlenecks of slow response or short-sighted decision-making in traditional single-architecture systems, enabling the production line to complete dynamic adaptation within 15 minutes (response time reduced by 70%). The strategy is verified by multi-dimensional indicators (equipment overall efficiency, energy consumption, and safety), and coupled with a potential automatic rollback mechanism, ensures zero-accident execution and stable, controllable operation. Simultaneously, the device can flexibly adapt to diverse scenarios such as sudden equipment failures and the introduction of new processes. Through topology adaptive algorithms and structured modeling mechanisms, it achieves agile adaptation from second-level intelligent regeneration to hour-level model reconstruction, significantly improving the production line's flexible manufacturing capabilities and intelligence level, providing reliable technical support for highly dynamic industrial environments.

[0088] It should be noted that the contents of the industrial production line topology adaptive reconfiguration device provided in this embodiment can be referred to in conjunction with the industrial production line topology adaptive reconfiguration method provided in the above embodiments, and the repeated parts will not be described again in this embodiment.

[0089] Furthermore, it is worth mentioning that all units involved in this embodiment are logic modules. In practical applications, a logic unit can be a physical unit, a part of a physical unit, or a combination of multiple physical units. In addition, to highlight the innovative aspects of this application, this embodiment does not introduce units that are not closely related to solving the technical problems proposed in this application; however, this does not mean that other units are absent from this embodiment.

[0090] Example 7: To verify the effectiveness and practicality of the industrial production line topology adaptive reconfiguration device, the following detailed explanation of the application effect of this device is provided through two specific implementation examples in typical industrial scenarios.

[0091] Scenario 1 is a scenario where a company's refining and chemical production line experiences a fault and can be intelligently regenerated within seconds.

[0092] One specific configuration of the industrial production line topology adaptive reconfiguration device is as follows: The edge computing node adopts an industrial gateway deployed locally on the refining and chemical production line. This gateway integrates a lightweight chemical equipment topology graph manager and a process knowledge graph engine, which can quickly respond to local disturbance handling needs; the cloud supercomputing center is deployed on a private cloud and runs a high-fidelity digital twin including a reactor thermodynamic model and a pipeline fluid dynamics model, which has full-element simulation and deduction capabilities; the communication network is built based on 5GTSN (Time-Sensitive Networking), forming a low-latency, high-reliability control channel to ensure the secure and efficient transmission of strategy data; the monitoring data covers multiple dimensions of information such as equipment health (e.g., vibration, temperature trends), process parameters (e.g., flow rate, pressure, temperature), and order urgency, providing data support for disturbance identification and strategy verification.

[0093] The specific implementation steps are as follows: First, the system detects a sudden drop in the health of the critical equipment—reactor R101—to 75% through real-time monitoring, triggering a fault warning and initiating the reconfiguration process. Subsequently, the edge node completes the topology reconfiguration operation within 1 minute, removing the failed R101 node from the production line topology model and automatically searching for feasible alternative paths based on the pre-stored process knowledge graph, activating the backup reactor R102, and generating a temporary scheduling strategy. Next, this temporary scheduling strategy is uploaded to the cloud supercomputing center, where the digital twin completes a full-element simulation within 3 minutes, verifying the strategy from three core dimensions: overall equipment efficiency (OEE ≥ 85%), energy consumption (≤ budget), and safety (no overpressure), confirming that it meets all preset requirements. Finally, the verified final strategy is distributed to the DCS system through the TSN network, and the production line completes a full reconfiguration within 15 minutes. At the same time, the system is set up with an automatic rollback mechanism; if the OEE drops by more than 5% within 1 hour, it automatically rolls back to the previous stable version to ensure production safety.

[0094] The device has significant effects: production line reconfiguration time has been reduced from 45 minutes in the traditional solution to 15 minutes, and response time has been shortened by 70%; each failure can avoid unplanned downtime of 2.5 hours, reducing economic losses by more than 500,000 yuan per incident; through the dual protection of cloud sandbox verification and automatic rollback mechanism, the reconfiguration process achieves zero accidents; at the same time, it provides support for the digital transformation of simulation training in the upstream, midstream and downstream of enterprises, and improves the efficiency of training resource generation by 15 times.

[0095] Scenario 2 is a scenario where a company reconstructs an hourly model to introduce a new process.

[0096] Another specific configuration of the industrial production line topology adaptive reconfiguration device is as follows: The edge lightweight unit is deployed in the ceramic enterprise workshop, mainly responsible for local data acquisition and preliminary disturbance identification, providing basic input for cloud processing; the cloud supercomputing inference engine runs a full-element digital twin model built based on the MBSE (Model-Based Systems Engineering) methodology, with efficient model integration and strategy optimization capabilities; the model library includes a physical entity library of equipment, a process behavior rule library, and a safety constraint library, providing standardized module support for new process adaptation; the input data covers key information such as new product introduction instructions, new equipment parameters, energy efficiency targets, and carbon emission limits, clarifying the reconfiguration direction and constraint requirements.

[0097] The specific implementation steps are as follows: First, the enterprise introduces a new energy-saving kiln. The system, through data monitoring and analysis, identifies the event as a disturbance of the process iteration change type, triggering the model reconstruction process. Second, the edge lightweight unit collects the parameters of the new equipment and the process flow diagram, and uploads them to the cloud supercomputing simulation engine. Third, based on the MBSE methodology, the cloud automatically integrates the new equipment into the full-element digital twin, embedding the mechanical dynamics model of the equipment and the corresponding process constraint rules to complete the model adaptation. Fourth, the system completes the new production line topology modeling, reconstruction strategy generation, and multi-dimensional verification within 2 hours, and distributes the final strategy to the production control system. Fifth, the need for manual rule configuration is reduced by 70% throughout the process. Engineers only need to confirm the final strategy and do not need to carry out manual modeling and complex parameter configuration work.

[0098] The device has achieved remarkable results: the topology model reconstruction cycle during the introduction of new processes has been shortened from the traditional monthly level to the hourly level, significantly improving enterprises' ability to quickly launch production; the need for manual rule configuration has been reduced by 70%, effectively releasing engineers' productivity and reducing labor costs; in terms of economic benefits, it helps enterprises achieve the goals of "five reductions and three promotions," saving more than 12 million kWh of electricity annually, with related business contracts exceeding 100 million yuan; in terms of social benefits, it forms a new paradigm for energy consumption optimization in industrial enterprises, providing strong technical support for the realization of dual-carbon goals.

[0099] The two scenarios described above respectively cover two core industrial scenarios: high-risk process safety assurance and dual-carbon energy efficiency management. Starting from two key needs—emergency handling of sudden equipment failures and adaptation to new process introductions—the invention fully demonstrates its significant technical advantages in areas such as second-level intelligent regeneration, hour-level model reconstruction, and safety closed-loop verification. Implementation results show that this device possesses high feasibility and broad application value, effectively addressing the technical pain points of traditional industrial production line reconstruction, such as slow response time, reliance on manual labor, and insufficient safety. It provides solid and reliable technical support for industrial enterprises to cope with dynamic and complex production environments.

[0100] Example 8: Another embodiment of this application relates to an electronic device, such as... Figure 3 As shown, it includes: at least one processor 301; and a memory 302 communicatively connected to at least one processor 301; wherein the memory 302 stores instructions executable by at least one processor 301, which are executed by at least one processor 301 to enable at least one processor 301 to perform the steps of the industrial production line topology adaptive reconfiguration method in the above embodiments.

[0101] The memory and processor are connected via a bus, which can include any number of interconnecting buses and bridges, connecting various circuits of one or more processors and memories. The bus can also connect various other circuits, such as peripheral devices, voltage regulators, and power management circuits, which are well known in the art and will not be described further herein. The bus interface provides an interface between the bus and the transceiver. The transceiver can be a single element or multiple elements, such as multiple receivers and transmitters, providing a unit for communicating with various other devices over a transmission medium. Data processed by the processor is transmitted over the wireless medium via an antenna, which further receives data and transmits it to the processor.

[0102] The processor manages the bus and general processing, and also provides various functions, including timing, peripheral interfaces, voltage regulation, power management, and other control functions. Memory is used to store data used by the processor during operation.

[0103] Example 9: Another embodiment of this application relates to a computer-readable storage medium storing a computer program. When executed by a processor, the computer program implements the method embodiments described above.

[0104] That is, those skilled in the art will understand that all or part of the steps in the methods of the above embodiments can be implemented by a program instructing related hardware. This program is stored in a storage medium and includes several instructions to cause a device (which may be a microcontroller, chip, etc.) or processor to execute all or part of the steps of the methods of the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as a USB flash drive, a portable hard drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0105] Those skilled in the art will understand that the above embodiments are specific embodiments for implementing this application, and in practical applications, various changes can be made to them in form and detail without departing from the spirit and scope of this application.

Claims

1. An adaptive topology reconstruction method for industrial production lines, characterized in that, include: Initiate disturbance identification for industrial production lines; if a preset disturbance is detected, determine the disturbance information. The edge computing nodes are scheduled to remove disturbances based on the disturbance information, and feasible alternative paths are searched through pre-stored process association data to generate a temporary reconfiguration strategy. The cloud supercomputing node receives the temporary reconstruction strategy and calls the full-element digital twin model to perform global simulation verification of the temporary reconstruction strategy. If the verification fails, the preset optimization engine is started to adjust and optimize the temporary reconstruction strategy, and the final reconstruction strategy that passes the verification is generated and output. The final reconfiguration strategy is sent to the production line execution control system to execute the final reconfiguration strategy.

2. The industrial production line topology adaptive reconstruction method according to claim 1, characterized in that, The scheduling edge computing node removes disturbances based on the disturbance information, including: The edge computing node is controlled to call the current production line topology model to match associated nodes based on the disturbance information; the production line topology model is constructed with production line equipment as nodes and process dependencies as edges; the disturbance information includes: disturbance type and associated impact data; Perform a topology adjustment operation on the associated node that is adapted to the disturbance type.

3. The industrial production line topology adaptive reconstruction method according to claim 2, characterized in that, The step of performing a topology adjustment operation on the associated nodes that is adapted to the disturbance type includes: If the disturbance type is a sudden equipment failure, then the associated node is the failed equipment node corresponding to the failure, and a removal operation is performed on the failed equipment node. If the disturbance type is a process iteration change, then the associated node is the target node involved in the process change. An attribute update operation is performed on the target node, or an add / delete operation is performed on the associated edges of the target node.

4. The industrial production line topology adaptive reconstruction method according to claim 2, characterized in that, Also includes: Real-time monitoring of equipment additions / reductions and process adjustments in the industrial production line, generating monitoring data that includes equipment node additions / reductions and process dependency adjustments; The monitoring data is input into a graph neural network, and the graph neural network is invoked to identify the changes in the dependency relationships between equipment nodes in the production line topology model based on the monitoring data. Based on the changes in the dependencies between the device nodes, the node information and edge relationships of the production line topology model are deduced and updated.

5. The industrial production line topology adaptive reconfiguration method according to claim 1, characterized in that, The step of calling the full-element digital twin model to perform global simulation verification of the temporary reconstruction strategy includes: The simulated production line operation status is verified based on preset verification indicators; the preset verification indicators include: the overall efficiency of the equipment meets the standards, the energy consumption cost control status, and the safety constraint compliance status. If the temporary reconstruction strategy meets all the preset verification indicators, it is determined that the verification is successful; If any of the preset verification indicators are not met, the verification is deemed to have failed.

6. The industrial production line topology adaptive reconfiguration method according to claim 1, characterized in that, The identification of disturbances during the start-up of industrial production lines includes: Real-time acquisition of multi-source heterogeneous data from industrial production lines; the multi-source heterogeneous data includes equipment layer data, control layer data, resource layer data, and business layer data; the equipment layer data includes: equipment health data and equipment operating status data; the control layer data includes: production line topology data and process logic relationship data; the resource layer data includes: work-in-process inventory data and energy supply capacity data; the business layer data includes: order urgency data and delivery window data. The system calls a preset rule engine and a lightweight machine learning model to process the multi-source heterogeneous data and identify whether there are preset production line disturbances.

7. The industrial production line topology adaptive reconfiguration method according to any one of claims 1 to 6, characterized in that, After the final reconfiguration strategy is issued to the production line execution control system, the following is also included: Continuously collect preset indicators of production line operation and compare them with preset thresholds; When the preset indicator deviates from the corresponding preset threshold, the historically stored stable strategy version is invoked, and a rollback command is sent to the production line execution control system to perform a strategy rollback operation.

8. An industrial production line topology adaptive reconfiguration device, characterized in that, include: The disturbance identification unit is used to initiate disturbance identification on the industrial production line. If a preset disturbance is detected, the disturbance information is determined. The edge response unit is used to schedule edge computing nodes, remove disturbances based on the disturbance information, search for feasible alternative paths through pre-stored process association data, and generate a temporary reconfiguration strategy. The remote adjustment unit is used to control the cloud supercomputing node to receive the temporary reconstruction strategy and call the full-element digital twin model to perform global simulation verification of the temporary reconstruction strategy. If the verification fails, the preset optimization engine is started to adjust and optimize the temporary reconstruction strategy, and generate and output the final reconstruction strategy that passes the verification. The strategy execution unit is used to send the final reconfiguration strategy to the production line execution control system to execute the final reconfiguration strategy.

9. An electronic device, characterized in that, include: At least one processor; And, a memory communicatively connected to the at least one processor; The memory stores instructions that can be executed by the at least one processor, which are executed by the at least one processor to enable the at least one processor to perform the industrial production line topology adaptive reconfiguration method as described in any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the industrial production line topology adaptive reconfiguration method as described in any one of claims 1 to 7.

Images