Industrial product production management system based on industrial internet

Abstract

The invention relates to the technical field of industrial internet, in particular to an industrial product production management system based on the industrial internet, which comprises a multi-source sensing module, a dynamic topology module and an intelligent decision module, abnormal events are detected through an event pattern recognition engine, and structured event data are output; the dynamic topology module maps the production element entities into reconfigurable topology nodes and marks a real-time attribute set; activating a space-time constraint generator according to the structured event data, and creating a dynamic constraint edge; when the change of the dynamic constraint edge exceeds a preset threshold value, generating a global topology reconstruction signal; otherwise, generating a local constraint adjustment signal; the intelligent decision-making module responds to the signal and triggers multi-objective optimization calculation based on transfer learning or executes incremental rescheduling with a memory function in a process correlation subgraph, so that efficient production management is realized.

Description

Technical Field

[0001] The present invention relates to the field of industrial Internet technology, and specifically to an industrial product production management system based on the industrial Internet. Background Art

[0002] With the rapid development of industrial Internet technology, traditional industrial product production management systems have gradually exposed the bottleneck of insufficient multi-dimensional collaboration capabilities. Existing systems mostly use a distributed data collection architecture, which makes it difficult to effectively integrate heterogeneous data from equipment sensors, order systems, and material tracking systems, resulting in a lack of dynamic correlation between production factor entities. In complex production scenarios, static topology modeling methods cannot respond in real time to dynamic disturbance events such as equipment energy efficiency fluctuations, order priority changes, and logistics path blockages. They often rely on manual experience for passive decision-making, and there are problems such as delayed response, local optimization, and conflicts between global goals.

[0003] In addition, the flexible production needs of multiple varieties and small batches have put forward higher requirements for the system's real-time perception, adaptive topology reconstruction and intelligent decision-making. Traditional solutions lack systematic technical support in abnormal event analysis, dynamic evolution of cross-process constraint networks and incremental scheduling optimization, and are difficult to meet the comprehensive optimization needs of modern industrial production for efficiency, energy consumption and delivery cycle. Summary of the invention

[0004] The purpose of the present invention is to provide an industrial product production management system based on the Industrial Internet to solve the problems raised in the above-mentioned background technology.

[0005] To achieve the above purpose, the present invention provides the following technical solution: an industrial product production management system based on the industrial Internet, comprising a multi-source perception module, a dynamic topology module and an intelligent decision-making module, wherein: The multi-source perception module is used to collect multi-dimensional industrial product data in real time through the industrial Internet protocol, and associate discrete production factor entities in the multi-dimensional industrial product data based on device fingerprint coding; and analyze abnormal events in the multi-dimensional industrial product data through an event pattern recognition engine, and output structured event data; The dynamic topology module establishes an event subscription channel with the multi-source perception module, which is used to map the production factor entity into a reconfigurable topology node and annotate the real-time attribute set; and: activating a corresponding spatiotemporal constraint generator according to the structured event data to create a dynamic constraint edge; When the dynamic constraint edge variation range exceeds a preset threshold, a global topology reconstruction signal is generated; otherwise, a local constraint adjustment signal is generated; The intelligent decision-making module establishes a data control connection with the dynamic topology module to respond to the global topology reconstruction signal and trigger the multi-objective optimization calculation based on transfer learning; respond to the local constraint adjustment signal and perform incremental rescheduling with memory function in the process-related subgraph; The modules of the system use an event-driven cascade response mechanism to convert the dynamic changes of production factors into computable constraint network evolution in real time, which is used to enhance the multi-dimensional dynamic coordination capability of the production management system.

[0006] As a further improvement of the technical solution, the collection of multi-dimensional industrial product data specifically includes: Obtain heterogeneous data from equipment sensors, production order systems, and material tracking systems through industrial equipment communication gateways and proxy services; Equipment operation data, order data and material data are standardized and packaged through the Industrial Internet protocol to form data messages with a unified time reference.

[0007] As a further improvement of the technical solution, the construction of the device fingerprint code includes: Integrate the hardware characteristics of the equipment controller serial number, the business attributes of the order number, and the spatial and temporal dimensions of the production line station coordinates; The encoding and parsing engine is used to materialize and anchor vibration sensor data, process parameters, and material batch information.

[0008] As a further improvement of the technical solution, the event pattern recognition engine includes: The feature extraction layer uses dynamic time windows to perform equipment data statistics, order data rule analysis, and spatiotemporal feature calculations on material data. The pattern analysis layer deploys three types of detection architectures: equipment anomaly detection model, order emergency processing model, and material blockage detection model; The event structured processing layer generates standard event data including event classification labels, entity identifiers and timestamps.

[0009] As a further improvement of the technical solution, the real-time attribute set annotated by the topological node includes: Equipment attributes include operating parameters, working status, and maintenance records; Order attributes include remaining delivery time, priority, and process requirements; Material attributes include shipping location, batch status, and inventory changes.

[0010] As a further improvement of the technical solution, the generation of the dynamic constraint edge includes: Equipment energy efficiency weight constraint edge, constructed based on equipment working efficiency and energy consumption level indicators; The capacity constraint edge of process connection is set according to the maximum task volume and the shortest time interval between processes; Logistics path constraint edges are generated through material retention node identification and path topology analysis.

[0011] As a further improvement of the technical solution, the process of the dynamic topology module determining the global topology reconstruction signal and the local constraint adjustment signal according to the dynamic constraint edge change range specifically includes: Set preset thresholds for each type of dynamic constraint edge, including equipment energy efficiency weight change threshold, process connection capacity change threshold, and logistics path change threshold If the change range of multiple dynamic constraint edges exceeds the corresponding preset threshold at the same time, and the change range of a single dynamic constraint edge exceeds half of the corresponding preset threshold, a global topology reconstruction signal is generated, where multiple dynamic constraint edges represent two or more; otherwise, a local constraint adjustment signal is generated.

[0012] As a further improvement of the technical solution, the intelligent decision-making module includes a global unit, which is used to respond to the global topology reconstruction signal and trigger the multi-objective optimization calculation based on transfer learning, wherein the multi-objective optimization calculation of transfer learning includes: Achieve energy efficiency strategy migration across production line equipment through feature alignment technology; Initializing the population based on the Pareto frontier solution set accelerates convergence; Dynamically adjust the weight coefficients of energy consumption targets and production efficiency.

[0013] As a further improvement of the technical solution, the intelligent decision module includes a local unit, which is used to respond to the local constraint adjustment signal and perform incremental rescheduling with memory function in the process association subgraph, wherein the incremental rescheduling includes: Avoid duplication of local optimal solutions through tabu search algorithm; The Lagrange multiplier method is used to adjust the constraint edge relaxation; The original scheduling parameters of nodes outside the process-related subgraph are retained.

[0014] As a further improvement of the technical solution, the process of generating the process-related subgraph specifically includes: According to the dynamic constraint edge type in the local constraint adjustment signal, the topological nodes are located; starting from the trigger node, the strongly associated nodes are tracked along the dynamic constraint edge through the graph traversal algorithm; only the nodes and edges that directly interact with the trigger constraint edge are included to generate a process-related subgraph.

[0015] Compared with the prior art, the present invention has the following beneficial effects: This industrial product production management system based on the Industrial Internet achieves accurate anchoring of production factors and structural analysis of abnormal events through a multi-source perception module, effectively breaking through the barriers of data islands; the dynamic topology module builds reconfigurable topological nodes, converts the real-time status changes of equipment, orders, and materials into dynamic adjustments of topological edge weights, and significantly improves the system's ability to characterize production disturbances; In addition, a hierarchical optimization mechanism combining transfer learning and incremental scheduling is adopted to achieve rapid rescheduling of local process subgraphs while ensuring global multi-objective optimization; through an event-driven cascade response mechanism, the dynamic changes of production factors are mapped in real time to the computable constraint network evolution process, comprehensively enhancing the comprehensive performance of the production management system in terms of abnormal response speed, multi-objective collaborative optimization and dynamic resource allocation, providing a systematic solution for intelligent production management in the industrial Internet environment. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 It is a schematic diagram of the overall module of the present invention; Figure 2 It is a schematic diagram of the intelligent decision-making module unit of the present invention.

[0017] In the figure: 100, multi-source perception module; 200, dynamic topology module; 300, intelligent decision-making module; 301, global unit; 302, local unit. DETAILED DESCRIPTION

[0018] The following will be combined with the accompanying drawings in the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0019] Next, see Figure 1 The present invention provides a technical solution: an industrial product production management system based on the industrial Internet, including a multi-source perception module 100, a dynamic topology module 200 and an intelligent decision-making module 300.

[0020] The multi-source perception module 100 is used to collect multi-dimensional industrial product data in real time through the industrial Internet protocol, and associate discrete production factor entities in the multi-dimensional industrial product data based on device fingerprint coding; and parse abnormal events in the multi-dimensional industrial product data through an event pattern recognition engine, and output structured event data; wherein the process of the multi-source perception module 100 for collecting multi-dimensional industrial product data in real time through the industrial Internet protocol specifically includes: The multi-source perception module 100 uses standard industrial Internet protocols (such as proxy services supporting industrial communication protocols and industrial equipment communication gateways) to obtain heterogeneous data of industrial products from various sources to form multi-dimensional industrial product data; the different sources include equipment sensors, production order systems, and material tracking systems; The multi-dimensional industrial product data comes from a wide range of sources, including equipment operation data (such as temperature, pressure, speed, etc.), order data (including order number, customer requirements, delivery date, etc.) and material data (including material location, status, quantity changes, etc.); the multi-dimensional industrial product data is standardized and packaged through the Industrial Internet protocol to form a data message with a unified time base. This design enables the system to fully grasp the situation on the production line.

[0021] The process of the multi-source perception module 100 associating discrete production factor entities in multi-dimensional industrial product data based on device fingerprint coding specifically includes: Based on the association mechanism of device fingerprint coding, by creating a digital identity code with unique, multi-dimensional identification for each physical device, each production task and each material unit, the heterogeneous data streams scattered in the device sensor, production order system and material tracking system are physically integrated; the coding system deeply integrates the equipment hardware characteristics (such as controller serial number), business attributes (such as order number) and time-space dimension coordinates (such as production line workstation coordinates), so that the waveform data collected by the vibration sensor, the process parameters in the enterprise resource planning system and the material batch information recorded by the warehouse management system, which were originally isolated data units, can be accurately anchored to specific equipment entities (such as CNC machine tools), order entities (such as emergency production tasks) and material entities (such as special alloy raw material batches) through the coding parsing engine, forming a data aggregation network with production factor entities as the core as entity identification.

[0022] The process of the multi-source perception module 100 parsing abnormal events in multi-dimensional industrial product data through an event pattern recognition engine and outputting structured event data specifically includes: The event pattern recognition engine includes a three-layer detection architecture: Feature extraction layer: Use dynamic time windows (such as 10-second cycles) to calculate the features of multi-dimensional industrial product data, including statistical features of equipment data (such as fluctuation amplitude), business rules of order data (such as remaining delivery time), and spatiotemporal features of material movement (such as transportation line blockage locations); Pattern analysis layer: deploy three types of detection models, including equipment anomaly detection model, order emergency processing model and material blockage detection model. The equipment anomaly detection model uses a neural network with temporal memory function to learn the multi-parameter association law of the equipment under normal conditions. When the deviation between the real-time data and the predicted value exceeds three times the standard deviation, an early warning is triggered and an abnormal equipment fluctuation event is generated. The order emergency processing model analyzes the emergency task identification in the production planning system, and comprehensively evaluates the feasibility of the task based on the production line load rate, and generates an abnormal event of order priority change. The material blockage detection model builds a logistics route map based on material location tracking data, identifies overtime detention nodes, and generates abnormal events of material flow blockage. Event structuring processing layer: converts abnormal events into standard event data containing multiple elements as structured event data, which includes event classification labels, entity identifiers and timestamps.

[0023] The dynamic topology module 200 establishes an event subscription channel with the multi-source perception module 100 to map the production factor entity into a reconfigurable topology node and annotate the real-time attribute set, specifically including: Establish an event subscription channel: The dynamic topology module 200 communicates with the multi-source perception module 100 through the event subscription mechanism. When an abnormal event is detected or structured event data is generated, the relevant information will be pushed to the dynamic topology module 200 through this channel; the event subscription channel ensures that the dynamic topology module 200 can receive the data output by the multi-source perception module 100 in real time, including equipment operation data, order data and material data.

[0024] Production factor entities are mapped as topological nodes: Production factor entities include but are not limited to equipment (such as CNC machine tools), orders (such as urgent production tasks) and materials (such as batches of special alloy raw materials). These production factor entities are mapped as topological nodes in the network. Each node represents a specific production unit and has a unique identifier. The identifier of each node is based on the device fingerprint code to ensure its uniqueness and traceability. For example, the node identifier of a CNC machine tool may contain information such as the controller serial number and location coordinates.

[0025] Labeling real-time attribute sets: Each topological node is accompanied by a real-time attribute set, which reflects the current status and characteristics of the production factor entity, including equipment attributes, order attributes and material attributes; equipment attributes include equipment operating parameters (such as temperature, pressure, speed), working status (such as running, shut down), maintenance records, etc.; order attributes include order number, customer requirements, delivery date, remaining delivery time, priority, etc.; material attributes include material location, status (such as in-warehouse, out-of-warehouse, in transit), quantity changes, batch information, etc.

[0026] Real-time update: The dynamic topology module 200 continuously receives real-time data from the multi-source perception module 100 and promptly updates the attribute set of each node, which ensures that the topology map always reflects the latest situation on the production line.

[0027] The dynamic topology module 200 activates the corresponding spatiotemporal constraint generator according to the structured event data and creates dynamic constraint edges, specifically including: According to the event classification labels in the structured event data (such as equipment abnormality, urgent order processing, material blockage), specific abnormal events are identified. According to the abnormal events, the corresponding spatiotemporal constraint generator is activated. When the abnormal event is an abnormal fluctuation of the equipment, the equipment energy efficiency weight generator is activated; when the abnormal event is a change in order priority, the process connection capacity generator is activated; when the abnormal event is a material flow blockage, the logistics path generator is activated; For abnormal fluctuation events of equipment, the equipment energy efficiency weight generator will calculate the energy efficiency weight of the equipment in the current state, including the equipment's working efficiency and energy consumption level indicators, and attach it as a constraint condition to the connection edge between related nodes as the equipment energy efficiency weight constraint edge; For urgent order processing events, the process connection capacity generator will evaluate the connection capacity between each process, that is, the maximum amount of tasks that can be processed or the shortest time interval between different processes, and attach these capacity restrictions as constraints to the connection edges between related nodes as process connection capacity constraint edges; For material blocking events, the logistics path generator constructs a logistics path graph based on the material location tracking data, identifies the timeout nodes, and generates logistics path constraint edges; Dynamic constraint edges include equipment energy efficiency weight constraint edges, process connection capacity constraint edges, and logistics path constraint edges.

[0028] The dynamic topology module 200 sets preset thresholds for each type of dynamic constraint edge, including equipment energy efficiency weight change thresholds, process connection capacity change thresholds, and logistics path change thresholds, where: If the change ranges of multiple dynamic constraint edges exceed the corresponding preset thresholds at the same time, and the change range of a single dynamic constraint edge exceeds half of the corresponding preset threshold, a global topology reconstruction signal is generated, where multiple dynamic constraint edges represent two or more; otherwise, a local constraint adjustment signal is generated, and the affected topological nodes are located according to the dynamic constraint edge type in the local constraint adjustment signal; starting from the trigger node, strongly associated nodes are tracked along the dynamic constraint edges through a graph traversal algorithm (such as breadth-first search); only nodes and edges that directly interact with the trigger constraint edges and have significant weight changes are included to form a process-associated subgraph.

[0029] The intelligent decision-making module 300 establishes a data control connection with the dynamic topology module 200 to respond to the global topology reconstruction signal and trigger a multi-objective optimization calculation based on transfer learning, specifically including: The intelligent decision module 300 and the dynamic topology module 200 establish a real-time data control connection through a high-throughput message middleware, supporting two-way transmission of signals and data; Use standardized data formats to encapsulate data such as signal types, constraint edge state parameters, and process subgraph topology structures; The global topology reconstruction signal and the local constraint adjustment signal are transmitted through event channels of different priorities to ensure a fast response to the high-priority signal, where the high-priority signal is the global topology reconstruction signal; See also Figure 2 , when receiving the global topology reconstruction signal sent by the dynamic topology module 200, the global unit 301 in the intelligent decision module 300 parses the signal content and extracts key parameters (such as the type of the super-threshold constraint edge, the amplitude of change, and the list of associated equipment / processes); Using transfer learning technology, similar scenarios (such as a combination of equipment energy efficiency drop and process capacity overrun) are retrieved from the historical optimization case library, and the corresponding optimization model parameters are loaded; Initialize the population based on the Pareto frontier solution set of historically similar scenarios, skipping the random initialization stage to accelerate convergence; Map the equipment energy efficiency optimization strategies of other production lines or historical cycles to the current constraint space (e.g., transfer logistics path optimization experience through feature alignment technology); Dynamically adjust the weights of multiple objectives according to the change range of the current constraint edge (for example, when the weight of equipment energy efficiency drops sharply, increase the penalty coefficient of energy consumption target); Generate a set of Pareto optimal scheduling solutions and select the solution with the highest comprehensive score based on real-time working conditions; The reconstructed global scheduling plan (such as equipment start and stop sequence, process connection rules, and logistics path planning) is pushed to the dynamic topology module 200 to trigger the update of the entire network topology.

[0030] The local unit 302 in the intelligent decision module 300 responds to the local constraint adjustment signal and performs incremental rescheduling with memory function in the process-related subgraph, specifically including: When receiving a local constraint adjustment signal, locating the affected process-related subgraph; Determine the subgraph boundaries through graph traversal algorithms (such as breadth-first search) to ensure that the adjustment range is minimized; Match similar local scenarios and verified feasible solutions from the local optimization memory library, retain the current scheduling status of nodes outside the process-related subgraph (such as other equipment operating parameters remain unchanged); only relax and iterate the affected constraint edges in the process-related subgraph (such as locally adjusting the equipment energy efficiency weight through the Lagrange multiplier method); use a local search algorithm with a taboo table in the process-related subgraph to avoid repeated invalid adjustments; The adjusted process-related subgraph scheduling parameters (such as the new energy efficiency coefficient of a certain equipment and the buffer capacity between processes) are pushed to the dynamic topology module 200 to trigger the update of local edge weights.

[0031] The above shows and describes the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited by the above embodiments. The above embodiments and descriptions are only preferred examples of the present invention and are not intended to limit the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, which fall within the scope of the present invention. The scope of protection of the present invention is defined by the attached claims and their equivalents.

Claims

1. The industrial product production management system based on the industrial Internet is characterized by: It comprises a multi-source perception module (100), a dynamic topology module (200) and an intelligent decision-making module (300), wherein: The multi-source perception module (100) is used to collect multi-dimensional industrial product data in real time through an industrial Internet protocol, and to associate discrete production factor entities in the multi-dimensional industrial product data based on device fingerprint coding; and to parse abnormal events in the multi-dimensional industrial product data through an event pattern recognition engine, and to output structured event data; The dynamic topology module (200) establishes an event subscription channel with the multi-source perception module (100) for mapping production factor entities into reconfigurable topology nodes and marking real-time attribute sets; and, activating a corresponding spatiotemporal constraint generator according to the structured event data to create a dynamic constraint edge; When the dynamic constraint edge variation range exceeds a preset threshold, a global topology reconstruction signal is generated; otherwise, a local constraint adjustment signal is generated; The intelligent decision-making module (300) establishes a data control connection with the dynamic topology module (200) to respond to a global topology reconstruction signal to trigger a multi-objective optimization calculation based on transfer learning; respond to a local constraint adjustment signal to perform incremental rescheduling with a memory function in a process-related subgraph; The modules of the system use an event-driven cascade response mechanism to convert the dynamic changes of production factors into computable constraint network evolution in real time, which is used to enhance the multi-dimensional dynamic coordination capability of the production management system.

2. The industrial product production management system based on the industrial Internet according to claim 1 is characterized in that: The collection of multi-dimensional industrial product data specifically includes: Obtain heterogeneous data from equipment sensors, production order systems, and material tracking systems through industrial equipment communication gateways and proxy services; Equipment operation data, order data and material data are standardized and packaged through the Industrial Internet protocol to form data messages with a unified time reference.

3. The industrial product production management system based on the industrial Internet according to claim 1 is characterized in that: The construction of the device fingerprint code includes: Integrate the hardware characteristics of the equipment controller serial number, the business attributes of the order number, and the spatial and temporal dimensions of the production line station coordinates; The encoding and parsing engine is used to materialize and anchor vibration sensor data, process parameters, and material batch information.

4. The industrial product production management system based on the industrial Internet according to claim 1 is characterized in that: The event pattern recognition engine comprises: The feature extraction layer uses dynamic time windows to perform equipment data statistics, order data rule analysis, and spatiotemporal feature calculations on material data. The pattern analysis layer deploys three types of detection architectures: equipment anomaly detection model, order emergency processing model, and material blockage detection model; The event structured processing layer generates standard event data including event classification labels, entity identifiers and timestamps.

5. The industrial product production management system based on the industrial Internet according to claim 1 is characterized in that: The real-time attribute set annotated by the topological node includes: Equipment attributes include operating parameters, working status and maintenance records; Order attributes include remaining delivery time, priority, and process requirements; Material attributes include shipping location, batch status, and inventory changes.

6. The industrial product production management system based on the industrial Internet according to claim 1 is characterized in that: The generation of the dynamic constraint edge includes: Equipment energy efficiency weight constraint edge, constructed based on equipment working efficiency and energy consumption level indicators; The capacity constraint edge of process connection is set according to the maximum task volume and the shortest time interval between processes; Logistics path constraint edges are generated through material retention node identification and path topology analysis.

7. The industrial product production management system based on the industrial Internet according to claim 1 is characterized in that: The process of the dynamic topology module (200) determining the global topology reconstruction signal and the local constraint adjustment signal according to the dynamic constraint edge variation range specifically includes: Set preset thresholds for each type of dynamic constraint edge, including equipment energy efficiency weight change threshold, process connection capacity change threshold, and logistics path change threshold If the change range of multiple dynamic constraint edges exceeds the corresponding preset threshold at the same time, and the change range of a single dynamic constraint edge exceeds half of the corresponding preset threshold, a global topology reconstruction signal is generated, where multiple dynamic constraint edges represent two or more; otherwise, a local constraint adjustment signal is generated.

8. The industrial product production management system based on the industrial Internet according to claim 1 is characterized in that: The intelligent decision-making module (300) comprises a global unit (301), wherein the global unit (301) is used to respond to a global topology reconstruction signal and trigger a multi-objective optimization calculation based on transfer learning, wherein the multi-objective optimization calculation of the transfer learning comprises: Achieve energy efficiency strategy migration across production line equipment through feature alignment technology; Initializing the population based on the Pareto frontier solution set accelerates convergence; Dynamically adjust the weight coefficients of energy consumption targets and production efficiency.

9. The industrial product production management system based on the industrial Internet according to claim 1 is characterized in that: The intelligent decision module (300) comprises a local unit (302), wherein the local unit (302) is used to respond to a local constraint adjustment signal and perform incremental rescheduling with a memory function in a process-related subgraph, wherein the incremental rescheduling comprises: Avoid duplication of local optimal solutions through tabu search algorithm; The Lagrange multiplier method is used to adjust the constraint edge relaxation; The original scheduling parameters of nodes outside the process-related subgraph are retained.

10. The industrial product production management system based on the industrial Internet according to claim 9 is characterized in that: The generation process of the process-related subgraph specifically includes: According to the dynamic constraint edge type in the local constraint adjustment signal, the topological nodes are located; starting from the trigger node, the strongly associated nodes are tracked along the dynamic constraint edge through the graph traversal algorithm; only the nodes and edges that directly interact with the trigger constraint edge are included to generate a process-related subgraph.

Images