Heterogeneous data optimization management method and system based on AI drive

Abstract

The invention provides an AI-driven heterogeneous data optimization management method and system, and the method comprises the steps: extracting the production dynamics, equipment association and data demand dimensions in a flexible production scene through AI modeling, constructing a scene knowledge graph, carrying out the analysis of data from a heterogeneous terminal, and converting the data into a standard format matched with a current production scene; according to a multi-dimensional association relationship in the scene knowledge graph, performing exception recognition, deletion repair or redundancy elimination on the data in the standardized format to realize quality optimization, allocating storage resources for the data subjected to quality optimization, and pre-loading associated data before nodes are produced; according to the method, decision information for production scheduling, equipment predictive maintenance and product quality tracing is output in combination with the incidence relation among equipment, processes and data defined in the scene knowledge graph, and the timeliness and accuracy of core services such as production scheduling, equipment maintenance and quality tracing are improved by constructing and updating the scene knowledge graph in real time.

Description

Technical Field

[0001] This invention relates to the field of heterogeneous data optimization technology, and in particular to an AI-driven heterogeneous data optimization management method and system. Background Technology

[0002] In the flexible production scenarios of smart manufacturing plants, production tasks, process routes, and equipment combinations are frequently and dynamically adjusted. Existing heterogeneous data management systems rely on predefined data parsing rules, static quality processing models, and fixed storage strategies, which cannot adapt to the real-time changes in data generation logic, relationships, and usage requirements during the production process.

[0003] This leads to data access delays when new scenarios emerge, a disconnect between processing results and actual working conditions, and a mismatch between storage resources and business needs, severely impacting the timeliness and accuracy of core business processes such as production scheduling, equipment maintenance, and quality traceability. Summary of the Invention

[0004] This invention provides an AI-driven method and system for optimizing and managing heterogeneous data, which addresses the shortcomings of existing technologies in terms of timeliness and accuracy in core business processes such as production scheduling, equipment maintenance, and quality traceability.

[0005] In a first aspect, the present invention provides an AI-driven method for optimizing and managing heterogeneous data, comprising: By using AI modeling to extract production dynamics, equipment associations, and data demand dimensions in flexible production scenarios, a scenario knowledge graph representing multidimensional relationships is constructed and updated in real time. By utilizing the device combination and communication protocol information in the scenario knowledge graph, data from heterogeneous terminals is parsed and converted into a standardized format adapted to the current production scenario; Based on the multidimensional relationships in the scene knowledge graph, anomaly identification, missing data repair, or redundancy removal are performed on the standardized data to achieve quality optimization. In response to the annotations on data real-time performance, storage period and access frequency in the scenario knowledge graph, storage resources are allocated for the data that has undergone quality optimization, and related data is preloaded before the production node; Based on the relationships between equipment, processes, and data defined in the scenario knowledge graph, decision information is output for production scheduling, predictive maintenance of equipment, and product quality traceability.

[0006] According to the present invention, an AI-driven heterogeneous data optimization management method is provided, wherein the method extracts production dynamics, equipment associations, and data demand dimensions in a flexible production scenario through AI modeling, and constructs and updates a scenario knowledge graph representing multidimensional relationships in real time, including: Extract production plans, process documents, equipment ledgers, and real-time production progress data from the factory information system; Based on the extracted data, AI modeling is used to automatically identify the current production batch, process route, and equipment combination, forming a dynamic dimension of production. Based on the equipment ledger and process information, AI modeling is used to automatically construct the linkage logic and parameter matching relationship between different equipment, forming the equipment association dimension. Based on business rules and historical scenarios, AI modeling is used to automatically determine the real-time and storage cycle requirements of data for each business scenario, forming a data demand dimension. By linking and integrating the aforementioned production dynamics dimension, equipment association dimension, and data demand dimension, an initial scenario knowledge graph is constructed.

[0007] According to the present invention, an AI-driven heterogeneous data optimization management method is provided, wherein the real-time updating of the scene knowledge graph representing multidimensional relationships includes: Continuously monitor dynamic changes in the production scenario, including the addition of new production batches, adjustments to process routes, and replacement of equipment combinations; When the dynamic changes are detected, the corresponding equipment nodes and process nodes are automatically added or updated in the scene knowledge graph. Based on the changed scenario, automatically establish data interaction rules and relationships between new nodes and existing nodes; The newly added nodes are labeled with corresponding data real-time and storage cycle requirements to complete the real-time update of the scene knowledge graph.

[0008] According to the AI-driven heterogeneous data optimization management method provided by the present invention, the method further includes labeling the newly added nodes with corresponding data real-time performance and storage cycle requirements tags: Based on the urgency of production tasks, the priority of different data requirement tags in the same scenario is dynamically sorted. The results of the dynamic sorting are associated with and stored in the corresponding production nodes and data demand dimension nodes in the scenario knowledge graph; When the production scenario or task priority changes, the priority ranking of the data requirement tags is recalculated and updated.

[0009] According to the present invention, an AI-driven heterogeneous data optimization management method is provided, wherein the step of performing anomaly identification, missing data repair, or redundancy removal on the standardized data based on the multidimensional relationships in the scene knowledge graph includes: The predefined equipment operating conditions and data association rules in the scenario knowledge graph are invoked to determine whether the current data value matches the corresponding production scenario stage and to identify scenario mismatch anomalies. For the identified periods of missing data, a data sequence that fits the current working conditions is generated and repaired based on the similar historical scenarios and parameters associated in the scene knowledge graph. Based on the data demand priority and node association rules defined in the scenario knowledge graph, duplicate data records or debugging data unrelated to production generated in different processes are identified and eliminated.

[0010] According to the present invention, an AI-driven heterogeneous data optimization management method is provided, wherein determining whether the current data value matches the corresponding production scenario stage includes: Obtain the normal numerical range of the current scene stage of the device from the scene knowledge graph; The reported data is compared with the normal value range, and the device status data is correlated to determine whether there is an anomaly.

[0011] According to the present invention, an AI-driven heterogeneous data optimization management method is provided, wherein the step of generating a data sequence that fits the current working condition for repair includes: Match historical scenarios with the same process route, material material, and similar equipment load as the current production batch from the scenario knowledge graph, and generate repair data based on the historical data; The process of identifying and removing duplicate data records or debugging data unrelated to production generated in different processes includes: Based on the process priorities marked in the scenario knowledge graph, the quality inspection data of high-priority processes are retained, and non-production data generated by the equipment during the commissioning phase is filtered out according to the production scenario tags.

[0012] According to the present invention, an AI-driven heterogeneous data optimization management method is provided, wherein responding to the annotations of data real-time performance, storage period, and access frequency in the scene knowledge graph, allocating storage resources for the quality-optimized data, and preloading related data before the production node includes: Analyze the real-time tags and storage cycle tags used to annotate production data in the aforementioned scenario knowledge graph; If data is marked as having high real-time requirements, high access frequency, and short storage cycle, it will be allocated to the high-speed solid-state storage area. If data is marked as having low access frequency, long storage period, and high security, it will be allocated to a low-cost distributed storage area. When the scenario knowledge graph detects that the production batch status has changed from in progress to completed, it automatically migrates the storage strategy of the corresponding data from the high-speed solid-state storage area to the distributed storage area.

[0013] According to the present invention, an AI-driven heterogeneous data optimization management method is provided, wherein the method combines the relationships between equipment, processes, and data defined in the scenario knowledge graph to output decision information for production scheduling, predictive equipment maintenance, and product quality traceability, including: In the scenario of predictive maintenance of equipment, the historical failure probability model of the target equipment in a specific production scenario is queried in the scenario knowledge graph, and maintenance suggestions with execution time windows are generated by combining real-time operation data and batch gaps in subsequent production plans. In the dynamic production scheduling scenario, when the scenario knowledge graph detects an abnormal status of a device node, it matches alternative devices based on the process connection rules and the equivalence of device functions in the graph, and generates a rearranged production scheduling scheme based on the current load data of each alternative device and the priority tags of the orders to be processed. In the product quality traceability scenario, based on the unique identifier of the defective product, the associated processing equipment nodes, process parameter nodes, material batch nodes, and quality inspection record nodes of each process are traced in reverse in the scenario knowledge graph to locate and output the source link that caused the defect and the associated data chain.

[0014] Secondly, the present invention provides an AI-driven heterogeneous data optimization management system, comprising: The module is used to extract production dynamics, equipment associations and data demand dimensions in flexible production scenarios through AI modeling, and to build and update a scenario knowledge graph that represents multi-dimensional relationships in real time. The parsing module is used to parse data from heterogeneous terminals using the device combination and communication protocol information in the scenario knowledge graph, and convert it into a standardized format that is compatible with the current production scenario. The optimization module is used to perform anomaly identification, missing data repair, or redundancy removal on the standardized data based on the multidimensional relationships in the scene knowledge graph, thereby achieving quality optimization. The response module is used to respond to the annotations on data real-time performance, storage period and access frequency in the scene knowledge graph, allocate storage resources for the data after quality optimization, and preload related data before the production node; The output module is used to combine the relationships between equipment, processes and data defined in the scenario knowledge graph to output decision information for production scheduling, predictive maintenance of equipment and product quality traceability.

[0015] Thirdly, the present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the AI-driven heterogeneous data optimization management method as described above.

[0016] Fourthly, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the AI-driven heterogeneous data optimization management method as described above.

[0017] Fifthly, the present invention also provides a computer program product, including a computer program that, when executed by a processor, implements the AI-driven heterogeneous data optimization management method as described above.

[0018] This invention provides an AI-driven heterogeneous data optimization management method and system. By constructing and updating a scenario knowledge graph in real time, it automatically captures changes in the production scenario and drives adaptive adjustments throughout the entire process of subsequent data parsing, quality optimization, storage scheduling, and decision generation. This solves the fundamental problems of traditional systems that rely on manual presets and have delayed responses. Through adaptive parsing and standardization, it ensures that multi-source heterogeneous data can be quickly and accurately converted into a format consistent with the current business scenario. Furthermore, by combining the association rules of the scenario graph for quality optimization, it makes anomaly identification, missing data repair, and redundancy removal more in line with actual production logic, significantly improving the timeliness and accuracy of core business operations such as production scheduling, equipment maintenance, and quality traceability. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0020] Figure 1 This is a flowchart illustrating the AI-driven heterogeneous data optimization management method provided in this embodiment; Figure 2 This is the structural intent of the AI-driven heterogeneous data optimization management system provided in this embodiment; Figure 3 This is a schematic diagram of the structure of the electronic device provided in this embodiment. Detailed Implementation

[0021] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0022] Figure 1 This is a flowchart illustrating the AI-driven heterogeneous data optimization management method provided in this embodiment.

[0023] like Figure 1 As shown in the figure, an AI-driven heterogeneous data optimization management method provided by this invention mainly includes the following steps: 101. Extract production dynamics, equipment associations, and data demand dimensions from flexible production scenarios through AI modeling, and construct and update a scenario knowledge graph that represents multi-dimensional relationships in real time.

[0024] AI modeling accurately captures the dynamic characteristics of flexible production scenarios, providing precise scenario-based information for subsequent data processing.

[0025] Specifically, AI automatically integrates with the factory's MES system, extracting production plans, process documents, equipment ledgers, and real-time production progress data from the factory information system. Production plans include production batch arrangements corresponding to orders; process documents cover the processing flow and parameter standards for each product; equipment ledgers record static information such as equipment models, functions, and installation locations; and real-time production progress data reflects the currently executing processes and the amount of production tasks completed. Based on the extracted data, AI automatically identifies the current production batch, process route, and equipment combination using deep learning algorithms, forming a dynamic production dimension and clarifying the core execution elements of the current production task. Based on the equipment ledger and process information, AI modeling automatically constructs the linkage logic and parameter matching relationships between different devices, forming an equipment association dimension and clarifying the data interaction requirements and parameter adaptation standards for equipment during process connections. Based on business rules and historical scenarios, AI modeling automatically determines the real-time and storage cycle requirements for data in each business scenario, forming a data demand dimension and clarifying the core indicators for data use in different business scenarios.

[0026] AI integrates and correlates the dimensions of production dynamics, equipment association, and data demand to construct an initial scenario knowledge graph. In the graph, the production dynamics dimension corresponds to "production batch node", "process route node", "equipment combination node" and "work order priority node", the equipment association dimension corresponds to "equipment node", "linkage logic node" and "parameter matching rule node", and the data demand dimension corresponds to "real-time demand node" and "storage cycle demand node". Each node is connected through association rules to form a complete scenario-data association system.

[0027] AI continuously monitors dynamic changes in the production scenario, including the addition of new production batches, adjustments to process routes, and replacement of equipment combinations. When a dynamic change is detected, the system automatically adds or updates corresponding equipment and process nodes in the scenario knowledge graph. For example, when adding a "customized batch of aerospace parts," a "ultrasonic testing equipment node" and an "ultrasonic testing process node" are added simultaneously. Based on the changed scenario, the system automatically establishes data interaction rules and relationships between new and existing nodes. For instance, it establishes a relationship between the "ultrasonic testing equipment node" and the "machine tool node," clarifying the interaction rules for the machine tool to transmit processing parameter data to the ultrasonic testing equipment and for the ultrasonic testing equipment to transmit test result data to the quality inspection system. New nodes are labeled with corresponding data real-time and storage cycle requirements. For example, nodes related to the "customized batch of aerospace parts" are labeled with "high real-time performance, high precision, and long storage cycle," completing the real-time update of the scenario knowledge graph.

[0028] Simultaneously, based on the urgency of production tasks, the priority of different data requirement tags within the same scenario is dynamically sorted. For example, the production scheduling data for urgent orders is prioritized higher than the historical archived data for regular orders, and the operational data of critical equipment is prioritized higher than the status data of auxiliary equipment. The results of this dynamic sorting are associated with and stored in the corresponding production nodes and data requirement dimension nodes in the scenario knowledge graph. When the production scenario or task priority changes, the priority ranking of the data requirement tags is recalculated and updated to ensure that the priority of data requirements in the graph always remains consistent with the core needs of the current production task.

[0029] The entire process can capture scene features and update the map without human intervention, solving the problem that traditional systems cannot adapt to scene changes in a timely manner, and providing accurate basis for subsequent data processing that fits the actual working conditions.

[0030] 102. Utilize the device combination and communication protocol information in the scenario knowledge graph to parse data from heterogeneous terminals and convert it into a standardized format that is compatible with the current production scenario.

[0031] By leveraging scenario-based knowledge graphs, heterogeneous data can be dynamically accessed with zero manual configuration.

[0032] Specifically, AI extracts the communication protocol type and data dimension requirements of the current equipment combination from the scene knowledge graph. The communication protocols include Modbus, Profinet, OPC UA, TCP / IP, etc., and the data dimension requirements are determined based on the current production batch and process route. For example, the combination of "machine tool + vision inspection equipment" corresponds to dimensions such as cutting parameter time-series data and defect image feature data. Based on the protocol type, AI automatically loads the corresponding parsing algorithm to extract the core data of each device. Simultaneously, according to the process association rules in the graph, it adds a unified work order ID and process node identifier to the parsed data of different devices, achieving accurate cross-device data association. For example, for the combination of machine tool and vision inspection equipment, AI identifies the Profinet protocol of the machine tool and the TCP / IP protocol of the vision equipment, extracting cutting parameter time-series data (such as spindle speed and cutting temperature) and defect image feature data (such as defect size and location) respectively, and adding the same "work order ID + process node identifier" to both types of data to ensure rapid association and querying in subsequent business scenarios.

[0033] When a factory introduces new types of industrial robots or adds special processes, there is no need to halt production for manual adjustments. AI retrieves equipment adaptation logic and process data requirements for similar scenarios from a scene knowledge graph, and automatically generates data parsing rules for new equipment and processes through transfer learning algorithms. For example, when adding a collaborative robot, AI obtains the equipment linkage rules for the assembly process, automatically analyzes the robot's joint angles and load data, and establishes a correlation with the machining accuracy data of upstream machine tools. When adding a 3D printing process, it automatically analyzes core data dimensions such as layer height, temperature, and printing speed during the printing process, ensuring that data from new equipment and processes are quickly integrated into the system.

[0034] AI uses business requirement tags from a scenario knowledge graph to convert heterogeneous data into a standardized format suitable for the current scenario. For unstructured quality inspection voice recordings, it transforms them into structured data with defect type, location, and severity, meeting the query and analysis needs of quality traceability scenarios. For less structured sensor time-series data, it converts it into aggregated data with device ID, process node, time window, and mean / peak value, meeting the need for quickly viewing equipment operating status in equipment maintenance scenarios. For semi-structured supply chain data, it transforms it into standardized data with "supplier ID-material batch-logistics node-timestamp," meeting the needs of supply chain traceability scenarios.

[0035] By standardizing the format, heterogeneous data can be quickly and automatically integrated, flexibly adapting to new equipment and processes, eliminating data format differences, and significantly improving data integration efficiency.

[0036] 103. Based on the multidimensional relationships in the scene knowledge graph, perform anomaly identification, missing data repair, or redundancy removal on standardized data to achieve quality optimization.

[0037] Specifically, the AI ​​invokes predefined equipment operating conditions and data association rules from the scene knowledge graph to obtain the normal numerical range for the current scene stage of the equipment. For example, when the equipment is in the "warm-up stage," the normal range for spindle speed is the low speed range; when it is in the "high-intensity cutting stage," the normal range for spindle speed is the high speed range. The data reported by the sensors is compared with the normal numerical range, and associated with equipment status data such as the current operating mode and load condition to determine if there are any anomalies. If the data value exceeds the normal range and does not match the current scene stage, it is determined to be a scene mismatch anomaly, and possible causes such as sensor false alarms or operating condition identification errors are marked.

[0038] For example, when the equipment is in the preheating stage, the spindle speed should be in the low range according to the graph. If the sensor reports high-frequency, high-speed data, the AI ​​will directly determine it as abnormal and mark it as a possible cause of "sensor false alarm". If the equipment is in the high-intensity cutting stage, the spindle speed data is lower than the normal range, and the equipment load data shows that it is currently operating at full load, then it will mark it as a possible cause of "operating condition identification error".

[0039] For identified periods of missing data, such as data interruptions caused by batch changes or equipment debugging, AI matches historical scenarios with the same process route, material material, and similar equipment load as the current production batch from the scenario knowledge graph, retrieves similar data in that scenario, and combines it with real-time parameters of the current scenario, such as current equipment load, material temperature, and ambient humidity, to generate a data sequence that fits the current working conditions to fill the missing data through a time-series prediction algorithm.

[0040] For example, when a batch change results in a 5-minute loss of cutting temperature data, AI generates a temperature change curve that fits the current operating conditions based on historical batch temperature data with "same process route + same materials + similar equipment load," ensuring the accuracy of the repaired data and avoiding data distortion caused by using simple linear interpolation.

[0041] AI identifies and eliminates duplicate data records or debugging data unrelated to production generated in different processes based on the data demand priority and node association rules defined in the scene knowledge graph. For duplicate defect records generated by multiple quality inspection processes of the same product, the quality inspection data of high-priority processes is retained and duplicate records of intermediate processes are eliminated according to the process priority marked in the graph. For non-production data generated by equipment during the debugging phase, it is automatically filtered according to production scene tags to avoid occupying effective storage resources. For duplicate data of the same dimension collected by different equipment, the data collected by key equipment is retained and duplicate data collected by auxiliary equipment is eliminated according to the data demand priority.

[0042] By improving the accuracy of anomaly data identification and the fit of missing data repair, redundant data can be effectively eliminated, avoiding data distortion caused by common processing methods, and providing accurate and reliable data support for business decisions.

[0043] 104. In response to the annotation of data real-time performance, storage period and access frequency in the knowledge graph, storage resources are allocated to the data that has been optimized for quality, and related data is preloaded before the production node.

[0044] Specifically, the AI ​​analyzes the real-time and storage cycle tags labeled for production data in the scenario knowledge graph, and dynamically allocates storage resources based on the data's access frequency. Data marked as high real-time, high access frequency, and short storage cycle, such as production scheduling data for urgent orders, is allocated to high-speed solid-state storage to ensure low latency during data retrieval. Data marked as low access frequency, long storage cycle, and high security, such as quality traceability data for core products, is allocated to low-cost distributed storage with added encryption and backup strategies to ensure the security and integrity of long-term data storage. When the scenario knowledge graph detects that a production batch status changes from "in progress" to "completed," the AI ​​automatically migrates the corresponding data's storage strategy from high-speed solid-state storage to distributed storage, adapting to the change from real-time access needs to traceability query needs.

[0045] Based on the production plan nodes in the scene's knowledge graph, AI identifies and preloads the relevant data needed for subsequent production stages. For example, if the graph shows that a batch of high-precision parts processing will start in one hour, AI preloads the process parameter data for that batch, historical fault data of the equipment used, processing quality data of similar batches, and inspection data of material batches. If the production plan is temporarily adjusted, such as an earlier start of a batch or a change in the process route, AI updates the preloaded list in real time, deleting unnecessary data and adding new required data to ensure that the required data can be directly accessed when production starts, reducing data access latency.

[0046] AI continuously monitors the data volume trends in the current production scenario, analyzing data such as production plans, order quantities, and equipment operating status to predict data volume changes. When a large customized order is received, and a significant increase in data volume is predicted, the corresponding storage resources are automatically expanded in advance to avoid insufficient storage affecting data storage. When orders are completed and production enters the off-season, and data volume decreases significantly, redundant storage resources are automatically released to reduce storage costs and achieve on-demand allocation of storage resources.

[0047] By improving storage resource utilization, reducing storage costs, and minimizing data retrieval latency, the timeliness requirements of flexible production can be met.

[0048] 105. Based on the relationships between equipment, processes and data defined in the scenario knowledge graph, output decision information for production scheduling, predictive maintenance of equipment and product quality traceability.

[0049] Specifically, in predictive maintenance scenarios, AI queries the scenario knowledge graph to identify historical failure probability models of the target equipment under specific production scenarios. Examples include association rules such as "the failure probability of a machine tool spindle surges after 100 hours of continuous operation under high-intensity cutting conditions" and "the wear probability of an industrial robot joint increases after 500 cycles of frequent start-stop operations." Combined with the equipment's current real-time operating data, such as cumulative operating time, spindle temperature, vibration frequency, joint operation counts, and batch intervals in subsequent production plans, maintenance recommendations with clearly defined execution time windows are generated.

[0050] For example, the scenario knowledge graph shows that the probability of failure of a machine tool spindle increases dramatically after continuous operation for a certain period of time in a high-intensity cutting scenario. The current machine tool spindle has been running continuously for 98 hours. The subsequent production plan shows that there is a batch gap in 2 hours. The AI ​​suggests that the spindle lubrication maintenance be performed during the batch gap in the next 2 hours to avoid the maintenance operation from conflicting with the production task and reduce production interruption.

[0051] In dynamic production scheduling scenarios, when the scenario knowledge graph detects abnormal equipment node status, such as a sudden machine tool failure or a robot shutdown, AI quickly matches alternative equipment based on the process connection rules and equipment functional equivalence relationships in the graph. This includes backup machine tools for the same process and other robots with similar functions. The system then retrieves the current load data and priority tags of pending orders for each alternative device to generate a rearranged production scheduling plan. This plan clarifies the production task allocation, process connection sequence, and completion time nodes for each device, ensuring timely delivery of core orders and minimizing losses caused by production interruptions.

[0052] In product quality traceability scenarios, when a defective product is discovered, AI uses the product's unique identifier to trace back to related processing equipment nodes, process parameter nodes, material batch nodes, and quality inspection record nodes for each process within the scenario's knowledge graph. Through comprehensive analysis of the associated data, it locates and outputs the source of the defect and the associated data chain. For example, it determines that the hardness of a batch of materials is substandard, leading to a processing defect, or that a machine tool's cutting parameters are off, causing excessive surface roughness, providing precise evidence for subsequent rectification.

[0053] By ensuring that decision-making information is highly aligned with actual production, production interruptions can be reduced, on-time delivery rates can be improved, rework costs can be lowered, and production and operational efficiency can be enhanced.

[0054] Figure 2 This is a schematic diagram of the structure of the AI-driven heterogeneous data optimization management system provided in this embodiment.

[0055] like Figure 2 As shown in the figure, this embodiment provides an AI-driven heterogeneous data optimization management system, including: Module 201 is used to extract the dimensions of production dynamics, equipment association and data demand in flexible production scenarios through AI modeling, and to build and update the scenario knowledge graph that represents multi-dimensional relationships in real time. The parsing module 202 is used to parse data from heterogeneous terminals using device combination and communication protocol information in the scenario knowledge graph, and convert it into a standardized format that is compatible with the current production scenario. Optimization module 203 is used to perform anomaly identification, missing data repair or redundancy removal on standardized format data based on the multi-dimensional relationships in the scene knowledge graph, so as to achieve quality optimization. The response module 204 is used to annotate the data real-time performance, storage period and access frequency in the response scenario knowledge graph, allocate storage resources for the quality-optimized data, and preload related data before the production node; The output module 205 is used to combine the relationships between equipment, processes and data defined in the scenario knowledge graph to output decision information for production scheduling, predictive maintenance of equipment and product quality traceability.

[0056] Figure 3 This is a schematic diagram of the structure of the electronic device provided in this embodiment.

[0057] like Figure 3As shown, the electronic device may include a processor 301, a communications interface 302, a memory 303, and a communication bus 304. The processor 301, communications interface 302, and memory 303 communicate with each other via the communication bus 304. The processor 301 can call logical instructions from the memory 303 to execute an AI-driven heterogeneous data optimization management method.

[0058] Furthermore, the logical instructions in the aforementioned memory 303 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, essentially, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0059] On the other hand, the present invention also provides a computer program product, which includes a computer program that can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer is able to execute the AI-driven heterogeneous data optimization management method provided by the above methods.

[0060] In another aspect, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, is implemented to perform the AI-driven heterogeneous data optimization management method provided by the above methods.

[0061] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0062] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0063] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. An AI-driven method for optimizing and managing heterogeneous data, characterized in that, include: By using AI modeling to extract production dynamics, equipment associations, and data demand dimensions in flexible production scenarios, a scenario knowledge graph representing multidimensional relationships is constructed and updated in real time. By utilizing the device combination and communication protocol information in the scenario knowledge graph, data from heterogeneous terminals is parsed and converted into a standardized format adapted to the current production scenario; Based on the multidimensional relationships in the scene knowledge graph, anomaly identification, missing data repair, or redundancy removal are performed on the standardized data to achieve quality optimization. In response to the annotations on data real-time performance, storage period and access frequency in the scenario knowledge graph, storage resources are allocated for the data that has undergone quality optimization, and related data is preloaded before the production node; Based on the relationships between equipment, processes, and data defined in the scenario knowledge graph, decision information is output for production scheduling, predictive maintenance of equipment, and product quality traceability.

2. The method according to claim 1, characterized in that, The process involves extracting production dynamics, equipment relationships, and data demand dimensions from flexible production scenarios through AI modeling, constructing and updating a scenario knowledge graph representing multidimensional relationships in real time, including: Extract production plans, process documents, equipment ledgers, and real-time production progress data from the factory information system; Based on the extracted data, AI modeling is used to automatically identify the current production batch, process route, and equipment combination, forming a dynamic dimension of production. Based on the equipment ledger and process information, AI modeling is used to automatically construct the linkage logic and parameter matching relationship between different equipment, forming the equipment association dimension. Based on business rules and historical scenarios, AI modeling is used to automatically determine the real-time and storage cycle requirements of data for each business scenario, forming a data demand dimension. By linking and integrating the aforementioned production dynamics dimension, equipment association dimension, and data demand dimension, an initial scenario knowledge graph is constructed.

3. The method according to claim 2, characterized in that, The real-time updated scene knowledge graph representing multidimensional relationships includes: Continuously monitor dynamic changes in the production scenario, including the addition of new production batches, adjustments to process routes, and replacement of equipment combinations; When the dynamic changes are detected, the corresponding equipment nodes and process nodes are automatically added or updated in the scene knowledge graph. Based on the changed scenario, automatically establish data interaction rules and relationships between new nodes and existing nodes; The newly added nodes are labeled with corresponding data real-time and storage cycle requirements to complete the real-time update of the scene knowledge graph.

4. The method according to claim 3, characterized in that, The method of labeling the newly added nodes with corresponding data real-time and storage cycle requirement tags also includes: Based on the urgency of production tasks, the priority of different data requirement tags in the same scenario is dynamically sorted. The results of the dynamic sorting are associated with and stored in the corresponding production nodes and data demand dimension nodes in the scenario knowledge graph; When the production scenario or task priority changes, the priority ranking of the data requirement tags is recalculated and updated.

5. The method according to claim 1, characterized in that, The step of performing anomaly identification, missing data repair, or redundancy removal on the standardized data based on the multidimensional relationships in the scene knowledge graph includes: The predefined equipment operating conditions and data association rules in the scenario knowledge graph are invoked to determine whether the current data value matches the corresponding production scenario stage and to identify scenario mismatch anomalies. For the identified periods of missing data, a data sequence that fits the current working conditions is generated and repaired based on the similar historical scenarios and parameters associated in the scene knowledge graph. Based on the data demand priority and node association rules defined in the scenario knowledge graph, duplicate data records or debugging data unrelated to production generated in different processes are identified and eliminated.

6. The method according to claim 5, characterized in that, The step of determining whether the current data value matches the corresponding production scenario stage includes: Obtain the normal numerical range of the current scene stage of the device from the scene knowledge graph; The reported data is compared with the normal value range, and the device status data is correlated to determine whether there is an anomaly.

7. The method according to claim 6, characterized in that, The repair process, which generates a data sequence that matches the current operating conditions, includes: Match historical scenarios with the same process route, material material, and similar equipment load as the current production batch from the scenario knowledge graph, and generate repair data based on the historical data; The process of identifying and removing duplicate data records or debugging data unrelated to production generated in different processes includes: Based on the process priorities marked in the scenario knowledge graph, the quality inspection data of high-priority processes are retained, and non-production data generated by the equipment during the commissioning phase is filtered out according to the production scenario tags.

8. The method according to claim 1, characterized in that, The response to the annotations on data real-time performance, storage period, and access frequency in the scenario knowledge graph, the allocation of storage resources for the quality-optimized data, and the preloading of associated data before the production node include: Analyze the real-time tags and storage cycle tags used to annotate production data in the aforementioned scenario knowledge graph; If data is marked as having high real-time requirements, high access frequency, and short storage cycle, it will be allocated to the high-speed solid-state storage area. If data is marked as having low access frequency, long storage period, and high security, it will be allocated to a low-cost distributed storage area. When the scenario knowledge graph detects that the production batch status has changed from in progress to completed, it automatically migrates the storage strategy of the corresponding data from the high-speed solid-state storage area to the distributed storage area.

9. The method according to claim 1, characterized in that, The process combines the relationships between equipment, processes, and data defined in the scenario knowledge graph to output decision information for production scheduling, predictive equipment maintenance, and product quality traceability, including: In the scenario of predictive maintenance of equipment, the historical failure probability model of the target equipment in a specific production scenario is queried in the scenario knowledge graph, and maintenance suggestions with execution time windows are generated by combining real-time operation data and batch gaps in subsequent production plans. In the dynamic production scheduling scenario, when the scenario knowledge graph detects an abnormal status of a device node, it matches alternative devices based on the process connection rules and the equivalence of device functions in the graph, and generates a rearranged production scheduling scheme based on the current load data of each alternative device and the priority tags of the orders to be processed. In the product quality traceability scenario, based on the unique identifier of the defective product, the associated processing equipment nodes, process parameter nodes, material batch nodes, and quality inspection record nodes of each process are traced in reverse in the scenario knowledge graph to locate and output the source link that caused the defect and the associated data chain.

10. An AI-driven heterogeneous data optimization management system, characterized in that, include: The module is used to extract production dynamics, equipment associations and data demand dimensions in flexible production scenarios through AI modeling, and to build and update a scenario knowledge graph that represents multi-dimensional relationships in real time. The parsing module is used to parse data from heterogeneous terminals using the device combination and communication protocol information in the scenario knowledge graph, and convert it into a standardized format that is compatible with the current production scenario. The optimization module is used to perform anomaly identification, missing data repair, or redundancy removal on the standardized data based on the multidimensional relationships in the scene knowledge graph, thereby achieving quality optimization. The response module is used to respond to the annotations on data real-time performance, storage period and access frequency in the scene knowledge graph, allocate storage resources for the data that has been optimized for quality, and preload related data before the production node; The output module is used to combine the relationships between equipment, processes and data defined in the scenario knowledge graph to output decision information for production scheduling, predictive maintenance of equipment and product quality traceability.

Images