Production traceability control method and system of digital intelligent chemical plant under Internet of Things

Abstract

The invention discloses a production traceability control method and system for a digital intelligent chemical plant under the Internet of Things, and relates to the technical field of production traceability control methods, and the method comprises the steps: determining a current production path according to the current working state of the digital intelligent chemical plant and the model of a workpiece; and the abnormal production area of the workpiece is determined based on the current production path and the abnormal production path, so that the accuracy of the abnormal production area of the workpiece is improved. Therefore, the repair time node of each abnormal factory component is determined based on the abnormal factory component corresponding to the production abnormal area, the Internet of Things and the current production position of the workpiece, and the repair workload of each abnormal factory component is determined according to the repair time node of each abnormal factory component and the production process table corresponding to the workpiece; according to the repair workload of each abnormal factory part and the Internet of Things, the autonomous repair strategy of the digital intelligent chemical plant for the workpiece is determined, and the accuracy of the autonomous repair strategy of the digital intelligent chemical plant for the workpiece is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of production traceability control methods, and in particular to a production traceability control method and system for a digital factory under the Internet of Things. Background Art

[0002] With the development of science and technology, digital factories are gradually being applied to people's lives and can produce workpieces in an unmanned state. Digital factories deeply integrate the physical production environment with the digital virtual environment, and comprehensively digitally transform the factory's production, management, and operations. This is a modern factory model. In existing technologies, workpieces are autonomously processed in digital factories and produced in corresponding factory components according to the various process items in the production process table. However, the production of workpieces in digital factories contains some abnormal production data, which triggers the digital factory to stop production of the workpiece based on the abnormal production data. It is impossible to trace and control the abnormal production data, which affects the accuracy of the digital factory's autonomous repair strategy for the workpiece. Summary of the Invention

[0003] The purpose of the present invention is to overcome the shortcomings of the existing technology. The present invention provides a production traceability control method and system for a digital factory under the Internet of Things.

[0004] An embodiment of the present invention provides a production traceability control method for a digital factory under the Internet of Things, including: collecting multiple production data combinations of the digital factory at different production nodes, and constructing a digital database based on each production data combination and the Internet of Things; tracing the production of multiple abnormal production data in each production data combination based on the digital database to determine the corresponding abnormal production node, and determining the abnormal production path according to the position of each abnormal production node and the abnormal range of the abnormal production data; determining the current production path according to the current working status of the digital factory and the model of the workpiece, and determining the production abnormality area of ​​the workpiece based on the current production path and the abnormal production path; determining the repair time node of each abnormal factory component based on the abnormal factory component corresponding to the production abnormality area, the Internet of Things and the current production position of the workpiece, and determining the repair workload of each abnormal factory component according to the repair time node of each abnormal factory component and the production process table corresponding to the workpiece; determining the autonomous repair strategy of the digital factory for the workpiece based on the repair workload of each abnormal factory component and the Internet of Things, without affecting the normal production of the workpiece, so as to repair each abnormal factory component in advance.

[0005] An embodiment of the present invention provides a production traceability control system for a digital and intelligent factory under the Internet of Things. The production traceability control system for a digital and intelligent factory under the Internet of Things is applied to the above-mentioned production traceability control method for a digital and intelligent factory under the Internet of Things. The production traceability control system for a digital and intelligent factory under the Internet of Things includes: The digital and intelligent database module is used to collect multiple production data combinations at different production nodes of the digital and intelligent factory and build a digital and intelligent database based on each production data combination and the Internet of Things; The abnormal production path module is used to trace the production source of multiple abnormal production data in each production data combination based on the digital database to determine the corresponding abnormal production nodes. The abnormal production path is determined according to the location of each abnormal production node and the abnormal range of the abnormal production data; The production anomaly area module is used to determine the current production path based on the current working status of the digital factory and the model of the workpiece, and to determine the production anomaly area of ​​the workpiece based on the current production path and the abnormal production path; A repair workload module is used to determine the repair time node of each abnormal factory component based on the abnormal factory component corresponding to the production abnormality area, the Internet of Things, and the current production location of the workpiece, and to determine the repair workload of each abnormal factory component based on the repair time node of each abnormal factory component and the production process table corresponding to the workpiece; The autonomous repair strategy module is used to determine the autonomous repair strategy of the digital factory for each abnormal factory component based on the repair workload of the component and the Internet of Things, without affecting the normal production of the component, so as to repair each abnormal factory component in advance.

[0006] Compared with the prior art, the present invention has the following beneficial effects: In an embodiment of the present invention, through the method in the embodiment of the present invention, the abnormal production path is determined according to the position of each abnormal production node and the abnormal range of the abnormal production data; the current production path is determined according to the current working status of the digital factory and the model of the workpiece, and the production abnormality area of ​​the workpiece is determined based on the current production path and the abnormal production path. Various abnormal production nodes are introduced, and the overall consideration of the current production path and the abnormal production path is compatible, thereby improving the accuracy of the production abnormality area of ​​the workpiece and performing traceability control on the production of the digital factory.

[0007] Therefore, based on the abnormal factory components, the Internet of Things and the current production position of the workpiece corresponding to the production abnormality area, the repair time node of each abnormal factory component is determined, and the repair workload of each abnormal factory component is determined according to the repair time node of each abnormal factory component and the production process table corresponding to the workpiece; according to the repair workload of each abnormal factory component and the Internet of Things, the autonomous repair strategy of the digital factory for the workpiece is determined, and the repair time node of each abnormal factory component is introduced, and the overall consideration of the repair workload of each abnormal factory component and the Internet of Things is realized, which improves the accuracy of the autonomous repair strategy of the digital factory for the workpiece without affecting the normal production of the workpiece, so as to repair each abnormal factory component in advance. BRIEF DESCRIPTION OF THE DRAWINGS

[0008] Figure 1 1 is a flow chart of a production traceability control method for a digital factory under the Internet of Things in an embodiment of the present invention; Figure 2 1 is a flow chart of step S11 in the production traceability control method of a digital factory under the Internet of Things in an embodiment of the present invention; Figure 3 1 is a flow chart of step S12 in the production traceability control method of a digital factory under the Internet of Things in an embodiment of the present invention; Figure 4 1 is a flow chart of step S13 in the production traceability control method of a digital factory under the Internet of Things in an embodiment of the present invention; Figure 5 1 is a flow chart of step S14 in the production traceability control method of a digital factory under the Internet of Things in an embodiment of the present invention; Figure 6 1 is a flow chart of step S15 in the production traceability control method of a digital factory under the Internet of Things in an embodiment of the present invention; Figure 7 This is a schematic diagram of the structural composition of the production traceability control system of the digital factory under the Internet of Things in an embodiment of the present invention. DETAILED DESCRIPTION

[0009] The technical solutions in the embodiments of the present invention will be described clearly and completely below with reference to the accompanying drawings in the embodiments of the present invention.

[0010] See also Figures 1 to 7 A production traceability control method for a digital factory under the Internet of Things is applied to production traceability control scenarios. The production traceability control method for a digital factory under the Internet of Things includes: Step S11: Collect multiple production data combinations from different production nodes of the digital factory, and build a digital database based on each production data combination and the Internet of Things; Step S12: Multiple abnormal production data in each production data combination are traced based on the digital database to determine the corresponding abnormal production nodes, and the abnormal production path is determined according to the location of each abnormal production node and the abnormal range of the abnormal production data; Step S13: Determine the current production path based on the current working state of the digital factory and the model of the workpiece, and determine the production abnormality area of ​​the workpiece based on the current production path and the abnormal production path; Step S14: Determine the repair time node of each abnormal factory component based on the abnormal factory component corresponding to the production abnormality area, the Internet of Things, and the current production location of the workpiece, and determine the repair workload of each abnormal factory component based on the repair time node of each abnormal factory component and the production process table corresponding to the workpiece; Step S15: Determine the autonomous repair strategy of the digital factory for the workpiece based on the repair workload of each abnormal factory component and the Internet of Things, without affecting the normal production of the workpiece, so as to repair each abnormal factory component in advance.

[0011] refer to Figure 2 In step S11, the specific steps are: S111: When the digital intelligent factory performs unmanned processing on a workpiece, the digital intelligent factory collects the digitized process of the workpiece, determines multiple autonomous production processes based on the division of the digitized process, and determines multiple production nodes based on the multiple autonomous production processes and the model of the workpiece; S112: Multiple production nodes are distributed in different locations of the digital factory and are targeted at different processing steps of the workpiece. The multiple production nodes are monitored in real time, and multiple production data of each production node are collected. In each production node, a corresponding production data combination is determined based on the synthesis of the multiple production data. S113: Construct a digital database based on the node position of the production node, the corresponding production data combination and the synthesis of the Internet of Things corresponding to the digital factory. The digital database is dynamically updated as the production data increases.

[0012] In an embodiment of the present application, when a digital factory performs unmanned processing on a workpiece, the digital factory introduces a digitized process for the workpiece. The digitized process refers to detailed digital instructions and blueprints that describe how to manufacture the workpiece, which usually come from CAD (computer-aided design) and CAM (computer-aided manufacturing) systems; the digitized process usually includes: geometric information, process information, and operation steps.

[0013] The purpose of division is to decompose a continuous or complex numerical process into smaller parts; the basis for division may include: physical equipment / station, operation type, logical stage, to output multiple autonomous production processes, which are relatively independent manufacturing units.

[0014] Multiple autonomous production processes and workpiece models have been introduced. Different workpiece models (even products from the same series) may have different sizes, materials, precision requirements or processing paths; therefore, the workpiece model information needs to be taken into consideration to ensure that the defined production nodes are targeted at the current specific workpiece.

[0015] A production node is the concrete manifestation of an autonomous production process in physical space or logical flow. It is usually a key point with a clear location and monitoring requirements. An autonomous production process corresponds to one or more production nodes. For example, the autonomous production process "CNC precision milling" corresponds to multiple production nodes on the CNC machine tool C, such as the "clamping and positioning completion point", "precision milling start point", and "precision milling end point". Each node requires the collection of different data (such as clamping force, vibration during processing, and dimensions after processing).

[0016] Furthermore, multiple production nodes are distributed in different locations of the digital factory, and each production node is associated with one or more specific processing procedures. For example, the "CNCA rough milling process point" corresponds to the rough milling processing procedure performed on the CNCA machine tool. This association ensures that the collected data can be correctly attributed to the corresponding manufacturing steps.

[0017] Real-time monitoring of multiple production nodes is achieved through IoT sensors and actuators deployed on each production node. Sensors convert physical world signals (such as temperature, vibration, and position) into digital signals. These devices are connected to edge computing nodes or directly to cloud platforms / data centers via wired or wireless networks. Each production node usually has more than one sensor or data source. For example, a CNC machine tool node simultaneously outputs multiple data streams such as spindle speed, feed speed, X / Y / Z axis position, current, and temperature. These data collectively describe the operating status of the node.

[0018] The synthesis of multiple production data is not a simple accumulation of data, but refers to the integration, association or calculation of multiple related data items from the same production node to form a more meaningful data structure or feature vector that can better reflect the overall status of the node; the synthesis method depends on the application requirements and analysis objectives; including: simple aggregation: such as calculating the average, maximum, minimum, standard deviation, etc. over a period of time; state description: combining multiple Boolean signals into a state description (such as "equipment is operating normally", "in processing", "waiting for materials"); feature extraction: using signal processing technology (such as FFT for frequency domain analysis) or statistical methods to extract key features from the original data; the synthesized "production data combination" should be more concise and more informative than the original data.

[0019] Therefore, a digital database is constructed based on the node position of the production node, the corresponding production data combination and the synthesis of the Internet of Things corresponding to the digital factory. The digital database is dynamically updated as the production data increases, and is compatible with the overall consideration of the node position of the production node, the corresponding production data combination and the synthesis of the Internet of Things corresponding to the digital factory, ensuring the accuracy of the construction of the digital database.

[0020] At this point, the node location of the production node, the corresponding production data combination and the Internet of Things corresponding to the digital factory are introduced. Each production node has its own unique physical location information, which is the coordinates, equipment number or process number within the workshop, etc.; the corresponding production data combination refers to the structured data set collected in real time from each production node, as mentioned above, including timestamps, process parameters, equipment status, etc. These data combinations need to be associated with specific production nodes to ensure data consistency and accuracy; Internet of Things technology provides the infrastructure for data collection, transmission and storage, which includes sensors, communication networks, data storage servers, etc.

[0021] Integrate node locations, production data combinations, and IoT infrastructure to build a unified digital database. This database needs to support real-time storage, query, and analysis of large-scale data. As production progresses, new data will be continuously generated and need to be added to the database. The database needs to support real-time updates to ensure the timeliness and accuracy of the data. Integrate scattered production data into a centralized, structured digital database.

[0022] refer to Figure 3 In step S12, the specific steps are: S121: Collecting various production data combinations, determining corresponding abnormal production detection patterns based on matching between each production data combination and the digital intelligence database, and triggering abnormality detection for each production data combination along the abnormal production detection pattern to output multiple abnormal production data; S122: Mark the abnormal locations corresponding to the multiple abnormal production data, determine the corresponding production traceability method based on the abnormal location, the corresponding abnormal production data, and the digital intelligent database, determine multiple traceability items based on the analysis of the production traceability method, cross-trace the workpieces with the multiple traceability items, and trigger production traceability of the multiple abnormal production data to determine multiple abnormal production nodes; S123: At each abnormal production node, determining an abnormal range of the abnormal production data based on a comparison of the corresponding plurality of abnormal production data, and determining an abnormal production path based on the position of each abnormal production node, the abnormal range of the abnormal production data, and the workpiece; In an embodiment of the present application, various production data combinations are collected, and the digital database not only stores real-time data, but also stores historical normal data, equipment parameters, process procedures, and even pre-defined abnormal pattern templates. The "matching" here refers to comparing the currently collected production data combination with the information stored in the database.

[0023] The matching process involves: checking whether the current data exceeds the normal parameter range of the node, workpiece model, and process stored in the database (for example, the normal range of spindle current is 10-15A, and the normal range of vibration amplitude is 0.05-0.20mm / s); comparing the current data pattern with the known abnormal patterns stored in the database for similarity (for example, historically, patterns of abnormal vibration and current usually correspond to tool wear); applying the rule engine stored in the database, which are generated based on expert knowledge or machine learning models (for example, "If the vibration amplitude exceeds 0.25mm / s three times in a row and the spindle current exceeds 13A, an abnormal alarm is triggered"); using statistical models (such as control charts, time series models) or machine learning models (such as model parameters of anomaly detection algorithms) stored in the database to determine whether the current data meets expectations.

[0024] Once the appropriate anomaly detection pattern is determined, the system conducts in-depth analysis of the collected production data combination according to the pattern's definition to identify specific anomalies or abnormal periods. Detection is performed based on the determined pattern: if it is a parameter range match, it checks which data points are out of range; if it is a historical pattern match, it calculates similarity to determine whether the anomaly threshold has been reached; if it is a rule engine match, it checks whether the rule conditions are met; if it is a model match, it uses the model to predict normal values, calculates the deviation between the actual value and the predicted value, and determines whether it is an anomaly.

[0025] Furthermore, the abnormal positions corresponding to multiple abnormal production data are marked, and the marks will contain key information of the abnormal data itself, such as the abnormality type, occurrence time, specific parameter values ​​involved, the workpiece ID to which it belongs, and most importantly - the production node location where it occurs; optionally, the system will update the records in the database, or maintain an abnormality list in the memory, clearly stating that "the workpiece W-2025-0715-001 at the CNCARoughMill node had an abnormality suspected of tool wear during the period of 2025-07-1514:14:00-14:15:00, with key indicators of vibration 0.32mm / s and current 13.5A".

[0026] The system needs to select the most appropriate traceability method based on the location of the exception (which node), the specific characteristics of the exception (what kind of exception it is and how severe it is), and the process knowledge, material flow information, equipment associations, etc. stored in the database.

[0027] Traceability methods include: forward traceability: starting from the abnormal node, tracing back the cause (for example, the processing quality of the previous node, problems with the material itself, the status of the tooling, etc.); backward traceability: starting from the abnormal node, tracing back the impact (for example, how the abnormality affects the processing accuracy of subsequent processes, the inspection pass rate of the final product, etc.); path traceability: tracing along the actual processing path of the workpiece, checking the data and status of all relevant nodes on the path; cause-effect traceability: trying to analyze the causal relationship between abnormal data and potential causes (for example, whether vibration abnormality will inevitably lead to dimensional deviation).

[0028] The system will analyze abnormal data. For example, "suspected tool wear" usually means that the problem originates from the tool itself (forward traceability) or the processing status of the previous step (forward traceability), and also affects the subsequent processing accuracy (backward traceability); the system will query the database to see if there is any related information about the CNCA rough milling process and the upstream clamping process, downstream fine milling process, tool change records, etc.; the database stores the rule that "rough milling quality directly affects the fine milling allowance", or records the fact that "tool T-12345 was last replaced on 2025-07-10"; the system determines that this traceability mainly adopts the method of "forward traceability" combined with "path traceability", focusing on the status of the CNCA clamping process point (CNCALoad) and tool T-12345.

[0029] After the traceability method is determined, it needs to be broken down into specific, executable operation items, namely traceability projects, which are the units for actually performing traceability analysis. Traceability projects include: checking the historical data of specific parameters: for example, checking the vibration and current data trends of tool T-12345 in the past few days; comparing data of different workpieces: for example, comparing the data of workpiece W-2025-0715-001 with other workpieces in the same batch at the CNCARoughMill node; querying the material / equipment status: for example, querying the clamping force data of the workpiece at the CNCALoad node, and querying the inventory record and wear record of tool T-12345; checking process parameters: for example, checking whether the cutting parameters (speed, feed rate) of the CNCA rough milling process are consistent with the process specifications.

[0030] The system executes the traceability projects determined by S122-3 in parallel or serially, collecting information from different angles; then, the system cross-compares and analyzes the information from different projects to verify hypotheses, eliminate interference, and ultimately determine which production nodes are directly related to the current anomaly.

[0031] Cross-traceability means that the results obtained from different traceability projects need to verify each other. For example, if the tool history data shows that its vibration has always been normal, but the clamping force data shows abnormalities, then the clamping problem is the cause; vice versa; if the data of other workpieces in the same batch are normal, then it points to a problem with the workpiece itself or its specific processing process.

[0032] Executing a traceability project is actually the process of extracting relevant information from the database and correlating and analyzing this information with the initial abnormal data; the ultimate goal is to find the production nodes that directly cause or significantly affect the current abnormality. These nodes include the node where the abnormality occurs, as well as upstream or downstream nodes.

[0033] Therefore, at each abnormal production node, the abnormal range of the abnormal production data is determined based on the comparison of the corresponding multiple abnormal production data, and the abnormal production path is determined based on the position of each abnormal production node, the abnormal range of the abnormal production data and the workpiece, which is compatible with the overall consideration of the position of each abnormal production node, the abnormal range of the abnormal production data and the workpiece, ensuring the accuracy of the abnormal production path.

[0034] At this point, at each abnormal production node, a comparison of multiple corresponding abnormal production data is implemented. The comparison usually includes the following: Time series comparison: observing the change trend of abnormal data over time; whether it is a sudden jump or a slow deterioration, and how long the abnormality lasts; Threshold comparison: comparing the abnormal data with the predefined threshold (such as that used in S121) to determine the amplitude and duration of the threshold exceeding the threshold; Historical data comparison: comparing the current abnormal data with the data of the same model of workpiece produced normally at this node in the past to see the amplitude of the deviation from the normal range; Same batch / same condition comparison: if multiple workpieces pass through this node in the same time period, compare their data to see whether the abnormality is specific to a certain workpiece or is universal; Parameter comparison: analyzing the correlation between different sensor data in the same node, for example, whether the vibration abnormality is accompanied by current abnormality or whether the temperature is increased. Through the above comparison, the specific manifestation of the abnormality (such as the vibration amplitude exceeds the normal range by 0.1mm / s for 30 seconds), the degree of impact (minor, moderate, severe) and the scope of impact (does it only affect a specific area of ​​the current workpiece or affect subsequent processes) are determined; the system will generate a detailed abnormality analysis report to quantify the degree of abnormality.

[0035] The system introduces the location of each abnormal production node, the abnormal range of abnormal production data and the workpiece, the node sequence: the order of the production process; the abnormal range: the severity of the abnormality and the nature of the impact (whether it directly affects the geometry of the workpiece or indirectly affects the status of subsequent processes); the workpiece characteristics: the process sensitivity of the workpiece itself; based on the analysis results, the system will mark the abnormal path on the visual interface (such as the production flow chart) or generate a text report to describe the path; the abnormal path describes how the problem affects the workpiece from the source, and at which subsequent nodes the impact appears or intensifies.

[0036] refer to Figure 4 In step S13, the specific steps are: S131: Collect response signals from each production node in the digital factory, determine the status coefficient of each production node based on the analysis of each response signal, and determine the current working status of the digital factory based on the node position, status coefficient, and the number of processed workpieces corresponding to each production node in the digital database; S132: Collect the workpiece model, determine the workpiece processing form based on the workpiece model and external form, and determine the current production path based on the workpiece processing form, the current working status of the digital factory, and multiple production nodes; S133: Comparing the current production path with the abnormal production path, determining an overlapping area based on the comparison between the current production path and the abnormal production path, determining multiple abnormal sub-areas based on the detection of the overlapping area, and determining the production abnormal area of ​​the workpiece based on the synthesis of the multiple abnormal sub-areas; In an embodiment of the present application, current status information of various production nodes (such as machine tools, robots, sensors, detection equipment, etc.) in a factory is collected in real time or quasi-real time, and this information is transmitted through an Internet of Things (IoT) network.

[0037] Data type examples: Status signal: "NodeA: Processing", "NodeB: Idle", "NodeD: Detection completed"; Performance signal: "NodeA: Spindle speed 8000 RPM", "NodeA: Feed rate 500 mm / min", "NodeC: Tool wear 0.05 mm"; Sensor signal: "NodeA: Cutting area temperature 85°C", "NodeC: Vibration sensor reading 0.8 g", "Ambient temperature and humidity"; Event signal: "NodeA: Alarm ALM-101 (Tool breakage) occurs", "NodeD: Out-of-tolerance workpiece detected"; Material signal: "NodeA: Workpiece loaded", "NodeB: Abrasive barrel liquid level low".

[0038] The collected raw response signals are usually discrete and heterogeneous, making them difficult to use directly for overall status assessment. Therefore, these signals need to be parsed and converted, mapping them to one or more "status coefficients" that can quantify the current operating status of the node. This parsing process involves: filtering, denoising, and normalizing the raw signals; comparing continuous signals (such as temperature and vibration) with preset thresholds to determine whether they are in normal, warning, or dangerous states. For example, if the vibration sensor reading of NodeC exceeds 1.0g, it is determined to be in a dangerous state; using machine learning models (such as SVM and neural networks) to comprehensively analyze a set of signals to determine whether the node is in a specific mode (such as normal processing, standby, overload, or impending failure); assigning weights to different types of signals or states, and calculating a comprehensive score as the status coefficient. For example, a simple scoring formula is: status coefficient = w1*status score + w2*performance score + w3*safety score, where each score comes from the analysis of the corresponding signal.

[0039] The status coefficients of individual nodes, combined with their position in the factory's production process and the factory's overall production progress, form a macro-assessment of the current working status of the entire factory; the digital database records the number of workpieces that have been successfully processed, which reflects the factory's production efficiency and current load.

[0040] The role of node location: Understanding node location helps understand production bottlenecks. For example, if the status coefficient of NodeD at the end of the process is low (such as a fault), the entire production process will stagnate even if the previous node is in good condition. If the status coefficient of an intermediate node (such as NodeB) is low, it will affect the input quality or speed of subsequent nodes. The role of the status coefficient: By combining the status coefficients of all nodes, the overall status index of the factory can be calculated (such as a weighted average). This index can reflect the overall health of the factory (whether it is operating smoothly, partially restricted, or close to paralysis).

[0041] The role of the number of processed workpieces: This number reflects the factory's output speed and load. If the number of processed workpieces is far lower than planned, combined with the node status coefficient, it can be judged that the factory is inefficient or stagnant. If the number is close to or reaches the plan, and the node status coefficient is generally high, it indicates that the factory is operating well. The system will combine the above information to generate a descriptive "current working status", which can be a text description (such as "The factory is running smoothly and the efficiency is up to standard. NodeC has a slight vibration anomaly that requires attention"), a comprehensive score (such as "Factory Status Index: 0.82"), or a more structured status report.

[0042] Furthermore, the model of the workpiece is collected, and the model and external form of the workpiece are introduced. The workpiece model information is used in combination with its actual physical state (external form) to determine the specific type and sequence of processing operations that need to be performed. This involves mapping general model information to specific processing strategies; "external form" refers to the current actual state of the workpiece, including: initial state: whether it is a blank, semi-finished product or a workpiece that needs to be reworked; processed features: which parts have been processed; which parts still need to be processed; physical state: whether there is deformation or damage; whether the size is within the allowable range.

[0043] The system will query the process database or digital twin model related to the model; the database stores recommended processing strategies for different models and different initial / current forms. For example, the processing form of a "TL-A2-V3" blank is "roughing + deburring + finishing"; and the processing form of a "TL-A2-V3" semi-finished product (roughing completed) is "deburring + finishing"; the system will combine sensor data (such as visual inspection and dimensional measurement) to confirm the current external form of the workpiece to ensure the accuracy of the processing form.

[0044] To generate a specific, executable movement and processing route for a workpiece within the factory, it is necessary to comprehensively consider: processing mode: determining which types of processes need to be processed (such as roughing, deburring, and finishing); current working status (from S131): whether each production node (machine tool, robot, etc.) is currently idle, busy, faulty, or has degraded performance; whether material handling equipment (such as AGV) is available; whether the energy supply is stable; multiple production nodes: which specific equipment can perform the required processes and where they are located; whether there is specific equipment suitable for processing this type of workpiece (such as a specific model of machine tool).

[0045] The system selects the appropriate process node based on the processing form; then, based on the status of each node determined in S131, it performs intelligent scheduling: giving priority to nodes that are in good condition (high status coefficient), idle, or about to be idle; if there are multiple optional nodes (for example, there are two roughing centers), the one with the lower current load, better location (reducing material handling time), or better condition will be selected; the physical connection between nodes and the material handling capacity (such as AGV path planning) need to be considered; and finally, a specific, chronologically arranged node sequence, namely the "current production path", is determined.

[0046] Therefore, the current production path and the abnormal production path are compared, and the overlapping area is determined based on the comparison between the current production path and the abnormal production path. Multiple sub-abnormal areas are determined based on the detection of the overlapping area. The production abnormal area of ​​the workpiece is determined based on the synthesis of multiple sub-abnormal areas. The overall consideration of the synthesis of multiple sub-abnormal areas is compatible to ensure the accuracy of the production abnormal area of ​​the workpiece. At the same time, various abnormal production nodes are introduced, which is compatible with the overall consideration of the current production path and the abnormal production path, improves the accuracy of the production abnormal area of ​​the workpiece, and performs traceability control on the production of the digital factory.

[0047] At this point, find out the similarities or overlaps between the processing flow of the current workpiece and the processing flow where the abnormality is known to have occurred. The "abnormal production path" here refers to the sequence of production nodes involved in the previously detected abnormal data; the comparison is not just to see whether the node names are the same, but also to consider whether the node type, processing content, workpiece status, etc. are related.

[0048] The system will match the current production path (a node sequence) with the historical production path (or abnormal path) related to the confirmed anomaly stored in the database. For example, the database records an abnormal path: AGV>CNCARoughMill>AGV>DeburrRobot>AGV>CNCARoughMill>AGV>DimensionInspection. When comparing, the system will check which nodes are the same in the current path and this abnormal path, or which nodes have different names but similar functions / locations. The current production path comes from the output of S132; the abnormal production path comes from the abnormal production nodes determined by S122, combined with the historical production process information related to these nodes stored in the database.

[0049] The overlapping area refers to the production nodes or continuous production node segments shared by the current production path and at least one abnormal production path; the overlapping area is determined to identify the key links where the current workpiece is affected by historical anomalies; through the comparison results of the previous step, the system marks the nodes contained in both the current path and the abnormal path, and these nodes and the path segments connected to them are the overlapping areas; if the nodes in the abnormal path do not exist in the current path, but they are very close to certain nodes of the current path in space or function, the system will also mark these nodes as potential overlapping areas; the overlapping area is usually represented in the form of a node list or node sequence, for example, if the current path is A>B>C>D, and the abnormal path is A>B>A>D, then the overlapping area is node B and node D.

[0050] The "sub-abnormal area" is a local area related to a specific abnormality that is further refined within the overlapping area. An overlapping area involves multiple processing steps or equipment, but not all parts are directly affected by the previous specific abnormality. For example, an overlapping area contains a processing node and a detection node, but the previous abnormality is only related to the processing node. The system will analyze each node or node segment in the overlapping area, combined with the specific type and location of the previous abnormality (from S122), to determine which parts are most affected by the abnormality, which involves: time correlation, spatial correlation, and process correlation. The system will mark these sub-areas as "high risk", "medium risk" or "low risk".

[0051] The sub-abnormal area can be represented as a sub-node within the overlapping area, an equipment component (such as a tool, a fixture), or a specific feature area on the workpiece. For example, in the overlapping areas B and D, if the previous abnormality was tool wear caused by rough milling (at node A that is not the current path), then the sub-abnormal area is mainly marked at node B (deburring, affecting subsequent processes) and node D (detection, used to verify the impact of the abnormality).

[0052] The multiple sub-abnormal areas identified in the previous step are integrated and comprehensively judged to eventually form a clear description of the abnormal impact range of the current production process of the workpiece. This is the final definition of the abnormal impact range; the system will consider the risk levels, mutual relationships and comprehensive impact of all sub-abnormal areas on the final quality of the workpiece. For example, if multiple sub-areas are marked as high risk, or they all point to a key feature on the workpiece, then the final "production abnormal area" will be defined more clearly and critically; the final "production abnormal area" is a location of problems in the workpiece production process or the workpiece itself. It is a specific equipment node, a specific processing step, a feature area on the workpiece, or a label indicating the risk level.

[0053] The production abnormality area is a comprehensive description, for example: "The production abnormality areas of the current workpiece TL-A2-V3 mainly involve the deburring process (NodeB) and the finishing process (NodeC). Special attention should be paid to the burr condition of the blade edge and the final dimensional accuracy, especially the thickness of the blade tip area." Or more specifically: "Production abnormality area: Tool path area #3 of finishing center C (NodeC), corresponding to the range of #45-#60 degrees on the blade leading edge."

[0054] Specifically, the current path is A>B>C>D; suppose there is an abnormal path in the database related to the previous CNCARoughMill (NodeA) tool wear anomaly, which is A>B>A>D; by comparison, it is found that node B and node D overlap; the overlapping area is determined to be node B (deburring robot) and node D (dimensional detection).

[0055] System analysis: Node B: deburring; although the previous abnormality occurred during rough milling (NodeA), the burr state after rough milling will affect the deburring effect; if the previous tool wear causes the workpiece edge to be irregular, then the deburring process (NodeB of the current path) is the sub-area that needs attention; Node D: dimensional inspection; inspection is the key to verifying the final quality, and any deviation introduced by the previous process will eventually be reflected in the inspection link; therefore, dimensional inspection (NodeD) is also a sub-abnormal area used to confirm the impact of the abnormality; Node C: finishing center; although the finishing center of the current path is NodeC', not NodeA where the abnormality occurred, finishing is a key process and a new tool is used; the system judges that the risk of NodeC' is low, but its processing results still need to be paid attention to; or, if the previous abnormality affects the overall morphology of the workpiece, then the processing of NodeC' is also affected; assuming that the system judges that the risk of NodeC' is low, it is not temporarily listed as a high-risk sub-area; therefore, the main sub-abnormal areas are NodeB and NodeD.

[0056] The system comprehensively determined that the previous anomaly (roughing tool wear) had affected the workpiece edge condition, and this impact would be transmitted to subsequent processes and ultimately confirmed during inspection. Therefore, the final production anomaly area determined was: "The current production anomaly area for workpiece TL-A2-V3 mainly involves the deburring process (NodeB) and the dimensional inspection process (NodeD). Special attention should be paid to the edge condition after deburring and the final dimensional tolerance, especially the tip thickness and blade body contour, to confirm whether it was potentially affected by the previous roughing anomaly."

[0057] refer to Figure 5 In step S14, the specific steps are: S141: collecting the abnormal production area, determining multiple component features based on the detection of the abnormal production area, and determining the abnormal factory component corresponding to the abnormal production area based on the multiple component features and the captured image corresponding to the abnormal production area; S142: Marking the IoT of the digital factory, and based on the IoT, performing abnormal management and control on each abnormal factory component, outputting the working data of each abnormal factory component, and determining the repair time node of each abnormal factory component based on the working data of each abnormal factory component and the current production location of the workpiece; S143: Collect the production process table corresponding to the workpiece, determine the production time of the workpiece in each production process based on the analysis of the production process table, and determine the repair workload of each abnormal factory component based on the repair time node of each abnormal factory component, the production time of the workpiece in each production process, and the workload of the abnormal factory component in the production process; In an embodiment of the present application, the system analyzes the collected data to look for patterns or features related to "abnormality", which involves: threshold comparison: comparing real-time data with preset normal working thresholds; trend analysis: observing the changing trend of data over time; statistical analysis: calculating the statistics of the data (such as mean, standard deviation) and comparing them with historical normal values; extracting key features that can characterize "abnormality" from the analysis results. These features should be able to point to specific physical components or subsystems, for example, "Z-axis deburring force is continuously higher than the threshold", "motor current fluctuations are abnormal", "machining time is 20% longer than the standard", etc.

[0058] Specifically, the system first locks on the "deburring process" (NodeB), and it begins to collect: force sensor data installed on the deburring tool (such as a rotating grinding head): X-axis force, Y-axis force, and Z-axis force; position data of the robot joint encoder; current and voltage data of the deburring motor and robot joint motor; NodeBPLC alarm log (for example, whether there are records of "tool overload" or "motion timeout"); real-time video stream captured by the surveillance camera installed next to the robot (displaying the contact between the tool and the blade); historical statistical data such as the average deburring force and average current value when NodeB processed TL-A2-V3 blades in the past 24 hours.

[0059] The system analyzed the collected data and found that the average Z-axis deburring force (force perpendicular to the blade surface) of the five most recently processed blades was 15% higher than the standard value, and the standard deviation was also large, indicating that the force value was unstable and on the high side. At the same time, the current driving the deburring motor also showed synchronous and periodic fluctuations, with the peak value about 10% higher than the normal value. There were no clear hard alarms in the PLC log, but the records showed that the average deburring time per piece increased by about 3 minutes compared to the standard 15 minutes.

[0060] The "component features" extracted by the system include: "abnormally high deburring force (+15%)"; "increased volatility of deburring force"; "abnormal fluctuation of deburring motor current"; and "extended single-piece processing time (+20%)".

[0061] Associate the "component features" extracted in the previous step with the various physical components contained in the production area (NodeB). For example, "abnormally high deburring force" is usually related to wear or installation problems of the deburring tool; "abnormal motor current fluctuation" is related to the motor itself, drive, or transmission mechanism.

[0062] Combined with previously collected visual images, for example, images taken by surveillance cameras directly show severe tool wear, or the blade is offset when clamped, resulting in abnormal deburring force; images can provide intuitive evidence to help confirm or correct judgments based on sensor data; comprehensive analysis results can ultimately determine which specific factory component has a problem, which is a specific hardware, such as a sensor, an actuator, a tool, a motor, or a small mechanical structure.

[0063] Specifically, the system correlated the extracted features: "abnormally high deburring force" + "increased force volatility" + "abnormal motor current fluctuation" + "extended processing time" - this combination of features strongly pointed to a problem with the deburring tool itself, the most common of which was tool wear or blunting. The system retrieved images from the NodeB surveillance camera; the images showed that when the deburring tool came into contact with the blade burr, the contact area was large, and obvious signs of wear were visible on the tool surface. Combining sensor data and visual images, the system ultimately determined that the "abnormal factory component" was the "grinding head tool installed on the NodeB deburring robot" (for example, model GrinderBlade-456).

[0064] Furthermore, the Internet of Things of the digital factory is marked, and based on the Internet of Things, each abnormal factory component is managed and controlled, and the working data of each abnormal factory component is output. The repair time node of each abnormal factory component is determined according to the working data of each abnormal factory component and the current production location of the workpiece, which is compatible with the overall consideration of the working data of each abnormal factory component and the current production location of the workpiece, and ensures the accuracy of the repair time node of each abnormal factory component.

[0065] At this point, marking the IoT of the digital factory does not just mean changing the state. It also requires associating the abnormal information with specific equipment (NodeB), components (GrinderBlade-456), abnormality type (wear), detection time, detection basis (sensor data, images), etc. to form a complete event record. This marking operation usually triggers a notification; the system will send alarms or update information to the relevant maintenance team, equipment operators, production schedulers and even MES systems to inform them that a problem has occurred with a component; the level of notification (for example, urgent, important, general) is automatically set according to the severity of the abnormality; on the factory's monitoring screen or the maintenance personnel's mobile terminal, the icon of the relevant equipment or the component's logo will change color (such as red) or pop up an alarm message, making the problem clear at a glance.

[0066] Depending on the severity of the anomaly, the system will automatically restrict or prohibit the component / equipment from performing certain operations. For example, if severe tool wear causes the workpiece to be scrapped or the equipment to be damaged, the system can automatically suspend the NodeB's processing task or set it to "maintenance mode" to prohibit automatic operation.

[0067] In some cases, if the anomaly allows, the system automatically adjusts equipment parameters to mitigate the impact. For example, if there is only slight wear, the system automatically reduces the force or speed of deburring to extend tool life or ensure processing quality until maintenance personnel arrive.

[0068] The system can adjust relevant resource allocation based on the abnormal information. For example, it can temporarily direct the workpiece that needs to be deburred to the backup equipment (if available), or adjust the priority of other processes. The system will isolate the component's subsequent operating data or increase the frequency of data collection to record the behavior during the abnormal period in more detail, providing a basis for subsequent analysis.

[0069] Specifically, the system found the device object representing the deburring robot NodeB on the factory's IoT platform (for example, a cloud-based device management platform); under this object, it found the component object representing GrinderBlade-456; the system updated the status of GrinderBlade-456 from "normal" to "abnormal wear (requires replacement)"; at the same time, the system recorded the abnormality type, detection time (for example, 2025-07-15 14:30), and detection basis (abnormal force and current, image confirmation of wear); the IoT platform automatically pushed an alarm message to the maintenance team's dedicated app: "Abnormal wear detected on the device NodeB (deburring robot) component GrinderBlade-456, please handle it as soon as possible", and on the factory's central monitoring screen, the NodeB icon turned flashing yellow, with a brief description of the abnormality.

[0070] Because the wear of the GrinderBlade-456 had caused abnormally high deburring forces, the system determined that continued use of the damaged workpiece would accelerate further tool wear. Therefore, the system sent a command to the NodeB controller via the IoT platform, setting the NodeB's status to "Maintenance Paused." The NodeB robot stopped operating automatically, and the display screen displayed the message "Maintenance Paused: Abnormal grinding head tool wear, please contact maintenance." The system also checked for backup deburring equipment (for example, NodeB'). If not, the system notified the production scheduling system to temporarily skip the deburring process for subsequent workpieces requiring deburring (such as TL-A2-V4 after TL-A2-V3) or to wait until the NodeB was repaired before processing.

[0071] The system needs to extract detailed operating data related to the abnormal component from the IoT platform, device logs, and sensor history databases. This includes but is not limited to: operating time, number of starts and stops, load history (current, force, temperature, etc.), error codes, alarm records, maintenance history, etc. This data needs to be output in a way that is easy to understand and analyze; the form includes: generating a detailed report, displaying key data charts (such as load change curves over time) on the maintenance personnel's interface, providing download links, etc. This data is an important basis for maintenance personnel to make diagnoses and decisions; by analyzing this data, maintenance personnel can gain a deeper understanding of the cause of the abnormality, the degree of wear of the component, and determine the best repair plan (for example, replacement or repair).

[0072] Maintenance personnel (or the auxiliary decision-making module in the system) analyze the output work data and, based on their understanding of the components, assess the urgency and time required for repair. For example, the data may show that a component is about to fail completely or that a specific spare part is required for repair. The system needs to know the location of the workpiece currently in production or about to enter the abnormal process. This can be obtained through the MES system or WMS (warehouse management system). For example, whether the workpiece is in a waiting area, being processed in other processes, or about to arrive at the abnormal equipment. The system combines the time required for repair with the workpiece location / plan information to determine the optimal repair time window.

[0073] The system combines the time required for repair with the workpiece location / plan information to determine the optimal repair time window; the system can calculate in which time period to perform the repair with the least impact on the overall production plan, for example, whether to repair immediately after the current workpiece is processed, or wait until the next shift, or wait for a critical workpiece to pass through before repairing; determining the time node also means coordinating the maintenance resources (personnel, spare parts, tools) to be available at that point in time; ultimately, one or more repair time recommendations are determined and recorded to guide subsequent maintenance actions.

[0074] Specifically, the system automatically generated a detailed working data report on the GrinderBlade-456 and pushed it to the engineer responsible for the equipment through the maintenance team's app; the report content includes: basic component information: model GrinderBlade-456, installation date, expected life; operation history: cumulative operating time (for example, 1200 hours), number of starts and stops (for example, 1500 times); load data: force sensor reading curve in the past 24 hours (showing that the force value gradually increased and reached an abnormal peak at 14:30), current sensor reading curve (showing increased current fluctuation); error record: alarm code recorded by the NodeB controller at 14:30 (for example, alarm code #E503: deburring force exceeded the limit); maintenance history: the time of the last tool replacement (for example, 2025-05-10, approximately 800 hours of operation).

[0075] The maintenance engineer reviewed the work data report of GrinderBlade-456 and, based on his experience, concluded that the tool was severely worn and needed to be replaced immediately. The estimated replacement time was 30-45 minutes (including obtaining the spare part and the replacement operation). At the same time, the MES system showed that the TL-A2-V3 blade currently being processed had completed rough machining and was on its way to the deburring process (NodeB), expected to arrive in 15 minutes. The TL-A2-V4 following closely behind would also arrive shortly after the TL-A2-V3 was completed. Considering that NodeB had already been After a pause, with the next workpiece about to arrive, the engineer decides to begin preparations for the replacement immediately. Based on the maintenance team's shift schedule and spare parts inventory information, the system confirms that an engineer is available around 3:00 PM and that the spare part (a new GrinderBlade-456) is available in the spare parts inventory, estimating that the replacement can begin at 3:10 PM. Therefore, the system determines the repair time for GrinderBlade-456 as "Estimated start time: 3:10 PM, Estimated completion time: 3:55 PM." This information is recorded and notified to the relevant engineers and production scheduler.

[0076] Therefore, the production process table corresponding to the workpiece is collected, and the production time of the workpiece in each production process is determined based on the analysis of the production process table. The repair workload of each abnormal factory component is determined according to the repair time node of each abnormal factory component, the production time of the workpiece in each production process and the workload of the abnormal factory department in the production process. This is compatible with the overall consideration of the repair time node of each abnormal factory component, the production time of the workpiece in each production process and the workload of the abnormal factory department in the production process, ensuring the accuracy of the repair workload of each abnormal factory component.

[0077] At this point, the system needs to extract the standard production plan for the workpiece currently being processed (such as TL-A2-V3) from the MES (manufacturing execution system) or ERP (enterprise resource planning system). This plan lists in detail all the production nodes (processes) that the workpiece needs to go through, the operations that should be performed at each node, the equipment used, the standard processing time, quality inspection requirements, etc. This process sheet is usually a structured data file that contains information such as process ID, process name, equipment / node name, standard working hours, operation description, required tools / fixtures, quality standards, etc.; after obtaining the process sheet, the system will perform some verification, such as checking whether the workpiece model matches, whether the process sheet is the latest version, etc., to ensure that subsequent calculations are based on accurate basic data.

[0078] The system extracts the standard production time for each process from the process table. These times are optimized and verified and represent the time required to complete the process under normal circumstances. In addition to the standard production time, the system also extracts other time information, such as preparation time, mold change time, quality inspection time, etc., but in this step, the main focus is on direct processing or handling time. The system organizes this extracted time information according to the process sequence to form a time series for subsequent calculation and analysis.

[0079] Specifically, for the TL-A2-V3 blade, the system retrieves its production process table from the MES; the table shows: Process 1: Rough machining (NodeA), standard time: 2 hours, equipment: CNC-Mill-001; Process 2: Deburring (NodeB), standard time: 15 minutes, equipment: RoboDeburr-023, tool: GrinderBlade-456; Process 3: Finishing (NodeC), standard time: 3 hours, equipment: CNC-Mill-002; Process 4: Dimension detection (NodeD), standard time: 10 minutes, equipment: CMM-005; Process 5: Surface treatment (NodeE), standard time: 4 hours, equipment: PlatingChamber-010; the system parses the process table of TL-A2-V3 and extracts the standard production time of each process: Process 1 (NodeA): 2 hours; Process 2 (NodeB): 15 minutes; Process 3 (NodeC): 3 hours; Process 4 (NodeD): 10 minutes; Process 5 (NodeE): 4 hours.

[0080] Compare the repair time node (15:10-15:55) of the previously identified abnormal component (GrinderBlade-456) with the current production location of the workpiece (TL-A2-V3 on its way to NodeB) and the standard time of each process. Determine whether the repair activity will affect the normal production of this workpiece or other subsequent workpieces. Analyze whether the repair time overlaps with the estimated time the workpiece will take in related processes. For example, if TL-A2-V3 is expected to arrive at NodeB at 15:30, then the repair time of GrinderBlade-456 (15:10-15:55) will overlap with the processing time of TL-A2-V3 at NodeB (15:30-15:45) by 15 minutes, which means that the repair work must be completed during this period, or the progress of the workpiece needs to be adjusted.

[0081] The "repair workload" here is a comprehensive indicator. It not only refers to the time required for physical repair, but also includes: Direct repair time: the actual operation time of replacing the GrinderBlade-456 (for example, 15:10-15:55, a total of 45 minutes), which includes removing the old blade, installing the new blade, preliminary calibration, etc.; Preparation time: the time for obtaining spare parts, preparing tools, and engineers going to the site; Adjustment / calibration time: After the repair is completed, the equipment needs to be additionally adjusted or calibrated to ensure that it returns to normal performance, and this part of time also needs to be included; Impact assessment on production: If the repair time is too long , causing the workpiece to wait at NodeB, then the waiting time should also be considered as part of the impact of the repair work, and additional operations (such as re-queuing and adjusting the time of subsequent processes) need to be added; resource requirements: Engineers with specific skills, specific tools, spare parts inventory, etc. are required, and the demand for these resources is also a reflection of the workload; combining the above factors, the total workload required to repair GrinderBlade-456 is calculated. For example, it is a data structure containing dimensions such as "total time (hours)", "required manpower (engineer level)", "required spare parts (quantity, type)", and "required tools (list)".

[0082] Specifically, the repair time is: 3:10 PM to 3:55 PM (45 minutes); the time for TL-A2-V3 to reach the NodeB is approximately 3:30 PM; the standard processing time for TL-A2-V3 at the NodeB is 15 minutes (3:30 PM to 3:45 PM); and the overlap time is: 3:30 PM to 3:45 PM (15 minutes).

[0083] System determination: Direct repair time: 45 minutes (included in 3:10 PM - 3:55 PM); Preparation time: Assuming an engineer needs to arrive 10 minutes in advance to prepare, and the spare part takes 5 minutes to retrieve from the warehouse, the total preparation time is approximately 15 minutes (assuming completion before 3:00 PM); Adjustment / calibration time: After replacement, calibration requires 5-10 minutes, assuming 10 minutes (3:55 PM - 4:05 PM); Impact on production: There is a 15-minute overlap, which means that TL-A2-V3 will wait at the NodeB, or the NodeB will need to process TL-A2-V3 immediately after repair, resulting in delays in the subsequent workpiece TL-A2-V4; Resource requirements: One senior robot maintenance engineer, a set of robot maintenance tools, and one new GrinderBlade-456 spare part are required.

[0084] The "repair workload" calculated by the system is expressed as follows: Total time: approximately 1 hour (including preparation, direct repair, calibration, and waiting / adjustment time); Required manpower: 1 senior engineer; Required spare parts: 1xGrinderBlade-456; Required tools: Standard robot maintenance tool kit; Assessment of impact on production: The TL-A2-V3 will be delayed by approximately 15 minutes at the NodeB, requiring notification to subsequent processes.

[0085] refer to Figure 6 In step S15, the specific steps are: S151: The IoT performs abnormality control on each abnormal factory component and collects repair data of each abnormal factory component at different time points. Simultaneously, the workpiece processing area relative to the abnormal factory component is collected, and a first repair event is determined based on the multiple repair data and the workpiece processing area. S152: Determine the autonomous repair strategy of the digital factory for the workpiece based on the first repair event, the repair workload of each abnormal factory component and the Internet of Things. At this time, the autonomous repair strategy is analyzed and the autonomous repair process is output. The early repair event of each abnormal factory component is determined based on the dynamic interaction between the autonomous repair process and the production process of the workpiece, so as to repair each abnormal factory component in advance.

[0086] In an embodiment of the present application, the Internet of Things performs abnormal management and control on each abnormal factory component, and collects the repair data of each abnormal factory component at different time nodes. At the same time, it collects the processing area of ​​the workpiece relative to the abnormal factory component, and determines the first repair event based on multiple repair data and the processing area of ​​the workpiece, which is compatible with the overall consideration of multiple repair data and the processing area of ​​the workpiece, ensuring the accuracy of the first repair event.

[0087] At this point, the IoT system first accurately identifies the GrinderBlade-456 and its deburring robot (NodeB) through sensors, controllers (such as PLCs), and device interfaces in its network. This usually requires the device itself to have a unique identifier (such as an RFID tag, MAC address, etc.).

[0088] The system will send instructions to the control unit of the NodeB to set its status to "under maintenance" or "downtime", which includes: prohibiting scheduling: the factory's MES or APS (Advanced Planning and Scheduling) system will receive a notification and will no longer assign new workpieces to the NodeB; physical isolation (optional): if a physical security door or area locking device is configured, the system will automatically lock it to prevent people from entering by mistake or equipment from accidentally starting; software isolation: at the software level, the NodeB's operating interface will display "maintenance mode", prohibiting operators from performing production-related operations.

[0089] In the control state, the IoT system will increase the monitoring frequency of the GrinderBlade-456 and its related components (such as motors and bearings). For example, it will collect vibration, temperature, current and other data more frequently to more accurately assess its current status and the urgency of repair.

[0090] Specifically, the IoT system identified the GrinderBlade-456 as an abnormal component through the sensor interface and built-in communication module on the NodeB control cabinet. The system immediately sent a message to the MES: "NodeB-GrinderBlade-456 entered maintenance status." The MES then updated the production plan and suspended the allocation of new blade processing tasks to the NodeB. At the same time, the NodeB's HMI (human-machine interface) popped up a red warning box, displaying "Maintenance Mode: Waiting for Tool Replacement." The system began collecting and continuously monitoring the tool's vibration and motor current data at a frequency of once per second.

[0091] Repair data comes from multiple sources and is collected through the IoT's sensor network and device interfaces: Component status data: as mentioned above, real-time data such as vibration, temperature, current, wear (if a wear sensor is installed or calculated through image recognition), and pressure; Historical data: The installation time, last replacement time, cumulative usage time, and past maintenance records of the GrinderBlade-456 are retrieved from the device's history records or maintenance database; Environmental data: Data that affects the repair process or component status, such as workshop temperature and humidity; Operational data: The NodeB's operation log before the abnormality occurs, such as processing parameters, workpiece type, and processing time.

[0092] Data collection is not a one-time process, but is carried out at different time points to form time series data. For example, data is collected at key time points such as the 1st minute, 5th minute, 10th minute, and 30th minute after the anomaly is confirmed, or continuously at a preset frequency (such as every minute). The collected data will be sent to the factory's data platform or edge computing node for preliminary cleaning and formatting, and then associated with the specific component GrinderBlade-456 and the abnormal event.

[0093] The system needs to determine which workpiece is currently being processed by the abnormal component. This is usually confirmed through the MES system or an RFID reader on the workstation, for example, confirming that the NodeB is currently processing workpiece TL-A2-V3. The system needs to determine which workpiece is currently being processed by the abnormal component. This is usually confirmed through the MES system or an RFID reader on the workstation, for example, confirming that the NodeB is currently processing workpiece TL-A2-V3. Process file: Extract the processing process file for the workpiece from the CAD / CAM system or process database to determine the specific location where the tool contacts the blade when the GrinderBlade-456 performs a deburring operation on the NodeB (such as a specific angle range on the blade edge). Real-time status: If supported by the system, the current position and posture of the tool, as well as information on the contact point with the workpiece, can be obtained from the NodeB's real-time motion control data. Visual information: If the workstation is equipped with a vision system, it can analyze the image to confirm the area of ​​the workpiece being processed by the tool. This "processing area" is usually described as a geometric area on the workpiece, such as "blade edge, from blade root to blade tip, with an angle range of 15° to 30°."

[0094] The system integrates and analyzes the collected repair data (component status, historical data, etc.) with the determined processing area information. For example, it analyzes the vibration and current data of the GrinderBlade-456, combines its historical wear rate with the current processing area (near the blade tip, which usually has larger burrs and wears faster), and evaluates the remaining life or damage of the tool. In combination with the processing area information, it evaluates the specific impact of continuing to use the abnormal component on the workpiece quality (for example, whether it will leave scratches on the blade edge or remove too much material). At the same time, combined with the repair workload calculated by S14, it estimates the time required for repair.

[0095] Based on preset rules or algorithms (for example, "If vibration exceeds threshold X and a critical area is being machined, the machine must be stopped for repair within Y minutes"), or based on an assessment of the impact on the production plan (for example, "If repair is not made now, TL-A2-V3 will be delayed beyond the allowable value, and the next workpiece TL-A2-V4 will arrive before repair, resulting in two consecutive workpiece delays"), the system determines a preliminary repair time point. This time point is the "first repair event," a relatively conservative or preliminary repair decision made based on all currently available information. It can be a specific date and time (for example, "2025-07-15 15:50"), a relative time (for example, "30 minutes before the next workpiece arrives"), or a condition (for example, "when the tool vibration exceeds 1.5g").

[0096] Furthermore, the autonomous repair strategy of the digital factory for the workpiece is determined based on the first repair event, the repair workload of each abnormal factory component and the Internet of Things. At this time, the autonomous repair strategy is analyzed and the autonomous repair process is output. The early repair event of each abnormal factory component is determined based on the dynamic interaction between the autonomous repair process and the production process of the workpiece, so as to repair each abnormal factory component in advance, realize the dynamic interaction between the autonomous repair process and the production process of the workpiece, and further accurately control the early repair events of each abnormal factory component. At the same time, the repair time node of each abnormal factory component is introduced, realizes the overall consideration of the repair workload of each abnormal factory component and the Internet of Things, improves the accuracy of the autonomous repair strategy of the digital factory for the workpiece, and does not affect the normal production of the workpiece, so as to repair each abnormal factory component in advance.

[0097] At this point, the system inputs the first repair event determined by S151 (preparation starts at 15:15, replacement starts at 15:20, and completion is at 15:45), the repair workload of GrinderBlade-456 (45 minutes), and the real-time information provided by the Internet of Things (such as the real-time progress of TL-A2-V3, the transportation status of TL-A2-V4, the load of other processes, the spare parts inventory status, the availability of maintenance personnel, etc.) into the advanced decision-making algorithm.

[0098] An optimal repair plan is developed to minimize the waiting time for workpiece TL-A2-V3 and delay the processing of the subsequent workpiece TL-A2-V4 as much as possible. It takes into account multiple factors: Time window: TL-A2-V3 must be replaced after the current operation is completed, and it must be completed before TL-A2-V4 arrives; Resource availability: Check whether there is a spare grinding head tool in stock; whether maintenance engineer A is free between 15:20 and 15:45; whether the robot arm that needs to be replaced requires additional debugging time; Parallel tasks: Whether engineer A can handle other low-priority maintenance tasks at the same time, or needs to focus entirely on replacing GrinderBlade-456; Risk factors: If engineer A cannot arrive immediately at 15:20, is there a backup plan (such as waiting for engineer B, which is more time-consuming).

[0099] Based on the above analysis, the system generates a preliminary autonomous repair strategy. For example, Strategy A strictly follows the first repair event. Engineer A prepares tools and spare parts at 3:15 PM and begins the replacement immediately after TL-A2-V3 completes the operation at 3:20 PM, with an estimated completion time of 3:45 PM. Strategy B: If Engineer A does have a brief conflict at 3:20 PM, the system will start preparations at 3:15 PM, but postpone the actual replacement to 3:30 PM, with an estimated completion time of 4:15 PM. This means that TL-A2-V4 will be delayed by 15 minutes. The system will select the strategy with the lowest risk and the least impact (Strategy A in this case).

[0100] The system breaks down the selected autonomous repair strategy (for example, Strategy A) and converts it into a series of specific, executable steps. This is not just "replacing GrinderBlade-456," but a more detailed process. The system generates a list of "autonomous repair procedures" and sends it to relevant execution units (such as maintenance terminals, robot control panels, material handling systems, etc.) via the Internet of Things. For the replacement of GrinderBlade-456, this list includes: 15:15: The material handling robot goes to the spare parts warehouse and takes out the spare grinding head tool numbered GB-456-BACKUP; 15:15: Engineer A's mobile terminal receives a notification: "Please go to NodeB immediately to prepare to replace GrinderBlade-456. The required tools are: wrench kit #12, torque wrench #3"; 15:15: The robot control system of NodeB receives the command: "Enter maintenance mode and prepare to release the locking device of GrinderBlade-456" ;15:20: TL-A2-V3 completes the deburring operation, and the system automatically suspends its subsequent processes;15:20: NodeB's robot control system receives the instruction: "Execute GrinderBlade-456 disassembly procedure";15:20: The material handling robot transports the old tool GB-456 to the waste area and transports the new tool GB-456-BACKUP to the designated location of NodeB;15:25: NodeB's robot control system receives the instruction: "Execute the GB-456-BACKUP installation program"15:35: NodeB's robot control system receives the instruction: "Execute the automatic calibration procedure of the new tool GB-456-BACKUP";15:45: NodeB's robot control system receives the instruction: "Exit maintenance mode and prepare to receive the next workpiece TL-A2-V4";15:45: TL-A2-V3 resumes subsequent processes;15:45: TL-A2-V4 is guided to NodeB to start the deburring process.

[0101] During the autonomous repair process, the system continuously monitors the actual situation; the IoT provides real-time feedback: the actual completion time of TL-A2-V3 (whether it is ahead of schedule or delayed); the actual arrival time of the material handling robot (whether it is on time); the actual arrival and operation time of Engineer A (whether it is as planned); the actual time required for robot disassembly, installation, and calibration (whether it is consistent with expectations); and the actual arrival time of TL-A2-V4 (whether it is on time).

[0102] The system compares and analyzes the real-time progress of the autonomous repair process with the production process of workpieces TL-A2-V3 and TL-A2-V4. Scenario 1 (as planned): If everything goes according to plan, TL-A2-V3 is completed at 15:20, the replacement is completed at 15:45, and TL-A2-V4 arrives at 15:45 and begins processing immediately. The "early repair event" would then be the replacement starting at 15:20, as originally planned. Scenario 2 (deviation): Suppose that due to a minor processing deviation, the deburring operation on TL-A2-V3 is actually completed at 15:18 and paused. Simultaneously, Engineer A arrives five minutes earlier than planned (15:10) and prepares the tool. The IoT captures these changes.

[0103] TL-A2-V3 has been completed, 2 minutes ahead of schedule; Engineer A is in place, 5 minutes ahead of schedule; the new tool is also in place (assuming the material robot is also ahead of schedule); although the original plan was to start the replacement at 3:20 PM, there is now a window to start earlier, such as 3:15 PM, which can further shorten the total waiting time for TL-A2-V3 (from 2 minutes to 0) and ensure that the NodeB is absolutely ready when TL-A2-V4 arrives at 15:45 PM; the system assesses the risk: whether starting the replacement early will affect Engineer A's other arrangements and whether there is enough time to complete the calibration. The conclusion is that it is feasible; the system generates an "early repair event": an instruction to advance the replacement of GrinderBlade-456 to 3:15 PM; based on this new "early repair event", the system dynamically adjusts the start time of the autonomous repair process and notifies the relevant execution units.

[0104] See also Figure 7 , Figure 7 : This is a schematic diagram of the structural composition of the production traceability control system of the digital factory under the Internet of Things in an embodiment of the present invention; the production traceability control system of the digital factory under the Internet of Things includes: The digital intelligence database module 21 is used to collect multiple production data combinations at different production nodes of the digital intelligence factory and build a digital intelligence database based on each production data combination and the Internet of Things; The abnormal production path module 22 is used to trace the production source of multiple abnormal production data in each production data combination based on the digital database to determine the corresponding abnormal production nodes, and determine the abnormal production path according to the location of each abnormal production node and the abnormal range of the abnormal production data; The production abnormality area module 23 is used to determine the current production path according to the current working status of the digital factory and the model of the workpiece, and determine the production abnormality area of ​​the workpiece based on the current production path and the abnormal production path; A repair workload module 24 is configured to determine a repair time node for each abnormal factory component based on the abnormal factory component corresponding to the production abnormality area, the Internet of Things, and the current production location of the workpiece, and to determine a repair workload for each abnormal factory component based on the repair time node for each abnormal factory component and the production process table corresponding to the workpiece; The autonomous repair strategy module 25 is used to determine the autonomous repair strategy of the digital factory for the workpiece based on the repair workload of each abnormal factory component and the Internet of Things, without affecting the normal production of the workpiece, so as to repair each abnormal factory component in advance.

[0105] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

Claims

1. A production traceability control method for a digital factory under the Internet of Things, characterized by: include: Collect multiple production data combinations from different production nodes of the digital factory and build a digital database based on each production data combination and the Internet of Things; Multiple abnormal production data in each production data combination are traced based on the digital database to determine the corresponding abnormal production nodes. The abnormal production path is determined based on the location of each abnormal production node and the abnormal range of the abnormal production data. Determine the current production path based on the current working status of the digital factory and the model of the workpiece, and determine the production abnormality area of ​​the workpiece based on the current production path and the abnormal production path; Determine the repair time node for each abnormal factory component based on the abnormal factory component corresponding to the production abnormality area, the Internet of Things, and the current production location of the workpiece; and determine the repair workload for each abnormal factory component based on the repair time node for each abnormal factory component and the production process table corresponding to the workpiece; Based on the repair workload of each abnormal factory component and the Internet of Things, the digital factory determines the autonomous repair strategy for the workpiece without affecting the normal production of the workpiece, so as to repair each abnormal factory component in advance.

2. The production traceability control method of a digital factory under the Internet of Things according to claim 1 is characterized in that: The method of collecting multiple production data combinations of the digital factory at different production nodes and building a digital database based on each production data combination and the Internet of Things includes: When the digital factory performs unmanned processing on a workpiece, the digital factory collects the digital process of the workpiece, and determines multiple autonomous production processes based on the division of the digital process. In addition, multiple production nodes are determined based on the multiple autonomous production processes and the model of the workpiece; Multiple production nodes are distributed in different locations of the digital factory and target different processing steps of the workpiece. Multiple production nodes are monitored in real time, and multiple production data of each production node are collected. In each production node, the corresponding production data combination is determined based on the synthesis of multiple production data. A digital database is constructed based on the node position of the production node, the corresponding production data combination and the synthesis of the Internet of Things corresponding to the digital factory. The digital database is dynamically updated as the production data increases.

3. The production traceability control method of a digital factory under the Internet of Things according to claim 1 is characterized in that: The multiple abnormal production data in each production data combination are traced based on the digital database to determine the corresponding abnormal production nodes, and the abnormal production path is determined according to the location of each abnormal production node and the abnormal range of the abnormal production data, including: Collect various production data combinations, determine corresponding abnormal production detection patterns based on the matching between each production data combination and the digital intelligence database, and trigger abnormality detection for each production data combination along the abnormal production detection pattern to output multiple abnormal production data; Mark the abnormal locations corresponding to multiple abnormal production data, determine the corresponding production traceability method based on the abnormal location, the corresponding abnormal production data, and the digital database, and determine multiple traceability items based on the analysis of the production traceability method. The multiple traceability items cross-trace the workpiece and trigger the production traceability of multiple abnormal production data to determine multiple abnormal production nodes; At each abnormal production node, the abnormal range of the abnormal production data is determined based on the comparison of the corresponding multiple abnormal production data, and the abnormal production path is determined based on the position of each abnormal production node, the abnormal range of the abnormal production data and the workpiece.

4. The production traceability control method of a digital factory under the Internet of Things according to claim 1 is characterized in that: The method of determining the current production path according to the current working state of the digital factory and the model of the workpiece, and determining the production abnormality area of ​​the workpiece based on the current production path and the abnormal production path includes: Collect the response signals of each production node in the digital factory, determine the status coefficient of each production node based on the analysis of each response signal, and determine the current working status of the digital factory based on the node position, status coefficient and the number of processed workpieces corresponding to the digital database of each production node.

5. The production traceability control method of a digital factory under the Internet of Things according to claim 4 is characterized in that: The method of determining a current production path according to the current working state of the digital factory and the model of the workpiece, and determining a production abnormality area of ​​the workpiece based on the current production path and the abnormal production path, further includes: Collect the workpiece model, determine the workpiece processing form based on the workpiece model and external form, and determine the current production path based on the workpiece processing form, the current working status of the digital factory, and multiple production nodes; The current production path is compared with the abnormal production path, and the overlapping area is determined based on the comparison between the current production path and the abnormal production path. Multiple sub-abnormal areas are determined based on the detection of the overlapping area, and the production abnormal area of ​​the workpiece is determined based on the synthesis of multiple sub-abnormal areas.

6. The production traceability control method of a digital factory under the Internet of Things according to claim 1 is characterized in that: The method includes determining a repair time node for each abnormal factory component based on the abnormal factory component corresponding to the production abnormality area, the Internet of Things, and the current production location of the workpiece, and determining a repair workload for each abnormal factory component according to the repair time node of each abnormal factory component and the production process table corresponding to the workpiece, including: Acquiring the abnormal production area, determining a plurality of component features based on the detection of the abnormal production area, and determining an abnormal factory component corresponding to the abnormal production area based on the plurality of component features and a captured image corresponding to the abnormal production area; Mark the Internet of Things of the digital factory, and based on the Internet of Things, perform abnormal management and control of each abnormal factory component, and output the working data of each abnormal factory component. Determine the repair time node of each abnormal factory component according to the working data of each abnormal factory component and the current production location of the workpiece.

7. The production traceability control method of a digital factory under the Internet of Things according to claim 6 is characterized in that: The method further includes determining a repair time node for each abnormal factory component based on the abnormal factory component corresponding to the production abnormality area, the Internet of Things, and the current production location of the workpiece, and determining a repair workload for each abnormal factory component based on the repair time node for each abnormal factory component and the production process table corresponding to the workpiece. Collect the production process table corresponding to the workpiece, determine the production time of the workpiece in each production process based on the analysis of the production process table, and determine the repair workload of each abnormal factory component according to the repair time node of each abnormal factory component, the production time of the workpiece in each production process and the workload of the abnormal factory department in the production process.

8. The production traceability control method of a digital factory under the Internet of Things according to claim 1 is characterized in that: The above-mentioned autonomous repair strategy of the digital factory for each abnormal factory component is determined based on the repair workload of each abnormal factory component and the Internet of Things, without affecting the normal production of the workpiece, so as to repair each abnormal factory component in advance, including: The Internet of Things conducts abnormal control on each abnormal factory component and collects the repair data of each abnormal factory component at different time nodes. At the same time, it collects the processing area of ​​the workpiece relative to the abnormal factory component and determines the first repair event based on multiple repair data and the processing area of ​​the workpiece.

9. The production traceability control method of a digital factory under the Internet of Things according to claim 8 is characterized in that: The method of determining the autonomous repair strategy of the digital factory for each abnormal factory component based on the repair workload of each abnormal factory component and the Internet of Things, without affecting the normal production of the workpiece, so as to repair each abnormal factory component in advance, also includes: Based on the first repair event, the repair workload of each abnormal factory component and the Internet of Things, the digital factory determines the autonomous repair strategy for the workpiece. At this time, the autonomous repair strategy is analyzed and the autonomous repair process is output. Based on the dynamic interaction between the autonomous repair process and the production process of the workpiece, the early repair event of each abnormal factory component is determined to repair each abnormal factory component in advance.

10. A production traceability control system for a digital factory under the Internet of Things, characterized by: The production traceability control system of the digital chemical factory under the Internet of Things is applied to the production traceability control method of the digital chemical factory under the Internet of Things as claimed in any one of claims 1 to 9. The production traceability control system of the digital chemical factory under the Internet of Things includes: The digital and intelligent database module is used to collect multiple production data combinations at different production nodes of the digital and intelligent factory and build a digital and intelligent database based on each production data combination and the Internet of Things; The abnormal production path module is used to trace the production source of multiple abnormal production data in each production data combination based on the digital database to determine the corresponding abnormal production nodes. The abnormal production path is determined according to the location of each abnormal production node and the abnormal range of the abnormal production data; The production anomaly area module is used to determine the current production path based on the current working status of the digital factory and the model of the workpiece, and to determine the production anomaly area of ​​the workpiece based on the current production path and the abnormal production path; A repair workload module is used to determine the repair time node of each abnormal factory component based on the abnormal factory component corresponding to the production abnormality area, the Internet of Things, and the current production location of the workpiece, and to determine the repair workload of each abnormal factory component based on the repair time node of each abnormal factory component and the production process table corresponding to the workpiece; The autonomous repair strategy module is used to determine the autonomous repair strategy of the digital factory for each abnormal factory component based on the repair workload of the component and the Internet of Things, without affecting the normal production of the component, so as to repair each abnormal factory component in advance.

Images