Monitoring method suitable for wooden furniture production line

Abstract

The invention discloses a monitoring method suitable for a wooden furniture production line, and relates to the technical field of wooden furniture production monitoring, and the method comprises the steps: S1, modular monitoring node deployment: deploying a plurality of modular monitoring nodes with sensing, processing and communication functions at each key station of processing equipment, generating unique identification information for each monitoring node, and transmitting the unique identification information to a server; the identification information is generated by a preset rule, and a registration request containing identity information, a node type and an equipment state is sent through a network. Monitoring nodes are arranged to send an automatic registration capability, a unique identifier is bound to a workpiece, state information of the whole process is collected and recorded, a complete product resume chain is constructed, quality traceability and problem positioning are facilitated, a knowledge graph modeling and graph structure reasoning method is introduced, potential faults are recognized in advance, and the quality of a product is improved. The multi-target linkage control is executed according to the risk level, and the real-time performance and the precision of the control are improved through the cooperative response of a standard interface, equipment and an alarm system.

Description

Technical Field

[0001] The present invention relates to the technical field of wood furniture production monitoring, and particularly relates to a monitoring method applicable to a wood furniture production line. Background Art

[0002] With the rapid development of the furniture manufacturing industry towards intelligence and automation, the requirements for safety monitoring and operation management of wood furniture production lines are increasing day by day. Problems such as equipment abnormalities, workpiece defects, and environmental risks that may occur during the production process, if not discovered and processed in time, are extremely likely to cause chain reactions, affecting product quality and personnel safety. Therefore, building an efficient, intelligent, and scalable production line monitoring system has become an important topic in the industry.

[0003] In the prior art, the disclosed patent CN106919130A proposes a zonal monitoring method for a wood furniture production line. By dividing the production site into multiple independent zones, detection devices, alarm devices, and isolation devices are deployed in each zone, and a control system is constructed by a processor to achieve status monitoring and isolation control of local areas. This solution has the advantages of simple structure, fast triggering, and strong regional control ability, and has a certain effect in improving the stability and safety of the system.

[0004] However, traditional zonal monitoring methods including this solution still have the following deficiencies: the monitoring systems of each zone are relatively independent, there is a lack of a coordination mechanism between systems, and it is impossible to flexibly adjust the deployment and logical configuration according to changes in the production line structure. Most monitoring events rely on alarms after sensor overrun or failure; it is impossible to predict potential risks or trend abnormalities in advance, focusing on regional status monitoring, unable to achieve status tracking and quality data recording of a single workpiece between multiple processes, and difficult to meet the requirements of product-level quality traceability. The control response is based on a preset program, making it difficult to achieve adaptive judgment and multi-target linkage; the data storage and management of each zone are decentralized, there are information islands, and it is difficult to support full-line-level data analysis and optimization decisions.

[0005] In summary, the existing monitoring technologies are unable to cope when faced with flexible manufacturing, complex process chains, and high-frequency change scenarios. There is an urgent need for a new monitoring method with modular deployment capabilities, supporting automatic identification and configuration, having the functions of full-process tracking of workpieces, intelligent prediction of abnormalities, and multi-target linkage control to better meet the comprehensive requirements of modern wood furniture production lines for intelligent, safe, and flexible monitoring. Summary of the Invention

[0006] The object of the present invention is to provide a monitoring method applicable to a wood furniture production line to solve the above deficiencies in the prior art.

[0007] To achieve the above object, the present invention provides the following technical solutions: A monitoring method applicable to a wooden furniture production line, the method comprising the following steps: S1. Modular monitoring node deployment: Multiple modular monitoring nodes with sensing, processing and communication functions are deployed at each key workstation of the processing equipment, and unique identification information is generated for each monitoring node. The identification information is generated by preset rules and a registration request containing identity information, node type and equipment status is sent through the network. S2. Automatic identification and process construction: When the monitoring node is powered on, it automatically sends registration information to the central monitoring platform through the network. The registration information includes the node's unique identifier, type, function configuration, etc. The central monitoring platform automatically configures the tasks of each node and constructs the production line monitoring flowchart according to the registration information, and updates the production line monitoring flowchart in real time when the monitoring node changes. S3. Workpiece process tracking: By attaching identification information to the workpiece and recording the status, parameters and event information related to the workpiece through each monitoring node, the workpiece can be tracked and data bound throughout different processes. S4. Knowledge graph modeling and reasoning: The central monitoring platform constructs a knowledge graph model based on process flow, equipment status, historical fault information, workpiece characteristics, and environmental data to establish the relationship between monitoring nodes and express the association pattern of potential faults. S5. Risk prediction: The central monitoring platform cleans and standardizes the real-time monitoring data uploaded by each monitoring node, and then performs feature matching or knowledge graph reasoning with the knowledge graph to determine whether there are potential risks and generate early warning signals. S6. Linkage control: If a fault is detected, an early warning message is generated and the downstream control system is linked to execute control actions, including shutdown control and alarm prompts. At the same time, a maintenance task order is generated and the early warning message is bound to the relevant workpiece for archiving.

[0008] Preferably, the modular monitoring node includes: a sensing unit for collecting information related to the workpiece or equipment, such as images, temperature and humidity, vibration, and current; The communication unit is used to interact with the central monitoring platform via wired or wireless means; The processing unit is used to perform preliminary analysis on the collected data and encapsulate it into registration requests or monitoring data.

[0009] Preferably, the unique identification information consists of the factory-programmed serial number, MAC address, or QR code identifier of the monitoring node, and the unique identifier is automatically read and a registration request is sent when the monitoring node is powered on or connected to the network for the first time.

[0010] Preferably, the registration information includes the unique identifier of the monitoring node, node type, workstation location, current firmware version, supported sensor types and their acquisition frequency parameters; After receiving the registration information, the central monitoring platform automatically assigns corresponding task configurations to the monitoring nodes according to the preset task template library. The task configurations include data collection cycle, upload path, early warning threshold and processing logic.

[0011] Preferably, the central monitoring platform automatically generates a monitoring flowchart representing the production line process flow based on the registration order, workstation location, and node function of each monitoring node using a graph structure algorithm. The flowchart is a directed graph, where nodes represent process units and edges represent process sequence or data flow direction. The flowchart is stored and maintained through a graph database or topology management module, and automatically adds, replaces or deletes flowchart nodes when monitoring nodes are added, disconnected or replaced due to faults.

[0012] Preferably, the identification information of the workpiece is a QR code label, RFID electronic tag or shape feature code extracted by image recognition attached to the surface of the workpiece, used to identify each workpiece. The identification information of each workpiece remains consistent during the process flow. The central monitoring platform integrates the workpiece status data, environmental parameters, event records, etc. uploaded by each monitoring node in chronological order to form the monitoring data generation timestamp and node information of each workpiece, and stores it in the workpiece history database for subsequent quality analysis.

[0013] Preferably, the central monitoring platform performs rule matching or graph traversal on the paths, node types, and attribute combinations in the knowledge graph to achieve reasoning analysis of potential fault causes and risk points; when constructing the knowledge graph, the central monitoring platform automatically extracts common fault patterns based on historical monitoring data and performs knowledge graph reasoning through graph structure for subsequent real-time data comparison and fault warning triggering.

[0014] Preferably, the knowledge graph reasoning includes rule reasoning and relational path reasoning. The rule reasoning is based on the logical relationships defined in the graph, and the path reasoning is based on the logical strength calculation of the relational chain between nodes. It is used to mine potential fault causal chains. If the knowledge graph reasoning or feature matching judgment has potential risks, the central monitoring platform generates an early warning data packet containing the risk level, predicted fault type, affected process, and associated workpiece number, and sends it to the human-machine interface and control system simultaneously.

[0015] Preferably, the linkage control is implemented through a preset control rule base, which includes multiple "condition-action" pairs. When the risk level, fault type, or workstation number contained in the warning information meets a certain rule condition, the corresponding control action is automatically executed.

[0016] Preferably, the linkage control action is relayed through an edge control gateway to achieve localized response. After receiving the control command from the central monitoring platform, the edge control gateway can cache, confirm, and perform secondary verification locally before executing the specific action.

[0017] In the above technical solution, the present invention provides a monitoring method applicable to wooden furniture production lines. It achieves the binding of unique identifiers to workpieces by deploying monitoring nodes to send automatic registration capabilities, collects and records the status information of the entire process, constructs a complete product history chain, which facilitates quality traceability and problem localization, and introduces knowledge graph modeling and graph structure reasoning methods to achieve early identification of potential faults, executes multi-objective linkage control according to risk level, and coordinates with equipment and alarm systems through standard interfaces to improve the real-time performance and accuracy of control.

[0018] It should be understood that the foregoing general description and the following detailed description are exemplary and illustrative only, and are not intended to limit this disclosure.

[0019] This application provides an overview of various implementations or examples of the technology described in this disclosure, and is not a full disclosure of the entire scope or all features of the disclosed technology. Attached Figure Description

[0020] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this invention. For those skilled in the art, other drawings can be obtained based on these drawings.

[0021] Figure 1 A monitoring flowchart provided for embodiments of the present invention; Figure 2 This is a schematic diagram of the knowledge graph construction process provided in an embodiment of the present invention. Detailed Implementation

[0022] To make the objectives, technical solutions, and advantages of the embodiments of this disclosure clearer, the technical solutions of the embodiments of this disclosure will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this disclosure. All other embodiments obtained by those skilled in the art based on the described embodiments of this disclosure without creative effort are within the scope of protection of this disclosure.

[0023] Reference Figure 1-2 As shown, the present invention provides a monitoring method suitable for wooden furniture production lines, the method comprising the following steps: S1. Modular monitoring node deployment: Multiple modular monitoring nodes with sensing, processing and communication functions are deployed at each key workstation of the processing equipment. Unique identification information is generated for each monitoring node. The identification information is generated by preset rules and a registration request containing identity information, node type and equipment status is sent through the network. S2. Automatic identification and process construction: When the monitoring node is powered on, it automatically sends registration information to the central monitoring platform through the network. The registration information includes the node's unique identifier, type, function configuration, etc. The central monitoring platform automatically configures the tasks of each node and constructs the production line monitoring flowchart based on the registration information, and updates the production line monitoring flowchart in real time when the monitoring node changes. S3. Workpiece process tracking: By attaching identification information to the workpiece and recording the status, parameters and event information related to the workpiece through each monitoring node, the workpiece can be tracked and data bound throughout different processes. S4. Knowledge graph modeling and reasoning: The central monitoring platform constructs a knowledge graph model based on process flow, equipment status, historical fault information, workpiece characteristics, and environmental data to establish the relationship between monitoring nodes and express the association pattern of potential faults. S5. Risk prediction: The central monitoring platform cleans and standardizes the real-time monitoring data uploaded by each monitoring node, and then performs feature matching or knowledge graph reasoning with the knowledge graph to determine whether there are potential risks and generate early warning signals. S6. Linkage control: If a fault is detected, an early warning message is generated and the downstream control system is linked to execute control actions, including shutdown control and alarm prompts. At the same time, a maintenance task order is generated and the early warning message is bound to the relevant workpiece for archiving.

[0024] Specifically, S1: Modular monitoring node deployment. Modular monitoring nodes with the following functions are deployed at multiple key workstations in the wooden furniture production line (e.g., cutting, drilling, milling, assembly, quality inspection, etc.): Sensing units include, but are not limited to, image acquisition devices (industrial cameras), environmental monitoring sensors (temperature, humidity, particulate matter, etc.), vibration sensors, current sensors, etc. Processing unit: Select an embedded processor that supports edge computing (such as an ARM or x86 architecture module) to perform preliminary calculations, filtering, and compression on the collected data; Communication unit: Supports industrial communication protocols (such as Modbus-TCP, MQTT, EtherCAT, etc.) to achieve bidirectional communication with the central monitoring platform.

[0025] Each monitoring node is pre-defined with a unique identifier during the manufacturing or initialization phase, such as the code (DeviceID), MAC address, or QR code for each processing device. Through the established coding rules, the identifiability and manageability of each monitoring node are ensured.

[0026] After the node is powered on, it encapsulates the device identity information, type, status parameters, firmware version, etc. into a registration request over the network and actively sends it to the central monitoring platform.

[0027] S2: Automatic Identification and Process Construction. After receiving a registration request, the central monitoring platform automatically identifies the node based on its identifier and functional information. The central monitoring platform has pre-set standard process flows and node function templates. Identification and monitoring node functions: such as image detection nodes, vibration monitoring nodes, visual recognition nodes, etc.; Automatic workstation binding: If the node number corresponds to "Station05", then it will be attached to the punching process; Automatic task configuration: Parameters such as data collection cycle, upload path, and early warning threshold are assigned to each monitoring node. The central monitoring platform automatically constructs a monitoring flowchart representing the process flow based on the deployment order and workstation location of each node, using a graph structure for management (e.g., a directed graph structure where nodes represent monitoring modules and edges represent the process flow sequence). When a node is replaced, disconnected, or added, the central monitoring platform automatically updates the flowchart based on changes in registration information, maintaining real-time accuracy of the system topology.

[0028] The "Production Line Monitoring Flowchart" refers to a structured, visual, and dynamically updated production line topology diagram automatically generated by the central monitoring platform based on the registration information, functional type, physical location, and process sequence of each modular monitoring node. It organizes the monitoring nodes in a graph structure, representing their positional relationships, data flow, and control dependencies throughout the entire production process. This is used for unified scheduling of monitoring tasks, presentation of monitoring status, and support for fault location and process management. It typically exists in the form of a directed graph or an ordered linked list of nodes, and consists of the following elements: Node: Represents an actual monitoring module, such as an image recognition node, vibration monitoring node, etc. Each node has a unique identifier and attributes, such as workstation name, function type, data status, etc. Edge: Represents the flow relationship of workpieces or information between nodes, that is, the sequence of the process flow, or the data dependency or linkage path between nodes; Attribute information: Each node and edge contains visual attributes, such as equipment operating status, workpiece number, acquisition frequency, abnormal status markers, etc. Legends and Layers: Flowcharts support multi-layer structures, such as process layer, alarm layer, and data layer, making it easy for users with different roles to view them.

[0029] S3: Workpiece process tracking. Each workpiece to be processed is assigned a unique identifier, such as an RFID tag, QR code, or visual code, which remains unchanged throughout the workpiece processing. Each monitoring node is equipped with a code reader (such as an RFID reader or industrial camera) to read the workpiece identifier passing through that station and bind the monitoring data to the corresponding workpiece.

[0030] The bound data includes: workpiece number, current workstation name, timestamp, collected environmental data (temperature, humidity), equipment operation data (current, vibration), and image detection results (such as defect detection images). The central monitoring platform automatically generates the production history chain of the workpiece, which can be used for quality traceability and statistical analysis.

[0031] S4: Knowledge graph modeling and reasoning. The central monitoring platform constructs a knowledge graph model based on information such as production process, equipment status, historical faults, environmental factors, and workpiece characteristics, and manages it using a graph database (such as Neo4j).

[0032] The central monitoring platform constructs and maintains a knowledge graph model based on graph structure. This model is used to integrate multi-source heterogeneous operating data in the production line, mine the implicit correlations between equipment status, process flow, workpiece characteristics, environmental factors and fault events, and realize intelligent expression and reasoning judgment of potential faults.

[0033] The knowledge graph model structure consists of the following two core elements: Entity nodes: Represent the basic identifiable objects in the production line, including but not limited to equipment nodes such as cutting machines, drilling machines, assembly machines, and testing units; Monitoring nodes include: image recognition units, vibration monitoring units, temperature and humidity sensing modules, etc. Process nodes: such as specific procedures like drilling, milling, and painting; Workpiece node: Each processed object, such as a door panel or drawer panel, has a unique identifier; Environmental nodes: represent external environmental conditions such as temperature, humidity, particle concentration, and illuminance; Fault nodes: such as abnormal events like "tool wear", "belt slippage", "abnormal feeding", and "blurred image"; Operation and maintenance nodes: such as maintenance records, alarm events, maintenance personnel and other auxiliary information.

[0034] Relationship edges; indicating the association between the above nodes, including "occurred at" relationships: such as "tool wear" occurring at "cutting equipment"; "Detected" relationship: such as "vibration monitoring node" detecting "abnormal vibration value"; "Influence" relationships: For example, "high ambient humidity" affects "gluing process quality"; "Cause" relationship: For example, "image blurriness" may lead to "visual recognition failure"; "Belongs to" relationship: For example, "workpiece WB20251015" belongs to "batch 5"; "Location association" relationship: such as "Node A" and "Node B" being located in the same process section.

[0035] The knowledge graph model is built based on the following data sources: real-time monitoring data, images uploaded by each node, sensor readings, workpiece status, etc. Historical fault data, alarm logs, maintenance records, downtime, and fault types; Process flow data, including process sequence, cycle time, and standard operating procedures; Environmental monitoring data, including factory temperature and humidity, dust concentration, and lighting conditions; Workpiece tracking data, including workpiece number, path, and status change history. The central monitoring platform uses graph databases (such as Neo4j and ArangoDB) for graph structure modeling and uses ETL processes to uniformly map structured and unstructured data into knowledge graph model entities and relation edges.

[0036] The central monitoring platform performs the following types of intelligent reasoning based on the knowledge graph model: Path similarity matching: It compares the "current status graph" constructed from the current real-time data with the historical fault paths. If the graph path structure and node features are highly similar, it is judged as a potential anomaly. Subgraph matching: Identifies whether a subgraph structure containing risk patterns exists; Graph Neural Network (GNN) Prediction: This method uses graph convolutional networks to embed node states for learning, enabling device health scoring and fault prediction. Rule-driven reasoning: Several rules are preset (such as "X + Y + Z → Fault A"). When the monitoring data meets the rule, an alert is triggered.

[0037] S5: Risk Prediction. The central monitoring platform, based on knowledge graphs and real-time data fusion, cleans, standardizes, and intelligently analyzes the monitoring data uploaded by each monitoring node. This enables the prediction of potential faults or abnormal trends and generates actionable early warning signals. This step includes the following key technical aspects: real-time data cleaning and standardization. The central monitoring platform receives data reporting packets from each monitoring node in time windows (e.g., every 10 seconds or every minute). The content may include image data, environmental parameters, process status, current / vibration readings, workpiece identification codes, etc.

[0038] To ensure the accuracy of the analysis results, when processing real-time data uploaded by monitoring nodes, if some data is found to be missing, methods such as linear interpolation, moving average, or using the previous valid data can be used to fill in the missing items to ensure data integrity and the stability of subsequent analysis, thus ensuring the continuity and availability of time-series data. The central monitoring platform needs to perform the following preprocessing steps on the data: handling missing values ​​by filling in the missing data using linear interpolation, moving average, or using the previous valid data to ensure continuity. In actual monitoring, due to network latency, node failures, sensor anomalies, etc., data at certain time points may be missing. For example, temperature data should be uploaded every second, but it is not uploaded in a certain second; the equipment vibration sensor suddenly loses power, generating null values; image acquisition loses frames, resulting in empty image data. To avoid affecting subsequent analysis or knowledge graph reasoning, these missing values ​​must be reasonably filled in to ensure the continuity and reliability of the data.

[0039] Linear interpolation calculates the missing value (e.g., time-temperature (°C)) linearly using two known values ​​before and after the missing value. 10:00:01 25.0 10:00:02 (Missing) 10:00:03 27.0 Missing value = 25.0 + (27.0 - 25.0) × 0.5 = 26.0℃ A moving average is used, taking data from several time points before and after the missing value, calculating the average, and filling it in; for example, given the value sequence: 24.8, 25.2, The sliding window has 3 points: 25.5, 25.1. Missing values ​​≈ (25.2 + 25.5 + 25.1) / 3 = **25.27**, which is suitable for situations where data fluctuations are relatively stable.

[0040] If the data loss period is short, the previous valid value can be used to replace the missing item; for example, time-current (A). 10:00:01 1.20 10:00:02

[0041] 10:00:03

[0042] 10:00:04 1.18 Then the missing points can be filled with 1.20.

[0043] Outlier removal is achieved by using IQR (interquartile range) or Z-score methods to identify and filter outlier values.

[0044] In actual industrial production monitoring data, sensors occasionally produce extreme values ​​(such as sudden increases in current or sudden drops in temperature) due to jitter, electromagnetic interference, network latency, or instantaneous fluctuations. If this data is not processed, it will severely interfere with fault prediction algorithms or knowledge graph inference results. Before performing real-time risk prediction, the central monitoring platform performs outlier detection and removal on various monitoring data. The following two mainstream statistical methods are mainly used: Method 1: IQR (Interquartile Range) is a commonly used outlier identification tool in statistics. It detects "extreme values" far from the main part of the data by analyzing the median and quartiles of the data.

[0045] It allows you to set a time window for certain types of sensor data (such as temperature readings within the last 5 minutes). Calculate the following for the dataset: First quartile Q1 (25th percentile), Third quartile Q3 (75th percentile), Interquartile Range (IQR) = Q3 - Q1; Set the abnormal threshold range: lower limit = Q1 - 1.5 × IQR, upper limit = Q3 + 1.5 × IQR, judgment: values ​​exceeding this range are judged as abnormal values ​​and are removed or replaced.

[0046] If Q1 = 25.0 and Q3 = 30.0, then IQR = 5.0, and the range = [25 - 7.5, 30 + 7.5] = [17.5, 37.5]. All temperature values ​​outside this range (such as 45.2 or 12.1) will be considered abnormal.

[0047] Method 2: Z-score standard deviation method, which represents the degree of deviation of a data point from the mean, expressed in standard deviation (σ). For normally distributed data, 99.7% fall within ±3σ; It calculates the mean (μ) and standard deviation (σ) of the data sequence: for each value, it calculates its Z-score: Set a threshold (usually ±3), that is: if the Z-score > 3 or < -3, it is considered an outlier.

[0048] S6: Linkage Control. After the central monitoring platform completes the map identification or feature matching of potential risks, if it determines that there is a fault or risk event, it will immediately activate the linkage control mechanism to coordinate the response control operations of the central monitoring platforms or equipment downstream of the production line, so as to ensure timely blocking of risks, reduction of losses, and protection of product quality and operational safety.

[0049] The linkage control process includes the following key steps: generation and structuring of early warning information. When the central monitoring platform identifies a risk event, it immediately generates an early warning data packet (also known as an "early warning information sheet") containing structured control instructions. The data packet includes: early warning number: a globally unique identifier; Risk types: such as tool wear, abnormal workpiece dimensions, and excessive vibration; Risk level: High / Medium / Low (based on inference confidence and scope of impact); Equipment or nodes involved: such as Station04 (milling station), Node-IR17; Trigger data summary: including key sensor readings, image scores, historical paths, etc.; Associated workpiece number: The workpiece that triggered this risk assessment, such as WB2025101506; Suggested response actions: such as shutdown, alarm activation, reduced speed operation, manual re-inspection, etc. Timestamp: The point in time when a risk was identified and an early warning was issued; Among them, the reasoning confidence level represents the degree of credibility of the knowledge graph or algorithm model for the current risk judgment result, and the value range is usually 0~1. For example: graph structure matching degree ≥ 95% → high confidence; Multiple feature nodes simultaneously satisfy historical path → high confidence; Graph embedding cosine similarity close to 1 → high confidence; A single outlier triggers the signal, with no other auxiliary features → low confidence.

[0050] Risk level Confidence Scope of influence Response urgency Example high ≥0.90 Multiple workstations / workpiece batches / high-value equipment Immediate shutdown required / Employee safety related Tool breakage tendency, excessive vibration, and severe image defects middle 0.60 - 0.89 Single workstation / Local processing available It can be delayed until the current process is completed. Material feeding deviation, insufficient lighting, and abnormal bonding trends Low < 0.60 Localized minor anomalies / Insufficient confidence Monitoring or manual re-inspection is recommended. Humidity threshold, slight image shift, minor vibration fluctuations Depending on the risk level, the central monitoring platform will adopt different control and response strategies: Risk level Linkage control Employee Response Workpiece processing Recording strategy high Immediate shutdown + alarm + dispatch It must be dealt with immediately. Workpiece isolation and warehousing / random inspection Forced recording and archiving middle Pop-up notification + recurring reminder Response within 3 hours Enter the sampling queue System tagging + tracking Low The system records internally and does not trigger an alarm. Ignore or pending Continue normally, mark Log recording The coordinated distribution of control commands, based on the risk type and the coordinated control rule base, determines which modules to send the warning command to: The central monitoring platform for equipment control sends commands to the corresponding controllers via industrial communication protocols (such as Modbus, OPC UA, EtherCAT, etc.): Stop command: such as immediately stopping the current cutting equipment; Pause for material: postpone the feeding of the next process; Reset control: restart the production process after the risk is eliminated.

[0051] The central monitoring platform for audible and visual alarms controls the alarm lights and buzzers at workstations or in the workshop: red light flashes + buzzer sounds for 3 seconds; the alarm LCD panel displays "risk type + handling suggestion".

[0052] If the risk level of the safety protection and isolation module is "high", the safety isolation device will be activated: the conveyor belt will be shut down; the physical barrier will be activated; and the electrical power failure protection device will be triggered.

[0053] The central monitoring platform for personnel dispatch pushes the following to maintenance terminals or repair workstations: maintenance task orders; risk details and suggested actions; and task acceptance function for repair personnel.

[0054] When a fault type requires manual intervention (such as tool replacement, image calibration, or cleaning), the central monitoring platform automatically generates a "Maintenance Task Order." This task order is then uploaded to the maintenance terminal's central monitoring platform, mobile industrial control tablet, or MES central monitoring platform for maintenance personnel to accept and execute. The maintenance task order includes: a task number, such as MT-20251015-01; Fault description: Abnormal vibration value + increased image defects, inferred to be tool wear. Suggested procedure: Stop the machine, check the cutting tools, and clean up debris. Assigned Personnel: Maintenance Technician ID, which can be set by the central monitoring platform or the team leader. Deadline: Response time is automatically generated based on equipment importance and cycle time. Attachments: Includes relevant images, monitoring data, workpiece numbers, etc. Workstation location: Positioned on a 2D factory map or by workstation number; By linking with the central monitoring platform for workpiece tracking, product-level quality traceability and responsibility allocation are achieved. The central monitoring platform binds the early warning information to the affected workpiece, forming a traceable chain of workpiece anomaly records. For example: Workpiece number: WB2025101506 Status: High image defect rate during assembly process + Abnormal vibration at Station 04 Warning type: Potential tool wear Recommended action: Proceed to manual re-inspection process. Archive path: Quality history central monitoring platform → Workpiece → Risk record item, which is used for customer complaint analysis; internal responsibility division; production optimization traceability; quality statistical analysis.

[0055] Fault closed-loop control and resolution: The central monitoring platform supports a complete closed loop of early warning → control → maintenance → recovery. The maintenance personnel completed the maintenance operation; The central monitoring platform received a "task completed" notification. The control center monitoring platform verifies that the risk has been eliminated (e.g., parameters have returned to normal). Automatically resume equipment operation or prompt for manual restart; Early warning records are archived.

[0056] In this invention, to achieve tiered response and refined control of risk events, the central monitoring platform classifies potential faults or risks into high, medium, and low levels based on multi-dimensional indicators such as the confidence level of the knowledge graph reasoning results, the scope of the anomaly's impact, and the urgency of the response. The central monitoring platform has a risk level determination rule base or scoring function that supports adaptive dynamic adjustment and selects different linkage control paths and early warning methods according to the risk level, significantly improving the accuracy of the monitoring system's response and its overall intelligence level.

[0057] In another embodiment of the present invention, the modular monitoring node includes: a sensing unit for collecting information such as images, temperature and humidity, vibration, and current related to the workpiece or equipment; The communication unit is used to interact with the central monitoring platform via wired or wireless means; The processing unit is used to perform preliminary analysis on the collected data and encapsulate it into registration requests or monitoring data.

[0058] The unique identification information consists of the factory-programmed serial number, MAC address, or QR code identifier of the monitoring node, and the unique identifier is automatically read and a registration request is sent when the monitoring node is powered on or connected to the network for the first time.

[0059] Registration information includes the unique identifier of the monitoring node, node type, workstation location, current firmware version, supported sensor types and their acquisition frequency parameters; After receiving the registration information, the central monitoring platform automatically assigns corresponding task configurations to the monitoring nodes according to the preset task template library. The task configurations include data collection cycle, upload path, early warning threshold and processing logic.

[0060] In a preferred embodiment of the present invention, the modular monitoring nodes are deployed at multiple key workstations on the wooden furniture production line to collect multi-source information related to equipment operating status, workpiece processing status, and the surrounding environment in real time, and upload it to the central monitoring platform. To achieve rapid deployment, automatic registration, and intelligent configuration management, each monitoring node is modular, intelligent, and communication-adaptive.

[0061] The modular monitoring node consists of: a sensing unit, which is used to sense and collect data on the physical state of the monitored object, and includes, but is not limited to, the following modules: Image acquisition module: Configured with an industrial camera or embedded vision module, used to acquire workpiece images and processing status screens; Environmental sensor module: Collects environmental parameters such as temperature, humidity, and particle concentration at the current workstation; Vibration and current monitoring module: Samples the vibration frequency and current fluctuations of key equipment (such as cutting machines, drilling machines, etc.) during operation to identify signals such as tool wear and abnormal operation; Workpiece identification module (optional): Supports reading unique workpiece identifiers in the form of QR codes, RFID, etc., for binding and tracking.

[0062] The communication unit is used for data transmission and interaction with the central monitoring platform. Communication methods include: wired methods such as Ethernet; wireless methods such as Wi-Fi, LoRa, 5G / 4G, Zigbee, etc. It supports automatic access mechanisms, such as MQTT protocol for node online notification and data publishing. The communication unit has a fault reconnection mechanism and supports strategies such as network disconnection retry, data caching and retransmission to ensure the integrity of monitoring data.

[0063] The processing unit is used to perform preliminary processing on the data collected by the sensing unit and execute the node's local control logic. Its functions include: format conversion of raw data, outlier removal, and feature extraction; generating registration requests and data upload packets according to the built-in configuration; executing preset early warning judgments locally (such as issuing an early warning when the temperature exceeds a threshold); and managing communication status and registration status.

[0064] The processing unit can be implemented by a microcontroller (MCU) or an edge computing board (such as Raspberry Pi or Jetson Nano), providing lightweight computing and control capabilities.

[0065] Unique Identification Information and Registration Process: To support automatic node identification and configuration distribution to the central monitoring platform, each monitoring node is pre-configured with unique identification information during the manufacturing or debugging phase. Sources include: Factory-programmed serial number: a globally unique serial number written into the control chip; MAC address: the hardware address integrated into the communication module; and QR code identification: a code affixed to the node's casing, which can be scanned and identified by the central monitoring platform. When a node powers on for the first time or connects to the network, the system automatically reads this unique identification information and constructs a registration request.

[0066] The central monitoring platform's response process is as follows: Upon receiving a registration request, the central monitoring platform automatically identifies the node type and workstation; it retrieves the corresponding configuration file from the preset task template library; it automatically assigns task configurations, including but not limited to: data acquisition cycle (e.g., once every 5 seconds); data upload path (e.g., upload to a specified process section directory); warning threshold (e.g., vibration value > 0.3g); local judgment logic (e.g., local reporting when temperature and humidity exceed set values); the configuration file is distributed to the node, and the node enters the working state; the central monitoring platform updates the "production line monitoring flowchart" in real time, adding the node to the corresponding workstation link.

[0067] Among them, the warning threshold refers to a set of numerical limits used to determine whether the monitoring data has reached the alarm or intervention conditions. Its form is usually as follows: single value comparison: such as temperature > 55℃, interval judgment: such as humidity ∉ [30%, 60%], rate of change judgment: such as image defect score growth rate > 20% / min, combined condition judgment: such as current > 1.5A and vibration > 0.25g.

[0068] The warning thresholds are set based on three sources for each monitoring node: Fixed configuration template (default value). For the same type of equipment or workstation, the central monitoring platform maintains a general threshold table; for example, milling station → vibration threshold: 0.3g, hot pressing station → temperature threshold: >130°, C gluing station → humidity threshold: <40%.

[0069] Historical Statistical Learning (Personalized Configuration): Statistical analysis of equipment operation history data: take the mean ± 2σ under normal operating conditions as the threshold; or set the 90th percentile as the alarm baseline; the threshold for each node can be personalized based on historical operating characteristics.

[0070] User manual settings (Operations and Maintenance Intervention): The system interface supports operations and maintenance engineers to manually set or adjust thresholds; it supports configuration on an item-by-item or batch settings.

[0071] Dynamic threshold adjustment: The central monitoring platform supports a threshold self-learning mechanism: daily / weekly calculation of the offset trend and dynamic optimization of the threshold; if the long-term fluctuations shift upward, the threshold is automatically increased to avoid false alarms; if a certain type of risk is frequently identified in the graph, the threshold can also be tightened.

[0072] Tiered early warning: Each monitoring indicator can be set with multiple threshold levels to trigger different levels of response: Example (taking vibration as an example): Normal: ≤0.25g; Slight abnormality: 0.25g–0.3g (low-level warning); Significant abnormality: 0.3g–0.4g (medium-level warning); Severe abnormality: >0.4g (high-level warning, stop immediately).

[0073] In this invention, the central monitoring platform issues a task configuration file to each successfully registered modular monitoring node, which includes multiple parameter settings, especially adaptively adjustable warning thresholds for various collected indicators. For example, the threshold for a vibration monitoring node can be initially set to 0.3g. If the detected value exceeds this threshold, a local warning is triggered or an abnormal event is uploaded. The central monitoring platform can also automatically adjust this threshold based on historical operating data and supports multi-level warning mechanisms, further improving the accuracy and intelligence of fault response.

[0074] In another embodiment of the present invention, the central monitoring platform automatically generates a monitoring flowchart representing the production line process flow based on the registration order, workstation location and node function of each monitoring node using a graph structure algorithm. The flowchart is a directed graph, where nodes represent process units and edges represent process sequence or data flow direction. The flowchart is stored and maintained through a graph database or topology management module, and automatically adds, replaces or deletes flowchart nodes when monitoring nodes are added, disconnected or replaced due to faults.

[0075] Specifically, the central monitoring platform possesses the capabilities of automatic identification, dynamic mapping, and real-time maintenance. It can automatically construct a monitoring flowchart representing the production line's technological process based on the registration information of each monitoring node, and continuously manage and optimize it based on a graph structure model. By constructing the monitoring flowchart, the following objectives are achieved: organizing monitoring nodes distributed across different workstations into a unified process structure according to the process sequence; forming a clear and visible data flow and logical transition path for the production line; providing a structural foundation for subsequent workpiece tracking, fault path analysis, and knowledge graph reasoning; and possessing dynamic topology awareness capabilities to support production line structure adjustments or node changes.

[0076] The flowchart generation mechanism includes: registration information parsing. After the modular monitoring nodes are powered on, each node will automatically send registration information to the central monitoring platform. The registration information includes: unique node identifier (such as MAC address, QR code, or burning number); node type (image node, temperature and humidity node, vibration node, etc.); workstation number or process sequence number (such as Station01, Station02); production line number (for multi-production line deployment); sensor type and sampling frequency configuration; and function description or role information (such as "detection", "processing monitoring", etc.).

[0077] Graph structure modeling: Based on the above information, a graph structure algorithm is used to generate a monitoring flowchart. The specific logic is as follows: each monitoring node is mapped to a node (vertex) in the graph, and its workstation number, function type, and other attributes are recorded; Based on the registration order or workstation order, a directed edge is automatically constructed to represent the flow relationship between the preceding and following process units; If a node collects multiple signal types (such as temperature and images), it can be considered a composite node and have multiple types of labels attached to it. Graph structures can be stored in the form of adjacency lists, adjacency matrices, etc., and support graph traversal, path search, and structure update operations.

[0078] A directed graph is defined as a graph with the following structural characteristics: nodes (V): representing each process unit and monitoring node; edges (E): representing the processing sequence, workpiece flow direction, or data flow path. Edge attributes: can include transmission delay, sensor type, upstream and downstream dependencies, etc.; weighted edge structures are supported, and the probability of fault propagation can be evaluated in conjunction with the inference engine.

[0079] Flowchart storage and dynamic maintenance: Graph database management efficiently supports the storage and management of graph structures. The central monitoring platform integrates graph database modules (such as Neo4j, JanusGraph, ArangoDB, etc.) or uses built-in topology management components to achieve the following functions: structured storage of nodes / edges; attribute updates, node retrieval, and path queries; support for node version control and graph status snapshots; and integration with a visual interface to facilitate operations and maintenance personnel in graphically viewing the production process structure.

[0080] The topology change awareness and graph update mechanism continuously monitors the online status and functional configuration of each monitoring node, supporting the following topology update operations: Topology event: Central monitoring platform response; Adding new nodes: Automatically adds new graph nodes, analyzes the relationship between upstream and downstream workstations, and establishes connecting edges; Node disconnection: Set the node's state to "offline" and disconnect the associated edge connection; Node failure replacement: Replace the original node with a new node while inheriting the original edge structure; The process section adjusts and reconstructs the subgraph, automatically updating the edge order and attributes; When the central monitoring platform performs the above-mentioned change operations, it will simultaneously update the monitoring flowchart and notify the upper-level modules (such as the workpiece tracking module, knowledge graph engine, etc.) to reconstruct the relevant dependencies to ensure the structural consistency and logical correctness of the monitoring system.

[0081] In summary, this implementation method, through the parsing of registration information of each monitoring node and the use of an automatic graph modeling algorithm, achieves structured modeling and visual representation of the entire production line monitoring process by the central monitoring platform. This monitoring flowchart not only dynamically senses changes in node status and adjusts the topology in real time, but also supports key functions such as workpiece tracking, knowledge graph reasoning, and risk linkage control in subsequent implementation steps of this invention. It possesses technical advantages such as flexible structure, rapid response, and clear data organization.

[0082] In another embodiment of the present invention, the identification information of the workpiece is a QR code label, RFID electronic tag or shape feature code extracted by image recognition attached to the surface of the workpiece, used to identify each workpiece. The identification information of each workpiece remains consistent during the process flow. The central monitoring platform integrates the workpiece status data, environmental parameters, event records, etc. uploaded by each monitoring node in chronological order to form the monitoring data generation timestamp and node information of each workpiece, and stores them in the workpiece history database for subsequent quality analysis.

[0083] In another embodiment of the present invention, the central monitoring platform performs rule matching or graph traversal on the paths, node types and attribute combinations in the knowledge graph to achieve reasoning analysis of potential fault causes and risk points; when constructing the knowledge graph, the central monitoring platform automatically extracts common fault patterns based on historical monitoring data and performs knowledge graph reasoning through graph structure for subsequent real-time data comparison and fault warning triggering.

[0084] Knowledge graph reasoning includes rule-based reasoning and relational path reasoning. Rule-based reasoning is based on the logical relationships defined in the graph, while path reasoning is based on the logical strength calculation of the relational chains between nodes. It is used to uncover potential fault causal chains. If knowledge graph reasoning or feature matching determines that there is a potential risk, the central monitoring platform generates an early warning data packet containing the risk level, predicted fault type, affected process, and associated workpiece number, and sends it to the human-machine interface and control system simultaneously.

[0085] Specifically, the central monitoring platform constructs an industrial knowledge graph model based on a graph structure. Through dynamic reasoning and matching of paths, node types, and attributes in the graph, it achieves intelligent identification and risk prediction and early warning of potential fault causes. Knowledge graph reasoning is the core intelligent analysis layer between data perception and coordinated control, providing data logic support and predictive decision-making capabilities for the entire intelligent monitoring system.

[0086] Knowledge graph modeling: The graph is constructed in the early stages of production through a central monitoring platform using the following methods: Entity node extraction: Based on historical production data and process knowledge, the following entity types are extracted from sensor data, equipment operation logs, and maintenance records: Process steps include: cutting, drilling, and applying skin. Equipment components: Cutting machine A, pressing machine B, robotic arm; Parameter nodes: current value, temperature value, vibration intensity; Fault type nodes: abnormal vibration, tool wear, overheating; Workpiece feature nodes: wood board type, size, texture matching degree; Environmental condition parameters: humidity, dust concentration, and light conditions.

[0087] Relationship edge modeling: By mining the historical co-occurrence, temporal sequence, and causal relationships between nodes, multiple types of edge relationships are constructed. Occurred at: Fault → Workstation / Equipment Cause of: Abnormal vibration → Tool wear Dependence: Process step → Upstream step Located at: Workpiece → Workstation Feature binding (HasFeature); Workpiece → Material / Dimension Pattern solidification: The central monitoring platform extracts high-frequency risk patterns from massive historical data, such as: "high workpiece density + long tool usage time + increased vibration frequency → tool wear"; "Abnormal humidity + change in adhesive type → unstable bonding quality." These empirical rules are solidified into traversable and reasonable subgraph patterns through graph structures.

[0088] Knowledge graph reasoning mainly includes the following two categories: rule-based reasoning. The central monitoring platform pre-sets several logical rules (IF-THEN) or subgraph pattern matching rules. When real-time monitoring data triggers these conditions, it is determined to be an early warning event. Example: IF: Node A (vibration) value > threshold and Node B (tool usage time) > X hours THEN: Activate node C (tool anomaly) → Risk level = Medium; Rule-based reasoning uses a knowledge graph engine for rapid matching and judgment, suitable for known patterns.

[0089] Path-based reasoning utilizes graph algorithms, such as shortest path, Path Ranking Algorithm (PRA), and graph neural network embedding vector similarity, through a central monitoring platform to uncover potential fault chains hidden between entity relationships. Example: Vibration node → Tool node → Image anomaly node → Product scrap node If the overall correlation strength of a certain path exceeds the threshold, it is considered a potential risk chain.

[0090] Path reasoning is suitable for identifying unknown risks or problems with multiple overlapping factors, and it has a certain learning ability.

[0091] Risk identification and early warning data generation: When the central monitoring platform determines that a potential risk exists through any of the above reasoning methods, the system will immediately generate a structured early warning data packet, the content of which includes: Risk level: High / Medium / Low, based on confidence level, scope of impact, and other indicators. Predicting fault types such as "tool wear", "overheating", and "assembly deviation" The affected process or equipment number indicates the workstation or equipment associated with the risk. If the risk can be associated with a specific workpiece, then bind its ID. Raw trigger data includes time, sensor readings, image summaries, etc. Graph path nodes and relational chains matched or activated by the inference path Timestamp risk identification time Data distribution and linkage control: Data is synchronously sent to the human-machine interface (HMI) for on-site display of risk details and response suggestions; Linkage control module: used to determine whether to perform a shutdown, alarm, or scheduled maintenance; Quality tracking system: used to link the risk to a specific workpiece; Operations and maintenance system: Generates maintenance task orders and drives subsequent processing.

[0092] This implementation method constructs an industrial knowledge graph model with rich entity and relational structures, and combines rule-based reasoning and path-based reasoning mechanisms to achieve intelligent judgment and response control of potential faults and risks in the production process. Through the structured expression of information such as risk level, fault type, and influencing nodes, this invention significantly improves the early warning accuracy, intelligent response capability, and quality traceability of wooden furniture production lines, and constructs a novel intelligent monitoring system driven by semantic association and causal reasoning.

[0093] In another embodiment of the present invention, the linkage control is implemented through a preset control rule base. The rule base includes multiple "condition-action" pairs. When the risk level, fault type or workstation number contained in the warning information meets a certain rule condition, the corresponding control action is automatically executed.

[0094] Linked control actions are relayed through the edge control gateway to achieve localized response. After receiving control commands from the central monitoring platform, the edge control gateway can cache, confirm, and perform secondary verification locally before executing the specific action.

[0095] After identifying and assessing potential risks, the central monitoring platform automatically matches the corresponding control strategy based on the preset linkage control rule library, and issues control commands through the edge control gateway to achieve localized linkage response on the production site.

[0096] Control Rule Base Definition: The central monitoring platform has a built-in Control Rule Base used to define the logical mapping relationship between risk warnings and control actions. This rule base is organized in the form of "condition-action" pairs, and each rule includes the following fields: Risk level criteria: High / Medium / Low, supporting multi-level matching. Fault type conditions such as: abnormal vibration, tool wear, and excessive temperature. The workstation number condition specifies the trigger range, such as Station03~05. Control actions such as: shutdown, buzzer alarm, flashing light bar, and speed limit operation. Set the delay time for the action (e.g., response within 5 seconds). Execution order when multiple priority rules conflict Does the local verification requirement necessitate that the edge gateway re-verify the sensor status? Matching and Triggering Logic: When the central monitoring platform identifies a potential fault based on knowledge graph reasoning or data feature matching and generates an early warning information data packet, the system will automatically retrieve control rules from the control rule base that match the fields of the early warning information (such as risk level, workstation number, and fault type). If a rule is matched, the platform immediately generates the corresponding control command and enters the linkage response process.

[0097] The issuance and execution process of linkage control commands: Control commands are encapsulated in a structured format, typically including: Control target ID: Target equipment number, workstation number; Command types include shutdown, speed limit, alarm activation, and power cut-off; Triggering reason: corresponding warning number and risk information summary Effective duration: The duration of the controlled action or the condition for its release. Security identifiers: Command signatures and encrypted fields used for authentication and verification. Receipt Requirement: Is it necessary to execute a receipt or status report? Control link path: The control link adopts a three-level structure of "central monitoring platform → edge control gateway → execution device" to ensure that the command has a clear path and relay management capabilities from recognition to execution.

[0098] Functional modules and responses of the edge control gateway: To achieve efficient and localized execution of control commands, this invention deploys an edge control gateway in key areas of the production line to receive platform commands and perform action issuance, status feedback, and local secondary verification.

[0099] The main functional modules of the edge gateway include: a communication receiving module that supports protocols such as TCP / IP, MQTT, and Modbus to receive control commands issued by the central platform; Local caching module: temporarily stores received instructions locally to support automatic retry or delay control in case of connection failure; Command verification module: Supports local sensor status comparison; if the command is of high risk level, it can be confirmed twice, such as: read the vibration value again to confirm that it exceeds the threshold before execution; Action execution module: Operates downstream devices according to the control type, such as controlling relay closure, PLC interface switching, and triggering IO signals; Feedback and reporting module: Generates execution result receipts after execution and sends them back to the central platform to achieve closed-loop tracking.

[0100] Linkage Response: Based on the control rules configuration, the system can achieve the following typical linkage control responses: If the vibration intensity is > 0.4g and the tool life is > 90%, the current workstation will stop and an alarm will be triggered locally. If the workpiece image defect score is > 0.9, the downstream station will limit the speed and the workpiece will be marked as a two-station random inspection. Multi-node current anomaly linkage; whole-line early warning + HMI pop-up prompts; This implementation method constructs a linkage control rule base containing multiple "condition-action" pairs, enabling the central monitoring platform to quickly match and execute control strategies after identifying potential risks. Combined with edge control gateways for localized caching, verification, and triggering of control commands, it significantly improves system response efficiency, security control capabilities, and production flexibility, thus constructing a complete "identification-response-closed-loop" intelligent monitoring and control system.

[0101] The foregoing has only described certain exemplary embodiments of the present invention by way of illustration. Undoubtedly, those skilled in the art can modify the described embodiments in various ways without departing from the spirit and scope of the present invention. Therefore, the foregoing drawings and descriptions are illustrative in nature and should not be construed as limiting the scope of protection of the claims of the present invention.

Claims

1. A monitoring method suitable for wooden furniture production lines, characterized in that, The method includes the following steps: S1. Modular monitoring node deployment: Multiple modular monitoring nodes with sensing, processing and communication functions are deployed at each key workstation of the processing equipment, and unique identification information is generated for each monitoring node. The identification information is generated by preset rules and a registration request containing identity information, node type and equipment status is sent through the network. S2. Automatic identification and process construction: After the monitoring node is powered on, it automatically sends registration information to the central monitoring platform through the network. The registration information includes the node's unique identifier, type, function configuration, etc. The central monitoring platform automatically configures the tasks of each node and constructs the production line monitoring flowchart according to the registration information, and updates the production line monitoring flowchart in real time when the monitoring node changes. S3. Workpiece process tracking: By attaching identification information to the workpiece and recording the status, parameters and event information related to the workpiece through each monitoring node, the workpiece can be tracked and data bound throughout different processes. S4. Knowledge graph modeling and reasoning: The central monitoring platform constructs a knowledge graph model based on process flow, equipment status, historical fault information, workpiece characteristics, and environmental data to establish the relationship between monitoring nodes and express the association pattern of potential faults. S5. Risk prediction: The central monitoring platform cleans and standardizes the real-time monitoring data uploaded by each monitoring node, and then performs feature matching or knowledge graph reasoning with the knowledge graph to determine whether there are potential risks and generate early warning signals. S6. Linkage control: If a fault is detected, an early warning message is generated and the downstream control system is linked to execute control actions, including shutdown control and alarm prompts. At the same time, a maintenance task order is generated and the early warning message is bound to the relevant workpiece for archiving.

2. The monitoring method for a wooden furniture production line according to claim 1, characterized in that, The modular monitoring node includes: a sensing unit, used to collect information related to the workpiece or equipment, such as images, temperature and humidity, vibration, and current; The communication unit is used to interact with the central monitoring platform via wired or wireless means; The processing unit is used to perform preliminary analysis on the collected data and encapsulate it into registration requests or monitoring data.

3. The monitoring method for a wooden furniture production line according to claim 1, characterized in that, The unique identification information consists of the factory-programmed serial number, MAC address, or QR code identifier of the monitoring node, and the unique identifier is automatically read and a registration request is sent when the monitoring node is powered on or connected to the network for the first time.

4. The monitoring method for a wooden furniture production line according to claim 1, characterized in that, The registration information includes the unique identifier of the monitoring node, node type, workstation location, current firmware version, supported sensor types and their acquisition frequency parameters; After receiving the registration information, the central monitoring platform automatically assigns corresponding task configurations to the monitoring nodes according to the preset task template library. The task configurations include data collection cycle, upload path, early warning threshold and processing logic.

5. A monitoring method for a wooden furniture production line according to claim 1, characterized in that, The central monitoring platform automatically generates a monitoring flowchart representing the production line process flow based on the registration order, workstation location, and node function of each monitoring node using a graph structure algorithm. The flowchart is a directed graph, where nodes represent process units and edges represent process sequence or data flow direction. The flowchart is stored and maintained through a graph database or topology management module, and automatically adds, replaces or deletes flowchart nodes when monitoring nodes are added, disconnected or replaced due to faults.

6. The monitoring method for a wooden furniture production line according to claim 1, characterized in that, The identification information of the workpiece is a QR code label, RFID electronic tag or shape feature code extracted by image recognition attached to the surface of the workpiece, used to identify each workpiece. The identification information of each workpiece remains consistent during the process flow. The central monitoring platform integrates the workpiece status data, environmental parameters, event records, etc. uploaded by each monitoring node in chronological order to form the monitoring data generation timestamp and node information of each workpiece, and stores it in the workpiece history database for subsequent quality analysis.

7. A monitoring method for a wooden furniture production line according to claim 1, characterized in that, The central monitoring platform performs rule matching or graph traversal on the paths, node types, and attribute combinations in the knowledge graph to perform reasoning analysis on the causes and risk points of potential faults. When constructing the knowledge graph, the central monitoring platform automatically extracts common fault patterns based on historical monitoring data and performs knowledge graph reasoning through the graph structure for subsequent real-time data comparison and fault warning triggering.

8. A monitoring method for a wooden furniture production line according to claim 1, characterized in that, The knowledge graph reasoning includes rule reasoning and relational path reasoning. The rule reasoning is based on the logical relationships defined in the graph, and the path reasoning is based on the logical strength calculation of the relational chain between nodes. It is used to mine potential fault causal chains. If the knowledge graph reasoning or feature matching judgment has potential risks, the central monitoring platform generates an early warning data packet containing the risk level, predicted fault type, affected process, and associated workpiece number, and sends it to the human-machine interface and control system simultaneously.

9. A monitoring method for a wooden furniture production line according to claim 1, characterized in that, The linkage control is achieved through a preset control rule base, which includes multiple "condition-action" pairs. When the risk level, fault type, or workstation number contained in the warning information meets a certain rule condition, the corresponding control action is automatically executed.

10. A monitoring method for a wooden furniture production line according to claim 9, characterized in that, The linkage control actions are relayed through the edge control gateway to achieve localized response. After receiving control instructions from the central monitoring platform, the edge control gateway can cache, confirm, and perform secondary verification locally before executing the specific action.

Images