A processing information analysis and exception handling method and system for customized furniture production

Abstract

The application discloses a processing information analysis and abnormality processing method and system for customized furniture production, relates to the technical field of furniture manufacturing, and realizes complete traceability, improves abnormality identification precision and response efficiency, reduces production failure and quality risk, and is suitable for intelligent management and control of large-scale customized furniture production by collecting order information, generating a unique traceability identifier, collecting and standardizing processing of full-link data such as order splitting, raw materials, processing, assembly, detection, logistics and installation, and constructing a complete traceability link; based on an improved LSTM network and the introduction of an attention mechanism, an abnormality early warning model is established to realize online monitoring and abnormality discrimination of production data, hierarchical early warning and rapid disposal. The system comprises an order information collection module, a full-process data collection module, a data processing module, an abnormality early warning module, a traceability disposal module, a database and a terminal interaction module.

Description

Technical Field

[0001] This invention relates to the field of furniture manufacturing technology, and more specifically to a method and system for processing information analysis and anomaly handling in the production of customized furniture. Background Technology

[0002] With the upgrading of residents' consumption and the growth of personalized needs, the customized furniture industry has developed rapidly. It is characterized by strong personalization, many varieties, small batches, and complex production processes. The entire process of ordering, from order placement, order splitting, production, assembly to delivery and installation, involves multiple links, multiple equipment and multiple operators. Any abnormality in any link will affect the order delivery cycle and product quality.

[0003] Currently, custom furniture companies mostly use traditional methods for order management and production control, which has the following problems: Order data is scattered throughout the entire process, and data from different stages is not effectively linked. When quality problems or delivery delays occur, it is impossible to quickly trace the abnormal process and its root cause, resulting in low traceability efficiency, poor accuracy, and difficulty in assigning responsibility. Abnormalities in the production process are mainly detected through manual monitoring, leading to delayed responses, which can easily cause abnormalities to escalate, increase production costs, and affect production schedules. There is a lack of scientific abnormality early warning mechanisms, making it impossible to predict abnormal risks in advance based on production data, and hindering proactive control of the production process. Furthermore, data at each stage is not standardized, resulting in low data utilization and an inability to provide data support for production process optimization and capacity improvement.

[0004] To address the aforementioned issues, some existing technologies have emerged related to early warning of abnormalities in furniture production. However, most of these technologies have limitations: some solutions only trace or warn for a single production stage, failing to cover the entire order process; some solutions employ simple data collection and comparison methods, resulting in low warning accuracy and a high false alarm rate, making them unsuitable for the diverse and variable batch production characteristics of customized furniture; and some solutions fail to achieve coordinated linkage between traceability and abnormal warning, making it difficult to update traceability data promptly after abnormality handling and thus hindering the formation of closed-loop management.

[0005] Therefore, it is necessary to propose a method and system for processing information analysis and anomaly handling in customized furniture production to solve the above problems. Summary of the Invention

[0006] (a) Technical problems to be solved

[0007] The purpose of this invention is to address the problems of fragmented data throughout the order process, delayed response, lack of a scientific anomaly warning mechanism, low data utilization, and inability to provide data support for production process optimization and capacity improvement. This invention provides a method and system for processing information analysis and anomaly handling in customized furniture production.

[0008] (II) Technical Solution

[0009] To achieve the above objectives, the present invention specifically adopts the following technical solution:

[0010] A method and system for processing information analysis and anomaly handling in customized furniture production, comprising the following steps:

[0011] Step S1: Initialize order information. Collect basic information, customization parameters, and customer requirements for customized furniture orders. Standardize and encode the order information to generate a unique order traceability identifier and establish an order information database.

[0012] Step S2: Full-process data collection. Based on the order traceability identifier, process data, parameter data, and status data of customized furniture are collected in real time from each stage of order splitting, raw material procurement, board processing, parts assembly, quality inspection, packaging and warehousing, logistics and distribution, and installation and acceptance, forming a full-process traceability dataset.

[0013] Step S3: Data processing and association. The data in the full-process traceability dataset is cleaned, deduplicated, and standardized. An association mapping relationship between order traceability identifiers and data in each link is established to construct the order full-process traceability link.

[0014] Step S4: Production anomaly early warning modeling. Based on historical order traceability data and production anomaly case data, an anomaly early warning model is constructed. Anomaly judgment threshold and early warning level are set for each production link. The anomaly early warning model uses an improved LSTM algorithm and combines production process standards to identify abnormal data.

[0015] Step S5: Real-time monitoring and anomaly warning. The collected real-time production data is dynamically monitored through the anomaly warning model. When the data exceeds the anomaly judgment threshold of the corresponding link, an alarm message is automatically generated according to the alarm level and pushed to the corresponding management terminal and production workstation. At the same time, the anomaly information is recorded to the order traceability link.

[0016] Step S6: Traceability Inquiry and Anomaly Handling. Receive traceability inquiry requests, retrieve full-process traceability data and anomaly handling records based on order traceability identifiers, and generate a traceability report; receive anomaly handling instructions, record the handling process and results, form an anomaly handling closed loop, and update the corresponding data in the order traceability chain.

[0017] Preferably, in step S1, the basic order information includes order number, customer information, order time, and delivery deadline; the customized parameter information includes furniture type, size specifications, material parameters, process requirements, and color parameters; the standardized code adopts the format of "order type code + timestamp + random verification code"; the order traceability identifier corresponds one-to-one with the order number and is associated with the raw material batch number, equipment number, and operator ID.

[0018] Preferably, in step S2, the process data includes the start time, end time, operation procedure, operator, and equipment operating status of each stage; the parameter data includes processing speed, cutting accuracy, adhesive coating thickness, assembly gap, and inspection data; the data acquisition adopts a combination of IoT sensors, PLC controllers, vision inspection equipment, and manual input terminals, and the acquisition frequency is set to 1-10Hz according to the production stage requirements, wherein the acquisition frequency of the processing stage and the inspection stage is not less than 5Hz.

[0019] Preferably, in step S3, data processing includes removing invalid data, correcting abnormal deviation data, and unifying data format and units; the association mapping relationship is stored in a tree structure, with the order traceability identifier as the root node, each production link as a child node, and the data of each link as the child node attribute, to achieve a three-level association traceability of "order-link-data".

[0020] Preferably, in step S4, the construction process of the anomaly early warning model includes:

[0021] S41: Collect historical order process data and production anomaly cases, including types such as unqualified raw materials, excessive processing precision, equipment failure, delay in delivery, and unqualified quality inspection. Classify and label the anomaly cases to generate a training dataset.

[0022] S42: Normalize the training dataset, divide it into training, validation and test sets, build a model using an improved LSTM algorithm, and introduce an attention mechanism to optimize the model's recognition accuracy of key production parameters.

[0023] S43: Adjust the model parameters through the validation set and set the anomaly judgment threshold corresponding to the warning level. The warning levels are divided into Level 1 warning, Level 2 warning and Level 3 warning. Different levels correspond to different push ranges and handling time limits.

[0024] S44: Use a test set to validate the model, ensuring that the model's anomaly identification accuracy is not less than 95% and the warning response time does not exceed 30 seconds.

[0025] Preferably, in step S5, the warning information includes order traceability identifier, abnormal process, abnormal parameter, abnormal value, warning level, handling suggestion and push time; the push method includes terminal pop-up window, SMS notification and voice reminder.

[0026] Preferably, in step S6, the traceability query request can be initiated through any of the following methods: order number, customer information, or order traceability identifier. The traceability report includes basic order information, full-process production data, anomaly records, anomaly handling process and results, and delivery status. The anomaly handling closed loop includes five stages: anomaly reception, cause analysis, handling implementation, effect verification, and record archiving. After the handling is completed, the model automatically updates the anomaly judgment threshold of the corresponding stage to optimize the early warning accuracy.

[0027] A system for analyzing processing information and handling anomalies in custom furniture production, characterized by comprising:

[0028] The order information collection module is used to collect basic information, customization parameter information and customer demand information of customized furniture orders, standardize the order information, generate a unique order traceability identifier, and establish and store the order information database.

[0029] The end-to-end data acquisition module includes an IoT sensor unit, a PLC control unit, a vision inspection unit, and a manual data entry unit. It is used to collect process data, parameter data, and status data of each production, logistics, and installation stage of customized furniture in real time based on order traceability tags, and transmit them to the data processing module.

[0030] The data processing and association module is used to clean, deduplicate, and standardize the collected data throughout the entire process, establish the association mapping relationship between order traceability identifiers and data at each stage, construct the order full-process traceability link, and store the processed data in the traceability database.

[0031] The anomaly warning module includes a model training unit, a real-time monitoring unit, an early warning generation unit, and an information push unit. It is used to build and train an anomaly warning model, dynamically monitor real-time production data, identify abnormal data, generate early warning information of corresponding levels, and push it to the corresponding terminal.

[0032] The traceability and handling module includes a traceability query unit, an anomaly receiving unit, a handling record unit, and a report generation unit. It is used to receive traceability query requests, retrieve traceability data to generate traceability reports, receive anomaly handling instructions, record the handling process and results, form an anomaly handling closed loop, and update the traceability link data.

[0033] The database module includes an order information database, a traceability database, an anomaly case database, and a model parameter database, which are used to store order information, full-process traceability data, historical anomaly cases, model training parameters, and anomaly handling records.

[0034] The terminal interaction module includes terminals for management personnel, team leaders, and workstation operators. These terminals are used to receive early warning information, initiate traceability queries, submit abnormal handling instructions, and view traceability reports and handling records.

[0035] Preferably, the IoT sensor unit in the full-process data acquisition module includes a temperature sensor, a humidity sensor, a pressure sensor, and a speed sensor, which are used to collect corresponding parameters of the processing environment, equipment operation, and raw material processing, respectively; the visual inspection unit includes an industrial camera and an image recognition module, which are used to collect visual data such as plate cutting accuracy, component assembly gap, and surface quality, and extract feature parameters through image recognition algorithms; the PLC control unit is linked with the furniture production equipment and logistics conveying equipment to collect equipment operating parameters and production progress data.

[0036] Preferably, the system further includes a permission management module and a data encryption module.

[0037] (III) Beneficial Effects

[0038] The beneficial effects of this invention are as follows:

[0039] 1. This invention collects and links data at every stage, from order placement and raw material procurement to installation and acceptance. Through a unique order traceability identifier, the entire process data can be quickly retrieved, and abnormal links and root causes can be accurately located, improving traceability efficiency and accuracy, facilitating responsibility determination, and meeting the compliance requirements of the customized furniture industry for product quality traceability.

[0040] 2. This invention constructs an anomaly warning model based on an improved LSTM algorithm, introduces an attention mechanism to improve warning accuracy, and combines a hierarchical warning mechanism to achieve real-time monitoring, accurate warning, and hierarchical push of production anomalies. It solves the problems of delayed response and high false alarm rate of traditional manual monitoring, and can predict anomaly risks in advance, handle anomalies in a timely manner, avoid the expansion of anomalies, reduce production costs, and ensure production progress and product quality.

[0041] 3. This invention realizes the coordinated linkage of traceability and anomaly early warning. Anomaly information is recorded to the traceability link in real time, and the entire process and result of anomaly handling are traceable, forming a closed-loop management of "monitoring-early warning-handling-verification-archiving". At the same time, after the handling is completed, the early warning model parameters are automatically optimized to improve the system's adaptability and long-term use effect. Attached Figure Description

[0042] Figure 1 This is a flowchart of the present invention;

[0043] Figure 2 This is a flowchart illustrating the training process of the anomaly warning model of the present invention. Detailed Implementation

[0044] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0045] Please refer to Figure 1-2. A method and system for processing information analysis and anomaly handling in customized furniture production includes the following steps:

[0046] Step S1: Initialize order information. Collect basic information, customization parameters, and customer requirements for customized furniture orders. Standardize and encode the order information to generate a unique order traceability identifier and establish an order information database.

[0047] Step S2: Full-process data collection. Based on order traceability identifiers, collect process data, parameter data, and status data in real time for each stage of customized furniture production, from order splitting, raw material procurement, board processing, parts assembly, quality inspection, packaging and warehousing, logistics and distribution, and installation and acceptance, to form a full-process traceability dataset.

[0048] Step S3: Data processing and association. Clean, deduplicate, and standardize the data in the full-process traceability dataset, establish the association mapping relationship between order traceability identifiers and data in each link, and construct the order full-process traceability link;

[0049] Step S4: Production anomaly early warning modeling. Based on historical order traceability data and production anomaly case data, an anomaly early warning model is constructed, and anomaly judgment thresholds and early warning levels are set for each production link.

[0050] Step S5: Real-time monitoring and anomaly warning. The collected real-time production data is dynamically monitored through the anomaly warning model. When the data exceeds the anomaly judgment threshold of the corresponding link, the warning information is automatically generated according to the warning level and pushed to the corresponding management terminal and production workstation. At the same time, the anomaly information is recorded to the order traceability link.

[0051] Step S6: Traceability Inquiry and Anomaly Handling. Receive traceability inquiry requests, retrieve full-process traceability data and anomaly handling records based on order traceability identifiers, and generate a traceability report; receive anomaly handling instructions, record the handling process and results, form an anomaly handling closed loop, and update the corresponding data in the order traceability chain.

[0052] In step S1, the basic order information includes the order number, customer information, order time, and delivery deadline; the customized parameter information includes furniture type, size specifications, material parameters, process requirements, and color parameters; the standardized coding adopts the format of "order type code + timestamp + random check code", and the order traceability identifier corresponds one-to-one with the order number and is associated with the raw material batch number, equipment number, and operator ID to ensure that the traceability identifier of each order is unique and to achieve accurate association between the order and each production element.

[0053] In step S2, the process data includes the start time, end time, operation procedure, operators, and equipment operating status of each stage; the parameter data includes processing speed, cutting accuracy, adhesive thickness, assembly gap, and inspection data; data acquisition adopts a combination of IoT sensors, PLC controllers, vision inspection equipment, and manual input terminals. The acquisition frequency is set to 1-10Hz according to the production stage requirements, with the acquisition frequency for processing and inspection stages not lower than 5Hz, to ensure the real-time, comprehensive, and accurate data acquisition and meet the needs of traceability and early warning.

[0054] In step S3, data processing includes: removing invalid data, correcting abnormal deviation data, and unifying data format and units; the association mapping relationship is stored in a tree structure, with the order traceability identifier as the root node, each production link as a child node, and the data of each link as the attributes of the child node, realizing a three-level association traceability of "order-link-data", which facilitates quick retrieval of relevant data of any link and improves traceability efficiency.

[0055] In step S4, the construction process of the anomaly early warning model includes:

[0056] S41: Collect historical order process data and production anomaly cases. Anomaly cases include types such as unqualified raw materials, excessive processing precision, equipment failure, delay in delivery, and unqualified quality inspection. Classify and label the anomaly cases to generate a training dataset.

[0057] S42: Normalize the training dataset, divide it into training set, validation set and test set, use an improved LSTM algorithm to build the model, introduce an attention mechanism to optimize the model’s recognition accuracy of key production parameters, and solve the problems of insufficient attention to key parameters and low warning accuracy of traditional LSTM algorithm.

[0058] The improved LSTM algorithm model establishment process is as follows:

[0059] Step 1: Data Preprocessing

[0060] 1. Data filtering and labeling

[0061] Select the data needed for modeling from the order's entire process traceability data:

[0062] ① Input feature set: covering key parameters of each process, with a total of 15-20 core features selected;

[0063] ② Tag set: Mark abnormal / normal states;

[0064] ③ Data volume requirements: Collect at least 12 months of historical data to ensure that the data covers different furniture categories, production batches, and process parameters.

[0065] 2. Data Standardization and Enhancement

[0066] ① Min-Max normalization is used to map all feature values ​​to the [0,1] interval to eliminate dimensional differences;

[0067] ② To address the issue of insufficient outlier samples, the SMOTE algorithm is used to generate synthetic outlier samples, balancing the ratio of positive to negative samples;

[0068] ③ Remove invalid data and correct abnormal data that deviates from the process range.

[0069] 3. Time-series data reconstruction

[0070] ① Reconstruct discrete single-stage data into a time-series sequence according to the time order of production processes:

[0071] ② Using the order traceability identifier as the unit, splice the time sequence data of each step according to the process sequence of "material cutting → edge sealing → drilling → assembly → inspection";

[0072] ③ Divide the time sequence window: Dynamically set the window length according to the production cycle of different processes to generate a time sequence sample set.

[0073] Step 2: Improve the LSTM model structure

[0074] The model structure consists of four layers:

[0075] 1. Input layer: Dimensionally adapted to the time-series data features of customized furniture production (time-series window length dynamically adjusted according to process, feature dimensions are 15-20).

[0076] 2. Attention Layer: Introducing a self-attention mechanism to automatically assign weights to key process features such as cutting accuracy and hole position deviation, thereby strengthening the influence of core features;

[0077] 3. Improved LSTM layer: Set 128 hidden layer neurons, add dropout to prevent overfitting, receive weighted feature input, and extract temporal anomaly patterns;

[0078] 4. Output layer: The softmax activation function is used to output 7 types of prediction results (normal + 6 types of production abnormalities); the loss function adds an abnormal sample penalty term on the basis of mean squared error (MSE) to increase the penalty weight for the prediction deviation of abnormal samples.

[0079] Step 3: Model Training and Validation

[0080] 1. Dataset partitioning

[0081] The preprocessed time series sample set was divided into training set, validation set and test set in a ratio of 7:2:1 to ensure that the distribution ratio of abnormal samples in each set is consistent.

[0082] 2. Model Training

[0083] ① Training parameters: Batch size is set to 32, number of iterations is set to 50, and early stopping is used to prevent overfitting;

[0084] ② Optimizer: The Adam optimizer is selected, with adaptive adjustment of the learning rate.

[0085] 3. Model Validation and Optimization

[0086] ③ Core Indicators: Anomaly identification accuracy ≥ 95%, early warning response time ≤ 30s, false alarm rate ≤ 3%;

[0087] ④ Optimization method: If the accuracy of identifying a certain type of anomaly is low (such as anomalies in the edge sealing process), increase the attention weight of that type of feature and retrain the model; if the false alarm rate is high, adjust the time window length or the penalty coefficient of the loss function.

[0088] Step 4: Model Deployment and Iterative Optimization

[0089] 1. Lightweight model deployment

[0090] The trained model is converted to ONNX format and deployed to industrial computers in the production site to ensure real-time monitoring response speed;

[0091] 2. Online Iterative Optimization

[0092] The model incorporates real-time production data. Every 100 new orders accumulated trigger an incremental training iteration, updating attention weights and anomaly detection thresholds to adapt to adjustments in production processes.

[0093] S43: Adjust model parameters through validation set, set anomaly judgment threshold corresponding to warning level. Warning level is divided into Level 1 warning (minor anomaly, can be handled on-site), Level 2 warning (general anomaly, process needs to be adjusted), and Level 3 warning (serious anomaly, work stoppage and investigation required). Different levels correspond to different push ranges and handling time limits, realize hierarchical control of anomalies, and improve handling efficiency.

[0094] S44: The model is validated using a test set to ensure that the model’s anomaly identification accuracy is not less than 95% and the early warning response time is not more than 30 seconds, meeting the needs of real-time early warning in the production site.

[0095] In step S5, the early warning information includes order traceability identifier, abnormal process, abnormal parameters, abnormal values, early warning level, handling suggestions, and push time; the push methods include terminal pop-up window, SMS notification, and voice reminder. The first-level early warning is pushed to the corresponding workstation operator and team leader, the second-level early warning is pushed to the production supervisor, and the third-level early warning is pushed to the factory manager and quality control department to ensure that abnormal information can be quickly transmitted to the relevant responsible persons and achieve timely response to abnormalities.

[0096] In step S6, traceability query requests can be initiated using any of the following methods: order number, customer information, or order traceability identifier. The traceability report includes basic order information, full-process production data, anomaly records, anomaly handling process and results, and delivery status. The anomaly handling closed loop includes five stages: anomaly reception, cause analysis, handling implementation, effect verification, and record archiving. After the handling is completed, the model automatically updates the anomaly judgment threshold of the corresponding stage, optimizes the early warning accuracy, realizes adaptive optimization of the model, and improves the long-term use effect of the system.

[0097] This invention also provides a system for full-process traceability and production anomaly early warning of customized furniture orders, used to implement the above-mentioned method. The system includes:

[0098] The order information collection module is used to collect basic information, customization parameter information and customer demand information of customized furniture orders, standardize the order information, generate a unique order traceability identifier, and establish and store the order information database.

[0099] The end-to-end data acquisition module includes an IoT sensor unit, a PLC control unit, a vision inspection unit, and a manual data entry unit. It is used to collect process data, parameter data, and status data of each production, logistics, and installation stage of customized furniture in real time based on order traceability tags, and transmit them to the data processing module.

[0100] The data processing and association module is used to clean, deduplicate, and standardize the collected data throughout the entire process, establish the association mapping relationship between order traceability identifiers and data at each stage, construct the order full-process traceability link, and store the processed data in the traceability database.

[0101] The anomaly warning module includes a model training unit, a real-time monitoring unit, an early warning generation unit, and an information push unit. It is used to build and train an anomaly warning model, dynamically monitor real-time production data, identify abnormal data, generate early warning information of corresponding levels, and push it to the corresponding terminal.

[0102] The traceability and handling module includes a traceability query unit, an anomaly receiving unit, a handling record unit, and a report generation unit. It is used to receive traceability query requests, retrieve traceability data to generate traceability reports, receive anomaly handling instructions, record the handling process and results, form an anomaly handling closed loop, and update the traceability link data.

[0103] The database module includes an order information database, a traceability database, an anomaly case database, and a model parameter database, which are used to store order information, full-process traceability data, historical anomaly cases, model training parameters, and anomaly handling records.

[0104] The terminal interaction module includes terminals for management personnel, team leaders, and workstation operators. These terminals are used to receive early warning information, initiate traceability queries, submit abnormal handling instructions, and view traceability reports and handling records.

[0105] The IoT sensor unit in the end-to-end data acquisition module includes temperature sensors, humidity sensors, pressure sensors, and speed sensors, which are used to collect corresponding parameters of the processing environment, equipment operation, and raw material processing, respectively. The vision inspection unit includes an industrial camera and an image recognition module, which is used to collect visual data such as sheet cutting accuracy, component assembly gaps, and surface quality, and extract feature parameters through image recognition algorithms. The PLC control unit is linked with the furniture production equipment and logistics conveying equipment to collect equipment operating parameters and production progress data, ensuring the comprehensiveness and accuracy of data collection.

[0106] The system also includes a permission management module and a data encryption module. The permission management module is used to assign different operating permissions to personnel in different positions. The management personnel terminal has the right to view the entire process data, set the warning level, and review the abnormal handling. The team leader terminal has the right to view the data of their team and report abnormal handling. The workstation operator terminal has the right to view the data of their workstation, receive abnormalities, and provide feedback on handling. The data encryption module uses the AES encryption algorithm to encrypt the storage and transmission of order information, traceability data, and abnormal records to ensure data security and privacy and prevent data leakage or tampering.

[0107] Example 1:

[0108] This embodiment provides a method for full-process traceability and production anomaly early warning for customized furniture orders, including the following steps:

[0109] Step S1: Order information initialization. Collect basic information, customization parameters, and customer requirements for the customized furniture order through the order information collection terminal. The basic order information includes the order number, customer name, contact information, order time, and delivery deadline. The customization parameters include furniture type, size specifications, material parameters, process requirements, and color parameters. The customer requirements include zoning design.

[0110] The above order information is standardized and encoded using the format of "order type code + timestamp + random check code". The order type code is DZ, the timestamp is a millisecond-level encoding of the order time, and the random check code is a 3-digit random number. A unique order traceability identifier is generated, which corresponds one-to-one with the order number. It is also associated with the raw material batch number, equipment number, and operator ID. An order information database is established to store all the above order-related information. The database uses MySQL to support fast data query and update.

[0111] Step S2: Full-process data collection. Based on order traceability identification, a combination of IoT sensors, PLC controllers, visual inspection equipment and manual data entry terminals is used to collect data in real time at each stage. The collection frequency is set as follows: 1Hz for raw material procurement and warehousing, 8Hz for order splitting, sheet metal processing, parts assembly and quality inspection, and 2Hz for packaging, logistics and distribution and installation and acceptance.

[0112] The specific data collected includes:

[0113] (1) Order splitting process: order splitting start time, end time, splitting personnel ID, order splitting software version, and order splitting parts list;

[0114] (2) Raw material procurement process: raw material name, batch number, quantity procured, procurement time, supplier information, and raw material testing data;

[0115] (3) Sheet material processing stage: processing equipment number, operator ID, processing start time, end time, processing speed, cutting accuracy, sheet material size error, and residual material size;

[0116] (4) Component assembly process: assembly equipment number, operator ID, assembly start time, end time, adhesive thickness, assembly gap, and number of hardware accessories installed;

[0117] (5) Quality inspection process: Inspection equipment number, inspection personnel ID, inspection start time, inspection end time, surface flatness, edge sealing firmness, hole position accuracy, inspection results;

[0118] (6) Packaging and outbound process: Packaging personnel ID, packaging start time, end time, packaging materials, outbound time, and outbound personnel ID;

[0119] (7) Logistics and distribution process: tracking number, delivery personnel ID, shipping time, estimated delivery time, real-time location, and delivery status;

[0120] (8) Installation and acceptance process: installer ID, installation start time, end time, installation accuracy, customer acceptance result, and acceptance time.

[0121] All the data collected above is transmitted to the data processing module to form a full-process traceability dataset.

[0122] Step S3: Data processing and association. The data processing module processes the data in the full-process traceability dataset: removing invalid data, correcting abnormal deviation data, and unifying data format and units.

[0123] Establish a mapping relationship between order traceability identifiers and data at each stage, using a tree structure for storage. The order traceability identifier DZ is the root node, and eight stages, including order splitting, raw material procurement, and board processing, are first-level child nodes. The specific data of each stage is the attribute of the corresponding child node, realizing a three-level association traceability of "order-stage-data". The processed data is stored in the traceability database, which is linked with the order information database to support data association queries.

[0124] Step S4: Production anomaly early warning modeling, the specific process is as follows:

[0125] S41: Collect historical order process data and production anomaly cases from the past year. Anomaly cases are categorized and labeled as: unqualified raw materials, excessive processing precision, equipment failure, delay in delivery, and unqualified quality inspection. Generate a training dataset, in which abnormal data accounts for 30% and normal data accounts for 70%.

[0126] S42: Normalize the training dataset, mapping all parameter data to the [0,1] interval, and randomly divide the dataset into training set, validation set and test set; use an improved LSTM algorithm to build an anomaly warning model, introduce an attention mechanism, focus on key production parameters such as processing accuracy, coating thickness and hole position accuracy, and optimize the model's recognition accuracy.

[0127] S43: Adjust model parameters using the validation set, and set anomaly detection thresholds for each production stage. For example, the anomaly detection threshold for cutting accuracy in the sheet metal processing stage is ±0.15mm, the anomaly detection threshold for adhesive thickness is 0.2-0.4mm, and the anomaly detection threshold for hole position accuracy is ±0.1mm; set warning levels:

[0128] Level 1 warning: Data exceeds the threshold but the deviation is small, and on-site handling is possible within 1 hour, such as when the cutting accuracy is 0.16mm;

[0129] Level 2 warning: Data exceeds the threshold and the deviation is large, requiring process adjustment. The handling time limit is 4 hours, such as when the cutting accuracy is 0.2mm.

[0130] Level 3 warning: Data significantly exceeds the threshold, requiring work stoppage for investigation. The handling time limit is 24 hours, such as when the cutting accuracy is 0.3mm.

[0131] S44: The model was validated using a test set. The test results showed that the model's anomaly recognition accuracy was 96.8% and the early warning response time was 22 seconds, meeting the requirements for real-time early warning in the production site.

[0132] Step S5: Real-time monitoring and anomaly warning. The anomaly warning module dynamically monitors the collected real-time production data through the anomaly warning model. For example, if the collected cutting accuracy data is 0.2mm in the sheet metal processing stage, which exceeds the anomaly judgment threshold, the model will determine it as a level two warning.

[0133] Automatically generate early warning information. The warning information content is as follows: "Order traceability identifier: DZ, abnormal link: sheet metal processing link, abnormal parameter: cutting accuracy, abnormal value: 0.2mm, warning level: level two warning, handling suggestion: adjust the cutting parameters of the CNC cutting machine, reprocess, handling time limit: 4 hours;

[0134] Warning information is pushed through three methods: terminal pop-up windows, SMS notifications, and voice reminders. The information is pushed to the terminals of workstation operators, team leaders, and production supervisors. At the same time, the abnormal information is recorded in the board processing link of the order traceability chain, and associated with information such as the time of occurrence of the abnormality, abnormal parameters, and warning level.

[0135] Step S6: Traceability and Anomaly Handling. After receiving the warning information, the team leader arranges operator CZ001 to handle the anomaly. The operator adjusts the cutting parameters of the CNC cutting machine and reprocesses the board. After processing, the cutting accuracy data is collected as 0.12mm, which meets the anomaly judgment threshold.

[0136] Operators submit abnormal handling instructions through workstation terminals, enter the handling process and results, and the handling record unit records the handling information to form an abnormal handling closed loop; at the same time, the model automatically updates the cutting accuracy abnormal judgment threshold in the plate processing stage and optimizes the early warning model parameters.

[0137] When a customer initiates a traceability query request, the traceability query unit retrieves full-process traceability data and anomaly handling records based on the order number and the order traceability identifier DZ, and generates a traceability report. The report includes basic order information, production data for each stage, anomaly records, anomaly handling process and results, and delivery status, and is pushed to the customer's and relevant management personnel's terminals.

[0138] Example 2:

[0139] This embodiment provides a customized furniture order end-to-end traceability and production anomaly early warning system, used to implement the method described in Embodiment 1. The system includes:

[0140] 1. Order Information Collection Module: This module uses a touch-screen input terminal installed at the order processing station. It has functions for information input, modification, and query. It is used to collect basic information, customization parameters, and customer requirements for customized furniture orders. The module standardizes and encodes the order information to generate a unique order traceability identifier. It also establishes and stores an order information database using MySQL, which supports batch import and export of data.

[0141] 2. Full-process data acquisition module: including IoT sensor unit, PLC control unit, vision inspection unit and manual data entry unit;

[0142] IoT sensor unit: including temperature sensor, humidity sensor, pressure sensor, and speed sensor, which are installed in the processing workshop, storage area, and production equipment respectively, to collect the processing environment temperature and humidity, equipment operating pressure and speed, pressure parameters during raw material processing, etc.

[0143] PLC control unit: adopts Siemens S7-1200 series PLC, which is linked with CNC cutting machine, edge banding machine, assembly equipment and logistics conveying equipment to collect equipment operating parameters and production progress data, and can receive abnormal handling instructions and adjust equipment operating parameters.

[0144] The visual inspection unit includes an industrial camera and an image recognition module, which are installed at the quality inspection station. The industrial camera is used to collect visual data such as the cutting accuracy of sheet metal, the assembly gap of parts, and the surface quality. The image recognition module uses the OpenCV algorithm library to extract the feature parameters in the visual data, compare them with standard parameters, and generate inspection data.

[0145] Manual data entry unit: A portable data entry terminal is used to enter data related to manual operations. It supports offline data entry and automatically synchronizes data after the network is restored.

[0146] 3. Data Processing and Association Module: An industrial computer is used with data processing software to clean, deduplicate, and standardize the collected data throughout the entire process. This module establishes a mapping relationship between order traceability identifiers and data at each stage, constructs a full-process order traceability link, and stores the processed data in a traceability database. The traceability database uses MongoDB, which supports the storage of massive amounts of data and fast querying.

[0147] 4. Anomaly Early Warning Module: Includes a model training unit, a real-time monitoring unit, an early warning generation unit, and an information push unit;

[0148] Model training unit: A GPU server is used to install the TensorFlow framework to build and train anomaly warning models. An improved LSTM algorithm with an attention mechanism is introduced, and historical data can be imported periodically to update model parameters and optimize warning accuracy.

[0149] Real-time monitoring unit: Used to receive real-time data transmitted by the full-process data acquisition module and input it into the anomaly early warning model for anomaly identification;

[0150] Early warning generation unit: Used to generate early warning information of corresponding levels based on the model's recognition results, including order traceability identifier, abnormal process, abnormal parameters, early warning level, handling suggestions, etc.

[0151] Information push unit: It adopts a message push server and supports three push methods: terminal pop-up, SMS notification and voice reminder. The push range can be set according to the warning level to ensure that abnormal information is quickly transmitted to the relevant responsible persons.

[0152] 5. Traceability and Handling Module: Includes a traceability query unit, an anomaly receiving unit, a handling record unit, and a report generation unit;

[0153] Traceability Query Unit: Supports query requests initiated by order number, customer information, or order traceability identifier to quickly retrieve full-process traceability data;

[0154] Anomaly receiving unit: used to receive anomaly information and handling instructions submitted by the terminal, and update the anomaly status in real time;

[0155] The incident handling record unit is used to record the incident handling process, results, personnel involved, and time of handling, forming an incident handling file;

[0156] Report generation unit: Used to automatically generate traceability reports, supports PDF export, and can be used for customer inquiries, quality audits, and management decisions.

[0157] 6. Database Module: Includes Order Information Database (MySQL), Traceability Database (MongoDB), Exception Case Database (MySQL), and Model Parameter Database (MySQL), used to store order information, full-process traceability data, historical exception cases, model training parameters, and exception handling records. The databases are linked through interfaces to achieve data sharing and synchronous updates.

[0158] 7. Terminal Interaction Module: Includes terminals for management personnel, team leaders, and workstation operators, all using industrial tablets or computers with the system client installed, and possessing the following functions:

[0159] Management terminal: View full-process data, set warning levels and anomaly judgment thresholds, review anomaly handling results, and view traceability reports and statistical analysis data;

[0160] Team Leader Terminal: View production data and early warning information for each workstation in the team, receive abnormal handling tasks, and report abnormal handling results;

[0161] Workstation operator terminal: View the production parameters and task requirements of this workstation, receive early warning information, submit abnormal handling instructions and handling records, and enter the production data of this workstation.

[0162] 8. Access control module and data encryption module:

[0163] Access Control Module: Employs a Role-Based Access Control (RBAC) mechanism to assign different operation permissions to personnel in different positions, prevent unauthorized operations, and regularly update permission information;

[0164] Data encryption module: It adopts the AES-256 encryption algorithm to encrypt and store order information, traceability data and abnormal records. Data transmission adopts the HTTPS protocol to ensure data security and privacy and prevent data leakage or tampering.

[0165] The system operation process in this embodiment is as follows: the order information collection module collects order information and generates an order traceability identifier; the full-process data collection module collects data from each stage based on the order traceability identifier and transmits it to the data processing and association module; the data processing and association module processes the data and constructs a traceability link; the anomaly warning module monitors real-time data through an early warning model, generates and pushes early warning information; the traceability and handling module receives traceability query requests and anomaly handling instructions, generates a traceability report and forms a closed-loop handling mechanism; each terminal realizes information viewing and operation through the terminal interaction module, and the permission management module and data encryption module ensure system and data security.

[0166] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0167] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for analyzing processing information and handling anomalies in customized furniture production, characterized in that: Includes the following steps: Step S1: Initialize order information. Collect basic information, customization parameters, and customer requirements for customized furniture orders. Standardize and encode the order information to generate a unique order traceability identifier and establish an order information database. Step S2: Full-process data collection. Based on the order traceability identifier, process data, parameter data, and status data of customized furniture are collected in real time from each stage of order splitting, raw material procurement, board processing, parts assembly, quality inspection, packaging and warehousing, logistics and distribution, and installation and acceptance, forming a full-process traceability dataset. Step S3: Data processing and association. The data in the full-process traceability dataset is cleaned, deduplicated, and standardized. An association mapping relationship between order traceability identifiers and data in each link is established to construct the order full-process traceability link. Step S4: Production anomaly early warning modeling. Based on historical order traceability data and production anomaly case data, an anomaly early warning model is constructed. Anomaly judgment threshold and early warning level are set for each production link. The anomaly early warning model uses an improved LSTM algorithm and combines production process standards to identify abnormal data. Step S5: Real-time monitoring and anomaly warning. The collected real-time production data is dynamically monitored through the anomaly warning model. When the data exceeds the anomaly judgment threshold of the corresponding link, an alarm message is automatically generated according to the alarm level and pushed to the corresponding management terminal and production workstation. At the same time, the anomaly information is recorded to the order traceability link. Step S6: Traceability Inquiry and Anomaly Handling. Receive traceability inquiry requests, retrieve full-process traceability data and anomaly handling records based on order traceability identifiers, and generate a traceability report; receive anomaly handling instructions, record the handling process and results, form an anomaly handling closed loop, and update the corresponding data in the order traceability chain.

2. The method for processing information analysis and anomaly handling in customized furniture production according to claim 1, characterized in that: In step S1, the basic order information includes order number, customer information, order time, and delivery deadline; the customized parameter information includes furniture type, size specifications, material parameters, process requirements, and color parameters; the standardized code adopts the format of "order type code + timestamp + random check code"; the order traceability identifier corresponds one-to-one with the order number and is associated with the raw material batch number, equipment number, and operator ID.

3. The method for processing information analysis and anomaly handling in customized furniture production according to claim 1, characterized in that: In step S2, the process data includes the start time, end time, operation procedure, operators, and equipment operating status of each stage; the parameter data includes processing speed, cutting accuracy, adhesive coating thickness, assembly gap, and inspection data; the data acquisition adopts a combination of IoT sensors, PLC controllers, vision inspection equipment, and manual input terminals, and the acquisition frequency is set to 1-10Hz according to the production stage requirements, of which the acquisition frequency for the processing stage and the inspection stage is not less than 5Hz.

4. The method for processing information analysis and anomaly handling in customized furniture production according to claim 1, characterized in that: In step S3, data processing includes removing invalid data, correcting abnormal deviation data, and unifying data format and units; the association mapping relationship is stored in a tree structure, with the order traceability identifier as the root node, each production link as a child node, and the data of each link as the child node attribute, to achieve a three-level association traceability of "order-link-data".

5. The method for processing information analysis and anomaly handling in customized furniture production according to claim 1, characterized in that, In step S4, the construction process of the anomaly early warning model includes: S41: Collect historical order process data and production anomaly cases, including types such as unqualified raw materials, excessive processing precision, equipment failure, delay in delivery, and unqualified quality inspection. Classify and label the anomaly cases to generate a training dataset. S42: Normalize the training dataset, divide it into training, validation and test sets, build a model using an improved LSTM algorithm, and introduce an attention mechanism to optimize the model's recognition accuracy of key production parameters. S43: Adjust the model parameters through the validation set and set the anomaly judgment threshold corresponding to the warning level. The warning levels are divided into Level 1 warning, Level 2 warning and Level 3 warning. Different levels correspond to different push ranges and handling time limits. S44: Use a test set to validate the model, ensuring that the model's anomaly identification accuracy is not less than 95% and the warning response time does not exceed 30 seconds.

6. The method for analyzing processing information and handling anomalies in customized furniture production according to claim 1, characterized in that: In step S5, the warning information includes order traceability identifier, abnormal process, abnormal parameter, abnormal value, warning level, handling suggestions, and push time; the push methods include terminal pop-up window, SMS notification, and voice reminder.

7. The method and system for processing information analysis and anomaly handling in customized furniture production according to claim 1, characterized in that: In step S6, traceability query requests can be initiated using any of the following methods: order number, customer information, or order traceability identifier. The traceability report includes basic order information, full-process production data, anomaly records, anomaly handling process and results, and delivery status. The anomaly handling closed loop includes five stages: anomaly reception, cause analysis, handling implementation, effect verification, and record archiving. After the handling is completed, the model automatically updates the anomaly judgment threshold for the corresponding stage to optimize the accuracy of early warning.

8. The processing information analysis and anomaly handling system for customized furniture production according to claim 1, characterized in that: The system for implementing the method of any one of claims 1-7 comprises: The order information collection module is used to collect basic information, customization parameter information and customer demand information of customized furniture orders, standardize the order information, generate a unique order traceability identifier, and establish and store the order information database. The end-to-end data acquisition module includes an IoT sensor unit, a PLC control unit, a vision inspection unit, and a manual data entry unit. It is used to collect process data, parameter data, and status data of each production, logistics, and installation stage of customized furniture in real time based on order traceability tags, and transmit them to the data processing module. The data processing and association module is used to clean, deduplicate, and standardize the collected data throughout the entire process, establish the association mapping relationship between order traceability identifiers and data at each stage, construct the order full-process traceability link, and store the processed data in the traceability database. The anomaly warning module includes a model training unit, a real-time monitoring unit, an early warning generation unit, and an information push unit. It is used to build and train an anomaly warning model, dynamically monitor real-time production data, identify abnormal data, generate early warning information of corresponding levels, and push it to the corresponding terminal. The traceability and handling module includes a traceability query unit, an anomaly receiving unit, a handling record unit, and a report generation unit. It is used to receive traceability query requests, retrieve traceability data to generate traceability reports, receive anomaly handling instructions, record the handling process and results, form an anomaly handling closed loop, and update the traceability link data. The database module includes an order information database, a traceability database, an anomaly case database, and a model parameter database, which are used to store order information, full-process traceability data, historical anomaly cases, model training parameters, and anomaly handling records. The terminal interaction module includes terminals for management personnel, team leaders, and workstation operators. It is used to receive early warning information, initiate traceability queries, submit abnormal handling instructions, and view traceability reports and handling records.

9. The processing information analysis and anomaly handling system for customized furniture production according to claim 8, characterized in that: The IoT sensor unit in the full-process data acquisition module includes a temperature sensor, a humidity sensor, a pressure sensor, and a speed sensor, which are used to collect corresponding parameters of the processing environment, equipment operation, and raw material processing, respectively; the vision inspection unit includes an industrial camera and an image recognition module, which are used to collect visual data such as plate cutting accuracy, component assembly gap, and surface quality, and extract feature parameters through image recognition algorithms; the PLC control unit is linked with the furniture production equipment and logistics conveying equipment to collect equipment operating parameters and production progress data.

10. The processing information analysis and anomaly handling system for customized furniture production according to claim 8, characterized in that: The system also includes a permission management module and a data encryption module.

Images