Garment production management method and digital management system thereof

Abstract

The invention belongs to the technical field of production management, particularly relates to a garment production management method and a digital management system thereof, and solves the technical problems of extensive abnormal classification and grade judgment and no hierarchical logic in control signal generation. The abnormal positioning signal or the risk early warning signal is generated according to the abnormal type matching degree of the normal production line for the slight abnormality, and the emergency shutdown signal or the local adjustment signal is generated according to the historical record for the serious abnormality, so that the abnormality handling efficiency is improved, and the batch defect occurrence rate is reduced; during signal processing, a raw material batch report, an equipment operation curve and an operator training record are synchronously called, and a classification attribution and association verification mode is adopted, so that the traceability coverage rate of potential factors is improved, the recurrence rate of similar abnormalities is reduced, a four-dimensional quantitative model is constructed, and the subjective error of risk level judgment is reduced.

Description

A management method for garment production and its digital management system Technical Field

[0001] This invention belongs to the field of production management technology, specifically a management method for garment production and its digital management system. Background Technology

[0002] With the fast fashion trend in the apparel industry, production cycles are constantly shortening, while consumers' demands for product quality are continuously increasing. Apparel production management faces the core challenge of balancing efficiency and quality.

[0003] Patent application CN202210331544.0 discloses a management system and method for garment production, comprising: an order management module for generating order data; a production line data acquisition module for acquiring various data from the production line; a processing procedure generation module for obtaining the results of automated processing procedures; a process simulation and adjustment module for obtaining various simulated and adjusted production process data; a decision-making module for inputting the processing procedure results decided by the production decision-maker; and a process allocation processing module for allocating production line processes based on the results of automated processing procedures, or adjusting production line processes based on the results of decision-making processing procedures.

[0004] However, the current mainstream quality control methods in the industry still have the following limitations: the level determination often uses a one-size-fits-all threshold, which cannot match the different impacts of different defects on product quality; the same handling process is used for minor and serious abnormalities; the risk diffusion is not judged by combining the historical abnormality records of normal production lines; when handling abnormal signals, the focus is often on a single factor, without integrating multi-dimensional data such as raw material batches, operator skills, and production environment, resulting in incomplete tracing of potential factors, repeated occurrence of similar abnormalities, and risk level classification relying on the subjective judgment of management personnel without a clear multi-dimensional quantitative model. Furthermore, the response time limit is not linked to the risk level, resulting in the lag in resource allocation for high-risk issues and the excessive effort spent on low-risk issues. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides a management method and digital management system for garment production, which solves the problems of crude anomaly classification and level determination, and the lack of hierarchical logic in the generation of control signals.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a management method for garment production, comprising the following steps: Step 1: Assigning a unique identifier to each garment production line and calculating the pass rate of randomly sampled finished products, while simultaneously screening out abnormal production lines with pass rates lower than a preset pass rate; Step 2: Extracting defective products from all abnormal production lines and classifying them by layer, counting the number of abnormal products in each category and the total number, calculating the proportion of the same type of abnormality, and determining the abnormality level as minor or serious; Step 3: For minor abnormalities, determining whether the same situation exists in normal production lines, generating an abnormality location signal or risk warning signal; for serious abnormalities, acquiring all normal production lines and their abnormal records, if the current serious abnormality... If the anomaly type is consistent with past anomaly records, it is judged as a systemic serious anomaly; otherwise, it is judged as an occasional serious anomaly. Step four: Process the anomaly location signal, extract the anomaly information and locate it in combination with production management data. Use a classification attribution + correlation verification mode to conduct in-depth attribution analysis of potential factors and generate source tracing and processing information. Step five: Process the risk warning signal, clean the collected real-time data, and conduct vertical comparison and horizontal correlation with historical production data. If there is a risk, generate a risk assessment analysis signal. Step six: Process the risk assessment analysis signal and use a multi-dimensional quantitative model including anomaly type severity, anomaly occurrence frequency, anomaly duration, and production factor correlation to assess the risk level.

[0007] As a further aspect of the present invention, each garment production line is assigned a unique label, denoted as production line i, where i = 1, 2, ..., j, and j represents the number of production lines. A random sample of n finished products is taken from each production line i, and the finished products undergo quality inspection. The quality inspection covers at least three dimensions: size deviation, sewing defects, and appearance flaws. A defective product is defined as a finished product exhibiting any one of the three abnormalities. The number of defective products corresponding to each production line i is counted. The defect rate is calculated using the formula: Defect Rate = (Defective Products / Sampled Products) × 100%. This defect rate is then used to calculate the product qualification rate corresponding to production line i. If the qualification rate is greater than the preset qualification rate, the production line is marked as a normal production line; otherwise, it is marked as an abnormal production line.

[0008] As a further aspect of the present invention, defective products from all abnormal production lines are extracted and recorded as a total abnormal product set. The abnormal products in the set are then classified into three categories: size deviation, sewing defects, and appearance defects. The number of abnormal products in each category is counted, and the proportion of the same type of abnormality is calculated. The abnormality proportion is compared with a preset level threshold to determine whether the abnormality level is minor or severe.

[0009] As a further aspect of the present invention, for minor anomalies, all normal production lines and their abnormal products are acquired, and the abnormal products are classified. If a secondary category consistent with the current minor anomaly type exists after classification, an anomaly location signal is generated; otherwise, a risk warning signal is generated.

[0010] As a further aspect of the present invention, for serious anomalies, all normal production lines and their past production process anomaly records are obtained. If the current serious anomaly type is consistent with the past anomaly records and occurs ≥3 times in a similar production cycle, it is determined to be a systemic serious anomaly, an emergency shutdown and rectification signal is generated, and a deep investigation process is triggered. If the current serious anomaly type is inconsistent with the past anomaly records, it is determined to be an occasional serious anomaly, targeted improvement measures are initiated, and serious anomaly handling information is generated.

[0011] As a further aspect of this invention, the abnormal location signal is processed to analyze the abnormal production line number and abnormality type in the signal. Combined with production management system data, the specific workstation and time range where the abnormality occurred are identified. If the same minor abnormality occurs in two consecutive time periods at the same workstation, the time range is expanded to the entire day. Core data for the corresponding time period of the workstation is retrieved. The core data includes at least raw material batches and inspection reports, equipment operating parameters and inspection records, and operator skill levels and training records. A classification attribution + correlation verification model is used to analyze potential factors, and the attribution dimensions include at least raw material quality fluctuations, equipment parameter deviations, and operator non-standard operations. If the abnormality occurs in isolation and there are no clear related factors, it is determined to be an accidental minor deviation; otherwise, traceability processing information is generated.

[0012] As a further aspect of this invention, the risk warning signal is processed by collecting real-time production data, which includes at least the production line operating speed, equipment temperature, equipment pressure, and raw material supply batches. The real-time data is cleaned, invalid data is deleted, and ambiguous information is corrected. The frequency of anomalies is statistically analyzed by grouping by anomaly type, process, and collection time window to generate a real-time anomaly frequency statistics table. Historical production data of the corresponding production line is retrieved, and abnormal production data is extracted and grouped by anomaly type, process, and time period to generate a historical anomaly frequency set. Vertical comparison and horizontal correlation are performed. A risk assessment analysis signal is generated if any of the following conditions are met: the frequency of the same type of anomaly exceeds the normal threshold for two consecutive days; the daily frequency of a single type of anomaly increases by ≥50% compared to the previous day; or the same type of anomaly in the same process / equipment / person occurs ≥3 times on the same day.

[0013] As a further aspect of this invention, the risk assessment analysis signals are processed. For the quantitative assessment of anomaly types, minor appearance defects are assigned 2 points, minor sewing defects 4 points, minor dimensional deviations 6 points, functional defects 8 points, and safety hazard defects 10 points. For the quantitative assessment of anomaly frequency, ≤2 occurrences within 24 hours are assigned 1 point, 3-5 occurrences within 24 hours 3 points, 6-10 occurrences within 24 hours 5 points, 11-20 occurrences within 24 hours 7 points, and so on. 10 points are awarded for occurrences exceeding 20 times within an hour; for the quantitative assessment of the duration of anomalies, 1 point is awarded for duration ≤1 production batch, 3 points for duration 2-3 production batches, 5 points for duration 4-5 production batches, 7 points for duration 6-10 production batches, and 10 points for duration >10 production batches; for the quantitative assessment of production factors, 1 point is awarded for isolated factors, 3 points for local factors, 5 points for cross-line factors, 7 points for global factors, and 10 points for uncontrollable factors.

[0014] As a further aspect of the present invention, the multi-dimensional quantitative model assesses the risk level as follows: the total risk assessment score is calculated according to the formula: Total Risk Assessment Score = (Severity Score of Anomaly Type × Weight 1) + (Frequency Score of Anomaly Occurrence × Weight 2) + (Duration Score of Anomaly × Weight 3) + (Relevance Score of Production Factors × Weight 4). If the total score is 0-3 points, the level is low risk; if it is 3.01-6 points, the level is medium risk; and if it is 6.01-10 points, the level is high risk.

[0015] A digital management system for garment production includes: an abnormal production identification module, which assigns a unique number to each garment production line, calculates the pass rate of randomly sampled finished products, filters abnormal production lines with a pass rate lower than a preset pass rate, extracts defective products from all abnormal production lines and categorizes them hierarchically, counts the number of abnormal products in each category and the total number, calculates the proportion of the same type of abnormality, and determines the abnormality level as minor or serious; and an abnormality type processing module, which processes minor and serious abnormalities separately. For minor abnormalities, it determines whether the same situation exists in normal production lines, generates an abnormality location signal or a risk warning signal, and transmits them to the abnormality source analysis module and the risk warning analysis module, respectively. For serious abnormalities, it acquires all normal production lines and their abnormal records. If the current serious abnormality type is consistent with past abnormal records, it is determined to be a systemic serious abnormality. If the anomaly is normal, it is considered an occasional severe anomaly; the anomaly tracing and analysis module processes the anomaly location signal, extracts the anomaly information and combines it with production management data for location, uses a classification attribution + correlation verification mode to conduct in-depth attribution analysis of potential factors, and generates tracing and processing information; the risk warning analysis module processes the risk warning signal, cleans the collected real-time data, and conducts vertical comparison and horizontal correlation with historical production data. If a risk exists, a risk assessment analysis signal is generated, which is then processed. A multi-dimensional quantitative model including anomaly type severity, anomaly frequency, anomaly duration, and production factor correlation is used to assess the risk level, and the generated risk level is transmitted to the management information output module; the management information output module displays the acquired risk level to the corresponding management personnel.

[0016] Compared with existing technologies, the present invention has the following beneficial effects: For minor anomalies, the present invention generates anomaly location signals or risk warning signals by combining the anomaly type matching degree of normal production lines; for serious anomalies, it generates emergency stop signals or local adjustment signals by combining historical records to determine whether they are systematic or sporadic. This improves the efficiency of anomaly handling, reduces the occurrence rate of batch defects, and simultaneously retrieves raw material batch reports, equipment operation curves, and operator training records when processing signals. It adopts a classification attribution + correlation verification mode to improve the coverage of potential factor tracing, reduce the recurrence rate of similar anomalies, construct a four-dimensional quantitative model to reduce subjective errors in risk level judgment, and establish a time limit mechanism for different risk responses to improve the efficiency of resource allocation for high-risk problems, reduce the effort required for low-risk problems, and improve overall production efficiency. Attached Figure Description

[0017] Figure 1 is a step-by-step diagram of the present invention; Figure 2 is a system block diagram of the present invention. Detailed Implementation

[0018] 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.

[0019] Example 1 (Referring to Figure 1): This invention provides a management method for garment production, including the following steps: Step 1: Obtain all garment production lines and assign them unique labels, denoted as i, where i = 1, 2, ..., j, and j represents the number of production lines. Randomly sample n finished products from different production lines i and inspect the quality of the randomly sampled finished products, specifically checking for size deviations, sewing defects, and appearance flaws. Identify defective products in the random sampling, where defective products indicate any one of the abnormalities in size deviation, sewing defects, or appearance flaws. Calculate the defect rate using the formula: Defect Rate = (Defective Products / Sampled Products) × 100%. Calculate the product qualification rate corresponding to production line i based on this rate and compare it with a preset qualification rate. If the pass rate is greater than the preset pass rate, it indicates that the production quality of the corresponding production line is normal, and the production line is marked as a normal production line. Conversely, if the pass rate is less than the preset pass rate, it indicates that the production quality of the corresponding production line is abnormal, and the production line is marked as an abnormal production line. Step 2: Extract abnormal products that are determined to be defective products in the sampling inspection of all abnormal production lines, perform stratified classification processing on abnormal products to obtain abnormal products of the same type, specifically including size deviation, sewing defects, and appearance defects. Count the number of abnormal products in each category, calculate the proportion of abnormal products of the same type, and compare the proportion of abnormal products of the same type with the corresponding preset level threshold to determine the abnormal level corresponding to the abnormal type. The abnormal level includes minor abnormality and serious abnormality.

[0020] Step 3: Analyze cases with a minor anomaly level. Obtain all normal production lines and the corresponding abnormal products within those lines. Classify the abnormal products according to their anomaly type and determine if any of the classified anomaly types match the current minor anomaly level. If so, generate an anomaly location signal; otherwise, generate a risk warning signal. Analyze cases with a severe anomaly level. Obtain all normal production lines and their past anomaly records. Compare the current severe anomaly type with past anomaly records. If the same severe anomaly type exists and occurs multiple times within similar production cycles, it is classified as a systemic severe anomaly, generating an emergency shutdown and rectification signal and triggering a deep investigation process. If no similar severe anomaly type exists or has only occurred sporadically in the past, it is classified as an sporadic severe anomaly, generating a local adjustment signal and initiating targeted improvement measures, and generating severe anomaly handling information.

[0021] Step 4: Process the generated abnormal location signals, extract the abnormal production line number and abnormality type contained in the signals, and combine them with production management system data to accurately pinpoint the specific workstation and time range where the abnormality occurred, ensuring accurate positioning. If the same minor abnormality occurs at the same workstation in two consecutive time periods, the time range needs to be expanded to the entire day to investigate whether there is a persistent problem. Clarify the data retrieval list and synchronize core data from the corresponding system, including raw material information, equipment operation data, and operator information. Use a classification attribution + correlation verification mode to conduct in-depth attribution analysis of potential factors. For the raw material dimension, compare the raw material parameters of the batch in question with those of qualified batches to determine whether there are minor quality fluctuations. For the equipment dimension, check the equipment operation curve to investigate whether there are minor faults such as parameters temporarily deviating from the set value and not triggering alarms. For the operator dimension, observe the on-site operation video or supervise the operation to determine whether there are any deficiencies in the execution of standardized operations. If the analysis finds that the abnormality only occurs in isolation and there are no clear related factors, it is judged as an accidental minor deviation; otherwise, traceability and processing information is generated.

[0022] Step 5: Process the generated risk warning signals. Collect real-time data during the production process, including production line operating speed, equipment temperature, pressure parameters, and raw material supply. Clean the obtained real-time data, specifically by deleting invalid data (such as incorrect entries or duplicate records) and correcting ambiguous information. Based on the cleaned real-time data, group it by anomaly type, corresponding process, and collection time window, and statistically analyze the frequency of anomalies in each dimension to generate a real-time anomaly frequency statistics table. Define the criteria for judging a single anomaly, such as counting a temperature exceeding the set range for ≥3 minutes as one anomaly. Simultaneously, analyze the production data... We retrieve historical production data for the corresponding production line from the database for the past three months (the duration can be flexibly configured). We accurately extract abnormal production data and group it by three dimensions: abnormality type, process, and time period. This generates a standardized set of historical abnormality frequencies, including the historical mean, peak value, fluctuation range, and common causes for each dimension. We then perform longitudinal comparisons and horizontal correlation analyses. For longitudinal comparisons, based on the real-time abnormality frequency statistics table and the historical abnormality frequency set, we conduct trend comparisons, including the following comparison directions: Core comparison item: the real-time abnormality frequency of the day and the historical mean of the corresponding time period (e.g., today's...). The temperature exceeded the standard 3 times between 9:00 and 10:00, compared to the historical average of 1 time (and historical peak values); auxiliary analysis: calculate the increase / decrease of real-time anomaly frequency compared to the historical average, mark whether there are special trends such as breaking through historical peaks or continuous abnormal increases, and locate the time nodes of abnormal fluctuations; for horizontal correlation, investigate the correlation between anomalies and production factors, verify the correlation strength through cross-analysis tables, and show the anomaly type in the cross-analysis tables; the table columns are production factor dimensions, such as equipment number, operator ID, raw material batch, specifically analyzed from the following dimensions: equipment dimension: analyze whether the same equipment has a concentrated occurrence of a certain type of anomaly; Personnel dimension: Analyze whether the frequency of anomalies in the same operator's work process is significantly higher than that of others; Material dimension: Verify whether anomalies in the corresponding work process occur in a concentrated manner after a batch of raw materials is put in, generate a production horizontal correlation analysis report, and identify the core correlation elements; The preset thresholds need to be set in combination with historical data averages, industry standards, and equipment process requirements. Risk assessment will be initiated if any of the following conditions are met: The frequency of the same type of anomaly exceeds the normal threshold for two consecutive days; The daily frequency of a single type of anomaly increases by ≥50% compared to the previous day; The same type of anomaly in the same work process / equipment / person occurs ≥3 times on the same day, and a risk assessment analysis signal is generated.

[0023] Step Six: Further process the generated risk assessment analysis signals. Based on the preset risk assessment model, comprehensively consider multiple dimensions such as anomaly type, anomaly frequency, anomaly duration, and production factors to quantitatively assess the risk level. The risk level is divided into three levels: low risk, medium risk, and high risk. The specific risk level assessment method is as follows: For the quantitative assessment of anomaly type, minor appearance defects are assigned 2 points, minor sewing defects are assigned 4 points, minor dimensional deviations are assigned 6 points, functional defects are assigned 8 points, and safety hazard defects are assigned 10 points; For the quantitative assessment of anomaly frequency, ≤2 occurrences within 24 hours are assigned 1 point, 3-5 occurrences within 24 hours are assigned 3 points, 6-10 occurrences within 24 hours are assigned 5 points, and so on. For anomalies occurring 11-20 times within a 24-hour period, a score of 7 is assigned; for anomalies occurring more than 20 times within a 24-hour period, a score of 10 is assigned. Regarding the quantitative assessment of the duration of anomalies, ≤1 production batch is assigned 1 point, 2-3 production batches are assigned 3 points, 4-5 production batches are assigned 5 points, 6-10 production batches are assigned 7 points, and >10 production batches are assigned 10 points. Regarding the quantitative assessment of production factors, isolated factors are assigned 1 point, local factors 3 points, cross-line factors 5 points, global factors 7 points, and uncontrollable factors 10 points. Based on the obtained quantitative scores from different dimensions, the total risk assessment score is calculated using the formula: Total Risk Assessment Score = (Severity Score of Anomaly Type × Weight 1) + (Frequency Score of Anomaly Occurrence × Weight 2) + The total risk assessment score is calculated by adding (anomaly duration score × weight 3) and (production factor correlation score × weight 4), and the risk level is classified according to the total score. If the total score is 0-3, the level is low risk, and the system automatically generates risk warning information and pushes it to relevant production management personnel and corresponding workstation operators to remind them to pay attention to potential risks, strengthen production process monitoring, and record the risk warning information to the production management database. If the total score is 3.01-6, the level is medium risk. In addition to generating and pushing risk warning information, the system automatically triggers the production adjustment suggestion generation module. Based on the anomaly type and risk level, combined with historical production data and expert experience database, targeted production adjustment suggestions are generated, such as adjusting equipment parameters and changing raw material batches. The system optimizes personnel scheduling and other aspects, and pushes the adjustment suggestions to the production decision-making and execution levels. At the same time, it activates an adjustment effect tracking mechanism to monitor production data after the adjustment in real time and evaluate the effect of the adjustment. If the level is 6.01-10, it is considered high risk. The system immediately generates an emergency shutdown command to forcibly stop the operation of the relevant production line to prevent the risk from escalating further. At the same time, it automatically triggers the emergency plan, notifying the emergency response team to quickly arrive at the scene to carry out emergency investigation and handling work, including equipment maintenance, raw material replacement, personnel evacuation, etc. The emergency shutdown information and emergency handling progress are fed back to the production management and senior management in real time to ensure smooth information flow so as to make timely decisions and deployments. The production line can only be restarted after the risk has been eliminated and the system has assessed and confirmed that it is safe.

[0024] Example 2 (Referring to Figure 2): This invention provides a digital management system for garment production, including an abnormal production identification module, an abnormality type processing module, an abnormality source analysis module, a risk warning analysis module, and a management information output module. As shown in Figure 2, information between these functional modules is transmitted unidirectionally. The abnormal production identification module assigns a unique number to each garment production line, calculates the pass rate of randomly sampled finished products, filters abnormal production lines with pass rates lower than a preset pass rate, extracts defective products from all abnormal production lines, categorizes them hierarchically, counts the number of abnormal products in each category and the total number, calculates the proportion of the same type of abnormality, and determines the abnormality level as minor or severe. The specific processing method is the same as in steps one and two. The abnormality type processing module handles minor and severe abnormalities separately. For minor abnormalities, it determines whether the same situation exists in normal production lines, generates an abnormality location signal or a risk warning signal, and transmits them to the abnormality source analysis module and the risk warning analysis module respectively. For severe abnormalities, it obtains all normal production lines and their abnormal records. If the current severe abnormality type is similar to... If past anomaly records are consistent, it is determined to be a systemic serious anomaly; otherwise, it is determined to be an occasional serious anomaly. The specific handling method is the same as the process in step three. The anomaly tracing and analysis module processes the anomaly location signal, extracts anomaly information, and combines it with production management data for location. It uses a classification attribution + correlation verification mode to conduct in-depth attribution analysis of potential factors and generates tracing and processing information. The specific handling method is the same as the process in step four. The risk warning analysis module processes the risk warning signal, cleans the collected real-time data, and conducts vertical comparison and horizontal correlation with historical production data. If a risk exists, a risk assessment analysis signal is generated. The specific handling method is the same as the process in step five. The risk assessment analysis signal is processed, and a multi-dimensional quantitative model including anomaly type severity, anomaly frequency, anomaly duration, and production factor correlation is used to assess the risk level. The specific handling method is the same as the process in step six. The generated risk level is then transmitted to the management information output module. The management information output module displays the acquired risk level to the corresponding management personnel.

[0025] Some of the data in the above formulas are numerical calculations with dimensions removed, and the contents not described in detail in this specification are all prior art known to those skilled in the art.

[0026] The above embodiments are only used to illustrate the technical methods of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical methods of the present invention without departing from the spirit and scope of the technical methods of the present invention.

Claims

1. A management method for garment production, characterized in that, Includes the following steps: Step 1: Assign a unique identifier to each garment production line and calculate the pass rate of randomly sampled finished products. Simultaneously, screen out abnormal production lines with pass rates lower than a preset pass rate. Step 2: Extract defective products from all abnormal production lines and categorize them hierarchically. Count the number of abnormal products in each category and the total number of defective products, calculate the percentage of the same type of abnormality, and determine the abnormality level as minor or severe. Step 3: For minor abnormalities, determine if the same situation exists in normal production lines and generate an abnormality location signal or risk warning signal. For severe abnormalities, obtain all normal production lines and their abnormality records. If the current severe abnormality type is consistent with past abnormality records, it is determined to be a systemic severe abnormality. If the anomaly is normal, it is judged as an occasional serious anomaly; Step 4: Process the anomaly location signal, extract the anomaly information and locate it in combination with production management data, and use the classification attribution + correlation verification mode to conduct in-depth attribution analysis of potential factors and generate source tracing and processing information; Step 5: Process the risk warning signal, clean the collected real-time data, and conduct vertical comparison and horizontal correlation with historical production data. If there is a risk, a risk assessment analysis signal is generated; Step 6: Process the risk assessment analysis signal and use a multi-dimensional quantitative model including the severity of the anomaly type, the frequency of the anomaly, the duration of the anomaly, and the correlation of production factors to assess the risk level.

2. The management method for garment production according to claim 1, characterized in that, All garment production lines are assigned a unique label, denoted as production line i, where i = 1, 2, ..., j, and j represents the number of production lines. n finished products are randomly selected from each production line i, and their quality is inspected. The quality inspection covers at least three dimensions: size deviation, sewing defects, and appearance flaws. A defective product is defined as a finished product exhibiting any one of the three abnormalities. The number of defective products corresponding to each production line i is counted. The defect rate is calculated using the formula: Defect Rate = (Defective Products / Sampled Products) × 100%. This defect rate is then used to calculate the product pass rate corresponding to production line i. If the pass rate is greater than the preset pass rate, the production line is marked as a normal production line; otherwise, it is marked as an abnormal production line.

3. The management method for garment production according to claim 1, characterized in that, Extract all defective products from abnormal production lines and record them as the total abnormal product set. Classify the abnormal products in the set into three categories: size deviation, sewing defect, and appearance defect. Count the number of abnormal products in each category and calculate the proportion of the same type of abnormality. Compare the abnormality proportion with a preset level threshold to determine whether the abnormality level is minor or severe.

4. The management method for garment production according to claim 1, characterized in that, For minor anomalies, acquire all normal production lines and their abnormal products, classify the abnormal products, and generate an anomaly location signal if a secondary category consistent with the current minor anomaly type exists after classification; otherwise, generate a risk warning signal.

5. The management method for garment production according to claim 1, characterized in that, For serious anomalies, obtain all normal production lines and their past anomaly records. If the current serious anomaly type is consistent with the past anomaly records and has occurred ≥3 times in a similar production cycle, it is determined to be a systemic serious anomaly. An emergency shutdown and rectification signal is generated and a deep investigation process is triggered. If the current serious anomaly type is inconsistent with the past anomaly records, it is determined to be an occasional serious anomaly. Targeted improvement measures are initiated, and serious anomaly handling information is generated.

6. The management method for garment production according to claim 1, characterized in that, The abnormal location signal is processed, and the abnormal production line number and abnormality type in the signal are analyzed. Combined with the production management system data, the specific workstation and time range of the abnormality are located. If the same minor abnormality occurs in two consecutive time periods at the same workstation, the time range is expanded to the entire day. The core data of the corresponding time period of the workstation is retrieved. The core data includes at least the raw material batch and inspection report, equipment operating parameters and inspection records, and operator skill level and training records. A classification attribution + correlation verification mode is used to analyze potential factors, and the attribution dimensions include at least raw material quality fluctuations, equipment parameter deviations, and operator non-standard operation. If the abnormality occurs in isolation and there are no clear related factors, it is judged as an accidental minor deviation; otherwise, traceability processing information is generated.

7. The management method for garment production according to claim 1, characterized in that, The risk warning signals are processed by collecting real-time production data, which includes at least production line operating speed, equipment temperature, equipment pressure, and raw material supply batches. The real-time data is cleaned, invalid data is deleted, and ambiguous information is corrected. The frequency of anomalies is statistically analyzed by grouping by anomaly type, process, and collection time window, generating a real-time anomaly frequency statistics table. Historical production data of the corresponding production line is retrieved, and abnormal production data is extracted and grouped by anomaly type, process, and time period to generate a historical anomaly frequency set. Vertical comparison and horizontal correlation are carried out. A risk assessment analysis signal is generated if any of the following conditions are met: the frequency of the same type of anomaly exceeds the normal threshold for two consecutive days; the daily frequency of a single type of anomaly increases by ≥50% compared to the previous day; or the same type of anomaly in the same process / equipment / person occurs ≥3 times on the same day.

8. The management method for garment production according to claim 1, characterized in that, The risk assessment analysis signals are processed, and quantitative assessments are performed on the anomaly types: minor appearance defects are assigned 2 points, minor sewing defects 4 points, minor dimensional deviations 6 points, functional defects 8 points, and safety hazard defects 10 points. Quantitative assessments are also performed on the anomaly frequency: ≤2 occurrences within 24 hours are assigned 1 point, 3-5 occurrences within 24 hours 3 points, 6-10 occurrences within 24 hours 5 points, 11-20 occurrences within 24 hours 7 points, and >20 occurrences within 24 hours 0 points. 20 occurrences are assigned 10 points; for the quantitative assessment of the duration of abnormalities, 1 point is assigned for ≤1 production batch, 3 points for 2-3 production batches, 5 points for 4-5 production batches, 7 points for 6-10 production batches, and 10 points for >10 production batches; for the quantitative assessment of production factors, 1 point is assigned for isolated factors, 3 points for local factors, 5 points for cross-line factors, 7 points for global factors, and 10 points for uncontrollable factors.

9. The management method for garment production according to claim 1, characterized in that, The multi-dimensional quantitative model assesses risk levels as follows: The total risk assessment score is calculated using the formula: Total Risk Assessment Score = (Severity Score of Anomaly Type × Weight 1) + (Frequency Score of Anomaly Occurrence × Weight 2) + (Duration Score of Anomaly × Weight 3) + (Interrelationship Score of Production Factors × Weight 4). If the total score is 0-3, the level is low risk; if it is 3.01-6, the level is medium risk; and if it is 6.01-10, the level is high risk.

10. A digital management system for garment production, used to execute the garment production management method according to any one of claims 1-9, characterized in that, include: An abnormal production identification module is used to assign a unique number to the garment production line, calculate the pass rate of randomly sampled finished products, screen abnormal production lines with a pass rate less than a preset pass rate, extract defective products from all abnormal production lines and classify them into layers, count the number of abnormal products in each category and the total number, calculate the proportion of the same type of abnormality, and determine the abnormality level as a minor abnormality or a serious abnormality. An anomaly type processing module is used to process minor and serious anomalies separately. For minor anomalies, it determines whether the same situation exists in normal production lines, generates an anomaly location signal or a risk warning signal, and transmits them to the anomaly tracing and analysis module and the risk warning analysis module, respectively. For serious anomalies, it obtains all normal production lines and their anomaly records. If the current serious anomaly type is consistent with past anomaly records, it is determined to be a systemic serious anomaly; otherwise, it is determined to be an occasional serious anomaly. An anomaly tracing and analysis module is used to process the anomaly location signal, extract anomaly information and locate it in combination with production management data. It uses a classification attribution + association verification mode to perform in-depth attribution analysis of potential factors and generate tracing and processing information. The risk warning analysis module is used to process risk warning signals, clean the collected real-time data, and conduct vertical comparison and horizontal correlation with historical production data. If a risk exists, a risk assessment analysis signal is generated. The risk assessment analysis signal is processed, and a multi-dimensional quantitative model including the severity of the anomaly type, the frequency of the anomaly, the duration of the anomaly, and the correlation of production factors is used to assess the risk level. The generated risk level is then transmitted to the management information output module. The management information output module is used to display the acquired risk level to the corresponding management personnel.

Images