Abnormal processing method, device and storage medium for sewing product working time standard

Abstract

The present invention discloses an abnormality handling method, device and storage medium for sewing product working time standard, the method comprises: real-time collection of sewing product related data, the sewing product related data at least comprising: various business data of sewing processing, environmental data and worker on-the-job data; processing the collected sewing product related data; inputting the processed data into a pre-constructed sewing product working time abnormality detection model for detection; when the current sewing product working time abnormality is detected, reporting the abnormal information in a preset manner according to a predetermined process. The scheme of the present invention reduces the time for handling working time abnormalities, reduces waiting and waste of related resources, and improves work efficiency.

Description

Technical Field

[0001] The present invention relates to the technical field of sewing products, and in particular to an abnormality processing method, device and storage medium for sewing product working time standards. Background Art

[0002] At present, production efficiency is the core competitiveness of the sewing products manufacturing industry. Standard working hours are the basis for production scheduling and verification of production progress. Timely mastering the production data of the workshop is very important for eliminating obstacles and correcting plans. The production working hours of workers are the only factor used to calculate production efficiency. Abnormal working hours directly affect the accuracy of production efficiency calculation and are also an important indicator affecting enterprise costs.

[0003] Usually, the reasons for abnormal working hours include: abnormal material supply, abnormal equipment failure, abnormal quality and other abnormalities. At present, after abnormal working hours occur, they are judged and handled by reporting them step by step. This process is not only cumbersome, but also wastes the working hours of operators. Timely handling of abnormal working hours can quickly solve the bottlenecks encountered in production activities, reduce waiting and waste of related resources, and thus improve economic benefits. Summary of the invention

[0004] The main purpose of the present invention is to provide a method, device and storage medium for handling abnormal working hours of sewing products, so as to reduce the time for handling abnormal working hours, reduce waiting and waste of related resources, and improve efficiency.

[0005] To achieve the above object, the present invention provides a method for handling abnormalities in the working time standard of sewing products, comprising the following steps:

[0006] Collecting sewing product related data in real time, wherein the sewing product related data at least includes: various business data of sewing processing, environmental data and worker on-the-job data;

[0007] Processing the collected data related to sewing products;

[0008] The processed data is input into a pre-built sewing product labor time anomaly detection model for detection;

[0009] When the current sewing product working hours are detected to be abnormal, the abnormal information is reported and processed in a preset manner according to a predetermined process.

[0010] The step of reporting the abnormal information in a preset manner according to a predetermined process includes:

[0011] The exception information will be fed back to the scene management personnel in accordance with the predetermined process through one or more methods such as lighting, voice, mobile phone text messages, and electronic display screens. The time consumed in exception handling will be monitored through the background, and the exceptions that are not handled within the time limit will be reported step by step. For exceptions that require cross-departmental collaboration, the relevant personnel will be notified through group messaging. Each exception will be managed as a work order, and the frequency and processing cycle of various exceptions will be summarized regularly.

[0012] The processing of the collected sewing product related data includes:

[0013] Perform data cleaning and data integration on the collected sewing product related data.

[0014] The step of inputting the processed data into a pre-built sewing product work time anomaly detection model for detection includes:

[0015] Performing model operations on the processed data, wherein the model operations include: data reduction, data conversion and model operations;

[0016] The data after model operation is input into the pre-built sewing product labor time cross-validation model to perform redundancy and correlation analysis, data conflict processing, data transformation and discretization processing to obtain the processing results;

[0017] The processing result is input into a pre-built sewing product working time abnormality judgment model to obtain and output the current sewing product working time abnormality conclusion, and the abnormality judgment model at least includes: a work idleness abnormality judgment model, an equipment failure judgment model, a quality abnormality judgment model, and a discipline abnormality judgment model.

[0018] Wherein, before the step of inputting the processed data into a pre-built sewing product work time abnormality detection model for detection, the method further includes:

[0019] Sample data was collected to build a cross-validation model for sewing product labor hours and an abnormality judgment model.

[0020] The steps of collecting sample data to build a sewing product working time cross-validation model and an abnormality judgment model include:

[0021] Collecting a data source of sewing product samples, wherein the data source of sewing product samples includes at least: various business data, environmental data and worker on-the-job data of sewing processing;

[0022] The collected sewing product sample data source is cleaned to obtain feature data; the cleaned feature data is processed to generate training data. The processing process includes sample sampling, sample weight adjustment, outlier removal, feature normalization, feature change, and feature combination;

[0023] Based on the generated training data, the preset algorithm is used for model training to build a sewing product working time abnormality judgment model, and the training data is cross-validated with data from other sources to obtain a cross-validation model.

[0024] The steps of training the model based on the generated training data using a preset algorithm to construct a sewing product working time abnormality judgment model, and cross-validating the training data with data from other sources to obtain a cross-validation model include:

[0025] Based on the generated training data, the abnormal working time features are selected, and the static features are selected as the A-type features;

[0026] Feature construction: statistical analysis is performed on the basis of the A-type features to obtain the B-type relationship features;

[0027] The generated training data is analyzed and modeled according to processing speed, qualified rate, non-standard time, time series temperature and humidity, brightness, equipment failure record, support record and category. The weights are defined by presetting the influence of each factor on the standard working hours. The data is then compared with the counting data from other sources. After ensuring that the consistency meets the requirements, it is inserted into the model for calculation to obtain a cross-validation model.

[0028] A neural network model is constructed, and the selected class A features and class B relationship features are passed into the neural network model as input. The model is trained in combination with the corresponding training data to obtain a sewing product working time abnormality judgment model.

[0029] The data cleaning includes: completing incomplete data, correcting erroneous data, and filtering out duplicate data.

[0030] The present invention also proposes an exception handling device for sewing product working time standards, wherein the main board of the device is integrated with an exception handling program for sewing product working time standards, and the main board is connected to an industrial sewing machine or other auxiliary equipment, and the main board is used to collect sewing product related data in real time, and the sewing product related data at least includes: various business data of sewing processing, environmental data and worker on-the-job data; processing the collected sewing product related data; inputting the processed data into a pre-constructed sewing product working time anomaly detection model for detection; when the current sewing product working time anomaly is detected, the abnormal information is reported and processed in a preset manner according to a predetermined process.

[0031] The present invention also proposes a computer-readable storage medium, on which is stored an exception handling program for the labor time standard for sewing products. When the exception handling program for the labor time standard for sewing products is executed by a processor, the steps of the exception handling method for the labor time standard for sewing products as described above are implemented.

[0032] The exception handling method, device and storage medium for the standard of sewing product working hours proposed in the embodiment of the present invention are connected to the industrial sewing machine or other auxiliary equipment through a special mainboard, and various business data, environmental data and worker on-the-job data of the processing are collected in real time; the data cleaning, data integration, data reduction and data conversion are applied and stored in the local database or cloud database after pre-processing, and the existing data model is used to invert the production site situation, and the standard working hours are compared to judge the abnormal category, and the abnormal information is fed back to the scene management personnel in the first time through the methods such as lighting, voice, mobile phone text messages, and electronic display screens according to the predetermined process, and the background monitors the time consumed by the abnormal handling at the same time, and the abnormalities that are not handled within the time limit will be reported step by step; for the abnormalities that need to be handled by cross-departmental collaboration, the relevant personnel are notified by group sending, and a single abnormality is managed as a work order, and the frequency and processing cycle of various abnormalities are regularly summarized. Compared with the prior art, the scheme of the present invention reduces the time for handling abnormal working hours, reduces the waiting and waste of related resources, and improves work efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] Figure 1 It is a flow chart of the abnormality handling method for the working time standard of sewing products of the present invention;

[0034] Figure 2 This is a schematic diagram of the overall hardware architecture of the solution of the present invention;

[0035] Figure 3 and is a schematic diagram of the overall processing flow of the solution of the present invention;

[0036] Figure 4 Schematic diagram of the main factors that need to be considered for formulating accurate working hours for the scheme of the present invention;

[0037] Figure 5 A schematic diagram of a neural network model is constructed for the present invention. DETAILED DESCRIPTION

[0038] It should be understood that the specific embodiments described herein are only used to explain the present invention, and are not used to limit the present invention.

[0039] Reference Figure 1 , Figure 1 The figure is a flow chart of the abnormality handling method for the working time standard of sewing products according to the present invention.

[0040] The method for handling abnormalities in the working time standard for sewing products comprises the following steps:

[0041] S1, real-time collection of sewing product related data, wherein the sewing product related data at least includes: various business data of sewing processing, environmental data and worker on-the-job data;

[0042] S2, processing the collected data related to sewing products;

[0043] S3, inputting the processed data into a pre-built sewing product labor time anomaly detection model for detection;

[0044] S4, when the working hours of the current sewing product are detected to be abnormal, the abnormal information is reported and processed in a preset manner according to a predetermined process.

[0045] The step of reporting the abnormal information in a preset manner according to a predetermined process includes:

[0046] The exception information will be fed back to the scene management personnel in accordance with the predetermined process through one or more methods such as lighting, voice, mobile phone text messages, and electronic display screens. The time consumed in exception handling will be monitored through the background, and the exceptions that are not handled within the time limit will be reported step by step. For exceptions that require cross-departmental collaboration, the relevant personnel will be notified through group messaging. Each exception will be managed as a work order, and the frequency and processing cycle of various exceptions will be summarized regularly.

[0047] The processing of the collected sewing product related data includes:

[0048] Perform data cleaning and data integration on the collected sewing product related data.

[0049] The step of inputting the processed data into a pre-built sewing product work time anomaly detection model for detection includes:

[0050] Performing model operations on the processed data, wherein the model operations include: data reduction, data conversion and model operations;

[0051] The data after model operation is input into the pre-built sewing product labor time cross-validation model to perform redundancy and correlation analysis, data conflict processing, data transformation and discretization processing to obtain the processing results;

[0052] The processing result is input into a pre-built sewing product working time abnormality judgment model to obtain and output the current sewing product working time abnormality conclusion, and the abnormality judgment model at least includes: a work idleness abnormality judgment model, an equipment failure judgment model, a quality abnormality judgment model, and a discipline abnormality judgment model.

[0053] Wherein, before the step of inputting the processed data into a pre-built sewing product work time abnormality detection model for detection, the method further includes:

[0054] Sample data was collected to build a cross-validation model for sewing product labor hours and an abnormality judgment model.

[0055] The steps of collecting sample data to build a sewing product working time cross-validation model and an abnormality judgment model include:

[0056] Collecting a data source of sewing product samples, wherein the data source of sewing product samples includes at least: various business data, environmental data and worker on-the-job data of sewing processing;

[0057] The collected sewing product sample data source is cleaned to obtain feature data; the cleaned feature data is processed to generate training data. The processing process includes sample sampling, sample weight adjustment, outlier removal, feature normalization, feature change, and feature combination;

[0058] Based on the generated training data, the preset algorithm is used for model training to build a sewing product working time abnormality judgment model, and the training data is cross-validated with data from other sources to obtain a cross-validation model.

[0059] The steps of training the model based on the generated training data using a preset algorithm to construct a sewing product working time abnormality judgment model, and cross-validating the training data with data from other sources to obtain a cross-validation model include:

[0060] Based on the generated training data, the abnormal working time features are selected, and the static features are selected as the A-type features;

[0061] Feature construction: statistical analysis is performed on the basis of the A-type features to obtain the B-type relationship features;

[0062] The generated training data is analyzed and modeled according to processing speed, qualified rate, non-standard time, time series temperature and humidity, brightness, equipment failure record, support record and category. The weights are defined by presetting the influence of each factor on the standard working hours. The data is then compared with the counting data from other sources. After ensuring that the consistency meets the requirements, it is inserted into the model for calculation to obtain a cross-validation model.

[0063] A neural network model is constructed, and the selected class A features and class B relationship features are passed into the neural network model as input. The model is trained in combination with the corresponding training data to obtain a sewing product working time abnormality judgment model.

[0064] In the above scheme, the data cleaning includes: completing incomplete data, correcting erroneous data, and filtering out duplicate data.

[0065] The technical solution of the embodiment of the present invention is described in detail below.

[0066] The method for handling exceptions in the standard of sewing product working hours of the present invention can be integrated in a special motherboard in the form of software, and the motherboard is connected to an industrial sewing machine or other auxiliary equipment. Various business data, environmental data and worker on-the-job data for real-time collection and processing; data cleaning, data integration, data reduction and data conversion are applied and stored in a local database or a cloud database after pre-processing; the existing data model is used to invert the production site situation, and the standard working hours are compared to determine the abnormal category; the abnormal information is fed back to the scene management personnel in the first time according to the predetermined process through lighting, voice, mobile phone text messages, electronic display screens and other methods; the background monitors the time consumed by the abnormal processing at the same time, and the abnormalities that are not processed within the time limit will be reported step by step; for the abnormalities that need to be processed by cross-departmental collaboration, the relevant personnel are notified by group sending, and a single abnormality is managed as a work order, and the frequency and processing cycle of various abnormalities are regularly summarized.

[0067] The hardware structure involved in the solution of the present invention is as follows Figure 2 shown.

[0068] Among them, the core motherboard is connected with the industrial sewing machine or other auxiliary equipment to collect and process various business data, environmental data and worker on-the-job data in real time; collect and process various business data, environmental data and worker on-the-job data in real time; apply data cleaning, data integration, data reduction and data conversion and other content pre-processing and store them in the local database or cloud database, apply the existing data model to invert the production site situation, compare the standard working hours to judge the abnormal category, and feedback the abnormal information to the scene management personnel in the first time through the methods such as lighting, voice, mobile phone text messages, electronic display screens, etc. according to the predetermined process. The background also monitors the time consumed by the abnormal processing, and the abnormalities that are not processed within the time limit will be reported step by step; for the abnormalities that need to be processed by cross-departmental collaboration, the relevant personnel are notified by group sending, and a single abnormality is managed as a work order, and the frequency and processing cycle of various abnormalities are regularly summarized. Compared with the prior art, the scheme of the present invention reduces the time for handling abnormal working hours, reduces the waiting and waste of related resources, and improves work efficiency.

[0069] Specifically, the core motherboard is equipped with big data computing module, artificial intelligence module, data acquisition module, off-site sensing module, environmental sensing module, etc., and is designed with USB, SD card, UART, SPI, SDIO, I2C, LED PWM, Motor PWM, I2S, IR, pulse counter, GPIO, ADC, DAC, 802.11n 2.4GHz WIFI, Bluetooth and other interfaces.

[0070] Among them, data sources and collection methods include:

[0071] 1. Material feeding data, number of pieces produced, cycle data, machine working status data

[0072] By connecting with industrial sewing machines (ordinary lockstitch machines, medium- and high-speed lockstitch machines, high-speed lockstitch machines, semi-automatic lockstitch machines, fully automatic lockstitch machines, overlock machines, buttonhole machines; bartacking machines, button sewing machines, blind stitch machines, double needle machines, automatic bag opening machines, computer embroidery machines; zigzag sewing machines; crescent machines; computer pattern machines, etc.), real-time capture of production beat data, such as real-time power, action start time, action end time, etc. Face recognition, RFID, barcodes, keyboards and physical buttons are used as auxiliary input methods for identity recognition and work order input.

[0073] 2. Off-duty data

[0074] The off-post sensing module is composed of infrared sensors, sonar sensors and CCD sensors. It determines whether there is anyone in the work area by integrating the data of infrared sensors, sonar sensors and CCD sensors, and determines that the person's actions are in normal working state.

[0075] 3. Natural environment data and human interference data

[0076] The environmental control and measurement module consists of a temperature sensor, a humidity sensor, a PM2.5 sensor, a noise sensor, a light sensor, a carbon dioxide sensor, a light sensor and a meteorological sensor, and is used to collect environmental data inside and outside the workshop.

[0077] Human interference data is calculated to identify conversations not related to work, sounds from video programs, etc.

[0078] 4. Work order data

[0079] The work order information module receives work order instructions from the cloud / mobile media and divides the work order information into action instructions and work order information. The action instructions are compiled into machine-executable action commands to coordinate the actions of each action unit; the work order information is converted into basic work order information and target quantity, beat requirements, etc.

[0080] Work order information includes: model number, date, target quantity, hourly production requirement (beat requirement)

[0081] 5. Support data

[0082] Record events that require exception processing, including event category (pending for materials, fault repair, etc.), request time, processing completion time, person in charge, etc.

[0083] 6. Person counting data

[0084] Obtain worker output reporting records from MIS or MES.

[0085] The overall processing flow is as follows Figure 3 As shown:

[0086] 1. Data collection. After collecting this information, data extraction design can be carried out.

[0087] 1. Core motherboard data source processing method

[0088] The work order data, operation data, environment data, off-duty data and support data collected by the core motherboard are stored in the DBMS (SQLServer, Oracle). By establishing a direct link relationship between the database systems, you can write a Select statement to directly access them.

[0089] Specifically, the acquisition module uses USB, SD card, UART, SPI, SDIO, I2C, LED PWM, Motor PWM, I2S, IR, pulse counter, GPIO, ADC, DAC, 802.11n 2.4GHz WIFI, Bluetooth and other interfaces to connect with electrical industrial equipment (such as sewing machines, machine tools, engraving machines, injection molding machines, etc.) to capture production beat data in real time, such as real-time power, action start time, action end time, etc. Face recognition, RFID, barcode, keyboard and physical keys are used as auxiliary input methods for identity recognition and work order input; the data of infrared sensors, sonar sensors and CCD sensors are used to determine whether there are people in the work area, and the time series data of people's actions are recorded in real time when they are judged as normal working state or off-duty state; the environmental time series data inside and outside the workshop are collected through the environmental module composed of temperature sensors, humidity sensors, PM2.5 sensors, noise sensors, light sensors, carbon dioxide sensors, light sensors and meteorological sensors.

[0090] 2. Data Processing

[0091] Big data preprocessing includes data cleaning, data integration and other processes. Data cleaning technology includes operations such as data inconsistency detection, noise data identification, data filtering and correction. Data integration is the integration of data from acquisition modules, MIS or other system data sources to achieve a centralized and unified database.

[0092] 1. Data cleaning

[0093] (1) Incomplete data: Data with missing information is written into the database only after a value is assigned to it.

[0094] (2) Incorrect data: For example, numeric data is input as full-width numeric characters, a carriage return is placed after a string of data, the date format is incorrect, the date is out of range, etc. This type of data should also be classified, and pre-corrections should be made for problems such as full-width characters and invisible characters before and after the data. For example, the date format should be changed to yyyy-MM-DD hh:mm:ss.

[0095] (3) Duplicate data: Filter out useless duplicate data.

[0096] Model operations include data reduction, data transformation and model operations. Data reduction is to reduce the size of the data set and simplify it without compromising the accuracy of the analysis results. It includes dimensionality reduction, data reduction, data sampling and other technologies. This process is conducive to improving the value density of big data, that is, improving the value of big data storage.

[0097] The tasks of data conversion mainly include inconsistent data conversion, data granularity conversion, and calculation of some business rules.

[0098] (1) Inconsistent data conversion: Unify the encoding of the same type of data in different systems.

[0099] (2) Conversion of data granularity: Aggregate environmental data into granularity, such as high temperature (humidity>30); hot and humid (30>humidity>25); comfortable (25>humidity>15); cold (humidity<15);

[0100] (3) Calculation of rules: Processing time = end time - start time; Maintenance time = end time - start time; Processing quantity = sum (count).

[0101] Data conversion processing includes technologies such as rule-based or metadata-based conversion and model and learning-based conversion. Data unification can be achieved through conversion. This process is conducive to improving the consistency and availability of big data.

[0102] In addition, it also includes: data conflict handling

[0103] (1) Counting data outside the attendance records shall be eliminated.

[0104] (2) Counting data before the support is completed is eliminated.

[0105] (3) Records with a single duration significantly higher than the average duration will be eliminated.

[0106] The big data computing module integrates the spark big data framework and can connect to the local database sqlite and ODBC external data sources. The data model library is constructed through algorithms, and each model is repeatedly evaluated and revised to finally obtain empirical data for various problems. The collected data is applied to each model for calculation, and the results are written into the memory.

[0107] The specific implementation process is as follows Figure 4 As shown:

[0108] 1. Define the characteristics of abnormal working hours, and initially select static characteristics as Class A features, as shown in Table 1 below:

[0109]

[0110]

[0111]

[0112] Table 1

[0113] 2. Feature construction: Based on the A-type features, statistical analysis is performed to obtain the B-type relationship features, as shown in Table 2 below:

[0114]

[0115]

[0116]

[0117]

[0118] Table 2

[0119] 3. After cleaning the data obtained by the equipment (de-duplication, removal of impurities, completion, etc.), establish analysis models according to processing speed, qualified rate, non-standard time, time series temperature and humidity, brightness, equipment failure records, support records and categories, define weights by presetting the impact of various factors on standard working hours, and then compare with the counting data from other sources. After confirming that its consistency meets the requirements, put it into the model for calculation. The specific calculation formula is as follows:

[0120]

[0121] Among them, P(A) is the pre-selected value, which pre-sets the processing speed, qualified rate, non-local time, timing temperature and humidity, brightness, equipment failure record, support record and category, etc., and the degree of influence on the processing speed; P(A|B) is the re-evaluation of the preset value when the above conditions occur; P(B|A) / P(B) is the adjustment factor to make the estimated probability closer to the actual probability.

[0122] 4. Build a neural network model and pass the A-type features and B-type relationship features as input to the model, such as Figure 5 As shown in X1…X n ; The connecting lines are the calculated weight values, ∫0…∫ n is a neuron; ∫ is a bias term; ∫ a …∫ z is the bias term; Y is the prediction result;

[0123] ∫ 0＝ Calculate the number of people who work without effort: X1 = the first piece start time of the shift - the attendance time, X2 = the number of equipment failure reports during the attendance time, X3 = the frequency of defective equipment during the attendance time, and X4 = the number of non-standard times during the attendance time. Name the line connecting X1…X4 to ∫0 as W 01 …W 04 , name the line connecting X1…X4 and ∫1 W 11 …W 14 , ∫ is the bias term b, that is, the calculation model of working without effort is as follows:

[0124]

[0125] Model training is an iterative and fitting process, in which there is a parameter e representing the error, ∑e i 2 =Total global error. Next, we apply the gradient descent method to make W and b in the model approach the optimal position.

[0126]

[0127] In addition, an embodiment of the present invention also proposes an exception handling device for sewing product working time standards, wherein a main board of the device is integrated with an exception handling program for sewing product working time standards, and the main board is connected to an industrial sewing machine or other auxiliary equipment, and the main board is used to collect sewing product related data in real time, and the sewing product related data at least includes: various business data of sewing processing, environmental data and worker on-the-job data; processing the collected sewing product related data; inputting the processed data into a pre-constructed sewing product working time anomaly detection model for detection; when the current sewing product working time is detected to be abnormal, the abnormal information is reported and processed in a preset manner according to a predetermined process.

[0128] In addition, an embodiment of the present invention also proposes a computer-readable storage medium, on which is stored an exception handling program for the labor time standard for sewing products. When the exception handling program for the labor time standard for sewing products is executed by a processor, the steps of the exception handling method for the labor time standard for sewing products as described above are implemented.

[0129] For the principle of this embodiment, please refer to the above embodiments, which will not be described in detail here.

[0130] The exception handling method, device and storage medium for the standard of sewing product working hours proposed in the embodiment of the present invention are connected to the industrial sewing machine or other auxiliary equipment through a special mainboard, and various business data, environmental data and worker on-the-job data of the processing are collected in real time; the data cleaning, data integration, data reduction and data conversion are applied and stored in the local database or cloud database after pre-processing, and the existing data model is used to invert the production site situation, and the standard working hours are compared to judge the abnormal category, and the abnormal information is fed back to the scene management personnel in the first time through the methods such as lighting, voice, mobile phone text messages, and electronic display screens according to the predetermined process, and the background monitors the time consumed by the abnormal handling at the same time, and the abnormalities that are not handled within the time limit will be reported step by step; for the abnormalities that need to be handled by cross-departmental collaboration, the relevant personnel are notified by group sending, and a single abnormality is managed as a work order, and the frequency and processing cycle of various abnormalities are regularly summarized. Compared with the prior art, the scheme of the present invention reduces the time for handling abnormal working hours, reduces the waiting and waste of related resources, and improves work efficiency.

[0131] The above are only preferred embodiments of the present invention, and are not intended to limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made using the contents of the present invention specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. A method for handling abnormalities in the work time standard for sewing products, characterized in that: The following steps are involved: Collecting sewing product related data in real time, wherein the sewing product related data at least includes: various business data of sewing processing, environmental data and worker on-the-job data; Processing the collected data related to sewing products; The processed data is input into a pre-built sewing product labor time anomaly detection model for detection; When the current sewing product working hours are detected to be abnormal, the abnormal information is reported and processed in a preset manner according to a predetermined process; The step of inputting the processed data into a pre-built sewing product work time anomaly detection model for detection comprises: Performing model operations on the processed data, wherein the model operations include: data reduction, data conversion and model operations; The data after model operation is input into the pre-built sewing product labor time cross-validation model to perform redundancy and correlation analysis, data conflict processing, data transformation and discretization processing to obtain the processing results; The processing result is input into a pre-built sewing product working time abnormality judgment model to obtain and output the current sewing product working time abnormality conclusion, and the abnormality judgment model at least includes: a work idleness abnormality judgment model, an equipment failure judgment model, a quality abnormality judgment model, and a discipline abnormality judgment model.

2. The abnormality handling method for the work time standard of sewing products according to claim 1, characterized in that: The step of reporting the abnormal information in a preset manner according to a predetermined process includes: The exception information will be fed back to the scene management personnel in accordance with the predetermined process through one or more methods such as lighting, voice, mobile phone text messages, and electronic display screens. The time consumed in exception handling will be monitored through the background, and the exceptions that are not handled within the time limit will be reported step by step. For exceptions that require cross-departmental collaboration, the relevant personnel will be notified through group messaging. Each exception will be managed as a work order, and the frequency and processing cycle of various exceptions will be summarized regularly.

3. The abnormality handling method for the work time standard of sewing products according to claim 1, characterized in that: The processing of the collected sewing product related data includes: Perform data cleaning and data integration on the collected sewing product related data.

4. The abnormality handling method for the work time standard of sewing products according to claim 1, characterized in that: Before the step of inputting the processed data into a pre-built sewing product work time abnormality detection model for detection, the following steps are also included: Sample data was collected to build a cross-validation model for sewing product labor hours and an abnormality judgment model.

5. The abnormality handling method for the work time standard of sewing products according to claim 4, characterized in that: The steps of collecting sample data to build a sewing product working time cross-validation model and an abnormality judgment model include: Collecting a data source of sewing product samples, wherein the data source of sewing product samples includes at least: various business data, environmental data and worker on-the-job data of sewing processing; The collected sewing product sample data source is cleaned to obtain feature data; the cleaned feature data is processed to generate training data. The processing process includes sample sampling, sample weight adjustment, outlier removal, feature normalization, feature change, and feature combination; Based on the generated training data, the preset algorithm is used for model training to build a sewing product working time abnormality judgment model, and the training data is cross-validated with data from other sources to obtain a cross-validation model.

6. The abnormality handling method for the work time standard of sewing products according to claim 5, characterized in that: The steps of training the model based on the generated training data using a preset algorithm to construct a sewing product working time abnormality judgment model, and cross-validating the training data with data from other sources to obtain a cross-validation model include: Based on the generated training data, the abnormal working time features are selected, and the static features are selected as the A-type features; Feature construction: statistical analysis is performed on the basis of the A-type features to obtain the B-type relationship features; The generated training data is analyzed and modeled according to processing speed, qualified rate, non-standard time, time series temperature and humidity, brightness, equipment failure record, support record and category. The weights are defined by presetting the influence of each factor on the standard working hours. The data is then compared with the counting data from other sources. After ensuring that the consistency meets the requirements, it is inserted into the model for calculation to obtain a cross-validation model. A neural network model is constructed, and the selected class A features and class B relationship features are passed into the neural network model as input. The model is trained in combination with the corresponding training data to obtain a sewing product working time abnormality judgment model.

7. The abnormality handling method for the work time standard of sewing products according to claim 3 or 5, characterized in that: The data cleaning includes: completing incomplete data, correcting erroneous data, and filtering out duplicate data.

8. A device for handling abnormalities in the working time standard of sewing products, characterized in that: The main board of the device is integrated with an abnormality processing program for the standard of sewing product working hours. The main board is connected to an industrial sewing machine or other auxiliary equipment. The main board is used to collect sewing product related data in real time, and the sewing product related data at least includes: various business data of sewing processing, environmental data and worker on-the-job data; process the collected sewing product related data; input the processed data into a pre-constructed sewing product working hour abnormality detection model for detection; when the current sewing product working hour abnormality is detected, the abnormal information is reported and processed in a preset manner according to a predetermined process; The main board is also used to perform model operations on the processed data, and the model operations include: data reduction, data conversion and model operations; inputting the data after model operation into a pre-constructed sewing product working time cross-validation model to perform redundancy and correlation analysis, data conflict processing, data transformation and discretization processing to obtain processing results; inputting the processing results into a pre-constructed sewing product working time abnormality judgment model to obtain and output the current sewing product working time abnormality conclusion, and the abnormality judgment model at least includes: a work slack abnormality judgment model, an equipment failure judgment model, a quality abnormality judgment model, and a discipline abnormality judgment model.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores an exception handling program for the labor time standard for sewing products, and when the exception handling program for the labor time standard for sewing products is executed by the processor, the steps of the exception handling method for the labor time standard for sewing products as described in any one of claims 1-7 are implemented.

Images