MES-based industrial production data intelligent analysis and management method

Abstract

The invention discloses an MES-based industrial production data intelligent analysis and management method, and relates to the technical field of industrial automation and information, and the method comprises the following specific steps: S100, collecting data in an industrial production process, S200, determining each data feature and carrying out slicing operation, S300, screening key features to form a dynamic slice, S400, carrying out comprehensive analysis on the dynamic slice, and S400, carrying out comprehensive analysis on the dynamic slice. According to the method, the analysis efficiency and the pertinence of the production data are remarkably improved through the unique dynamic data slicing technology, and in the complex industrial production process, the analysis efficiency and the pertinence of the production data are greatly improved. The key data subset can be quickly focused according to the production task priority, the equipment key degree and the real-time data fluctuation characteristics, invalid analysis of mass irrelevant data in a traditional method is avoided, and the overall production benefit and the product quality are improved.

Description

Technical Field

[0001] The present invention relates to the field of industrial automation and information technology, and in particular to an intelligent analysis and management method for industrial production data based on MES. Background Art

[0002] In today's highly industrialized era, manufacturing companies are facing increasingly fierce market competition and the challenges of growing customer demand diversity. The complexity and scale of industrial production processes continue to rise, generating massive amounts of production data. These data cover every aspect from raw material procurement, equipment operating status, production process parameters to product quality testing. Effective management and analysis of these production data is crucial for companies to optimize production processes, improve product quality, reduce costs and enhance market competitiveness.

[0003] However, existing industrial production data management methods have significant defects when facing complex production scenarios. In terms of data slicing analysis, traditional methods use a fixed, preset slicing method, which cannot be flexibly adjusted according to the real-time priority of production tasks, the actual criticality of equipment, and the dynamic fluctuation characteristics of data. This static slicing method causes a large amount of irrelevant data to be included in the analysis scope, while the truly critical data is submerged, greatly reducing the efficiency and accuracy of the analysis. In terms of production decision support, due to the lack of accurate data analysis, it is difficult for enterprises to quickly locate the root causes of problems in the production process and make effective decision adjustments in time, which often leads to production delays, waste of resources and unstable quality, seriously restricting the development potential and profitability of enterprises.

[0004] In summary, the existing industrial production data management technology is difficult to meet the urgent needs of modern manufacturing enterprises for efficient and accurate production management. Therefore, there is an urgent need for an intelligent analysis and management method for industrial production data based on MES, which can dynamically slice and deeply analyze production data, provide enterprises with timely and accurate decision-making basis, and enhance the competitiveness and sustainable development capabilities of enterprises in a complex and changing market environment. Summary of the invention

[0005] The purpose of the present invention is to make up for the shortcomings of the prior art and to provide an intelligent analysis and management method for industrial production data based on MES. It can use dynamic data slicing technology in dynamic data analysis for complex production processes. It can automatically slice production data dynamically according to the priority of production tasks, the criticality of equipment and the fluctuation characteristics of real-time data, and focus on analyzing data subsets in specific time periods and specific production links, so as to quickly and accurately locate the key factors affecting production efficiency and product quality, and significantly improve the analysis efficiency and pertinence.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions: an intelligent analysis and management method for industrial production data based on MES, the specific steps of the method are: S100, collecting data from the production process in real time through the MES system, cleaning and integrating the data, and forming a data information database, wherein the data information database includes production task information, equipment status information, and quality inspection information; The production task information includes the urgency of the delivery time , Order Revenue Coefficient , the degree of close cooperation with customers , by weighted summing the parameters of different production tasks, the priority value of the production task is obtained , as the basis for data slicing operations; The device status information includes the frequency of failure , Maintenance cost ratio , Equipment operation stability indicators , the criticality index of the equipment is calculated by weighted calculation based on the parameters of the equipment status information , and use this indicator as the basis for data slicing operations; The quality detection information includes the coefficient of variation of quality data , Defective product growth rate , Critical Quality Characteristics Deviation By weighting and summing the various data fluctuation indicators of quality inspection information, a comprehensive evaluation value of the real-time data fluctuation characteristics is obtained. , used to identify and filter out data intervals with abnormal fluctuations; S200, automatically and dynamically slices production data according to the priority of production tasks, criticality indicators of equipment and real-time data fluctuation characteristics; The dynamic slicing operation includes time slicing, production link slicing and attribute slicing; S300, filter out high-priority production tasks, key equipment, and data within the abnormal data fluctuation range, that is, focus on analyzing a specific time period Specific production links and specific data attributes Data subsets that form dynamic data slices ; S400, use data analysis methods to conduct in-depth analysis of dynamic data slices and locate key factors; S500, the analysis results are sent to the MES system in a visual manner and stored in the data information database to provide a basis for production decisions.

[0007] Furthermore, in the S100 production task information: The urgency of the delivery time The remaining delivery days of the current production task Standard delivery days is measured by the difference between ; The order profit coefficient The expected revenue from the current production task The average revenue of the enterprise's production tasks during the same period The ratio is given, that is ; The degree of close cooperation with the client Based on the number of past cooperation with customers Total cooperation amount Comprehensive evaluation of the product of Calculate the priority of production tasks based on production task information for: ,in, , , They are the weight coefficients of delivery time urgency, order revenue coefficient, and customer cooperation closeness, and they meet .

[0008] Furthermore, in the S100 device status information: The frequency of the fault Count the number of times a device fails within a specific time period , ; The proportion of maintenance costs By calculating the average cost of a single equipment repair Equipment purchase cost The ratio of ; The equipment operation stability index It is determined by monitoring the fluctuation range of the equipment operating parameters and using the standard deviation of the operating parameters The average value of this parameter The ratio of ; Calculate the criticality index of the equipment based on the equipment status information for: ,in, , , They are the weight coefficients of failure frequency, maintenance cost ratio, and equipment operation stability index, and meet the requirements. .

[0009] Furthermore, in the S100 quality inspection information: The coefficient of variation of the quality data The sample standard deviation of the quality inspection data Divide by the sample mean Get, that is , used to measure the discreteness of quality data; The defective product growth rate Calculate the growth value of the number of defective products in the current time period Compared with the number of defective products in the previous period The ratio of ; The critical quality characteristic deviation Standard values ​​for key quality characteristics of products The actual detection value The difference of ; Calculate real-time data fluctuation characteristics through quality inspection information for: ,in, They are the coefficient of variation of quality data, the growth rate of defective products, and the weight coefficient of the deviation of key quality characteristics, and they meet .

[0010] Furthermore, when the S200 performs the dynamic slicing operation, for the time slice, the start time of the production task is used. and end time Based on the priority of production tasks and comprehensive evaluation of real-time data fluctuation characteristics To determine the specific slicing time range, that is, when Above priority threshold and Greater than the fluctuation threshold When , the time slice range is reduced to ,in and The time offset is dynamically adjusted according to the urgency of the production task and the data fluctuation. The calculation methods are as follows: ,in and is the time adjustment factor.

[0011] Furthermore, when the S200 performs slicing in the production process, according to the criticality index of the equipment Classify each link in the production process and set the key indicators Above criticality threshold The production link where the equipment is located is determined as the key production link, and the data of the key production link is extracted and analyzed. At the same time, for the production link that has a direct upper and lower relationship with the key production link, its data extraction weight is According to the correlation coefficient Adjust the data extraction weight to ,in Extract weights for basic data, The weight adjustment coefficient is used. The key production links and related links are extracted according to the adjusted weights, and the data of non-key links are screened out.

[0012] Furthermore, when performing attribute slicing in S200, an attribute association matrix is ​​constructed based on various attribute data in the production task information, equipment status information, and quality inspection information. , for production task priorities , Equipment criticality index and comprehensive evaluation of real-time data fluctuation characteristics Correlation is above the correlation threshold The attributes are extracted and retained first in the attribute slices. The attribute correlation is calculated using the Pearson correlation coefficient method. and comprehensive evaluation indicators , and its Pearson correlation coefficient The calculation formula is: ,in, For attributes In the The values ​​in the data samples are For attributes The sample mean of The comprehensive evaluation value of the corresponding production task priority, equipment criticality index and real-time data fluctuation characteristics is in the first The values ​​in the data samples are is its sample mean, is the number of data samples.

[0013] Furthermore, the S300 selects high-priority production tasks, corresponding key equipment, and data in the abnormal data fluctuation range from the data after dynamic slicing, and further integrates the selected data to form dynamic data slices for analysis, that is, the time slices are in the time interval The data within the production link is sliced ​​to determine the production link set , including key links and associated links, attribute slices determine attribute sets , the data subset of the dynamic data slice Indicates the time interval The first At a certain time point, the first The corresponding production link The data value of an attribute.

[0014] Furthermore, the S400 uses principal component analysis (PCA) of the data analysis method to analyze the data subset. Standardized processing to obtain , the standardization process is for each attribute , calculate its mean and standard deviation ,but ; Construct the covariance matrix Cov, whose elements in the variance matrix for: ,in is the number of data points, Is an attribute In the production stage The mean on Is an attribute In the production stage The mean on ; Perform eigenvalue decomposition on the covariance matrix to obtain the eigenvalue and the corresponding eigenvector , , is the number of attributes, sort the eigenvectors in descending order of eigenvalues, and select the first The eigenvectors form the projection matrix ,pass Get the reduced dimension dataset ; The dataset Consider it as a transaction database, where each row represents a transaction and each column represents an attribute. Determine the minimum support threshold and the minimum confidence threshold , scan the dataset , count the number of occurrences of each attribute, and select the frequent itemsets that meet the minimum support threshold Connect and generate candidate item sets , scan the dataset again ,statistics The support of each item in the set is frequently obtained , repeatedly search for frequent item sets, generate high-order candidate sets and frequent sets, until no new frequent sets can be generated; For the final , calculate its confidence ,in Representing Item Sets The support degree of is the proportion of the number of occurrences to the total number of transactions, so as to extract the association rules whose confidence is greater than the minimum confidence threshold.

[0015] Compared with the existing technology, this MES-based industrial production data intelligent analysis and management method has the following beneficial effects: 1. The present invention significantly improves the analysis efficiency and pertinence of production data through unique dynamic data slicing technology. In complex industrial production processes, it can quickly focus on key data subsets based on production task priority, equipment criticality and real-time data fluctuation characteristics, avoiding the ineffective analysis of massive irrelevant data by traditional methods, and improving overall production efficiency and product quality.

[0016] 2. The present invention conducts in-depth analysis of dynamic data slices to discover the potential patterns and correlations in production data, so that enterprise managers can clearly understand the overall picture of the production process and the mutual influence of various factors. This data-driven decision-making method effectively reduces the uncertainty and risk in the production process, enabling enterprises to respond to market changes more flexibly and optimize resource allocation.

[0017] Other advantages, objectives and features of the present invention will be set forth in part in the following description and, in part, will be apparent to those skilled in the art based on an examination of the following or may be taught from the practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention, and for ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0019] Figure 1 It is an operation diagram of an intelligent analysis and management method of industrial production data based on MES; Figure 2 This is an implementation flow chart of Example 2. DETAILED DESCRIPTION

[0020] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the specific implementation mode, structure, characteristics and effects of the present invention are described in detail below in combination with the accompanying drawings and preferred embodiments.

[0021] Embodiment 1 like Figure 1As shown, this embodiment shows in detail the application process of an intelligent analysis and management method for industrial production data based on MES, which collects production data in real time, cleans and integrates data to build a data information database, determines relevant indicators and weights based on production task information, equipment status information and quality inspection information, and then performs dynamic slicing operations. After screening out dynamic data slices, data analysis methods are used to deeply explore key factors, and the results are visualized to assist production decisions.

[0022] The MES system is used to comprehensively collect various data in the production process, including production task information, equipment status information and quality inspection information. Among the production task information, the urgency of delivery time is also included. Remaining delivery days through current production tasks Standard delivery days The difference is that , this difference directly reflects the time urgency of the task. The smaller the difference, the more urgent the task; order revenue coefficient The expected revenue from the current production task Divided by the average revenue of the enterprise's production tasks during the same period Get, that is The larger the coefficient, the higher the potential contribution of this task to the company's revenue; the closer the customer cooperation Based on the number of past cooperation with customers Total cooperation amount The more times of cooperation and the larger the amount, the higher the degree of closeness; according to the formula (in ) Calculate the priority of production tasks , weight coefficient , , Represents the importance that the enterprise attaches to each factor. It accurately quantifies the task priority through weighted summation, providing a basis for subsequent data screening. In terms of equipment status information, the frequency of faults Count the number of times a device fails within a specific time period ,Right now , the more failures there are, the worse the equipment stability is; the maintenance cost accounts for Average cost of equipment maintenance Equipment purchase cost Divide them and we get The higher the ratio, the heavier the equipment maintenance cost burden; equipment operation stability index Determined by monitoring the fluctuation range of the operating parameters of the equipment, the standard deviation of the operating parameters is used With the average The ratio calculation is The larger the ratio, the lower the running stability. Calculate the criticality index of the equipment , weight coefficient , , Represents the importance of the equipment in the production process and the company's maintenance strategy, so as to accurately determine the criticality of the equipment and guide the direction of data screening; in quality inspection information, the coefficient of variation of quality data The sample standard deviation of the quality inspection data Divide by the sample mean Calculate, that is It effectively measures the degree of dispersion of quality data. The larger the coefficient of variation, the more drastic the quality fluctuation. The growth rate of defective products is the growth value of the number of defective products in the current time period Compared with the number of defective products in the previous period The ratio of The higher the growth rate, the worse the production quality; the deviation of key quality characteristics Standard values ​​for key quality characteristics of products The actual detection value The difference of The larger the difference, the farther the key quality characteristics of the product deviate from the standard. ) to obtain the comprehensive evaluation value of the real-time data fluctuation characteristics , weight system , , The importance of the quality factors that represent their impact on production is determined, and the comprehensive evaluation value is used to accurately lock in the abnormal range of data fluctuations and screen key data. The collected raw data is cleaned to remove noise and erroneous data, and is integrated and stored in the data information database to provide basic data for subsequent analysis.

[0023] The collected basic data is used to perform dynamic slicing operations, which include time slicing, production link slicing, and attribute slicing. For time slicing, the production task start time is used as the and end time As the cornerstone, combined with production task priorities and comprehensive evaluation value of real-time data fluctuation characteristics Determine the slice range. Above priority threshold and Greater than the fluctuation threshold When, according to the formula and ( , Calculate the time offset for the time adjustment factor), narrowing the time slice range to By dynamically adjusting time slices, we can accurately focus on key periods and exclude irrelevant time interval data, greatly improving analysis efficiency and pertinence, ensuring that analysis is focused on key time nodes, and slicing production links based on equipment criticality indicators. Classify each link of the production process and Above criticality threshold The link where the equipment is located is defined as the key production link, and its data is extracted for in-depth analysis. For the links that are directly related to the key production links, the correlation coefficient is calculated based on the correlation coefficient. Adjust the data extraction weight, the formula is ( Extract weights for basic data, is the weight adjustment coefficient). This is because the data of key links and closely related links have a significant impact on production. By reasonably allocating weights, we can ensure that key data are fully extracted, avoid wasting analysis resources, focus the analysis on the core production area, and jointly constitute the production link slice data; for attribute slices, we construct an attribute association matrix based on various attribute data in production task information, equipment status information, and quality inspection information. , using the Pearson correlation coefficient method to calculate the attributes and production task priorities , Equipment criticality index and comprehensive evaluation value of real-time data fluctuation characteristics For correlations above the correlation threshold Attributes, priority extraction and retention in attribute slices, Pearson correlation coefficient The calculation formula is ( For attributes In the The values ​​in the data samples are For attributes The sample mean of For the corresponding , , In the The values ​​in the data samples are is its sample mean, is the number of data samples). In this way, we can screen out the attributes that are closely related to the key factors, discard the irrelevant attributes, further focus on the analysis, and lay the foundation for mining the deep laws of the data.

[0024] According to the priority weight value of the production task, the production data corresponding to the high-priority production task is screened out, and the relevant data interval in the time dimension is determined. Combined with the key level of the equipment, the data generated by the key equipment in the above data interval is further focused on. Referring to the real-time data fluctuation characteristics, the screened data is screened again, and the data subset in the data fluctuation abnormal interval is extracted, so as to integrate the time slice (in the time interval Data within), production link slices (a set of determined production links , including key and associated links) and attribute slices (a certain set of attributes As a result, a subset of data is formed for dynamic data slicing , which represents the specific attribute data value of a specific production link within a specific time period, and accurately focuses on the key data of the production process.

[0025] Data subsets for generated dynamic data slices Perform principal component analysis (PCA), first perform standardization, and then for each attribute Calculate the mean and standard deviation , through the formula Get a standardized dataset , eliminate the differences in the dimensions of different attributes, make the data comparable, and construct the covariance matrix Cov, the elements of which are According to the formula Calculate to reflect the linear relationship between attributes, then decompose the covariance matrix eigenvalues ​​to obtain the eigenvalues and the eigenvector ( , is the number of attributes), sort by eigenvalue in descending order, select the first The eigenvectors form the projection matrix , and finally Get the reduced dimension dataset , effectively reducing the data dimension, retaining key information, facilitating subsequent analysis and processing, and converting the reduced-dimensional data set Treat it as a transaction database and set a minimum support threshold and the minimum confidence threshold , scan the data set to count the number of occurrences of each attribute, and find the frequent itemsets that meet the minimum support threshold , connect to generate candidate item sets , scan again to get the frequent itemsets , repeat the operation until no new frequent itemsets can be generated, and calculate the confidence of the final frequent itemsets, such as , extract confidence-level super-threshold association rules, and mine potential associations between attributes, such as analyzing the correlation between abnormal data fluctuations of key equipment during the execution of high-priority production tasks and product quality defects, so as to quickly and accurately locate the key factors affecting production efficiency and product quality.

[0026] The key factors and correlations obtained from the analysis are sent to the MES system in an intuitive visual form, allowing production managers to clearly understand the overall production picture and accurately grasp the interactions between various factors. Based on the visualization results, managers can timely adjust production plans, optimize equipment maintenance strategies, and improve process parameters to achieve intelligent control and continuous optimization of the production process, thereby enhancing the company's market competitiveness.

[0027] Embodiment 2 like Figure 2 As shown, based on the first embodiment, this embodiment describes in detail a specific implementation process of an industrial production data intelligent analysis and management method based on MES: First, we enter the data collection and integration stage (S100). In an industrial production environment, the MES system undertakes the key task of data collection. The collected data includes production task information, equipment status information, and quality inspection information. These multi-source heterogeneous data are integrated and pre-processed to ensure the accuracy and completeness of the data and stored in the data information database.

[0028] Then, enter the dynamic data slice definition stage (S200), determine the priority evaluation index system of production tasks according to the characteristics of the production process and the needs of enterprise production management, comprehensively consider factors such as order delivery time, product profit, and customer importance, and assign corresponding priority weight values ​​to each production task; evaluate the criticality of production equipment, and divide the equipment into different critical levels according to factors such as the scope of impact of equipment failures, maintenance costs, and the degree of bottleneck constraints on the production process; monitor the fluctuation characteristics of production data in real time, identify abnormal changes in data, such as sudden changes, periodic fluctuation anomalies, etc., and record the time period of fluctuations.

[0029] Subsequently, the dynamic data slice generation stage (S300) is entered. According to the priority weight value of the production task, the production data corresponding to the high-priority production task is screened out, and its relevant data interval in the time dimension is determined; combined with the critical level of the equipment, the data generated by the key equipment in the above data interval is further focused on; with reference to the real-time data fluctuation characteristics, the filtered data is screened again, and the data subset in the data fluctuation abnormal interval is extracted to form the final dynamic data slice.

[0030] Next, the data subset analysis phase (S400) is entered. For the generated dynamic data slices, statistical analysis methods are used to conduct in-depth analysis on the data subsets of a specific time period and a specific production link, and to explore the potential patterns and correlations behind the data. For example, the correlation between abnormal data fluctuations of key equipment during the execution of high-priority production tasks and product quality defects is analyzed, so as to quickly and accurately locate the key factors affecting production efficiency and product quality.

[0031] Finally, the result presentation and decision support stage (S500) is entered, where the results of data analysis are presented to enterprise managers and production decision makers in an intuitive and easy-to-understand manner, providing them with accurate decision-making basis so that they can take corresponding measures in time to make production adjustments, equipment maintenance, quality improvement and other operations, thereby realizing intelligent management and optimization of industrial production processes.

[0032] In summary, this embodiment comprehensively demonstrates the implementation steps of the intelligent analysis and management method for industrial production data based on MES. By systematically collecting and integrating production data, performing multi-dimensional dynamic slicing operations, forming accurate dynamic data slicing, deeply analyzing the data and locating key factors, and finally visualizing the results to support production decisions, this method helps enterprises fully utilize the value of production data, optimize production processes, improve production efficiency and product quality, reduce costs and risks, and enhance the competitiveness of enterprises in the market.

[0033] The above description is only a preferred embodiment of the present invention and does not limit the present invention in any form. Although the present invention has been disclosed as a preferred embodiment as above, it is not used to limit the present invention. Any technical personnel in this field can make some changes or modify the technical contents disclosed above into equivalent embodiments without departing from the scope of the technical solution of the present invention. However, any brief modifications, equivalent changes and modifications made to the above embodiments based on the technical essence of the present invention without departing from the content of the technical solution of the present invention are still within the scope of the technical solution of the present invention.

Claims

1. A method for intelligent analysis and management of industrial production data based on MES, characterized in that: The specific steps of this method are: S100, collecting data from the production process in real time through the MES system, cleaning and integrating the data, and forming a data information database, wherein the data information database includes production task information, equipment status information, and quality inspection information; The production task information includes the urgency of the delivery time , Order Revenue Coefficient , the degree of close cooperation with customers , by weighted summing the parameters of different production tasks, the priority value of the production task is obtained , as the basis for data slicing operations; The device status information includes the frequency of failure , Maintenance cost ratio , Equipment operation stability indicators , the criticality index of the equipment is calculated by weighted calculation based on the parameters of the equipment status information , and use this indicator as the basis for data slicing operations; The quality detection information includes the coefficient of variation of quality data , Defective product growth rate , Critical Quality Characteristics Deviation By weighting and summing the various data fluctuation indicators of quality inspection information, a comprehensive evaluation value of the real-time data fluctuation characteristics is obtained. , used to identify and filter out data intervals with abnormal fluctuations; S200, automatically and dynamically slices production data according to the priority of production tasks, criticality indicators of equipment and real-time data fluctuation characteristics; The dynamic slicing operation includes time slicing, production link slicing and attribute slicing; S300, filter out high-priority production tasks, key equipment, and data within the abnormal data fluctuation range, that is, focus on analyzing a specific time period Specific production links and specific data attributes Data subsets that form dynamic data slices ; S400, use data analysis methods to conduct in-depth analysis of dynamic data slices and locate key factors; S500, the analysis results are sent to the MES system in a visual manner and stored in the data information database to provide a basis for production decisions.

2. According to claim 1, a method for intelligent analysis and management of industrial production data based on MES is characterized in that: In the S100 production task information: The urgency of the delivery time The remaining delivery days of the current production task Standard delivery days is measured by the difference between ; The order profit coefficient The expected revenue from the current production task The average revenue of the enterprise's production tasks during the same period The ratio is given, that is ; The degree of close cooperation with the client Based on the number of past cooperation with customers Total cooperation amount Comprehensive evaluation of the product of Calculate the priority of production tasks based on production task information for: ,in, , , They are the weight coefficients of delivery time urgency, order revenue coefficient, and customer cooperation closeness, and they meet .

3. The method for intelligent analysis and management of industrial production data based on MES according to claim 1, characterized in that: In the S100 device status information: The frequency of the fault Count the number of times a device fails within a specific time period , ; The proportion of maintenance costs By calculating the average cost of a single equipment repair Equipment purchase cost The ratio of ; The equipment operation stability index It is determined by monitoring the fluctuation range of the equipment operating parameters and using the standard deviation of the operating parameters The average value of this parameter The ratio of ; Calculate the criticality index of the equipment based on the equipment status information for: ,in, , , They are the weight coefficients of failure frequency, maintenance cost ratio, and equipment operation stability index, and meet the requirements. .

4. The method for intelligent analysis and management of industrial production data based on MES according to claim 1, characterized in that: In the S100 quality inspection information: The coefficient of variation of the quality data The sample standard deviation of the quality inspection data Divide by the sample mean Get, that is , used to measure the discreteness of quality data; The defective product growth rate Calculate the growth value of the number of defective products in the current time period Compared with the number of defective products in the previous period The ratio of ; The critical quality characteristic deviation Standard values ​​for key quality characteristics of products The actual detection value The difference of ; Calculate real-time data fluctuation characteristics through quality inspection information for: ,in, They are the coefficient of variation of quality data, the growth rate of defective products, and the weight coefficient of the deviation of key quality characteristics, and they meet .

5. The method for intelligent analysis and management of industrial production data based on MES according to claim 1, characterized in that: When the S200 performs the dynamic slicing operation, for the time slice, the starting time of the production task is used. and end time Based on the priority of production tasks and comprehensive evaluation of real-time data fluctuation characteristics To determine the specific slicing time range, that is, when Above priority threshold and Greater than the fluctuation threshold When , the time slice range is reduced to ,in and The time offset is dynamically adjusted according to the urgency of the production task and the data fluctuation. The calculation methods are as follows: ,in and is the time adjustment factor.

6. The method for intelligent analysis and management of industrial production data based on MES according to claim 1, characterized in that: When the S200 performs slicing in the production process, according to the criticality index of the equipment Classify each link in the production process and set the key indicators Above criticality threshold The production link where the equipment is located is determined as the key production link, and the data of the key production link is extracted and analyzed. At the same time, for the production link that has a direct upper and lower relationship with the key production link, its data extraction weight is According to the correlation coefficient Adjust the data extraction weight to ,in Extract weights for basic data, The weight adjustment coefficient is used. The key production links and related links are extracted according to the adjusted weights, and the data of non-key links are screened out.

7. The method for intelligent analysis and management of industrial production data based on MES according to claim 1, characterized in that: When performing attribute slicing, the S200 constructs an attribute association matrix based on various attribute data in the production task information, equipment status information, and quality inspection information. , for production task priorities , Equipment criticality index and comprehensive evaluation of real-time data fluctuation characteristics Correlation is above the correlation threshold The attributes are extracted and retained first in the attribute slices. The attribute correlation is calculated using the Pearson correlation coefficient method. and comprehensive evaluation indicators , and its Pearson correlation coefficient The calculation formula is: ,in, For attributes In the The values ​​in the data samples, For attributes The sample mean of The comprehensive evaluation value of the corresponding production task priority, equipment criticality index and real-time data fluctuation characteristics is in the first The values ​​in the data samples, is its sample mean, is the number of data samples.

8. The method for intelligent analysis and management of industrial production data based on MES according to claim 1, characterized in that: The S300 selects high-priority production tasks, corresponding key equipment, and data in the abnormal data fluctuation range from the data after dynamic slicing, and further integrates the selected data to form dynamic data slices for analysis, that is, the time slices are in the time interval The data within the production link is sliced ​​to determine the production link set , including key links and associated links, attribute slices determine attribute sets , the data subset of the dynamic data slice Indicates the time interval The first At a certain time point, the first The corresponding production link The data value of an attribute.

9. The method for intelligent analysis and management of industrial production data based on MES according to claim 1, characterized in that: The S400 uses principal component analysis (PCA) of the data analysis method to analyze the data subsets. Standardized processing to obtain , the standardization process is for each attribute , calculate its mean and standard deviation ,but ; Construct the covariance matrix Cov, whose elements in the variance matrix for: ,in is the number of data points, Is an attribute In the production stage The mean on Is an attribute In the production stage The mean on ; Perform eigenvalue decomposition on the covariance matrix to obtain the eigenvalue and the corresponding eigenvector , , is the number of attributes, sort the eigenvectors in descending order of eigenvalues, and select the first The eigenvectors form the projection matrix ,pass Get the reduced dimension dataset ; The dataset Consider it as a transaction database, where each row represents a transaction and each column represents an attribute. Determine the minimum support threshold and the minimum confidence threshold , scan the dataset , count the number of occurrences of each attribute, and select the frequent itemsets that meet the minimum support threshold Connect and generate candidate item sets , scan the dataset again ,statistics The support of each item in the set is frequently obtained , repeatedly search for frequent item sets, generate high-order candidate sets and frequent sets, until no new frequent sets can be generated; For the final , calculate its confidence ,in Representing Item Sets The support degree of is the proportion of the number of occurrences to the total number of transactions, so as to extract the association rules whose confidence is greater than the minimum confidence threshold.

Images