MES data intelligent management system and method based on data analysis

Abstract

The application discloses a kind of MES data intelligent management system and method based on data analysis, it is related to data analysis technical field.The system of the present application includes: historical production data acquisition and processing module, production efficiency evaluation and analysis module, trend feature calculation and anomaly detection module and real-time production data analysis and intelligent early warning module;Historical production data acquisition and processing module is responsible for obtaining data from system, calculates the waiting time between processes, and carries out the preliminary processing of data;Production efficiency evaluation and analysis module extracts features and evaluates production efficiency, analyzes the trend of historical data;Trend feature calculation and anomaly detection module is based on the trend feature calculation and analysis of production efficiency, carries out abnormal trend identification;Real-time production data analysis and intelligent early warning module handles real-time data, and carries out data comparison and early warning.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of data analysis, in particular to an MES data intelligent management system and method based on data analysis. BACKGROUND

[0002] With the continuous improvement of the informatization degree of manufacturing industry, the manufacturing execution system (MES) as an important part of enterprise informatization, is mainly used for managing production planning, scheduling, quality control and other links, helping enterprises to realize real-time monitoring and fine management of production process. The traditional MES system mainly focuses on real-time monitoring and data collection of production process, aiming to ensure that the production progress and quality of products from raw materials to finished products are effectively controlled.

[0003] However, most of the current MES systems on the market are limited to event tracking and work record reporting of the production process, usually by scanning barcodes or two-dimensional codes to register work orders and record the completion of each production process. Such systems often focus on post-processing and passive data collection of the production process, lacking real-time intelligent analysis and active decision support functions for production data. Therefore, although the existing MES system can basically monitor and collect data of the production process, its ability in data analysis is relatively limited, lacking accurate prediction of production bottlenecks and potential risks. This makes managers have to rely on manual experience to make adjustments when facing production delays or bottlenecks, failing to timely discover potential problems and take preventive measures. SUMMARY

[0004] The purpose of the present application is to provide an MES data intelligent management system and method based on data analysis to solve the problems raised in the background.

[0005] In order to solve the above technical problems, the present application provides the following technical solutions:

[0006] An MES data intelligent management method based on data analysis, comprising the following steps:

[0007] Step S100. Obtain all historical production records corresponding to production processes of a production workshop in a selected period, and define each production process as a production node; according to the production process sequence, obtain the waiting time between each adjacent two production nodes, defined as inter-process waiting time, aggregate each inter-process waiting time in the selected period, and construct an inter-process waiting time set;

[0008] Step S200. According to the historical production records, the historical production data is divided according to the corresponding production nodes, and is divided into several historical production data segments, and the corresponding node data set is constructed; the corresponding historical production data features are extracted from each node data set, and the production efficiency of the corresponding historical production data segment is evaluated according to the historical production data features;

[0009] Step S300. Each element in the inter-process waiting time set is corresponded to the production efficiency evaluation result of the corresponding historical production data segment, the correlation between the inter-process waiting time and the production efficiency evaluation result is analyzed, and the historical production data feature trend curve is obtained based on the correlation between the inter-process waiting time and the production efficiency evaluation result;

[0010] Step S400. Obtain the real-time production data corresponding to each production node, analyze the real-time production data, and extract the real-time production data features; analyze the real-time production data features, obtain the real-time production data feature trend curve, and compare the real-time production data feature trend curve with the historical production data feature trend curve, thereby generating the corresponding notification information.

[0011] Further, step S100 includes:

[0012] S101. Through the MES system, obtain the historical production records corresponding to all production processes of the production workshop in the selected period, and define each production process as a production node; the historical production record refers to the data information recorded in the production process of each production process, including process name, start time, end time, historical event record and historical production data; according to the production process sequence recorded in the MES system, the start time Ts and the end time Te of each production node are obtained in turn, thereby obtaining the time range of the corresponding production node as [ts, te]; according to the production process sequence, the waiting time ti between each adjacent two production nodes is calculated in turn, and ti = ts(i+1)-te(i), wherein ts(i+1) represents the start time of the production node i+1, and te(i) represents the end time of the production node i;

[0013] S102. Defining the waiting time ti between each two adjacent production nodes as an inter-process waiting time, collecting each inter-process waiting time in a selected period to build an inter-process waiting time set T, and T={t1, t2,..., tn-1}, wherein t1 represents the waiting time between production node 1 and production node 2, t2 represents the waiting time between production node 2 and production node 3, and tn-1 represents the waiting time between production node n-1 and production node n, n represents the total number of production nodes; wherein each inter-process waiting time set T corresponds to a production process sequence, and the number of inter-process waiting time sets T is equal to the number of production process sequences in the selected period, and the production task corresponding to each production process sequence is the same.

[0014] Further, step S200 comprises:

[0015] S201. For each production process sequence, obtaining the corresponding historical production record, dividing the historical production data corresponding to each production node in chronological order into several historical production data segments, and each historical production data segment has a one-to-one correspondence with the production node corresponding to the corresponding production process sequence; according to the production node number, collecting the historical production data segments corresponding to all the same production node numbers in the selected period, and sorting the historical production data segments in chronological order to form a node data set Ai, and Ai={ai1, ai2,..., aim}, wherein ai1 represents the historical production data segment corresponding to the first production process sequence of the i-th production node, ai2 represents the historical production data segment corresponding to the second production process sequence of the i-th production node, and aim represents the historical production data segment corresponding to the m-th production process sequence of the i-th production node, and m represents the number of production process sequences in the selected period;

[0016] S202. For each element aij in the node data set Ai, the corresponding historical production data features are extracted, and the historical production data features include production time, number of abnormal events, production yield, and rejection rate; the collected historical production data features are normalized to obtain the corresponding historical production data feature values, and each element aij in the node data set Ai is evaluated for production efficiency according to the historical production data feature values, thereby calculating the comprehensive production efficiency indicator Eij, and Eij=α×(Pij / Tij)-β×Nij-γ×Qij, wherein α, β and γ represent weight coefficients, and α+β+γ=1; Tij represents the feature value of the production time corresponding to the element aij, Pij represents the feature value of the production product quantity corresponding to the element aij, Nij represents the feature value of the number of abnormal events corresponding to the element aij, and Qij represents the feature value of the rejection rate corresponding to the element aij.

[0017] Since step S200 is analyzed for each production process sequence, the historical production record of each production process sequence is obtained, the corresponding historical production data is divided according to the corresponding production node, thereby obtaining a plurality of historical production data segments, and each production node corresponds to a historical production data segment; and because there are a plurality of production process sequences in the selected period, for each production node, the number of corresponding historical production data segments is equal to the number of production process sequences in the selected period.

[0018] Further, step S300 includes:

[0019] S301. For each node data set Ai, obtain the comprehensive production efficiency index Eij of all historical production data segments in the node data set Ai, match the corresponding inter-process waiting time tij according to the corresponding production process sequence number and the corresponding production node number, thereby forming a plurality of associated data pairs G, and G=(Eij, tij); for each node data set Ai, aggregate the corresponding associated data pairs G, and express the associated data pairs G as a data point in a plane rectangular coordinate system, and the horizontal axis of the plane rectangular coordinate system represents the comprehensive production efficiency index Eij, and the vertical axis represents the inter-process waiting time tij; according to the order in the node data set Ai, the data points are sequentially connected, thereby obtaining a curve Li;

[0020] S302. For each node data set Ai corresponding curve Li, the values of all data points on the curve Li are obtained, the mean and standard deviation of the inter-process waiting time and the comprehensive production efficiency index are calculated in sequence, and the trend characteristic index Rij is calculated according to the mean and standard deviation of the inter-process waiting time and the comprehensive production efficiency index, and the specific calculation formula is:

[0021] Rij=w1·[(Eij-μEi) / 3σEi]+w2·[(tij-μti) / 3σti];

[0022] Wherein, Rij represents the trend characteristic index corresponding to the jth production process sequence of the ith production node, μEi and σEi represent the mean and standard deviation of the horizontal axis value of the data points of the curve Li corresponding to the node data set Ai, μti and σti represent the mean and standard deviation of the vertical axis value of the data points of the curve Li corresponding to the node data set Ai, w1 and w2 represent weight coefficients, and w1+w2=1;

[0023] S303. The trend characteristic index Rij corresponding to each element in the node data set Ai is summarized, the trend characteristic index threshold R corresponding to the node data set Ai is defined in combination with the historical production record, the trend characteristic index Rij corresponding to each element in the node data set Ai is compared with the trend characteristic index threshold R, if Rij≤R, the data point on the corresponding curve Li is marked as a normal data point, if Rij>R, the data point on the corresponding curve Li is marked as an abnormal data point; according to the marked position of the abnormal data point on the curve Li, the curve Li is divided to obtain a plurality of continuous curve segments, and each continuous curve segment is composed of normal data points, each continuous curve segment is fitted to obtain the curve equation of each continuous curve segment, and the curve equation of each continuous curve segment is defined as the historical production data characteristic trend curve Q.

[0024] Further, step S400 comprises:

[0025] S401. Through the MES system, real-time production data corresponding to each production node is acquired, and the real-time production data is analyzed by referring to the analysis mode of the historical production data, so as to construct a corresponding node data set Bi for each production node; for each element in the node data set Bi of each production node, the corresponding real-time production data characteristics are extracted by referring to the analysis mode of the node data set Ai, and the comprehensive production efficiency index E’ij corresponding to each element in the node data set Bi is calculated according to the real-time production data characteristics; according to the production process sequence number and the corresponding production node number corresponding to each element in the node data set Bi, the inter-process waiting time t’ij of the corresponding element is matched; the comprehensive production efficiency index E’ij and the inter-process waiting time t’ij of each element in the node data set Bi are summarized to form a plurality of associated data pairs G’, and G’=(E’ij, t’ij);

[0026] S402. For each production node corresponding to the associated data pair G', it is represented as a data point in a plane rectangular coordinate system in turn, and the data points are connected in turn according to the order of the node data set Bi, so as to obtain the real-time production data characteristic trend curve S; according to the production node number corresponding to the real-time production data characteristic trend curve S, the historical production data characteristic trend curve Q with the same production node number is matched, the real-time production data characteristic trend curve S is moved on the historical production data characteristic trend curve Q in turn, the highest coincidence degree is selected as the matching result, and the real-time production data characteristic trend curve S is marked on the matched historical production data characteristic trend curve Q, the number of data points between the last data point of the real-time production data characteristic trend curve S and the last data point of the matched historical production data characteristic trend curve Q is counted, and the last data point of the real-time production data characteristic trend curve S and the last data point of the matched historical production data characteristic trend curve Q do not coincide; when the number of data points is equal to 0, the number of the current production node is output to the relevant personnel, and the relevant personnel carries out corresponding processing.

[0027] An MES data intelligent management system based on data analysis, comprising: a historical production data acquisition and processing module, a feature extraction and efficiency evaluation module, a trend characteristic analysis and abnormality detection module, and a real-time production data analysis and intelligent early warning module;

[0028] The historical production data acquisition and processing module acquires historical production records corresponding to all production processes of the production workshop in a selected period, and defines each production process as a production node; according to the production process sequence, the waiting time between each adjacent two production nodes is obtained, which is defined as the inter-process waiting time, the inter-process waiting time of each process in the selected period is summarized, and an inter-process waiting time set is constructed;

[0029] The feature extraction and efficiency evaluation module divides the historical production data according to the corresponding production nodes based on the historical production records, divides the historical production data into several historical production data segments, and constructs the corresponding node data set; the corresponding historical production data characteristics are extracted from each node data set, and the production efficiency of the corresponding historical production data segment is evaluated according to the historical production data characteristics;

[0030] The trend characteristic analysis and abnormality detection module corresponds each element in the inter-process waiting time set with the production efficiency evaluation result of the corresponding historical production data segment, analyzes the correlation between the inter-process waiting time and the production efficiency evaluation result; based on the correlation between the inter-process waiting time and the production efficiency evaluation result, the historical production data characteristic trend curve is obtained;

[0031] The real-time production data analysis and intelligent early warning module obtains real-time production data corresponding to each production node, analyzes the real-time production data, and extracts real-time production data features; the real-time production data features are analyzed to obtain a real-time production data feature trend curve, and the real-time production data feature trend curve is compared with a historical production data feature trend curve to generate corresponding notification information.

[0032] Further, the historical production data acquisition and processing module includes a data acquisition unit, a production node definition unit, and a inter-process waiting time calculation unit.

[0033] The data acquisition unit obtains historical production records corresponding to all production processes of the production workshop in a selected period from the MES system; the production node definition unit defines each production process as a production node and obtains the time range of each production node; the inter-process waiting time calculation unit calculates the waiting time between each adjacent two production nodes, aggregates the waiting time between each process in the selected period, and constructs an inter-process waiting time set.

[0034] Further, the feature extraction and efficiency evaluation module includes a data division and feature extraction unit and an efficiency evaluation unit.

[0035] The data division and feature extraction unit divides historical production data according to production nodes to form several historical production data segments, and extracts corresponding historical production data features; the efficiency evaluation unit evaluates the production efficiency of each node data set and calculates a comprehensive production efficiency index.

[0036] Further, the trend feature analysis and anomaly detection module includes a trend feature analysis unit and an anomaly detection unit.

[0037] The trend feature analysis unit analyzes the correlation between production efficiency and inter-process waiting time based on historical production data features, and calculates a trend feature index; the anomaly detection unit compares the trend feature index with a trend feature index threshold to mark normal data points and abnormal data points; and obtains a historical production data feature trend curve based on the abnormal data points.

[0038] Further, the real-time production data analysis and intelligent early warning module includes a real-time production data analysis unit and a trend comparison and early warning unit.

[0039] The real-time production data analysis unit obtains real-time production data corresponding to each production node from the MES system, analyzes the real-time production data, extracts real-time production data features, and obtains a real-time production data feature trend curve; the trend comparison and early warning unit compares the real-time production data feature trend curve with a historical production data feature trend curve to generate corresponding notification information.

[0040] Compared with the prior art, the beneficial effects of the present application are: by analyzing historical production data, extracting production efficiency, inter-process waiting time and other characteristics, and dynamically monitoring and analyzing real-time production data through trend characteristic indicators, potential risks and bottlenecks in the production process can be identified in advance, thereby realizing proactive early warning and decision support; compared with the post-processing of traditional MES systems, the present application provides real-time intelligent analysis and prediction functions, enhancing the proactive management capability of the production process. By correlating inter-process waiting time and production efficiency evaluation results, the present application can identify potential bottlenecks and low-efficiency links in the production process, and further provide optimization suggestions for managers, so as to take effective measures before problems occur, reduce production delays and quality problems. Through analysis of historical production data and real-time production data, the present application can evaluate the comprehensive efficiency of production nodes and provide specific optimization direction for managers, further improving production efficiency, reducing resource waste, optimizing production scheduling, and improving overall resource utilization. Unlike traditional methods that rely on manual experience and manual adjustment, the present application has automatic evaluation and prediction capabilities based on data analysis, which helps to reduce the decision-making pressure of managers, provides data-driven decision support, and reduces the risk of human error. The present application can compare and analyze the trend curve of production data characteristics, discover abnormal data points in production in a timely manner, mark them as abnormal and issue warnings, which can effectively avoid production losses due to delayed problem discovery. The present application refines data analysis for each production node and customizes corresponding production efficiency evaluation and process optimization strategies for each production node, thereby ensuring fine management of the entire production process, optimizing each process link, and improving the operation efficiency of the overall production line. BRIEF DESCRIPTION OF DRAWINGS

[0041] The accompanying drawings are included to provide a further understanding of the present application, and constitute a part of the specification, which together with the embodiments of the present application serve to explain the present application, and do not constitute a limitation of the present application. In the drawings:

[0042] Figure 1 is a module schematic diagram of a MES data intelligent management system based on data analysis of the present application. DETAILED DESCRIPTION

[0043] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the present application.

[0044] Please refer to Figure 1 , the present application provides technical solutions:

[0045] The MES data intelligent management system based on data analysis comprises a historical production data acquisition and processing module, a feature extraction and efficiency evaluation module, a trend feature analysis and anomaly detection module, and a real-time production data analysis and intelligent early warning module.

[0046] The historical production data acquisition and processing module acquires historical production records corresponding to all production processes of a production workshop in a selected period, and defines each production process as a production node; according to the production process sequence, the waiting time between each adjacent two production nodes is acquired, which is defined as the inter-process waiting time, the inter-process waiting time of each process in the selected period is summarized, and an inter-process waiting time set is constructed.

[0047] The feature extraction and efficiency evaluation module divides the historical production data according to the corresponding production nodes based on the historical production records, divides the historical production data into a plurality of historical production data segments, and constructs a corresponding node data set; the corresponding historical production data features are extracted from each node data set, and the production efficiency of the corresponding historical production data segment is evaluated according to the historical production data features;

[0048] The trend feature analysis and anomaly detection module corresponds each element in the inter-process waiting time set with the production efficiency evaluation result of the corresponding historical production data segment, analyzes the correlation between the inter-process waiting time and the production efficiency evaluation result, and obtains a historical production data feature trend curve based on the correlation between the inter-process waiting time and the production efficiency evaluation result.

[0049] The real-time production data analysis and intelligent early warning module acquires real-time production data corresponding to each production node, analyzes the real-time production data, and extracts real-time production data features; the real-time production data feature trend curve is obtained by analyzing the real-time production data features, and the real-time production data feature trend curve is compared with the historical production data feature trend curve, thereby generating corresponding notification information.

[0050] The historical production data acquisition and processing module comprises a data acquisition unit, a production node definition unit, and an inter-process waiting time calculation unit.

[0051] The data acquisition unit acquires historical production records corresponding to all production processes of a production workshop in a selected period from a MES system; the production node definition unit defines each production process as a production node, and acquires the time range of each production node; the inter-process waiting time calculation unit calculates the waiting time between each adjacent two production nodes, summarizes the waiting time of each inter-process in the selected period, and constructs an inter-process waiting time set.

[0052] The feature extraction and efficiency evaluation module comprises a data division and feature extraction unit and an efficiency evaluation unit.

[0053] The data partitioning and feature extraction unit divides the historical production data according to the production nodes, forming several historical production data segments, and extracts the corresponding historical production data features; the efficiency evaluation unit evaluates the production efficiency of each node dataset and calculates the comprehensive production efficiency index.

[0054] The trend feature analysis and anomaly detection module includes a trend feature analysis unit and an anomaly detection unit;

[0055] The trend feature analysis unit analyzes the correlation between production efficiency and waiting time between processes based on historical production data features, and calculates trend feature indicators; the anomaly detection unit compares the trend feature indicators with trend feature indicator thresholds to mark normal data points and abnormal data points; based on the abnormal data points, the historical production data feature trend curve is obtained.

[0056] The real-time production data analysis and intelligent early warning module includes a real-time production data analysis unit and a trend comparison and early warning unit;

[0057] The real-time production data analysis unit obtains real-time production data corresponding to each production node from the MES system, analyzes the real-time production data, extracts real-time production data features, and thus obtains the real-time production data feature trend curve; the trend comparison and early warning unit compares the real-time production data feature trend curve with the historical production data feature trend curve, and thus generates corresponding notification information.

[0058] A data-driven intelligent management method for MES (Manufacturing Execution System) includes the following steps:

[0059] Step S100. Obtain the historical production records corresponding to all production processes in the production workshop within the selected period, and define each production process as a production node; according to the production process sequence, obtain the waiting time between each pair of adjacent production nodes, define it as the inter-process waiting time, summarize the inter-process waiting time within the selected period, and construct the inter-process waiting time set.

[0060] Step S200. Based on historical production records, divide the historical production data into several historical production data segments according to the corresponding production nodes, and construct corresponding node datasets; extract the corresponding historical production data features from each node dataset, and evaluate the production efficiency of the corresponding historical production data segments based on the historical production data features.

[0061] Step S300. Corresponding each element in the inter-process waiting time set to the production efficiency evaluation result of the corresponding historical production data segment, analyze the correlation between the inter-process waiting time and the production efficiency evaluation result; based on the correlation between the inter-process waiting time and the production efficiency evaluation result, obtain the historical production data feature trend curve;

[0062] Step S400. Obtain the real-time production data corresponding to each production node, analyze the real-time production data to extract the real-time production data features; analyze the real-time production data features to obtain the real-time production data feature trend curve, and compare the real-time production data feature trend curve with the historical production data feature trend curve to generate the corresponding notification information.

[0063] Step S100 includes:

[0064] S101. Through the MES system, obtain the historical production records corresponding to all production processes of the production workshop in the selected period, and define each production process as a production node; the historical production record refers to the data information recorded in each production process during the production process, including process name, start time, end time, historical event record and historical production data; according to the production process sequence recorded in the MES system, the start time Ts and the end time Te of each production node are obtained in turn, so that the time range of the corresponding production node is [ts, te]; according to the production process sequence, the waiting time ti between each adjacent two production nodes is calculated in turn, and ti = ts(i+1)-te(i), wherein ts(i+1) represents the start time of the production node i+1, and te(i) represents the end time of the production node i;

[0065] S102. Define the waiting time ti between each adjacent two production nodes as the inter-process waiting time, and collect each inter-process waiting time in the selected period to construct the inter-process waiting time set T, and T = {t1, t2,..., tn-1}, wherein t1 represents the waiting time between the production node 1 and the production node 2, t2 represents the waiting time between the production node 2 and the production node 3, and tn-1 represents the waiting time between the production node n-1 and the production node n, n represents the total number of production nodes; wherein each inter-process waiting time set T corresponds to a production process sequence, and the number of inter-process waiting time sets T is equal to the number of production process sequences in the selected period, and the production tasks corresponding to each production process sequence are the same.

[0066] In this embodiment, in the MES system, a period is selected, for example, a day or a week, and the historical records of all production processes in the period are obtained. Each production process includes the following information: process name, start time (ts), end time (te), historical event record: records events during the process (such as equipment failure, production interruption, etc.), historical production data: such as yield, pass rate, quality inspection results, etc.

[0067] The waiting time between each adjacent two production nodes is calculated through the production process sequence of the MES system. For example, the end time te1 of the production node 1 in the ith production process sequence is 10:00, and the start time ts2 of the production node 2 is 10:05, and the corresponding waiting time t1 = ts2-te1 = 5 min; The waiting time between all adjacent production nodes in the ith production process sequence is summarized to form the corresponding inter-process waiting time set T, and T = {t1, t2,..., tn-1}.

[0068] Step S200 includes:

[0069] S201. For each production process sequence, the corresponding historical production record is obtained, and the historical production data corresponding to each production node is divided in time sequence to form a plurality of historical production data segments, and each historical production data segment has a one-to-one correspondence with the production node corresponding to the corresponding production process sequence; according to the production node number, the historical production data segments corresponding to the same production node number in the selected period are summarized, and the historical production data segments are sorted in time sequence, thereby forming a node data set Ai, and Ai = {ai1, ai2,..., aim}, wherein ai1 represents the historical production data segment corresponding to the first production process sequence of the ith production node, ai2 represents the historical production data segment corresponding to the second production process sequence of the ith production node, and so on, and aim represents the historical production data segment corresponding to the mth production process sequence of the ith production node, and m represents the number of production process sequences in the selected period;

[0070] S202. Extract the corresponding historical production data features respectively for each element aij in the node data set Ai, and the historical production data features include production time, number of abnormal events, production yield, and rejection rate; normalize the collected historical production data features to obtain corresponding historical production data feature values, and perform production efficiency evaluation on each element aij in the node data set Ai according to the historical production data feature values, thereby calculating a comprehensive production efficiency indicator Eij, and Eij = a x (Pij / Tij) - β x Nij - γ x Qij, where a, β, and γ are weight coefficients, and a + β + γ = 1; Tij represents the feature value of the production time corresponding to the element aij, Pij represents the feature value of the production product quantity corresponding to the element aij, Nij represents the feature value of the number of abnormal events corresponding to the element aij, and Qij represents the feature value of the rejection rate corresponding to the element aij.

[0071] Since step S200 is to analyze each production process sequence, the historical production records of each production process sequence are obtained, and the corresponding historical production data is divided according to the corresponding production nodes, thereby obtaining a plurality of historical production data segments, and each production node corresponds to a historical production data segment; and because there are a plurality of production process sequences in the selected period, for each production node, the number of corresponding historical production data segments is equal to the number of production process sequences in the selected period.

[0072] In this embodiment, it is assumed that the historical production data features of node 1 of the ith production process sequence are as follows:

[0073] Node 1: production time 10 hours, yield 1000 pieces, abnormal event 2 times, rejection rate 5%.

[0074] Normalize the collected historical production data features of node 1 to obtain corresponding historical production data feature values; it is assumed that the preset is:

[0075] The maximum value of the production time (T) is 12 hours, and the minimum value is 8 hours.

[0076] The maximum value of the yield (P) is 1200 pieces, and the minimum value is 800 pieces.

[0077] The maximum value of the number of abnormal events (N) is 3 times, and the minimum value is 0 times.

[0078] The maximum value of the rejection rate (Q) is 5%, and the minimum value is 3%.

[0079] According to the formula of minimum-maximum normalization: X_norm=(X-X_min) / (X_max-X_min), wherein X is the original data, X_min is the minimum value of the feature, and X_max is the maximum value of the feature; then the normalized historical production data feature values are Ti1=0.5, Pi1=0.5, Ni1=0.667, and Qi1=1.

[0080] According to the formula: Eij=α×(Pij / Tij)-β×Nij-γ×Qij, the corresponding comprehensive production efficiency index Ei1 is calculated, assuming that the weight coefficients are: α=0.5, β=0.3, and γ=0.2, then the corresponding Ei1=0.1.

[0081] Step S300 comprises:

[0082] S301. For each node data set Ai, the comprehensive production efficiency index Eij of all historical production data segments in the node data set Ai is obtained, the corresponding inter-process waiting time tij is matched according to the corresponding production process sequence number and the corresponding production node number, thereby forming a plurality of associated data pairs G, and G=(Eij, tij); for each node data set Ai, the corresponding associated data pairs G are summarized, and the associated data pairs G are represented as a data point in a plane rectangular coordinate system, wherein the horizontal axis of the plane rectangular coordinate system represents the comprehensive production efficiency index Eij, and the vertical axis represents the inter-process waiting time tij; the data points are sequentially connected in order according to the node data set Ai, thereby obtaining a curve Li;

[0083] S302. For each curve Li corresponding to each node data set Ai, the values of all data points on the curve Li are obtained, the mean and standard deviation of the inter-process waiting time and the comprehensive production efficiency index are sequentially calculated, and the trend characteristic index Rij is calculated according to the mean and standard deviation of the inter-process waiting time and the comprehensive production efficiency index, and the specific calculation formula is:

[0084] Rij=w1·[(Eij-μEi) / 3σEi]+w2·[(tij-μti) / 3σti];

[0085] Wherein, Rij represents the trend characteristic index corresponding to the jth production process sequence of the ith production node, μEi and σEi represent the mean and standard deviation of the horizontal axis values of the data points of the curve Li corresponding to the node data set Ai, μti and σti represent the mean and standard deviation of the vertical axis values of the data points of the curve Li corresponding to the node data set Ai, w1 and w2 represent weight coefficients, and w1+w2=1.

[0086] S303. The trend characteristic index Rij corresponding to each element in the node data set Ai is summarized, the trend characteristic index threshold R corresponding to the node data set Ai is defined in combination with the historical production record, the trend characteristic index Rij corresponding to each element in the node data set Ai is compared with the trend characteristic index threshold R, if Rij≤R, the data point on the curve Li corresponding thereto is marked as a normal data point, if Rij>R, the data point on the curve Li corresponding thereto is marked as an abnormal data point; the curve Li is divided according to the marked position of the abnormal data point on the curve Li, a plurality of continuous curve segments are obtained, and each continuous curve segment is composed of normal data points, each continuous curve segment is fitted, thereby obtaining the curve equation of each continuous curve segment, and the curve equation of each continuous curve segment is defined as the historical production data characteristic trend curve Q.

[0087] Step S400 comprises:

[0088] S401. Through the MES system, real-time production data corresponding to each production node is acquired, the real-time production data is analyzed by referring to the analysis mode of the historical production data, thereby constructing a corresponding node data set Bi for each production node; for each element in the node data set Bi of each production node, the corresponding real-time production data characteristic is extracted by referring to the analysis mode of the node data set Ai, and the comprehensive production efficiency index E'ij corresponding to each element in the node data set Bi is calculated according to the real-time production data characteristic; the inter-process waiting time t'ij of the corresponding element is matched according to the production process sequence number and the corresponding production node number corresponding to each element in the node data set Bi; the comprehensive production efficiency index E'ij and the inter-process waiting time t'ij of each element in the node data set Bi are summarized to constitute a plurality of associated data pairs G', and G'=(E'ij, t'ij);

[0089] S402. For each production node corresponding to the associated data pair G', it is represented as a data point in a plane rectangular coordinate system in turn, and the data points are connected in turn according to the order of the node data set Bi, so as to obtain a real-time production data characteristic trend curve S; according to the production node number corresponding to the real-time production data characteristic trend curve S, the historical production data characteristic trend curve Q with the same production node number is matched, the real-time production data characteristic trend curve S is moved on the historical production data characteristic trend curve Q in turn, the highest coincidence degree is selected as the matching result, and the real-time production data characteristic trend curve S is marked on the matched historical production data characteristic trend curve Q, the number of data points between the last data point of the real-time production data characteristic trend curve S and the last data point of the matched historical production data characteristic trend curve Q is counted, and the last data point of the real-time production data characteristic trend curve S and the last data point of the matched historical production data characteristic trend curve Q do not coincide; when the number of data points is equal to 0, the number of the current production node is output to the relevant personnel, and the relevant personnel carries out corresponding processing.

[0090] In the embodiment, the default real-time production data characteristic trend curve S does not have abnormal data points, so the number of data points of the real-time production data characteristic trend curve S is less than the number of data points of the historical production data characteristic trend curve Q with the same production node number; the specific operation process of moving the real-time production data characteristic trend curve S on the historical production data characteristic trend curve Q in turn is as follows:

[0091] The data points on the historical production data characteristic trend curve Q are defined as starting points in turn, the starting points of the real-time production data characteristic trend curve S are aligned on the starting points of the historical production data characteristic trend curve Q in turn, the first data point of S is aligned with the first data point of Q, the coincidence degree (such as correlation, similarity) between the data points is calculated, and the process of moving the curve is found to find the most matched starting point position until the coincidence degree of the starting point of S and a certain point in Q reaches the highest. This process can be completed by calculating the correlation, Euclidean distance or other matching metrics between the curves; after the S curve and the Q curve are aligned, the number of data points between the last data point of the real-time production data characteristic trend curve S and the last data point of the matched historical production data characteristic trend curve Q is calculated; when the number of data points is 0, it indicates that there may be an abnormal situation of low production efficiency or long waiting time between processes in the next production process of the current production node, so as to realize the accurate prediction of production bottlenecks and potential risks, and then a notification information is generated through the MES system to inform the relevant personnel of the current production node number and the measures to be taken (such as checking equipment, scheduling production, optimizing production process, etc.).

[0092] It is to be noted that, in the present text, relational terms such as first and second and the like can be used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Moreover, the terms "comprises", "comprising", or any other variations thereof, are intended to cover a non-exclusive inclusion such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but can include other elements not expressly listed or inherent to such process, method, article, or apparatus.

[0093] Finally, it should be noted that the above-mentioned only constitutes preferred embodiments of the present application and is not intended to limit the present application. Although the present application has been described in detail with reference to the foregoing embodiments, it will be apparent to those skilled in the art that modifications, equivalent replacements, improvements and the like of the technical solutions described in the foregoing embodiments can still be made. Any modifications, equivalent replacements, improvements and the like made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. A data-driven intelligent management method for MES (Manufacturing Execution System), characterized in that: The method includes the following steps: Step S100. Obtain the historical production records corresponding to all production processes in the production workshop within the selected period, and define each production process as a production node; according to the production process sequence, obtain the waiting time between each pair of adjacent production nodes, define it as the inter-process waiting time, summarize the inter-process waiting time within the selected period, and construct the inter-process waiting time set. Step S200. Based on historical production records, divide the historical production data into several historical production data segments according to the corresponding production nodes, and construct corresponding node datasets; extract the corresponding historical production data features from each node dataset, and evaluate the production efficiency of the corresponding historical production data segments based on the historical production data features. Step S300. Match each element in the inter-process waiting time set with the corresponding historical production data segment's production efficiency evaluation result, and analyze the correlation between inter-process waiting time and production efficiency evaluation result; based on the correlation between inter-process waiting time and production efficiency evaluation result, obtain the historical production data characteristic trend curve; Step S400. Obtain the real-time production data corresponding to each production node, and analyze the real-time production data in accordance with the analysis method of historical production data to extract the real-time production data features; analyze the real-time production data features to obtain the real-time production data feature trend curve, and compare the real-time production data feature trend curve with the historical production data feature trend curve to generate corresponding notification information. Step S200 includes: S201. For each production process sequence, obtain the corresponding historical production records. Divide the historical production data corresponding to each production node into several historical production data segments according to the time sequence. Each historical production data segment has a one-to-one correspondence with the production node corresponding to the corresponding production process sequence. According to the production node number, summarize all historical production data segments corresponding to the same production node number within the selected period, and sort the historical production data segments according to the time sequence to form a node dataset Ai, where Ai={ai1,ai2,...,aim}, where ai1 represents the historical production data segment corresponding to the first production process sequence of the i-th production node, ai2 represents the historical production data segment corresponding to the second production process sequence of the i-th production node, and so on. aim represents the historical production data segment corresponding to the m-th production process sequence of the i-th production node, and m represents the number of production process sequences within the selected period. S202. For each element aij in the node dataset Ai, extract the corresponding historical production data features, including production time, number of abnormal events, production output, and defect rate. Normalize the collected historical production data features to obtain the corresponding historical production data feature values. Evaluate the production efficiency of each element aij in the node dataset Ai based on the historical production data feature values ​​to calculate the comprehensive production efficiency index Eij, where Eij=α×(Pij / Tij)-β×Nij-γ×Qij, where α, β, and γ are all weight coefficients, and α+β+γ=1; Tij represents the feature value of production time corresponding to element aij, Pij represents the feature value of the number of products produced corresponding to element aij, Nij represents the feature value of the number of abnormal events corresponding to element aij, and Qij represents the feature value of the defect rate corresponding to element aij. Step S300 includes: S301. For each node dataset Ai, obtain the comprehensive production efficiency index Eij of all historical production data segments in node dataset Ai. Based on the corresponding production process sequence number and the corresponding production node number, match the corresponding inter-process waiting time tij to form several related data pairs G, where G=(Eij,tij). For each node dataset Ai, summarize the corresponding related data pairs G and represent the related data pairs G as a data point in a Cartesian coordinate system. The horizontal axis of the Cartesian coordinate system represents the comprehensive production efficiency index Eij, and the vertical axis represents the inter-process waiting time tij. Connect the data points sequentially according to the order in node dataset Ai to obtain curve Li. S302. For each node dataset Ai corresponding to curve Li, obtain the values ​​of all data points on curve Li, and calculate the mean and standard deviation of the inter-process waiting time and the comprehensive production efficiency index in turn. Calculate the trend characteristic index Rij based on the mean and standard deviation of the inter-process waiting time and the comprehensive production efficiency index, and the specific calculation formula is as follows: Rij=w1·[(Eij-μEi) / 3σEi]+w2·[(tij-μti) / 3σti]; Where Rij represents the trend feature index corresponding to the j-th production process sequence of the i-th production node, μEi and σEi represent the mean and standard deviation of the horizontal axis values ​​of the data points of the curve Li corresponding to the node dataset Ai, μti and σti represent the mean and standard deviation of the vertical axis values ​​of the data points of the curve Li corresponding to the node dataset Ai, and w1 and w2 represent the weight coefficients, and w1+w2=1. S303. Summarize the trend feature index Rij corresponding to each element in the node dataset Ai, define the trend feature index threshold R corresponding to the node dataset Ai in combination with historical production records, compare the trend feature index Rij corresponding to each element in the node dataset Ai with the trend feature index threshold R. If Rij≤R, then mark the data point on the corresponding curve Li as a normal data point. If Rij>R, then mark the data point on the corresponding curve Li as an abnormal data point. According to the marked position of the abnormal data points on the curve Li, divide the curve Li into several continuous curve segments, and each continuous curve segment is composed of normal data points. Fit each continuous curve segment to obtain the curve equation of each continuous curve segment, and define the curve equation of each continuous curve segment as the historical production data feature trend curve Q.

2. The MES data intelligent management method based on data analysis according to claim 1, characterized in that: Step S100 includes: S101. Using the MES system, obtain the historical production records corresponding to all production processes in the production workshop within the selected period, and define each production process as a production node; the historical production records refer to the data information of each production process during the production process, including process name, start time, end time, historical event records, and historical production data; according to the production process sequence recorded in the MES system, obtain the start time Ts and end time Te of each production node in sequence, thereby obtaining the time range of the corresponding production node as [ts,te]; according to the production process sequence, calculate the waiting time ti between each pair of adjacent production nodes in sequence, and ti=ts(i+1)-te(i), where ts(i+1) represents the start time of production node i+1, and te(i) represents the end time of production node i; S102. Define the waiting time ti between any two adjacent production nodes as the inter-process waiting time. Summarize the inter-process waiting times within the selected period to construct an inter-process waiting time set T, where T = {t1, t2, ..., tn-1}, where t1 represents the waiting time between production node 1 and production node 2, t2 represents the waiting time between production node 2 and production node 3, and so on, tn-1 represents the waiting time between production node n-1 and production node n, and n represents the total number of production nodes; each inter-process waiting time set T corresponds to a production process sequence, and the number of inter-process waiting time sets T is equal to the number of production process sequences within the selected period, and the production tasks corresponding to each production process sequence are the same.

3. The MES data intelligent management method based on data analysis according to claim 1, characterized in that: Step S400 includes: S401. Through the MES system, obtain the real-time production data corresponding to each production node, and analyze the real-time production data by referring to the analysis method of historical production data, thereby constructing a corresponding node dataset Bi for each production node; for each element in the node dataset Bi of each production node, extract the corresponding real-time production data features by referring to the analysis method of node dataset Ai, and calculate the comprehensive production efficiency index E'ij corresponding to each element in the node dataset Bi based on the real-time production data features; match the inter-process waiting time t'ij of the corresponding element according to the production process sequence number and the corresponding production node number of each element in the node dataset Bi; summarize the comprehensive production efficiency index E'ij and the inter-process waiting time t'ij of each element in the node dataset Bi to form several related data pairs G', and G'=(E'ij,t'ij); S402. For each production node's associated data pair G', represent it as a data point in a Cartesian coordinate system. Connect the data points sequentially according to the order of the node dataset Bi to obtain the real-time production data characteristic trend curve S. Based on the production node number corresponding to the real-time production data characteristic trend curve S, match the historical production data characteristic trend curve Q with the same production node number. Move the real-time production data characteristic trend curve S sequentially on the historical production data characteristic trend curve Q, select the one with the highest overlap as the matching result, and mark the real-time production data characteristic trend curve S on the matched historical production data characteristic trend curve Q. Count the number of data points between the last data point of the real-time production data characteristic trend curve S and the last data point of the matched historical production data characteristic trend curve Q, ensuring that the last data point of the real-time production data characteristic trend curve S and the last data point of the matched historical production data characteristic trend curve Q do not overlap. When the number of data points is 0, output the current production node number to the relevant personnel for appropriate processing.

4. A data analysis-based MES data intelligent management system, applied to the data analysis-based MES data intelligent management method according to any one of claims 1-3, characterized in that: The system includes: a historical production data acquisition and processing module, a feature extraction and efficiency evaluation module, a trend feature analysis and anomaly detection module, and a real-time production data analysis and intelligent early warning module. The historical production data acquisition and processing module obtains historical production records corresponding to all production processes in the production workshop within the selected period, and defines each production process as a production node; according to the production process sequence, it obtains the waiting time between each two adjacent production nodes, which is defined as the inter-process waiting time, summarizes the inter-process waiting time within the selected period, and constructs a set of inter-process waiting times. The feature extraction and efficiency evaluation module divides historical production data into several historical production data segments according to the corresponding production nodes based on historical production records, and constructs corresponding node datasets; it extracts the corresponding historical production data features from each node dataset, and evaluates the production efficiency of the corresponding historical production data segments based on the historical production data features. The trend feature analysis and anomaly detection module maps each element in the inter-process waiting time set to the production efficiency evaluation result of the corresponding historical production data segment, analyzes the correlation between inter-process waiting time and production efficiency evaluation result, and obtains the historical production data feature trend curve based on the correlation between inter-process waiting time and production efficiency evaluation result. The real-time production data analysis and intelligent early warning module acquires real-time production data corresponding to each production node, analyzes the real-time production data to extract real-time production data features, analyzes the real-time production data features to obtain a real-time production data feature trend curve, and compares the real-time production data feature trend curve with the historical production data feature trend curve to generate corresponding notification information.

5. The MES data intelligent management system based on data analysis according to claim 4, characterized in that: The historical production data acquisition and processing module includes a data acquisition unit, a production node definition unit, and an inter-process waiting time calculation unit. The data acquisition unit obtains historical production records corresponding to all production processes in the production workshop within the selected period from the MES system; the production node definition unit defines each production process as a production node and obtains the time range of each production node; the inter-process waiting time calculation unit calculates the waiting time between each pair of adjacent production nodes, summarizes the waiting time between each process within the selected period, and constructs an inter-process waiting time set.

6. The MES data intelligent management system based on data analysis according to claim 4, characterized in that: The feature extraction and efficiency evaluation module includes a data partitioning and feature extraction unit and an efficiency evaluation unit; The data partitioning and feature extraction unit divides the historical production data according to production nodes to form several historical production data segments and extracts the corresponding historical production data features; the efficiency evaluation unit evaluates the production efficiency of each node dataset and calculates the comprehensive production efficiency index.

7. The MES data intelligent management system based on data analysis according to claim 4, characterized in that: The trend feature analysis and anomaly detection module includes a trend feature analysis unit and an anomaly detection unit; The trend feature analysis unit analyzes the correlation between production efficiency and waiting time between processes based on historical production data features, and calculates trend feature indicators. The anomaly detection unit compares trend feature indicators with trend feature indicator thresholds to mark normal data points and abnormal data points; based on the abnormal data points, it obtains the historical production data feature trend curve.

8. The MES data intelligent management system based on data analysis according to claim 4, characterized in that: The real-time production data analysis and intelligent early warning module includes a real-time production data analysis unit and a trend comparison and early warning unit; The real-time production data analysis unit obtains the real-time production data corresponding to each production node from the MES system, analyzes the real-time production data, extracts the real-time production data features, and thus obtains the real-time production data feature trend curve. The trend comparison and early warning unit compares the real-time production data characteristic trend curve with the historical production data characteristic trend curve to generate corresponding notification information.

Images