MES-based digital visual production management system and method

Abstract

The invention discloses a digital visual production management system and method based on MES, and relates to the technical field of production equipment data analysis, the digital visual production management method comprises the following steps: establishing a distributed network architecture, collecting production process data, and ensuring that the data can be collected and transmitted to the MES in real time; the production process data comprises a production equipment state and sensor data; carrying out data preprocessing on the collected production process data; constructing an equipment running diagram by taking the preprocessed production process data as a base, identifying an abnormal state in operation of the production equipment by analyzing structural features of the equipment running diagram, and providing equipment abnormity early warning; and visual display of equipment abnormity early warning is supported. According to the invention, intelligent analysis and abnormity identification are carried out by constructing the equipment running diagram and introducing the diagram convolutional network, and real-time early warning and dynamic monitoring are provided.

Description

Technical Field

[0001] The present invention relates to the technical field of production equipment data analysis, and in particular to a digital visualization production management system and method based on MES. Background Art

[0002] The field of production equipment data analysis technology integrates the Industrial Internet of Things (IIoT), big data analysis, artificial intelligence (AI), machine learning (ML) and visualization technology, covering data collection and transmission, preprocessing and quality improvement, modeling and analysis, anomaly detection and fault diagnosis, and intelligent application.

[0003] In modern manufacturing, the operating efficiency, energy consumption and fault status of production equipment directly affect product quality and production costs; traditional equipment management methods usually rely on manual inspections and empirical judgments, and data processing relies on static analysis, lacking real-time and dynamic characteristics; existing production equipment management systems are mostly based on centralized architectures, with limited data collection coverage and lack of deep mining capabilities for multi-source heterogeneous data, making it difficult to effectively monitor equipment operating status and predict potential failures; in addition, visualization methods are mostly based on tables and simple graphical displays, making it difficult to provide intuitive dynamic monitoring and complex correlation analysis. Summary of the invention

[0004] The purpose of the present invention is to provide a digital visual production management system and method based on MES to solve the problems raised in the prior art.

[0005] To achieve the above object, the present invention provides the following technical solution: a digital visualization production management method based on MES, comprising: Establish a distributed network architecture to collect production process data and ensure that the data can be collected and transmitted to the MES system in real time; the production process data includes production equipment status and sensor data; Perform data preprocessing on the collected production process data; Based on the pre-processed production process data, the equipment operation diagram is constructed. By analyzing the structural characteristics of the equipment operation diagram, the abnormal state of the production equipment operation is identified and the equipment abnormality warning is provided; Supports visual display of equipment abnormality warnings.

[0006] Establish a distributed network architecture to collect production process data and ensure that the data can be collected and transmitted to the MES system in real time; the production process data includes production equipment status and sensor data, specifically: Establish a distributed network architecture, install sensors, PLCs and SCADA data acquisition equipment, realize distributed data access of production equipment status and sensor data, and transmit them to the MES system; Integrate a central database in the MES system to store received production equipment status and sensor data; The production equipment status includes operating status, production efficiency, failure rate and equipment switch status; The sensor data includes temperature, pressure, vibration and current; Among them, temperature includes internal temperature of the equipment, surface temperature and ambient temperature; Pressure includes hydraulic and pneumatic systems; Vibration is used for equipment fault detection; Current and voltage are used to monitor equipment energy consumption; Vibration data is collected from accelerometers, temperature is obtained from thermocouples, and current is monitored through current sensors.

[0007] The collected production process data is preprocessed, specifically: Identify outliers by calculating statistical indicators of the data; Among them, the outlier processing method includes: Time series data is filled by the previous and next values; Non-time series data are filled in using mean, mode, and median; Data denoising using wavelet transform; Data normalization was performed using Z-score; Set the time step alignment timestamp to align the time series data of different production equipment in chronological order.

[0008] The time step alignment timestamp alignment method is to align according to the same working time of different devices.

[0009] Based on the pre-processed production process data, the equipment operation diagram model is constructed, specifically: The device operation diagram is constructed in the following manner: The nodes in the equipment operation diagram include production equipment and sensors on the production equipment; the node attributes are composed of production equipment status and sensor data; Each node represents a production device or a sensor on the production device, and the node attributes consist of the production device status or sensor data.

[0010] Wherein, when the node is a production device, the node attribute represents all sensor data of the production device; when the node represents a sensor on the device, the node attribute represents the data of the sensor; The node attributes are specifically: ;x j ∈{x1,x2,…,x m}; Among them, x j is the measurement value of the jth sensor on the production equipment corresponding to node i; m represents the total number of sensors on the production equipment, m is a positive integer, R represents a real number set; i represents the node label, i is a positive integer; j represents the sensor number, j is a positive integer, 1≤j≤m; v i Represents the node attribute list of the i-th node; The edges in the equipment operation graph represent the interactions and dependencies between nodes, which are determined in the following way: Create edges based on temporal similarity relationships: Get the time series data list Q for each production equipment c =[Q c (1),Q c (2),…,Q c (T)]; where c represents the production equipment number and is a positive integer; Q c represents the time series data list of the production equipment numbered c; T represents the time step; Q c (T) represents the time series data corresponding to the production equipment numbered c at time step T; Calculate the correlation coefficient: ; Among them, P c,v represents the correlation coefficient between the production equipment numbered c and the production equipment numbered v; cov(Q c ,Q v ) represents the covariance between the time series data list of the production equipment numbered c and the time series data list of the production equipment numbered v; σ(Q c ) represents the standard deviation of the data in the time series data list of the production equipment numbered c; σ(Q v ) represents the standard deviation of the data in the time series data list of the production equipment numbered v; Similarity is calculated using Dynamic Time Warping (DTW): ; Among them, DTW(Q c ,Q v ) represents the time series data list Q c With the time series data list Q v The similarity between them; B represents the sum of the aligned time steps, B is a positive integer; b represents the aligned time step label, b is a positive integer, 1≤b≤B; Set the correlation coefficient threshold φ; The calculated correlation coefficient is judged according to the correlation coefficient threshold, the edge in the equipment operation graph is determined, and the similarity size is determined as the weight of the edge.

[0011] The correlation coefficient threshold φ is set based on experience and can be the average value of the covariance between the time series data lists of all production equipment in history.

[0012] Among them, no edge is established between production equipment or sensor data with a correlation coefficient less than the threshold φ.

[0013] By analyzing the structural characteristics of the equipment operation diagram, the abnormal state of the production equipment can be identified and an early warning of equipment abnormality can be provided, specifically: The analysis is performed through the graph convolutional network GCN, specifically: According to the constructed equipment operation graph, obtain the adjacency matrix and the initial node attribute matrix; The adjacency matrix A∈R N×N ; Where N represents the number of nodes, and N is a positive integer; The initial node attribute matrix X∈R N×m ; Wherein, each data in the adjacency matrix A indicates whether there is an edge between the production equipment of the corresponding two nodes; Each data in the initial node attribute matrix X represents the device status and sensor data of the corresponding production equipment; The feature update process of the GCN is specifically as follows: Step 1, feature aggregation: use the normalized adjacency matrix to perform weighted aggregation on node features; Step 2, feature transformation: Use a learnable weight matrix to linearly transform the features and learn higher-dimensional feature representations; Step 3, nonlinear activation: introduce nonlinearity through activation function to increase the model's expressiveness; Among them, the formula of the feature update process of the GCN is expressed as: ; Among them, H (l) represents the node attribute matrix of the lth layer; W (l) represents the learnable weight matrix of the lth layer; A* represents the normalized adjacency matrix, which is used to propagate the neighbor information of the node; σ represents the nonlinear activation function; H (l+1) Represents the node attribute matrix of the l+1th layer; Wherein, the nonlinear activation function may be ReLU(σ(x)=max(0,x)); wherein x represents feature update process data; Among them, H (0) =X, X represents the initial node attribute matrix; In the adjacency matrix, 1 indicates that there is an edge between nodes, and 0 indicates that there is no edge between nodes; Obtain the time series data list of production equipment in real time and calculate the abnormal status value of production equipment, specifically: ; Among them, y i represents the abnormal state value of the production equipment of the i-th node; C(v i ) is the clustering coefficient of node i; e represents the total number of neighbor nodes of node i; i' represents a neighbor node of node i; ΔDTW(i,i') represents the difference in similarity between the time series data list of production equipment corresponding to node i and the time series data list of production equipment corresponding to node i'; Set abnormal status thresholds for production equipment; The abnormal state threshold of the production equipment is set according to historical data, and can be set to the average value of the change difference of the similarity between the time series data lists of all abnormal production equipment within a period of time.

[0014] Compare the abnormal status value of the production equipment calculated in real time with the abnormal status threshold of the production equipment, issue an early warning for the production equipment whose value is greater than the abnormal status threshold of the production equipment, and arrange corresponding personnel to conduct inspections; Calculate the clustering coefficient: C(v i )=(2*s) / K; where C(v i ) is the clustering coefficient of node i, which indicates the importance of node i in the equipment operation graph; s indicates the degree of node i; K indicates the total number of edges in the equipment operation graph.

[0015] Specifically: The normalized adjacency matrix A* is obtained by normalizing the node degree; The calculation method of the normalized adjacency matrix A* is specifically as follows: ; Where A* represents the normalized adjacency matrix; D represents the node degree matrix; ; Among them, E represents the identity matrix; A' represents the self-connected adjacency matrix; A represents the adjacency matrix.

[0016] Supports visual display of equipment abnormality warnings, specifically: Mark the real-time node attributes of each node on the equipment operation diagram; Update the weight changes of edges in the equipment operation graph in real time; Highlight abnormal devices; Support playback viewing.

[0017] A digital visual production management system based on MES, the digital visual production management system includes a distributed network architecture module, a data processing and storage module, a data analysis module, an anomaly detection module and a visual display module; The distributed network architecture module is used to build a distributed network architecture and collect production equipment status and sensor data; Pre-process the collected production process data; The data analysis module is used to construct an equipment operation diagram based on the preprocessed production process data and analyze the structural characteristics of the equipment operation diagram; The anomaly detection module is used to analyze the adjacency matrix and node attribute matrix of the equipment operation graph through a graph convolution network to calculate the abnormal state value of the production equipment; The visualization display module is used to judge the abnormal state value of the production equipment according to the set abnormal state threshold of the production equipment, trigger an early warning for the equipment exceeding the threshold; and provide a user interaction interface to support visualization display.

[0018] Specifically: Real-time collection of production equipment status and sensor data, including operating status, production efficiency, failure rate, switch status, and temperature, pressure, vibration and current sensor data; Build a distributed network through sensors, PLC and SCADA equipment to achieve distributed data collection; The collected production process data is transmitted to the MES system in real time through a distributed network, providing basic data support for subsequent processing.

[0019] Specifically: Update node attributes and edge weight changes in real time to reflect the dynamic evolution of device status; Summarize abnormal status data, generate early warning reports, and assist managers in making decisions; The visual display provides playback.

[0020] Compared with the prior art, the beneficial effects of the present invention are as follows: the present invention realizes real-time collection and transmission of production equipment status and sensor data by establishing a distributed network architecture, improves data quality by using unified data preprocessing, and provides real-time early warning and dynamic monitoring by constructing an equipment operation diagram and introducing a graph convolutional network (GCN) for intelligent analysis and anomaly identification; combined with dynamic time warping (DTW) and correlation analysis, it reveals the dependency relationship and global operation mode between devices, supports real-time updating and visual display of equipment operation diagrams, intuitively presents equipment status and abnormal information, significantly improves the intelligence, real-time and reliability of production equipment management, reduces equipment failure rate and maintenance cost, and provides efficient and reliable technical support for intelligent manufacturing. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 A schematic diagram of the steps of a digital visualization production management method based on MES of the present invention; Figure 2 The present invention is a module schematic diagram of a digital visual production management system based on MES. DETAILED DESCRIPTION

[0022] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0023] Example: Figure 1-Figure 2 As shown, the present invention provides a technical solution, such as Figure 1 As shown, a digital visualization production management method based on MES includes: Establish a distributed network architecture to collect production process data and ensure that the data can be collected and transmitted to the MES system in real time; the production process data includes production equipment status and sensor data; Perform data preprocessing on the collected production process data; Based on the pre-processed production process data, the equipment operation diagram is constructed. By analyzing the structural characteristics of the equipment operation diagram, the abnormal state of the production equipment operation is identified and the equipment abnormality warning is provided; Supports visual display of equipment abnormality warnings.

[0024] Establish a distributed network architecture to collect production process data and ensure that the data can be collected and transmitted to the MES system in real time; the production process data includes production equipment status and sensor data, specifically: Establish a distributed network architecture, install sensors, PLCs and SCADA data acquisition equipment, realize distributed data access of production equipment status and sensor data, and transmit them to the MES system; Integrate a central database in the MES system to store received production equipment status and sensor data; The production equipment status includes operating status, production efficiency, failure rate and equipment switch status; The sensor data includes temperature, pressure, vibration and current; Among them, temperature includes internal temperature of the equipment, surface temperature and ambient temperature; Pressure includes hydraulic and pneumatic systems; Vibration is used for equipment fault detection; Current and voltage are used to monitor equipment energy consumption; Vibration data is collected from accelerometers, temperature is obtained from thermocouples, and current is monitored through current sensors.

[0025] The collected production process data is preprocessed, specifically: Identify outliers by calculating statistical indicators of the data; Among them, the outlier processing method includes: Time series data is filled by the previous and next values; Non-time series data are filled in using mean, mode, and median; Data denoising using wavelet transform; Data normalization was performed using Z-score; Set the time step alignment timestamp to align the time series data of different production equipment in chronological order.

[0026] The time step alignment timestamp alignment method is to align according to the same working time of different devices.

[0027] Based on the pre-processed production process data, the equipment operation diagram model is constructed, specifically: The device operation diagram is constructed in the following manner: The nodes in the equipment operation diagram include production equipment and sensors on the production equipment; the node attributes are composed of production equipment status and sensor data; Each node represents a production device or a sensor on the production device, and the node attributes consist of the production device status or sensor data.

[0028] Wherein, when the node is a production device, the node attribute represents all sensor data of the production device; when the node represents a sensor on the device, the node attribute represents the data of the sensor; The node attributes are specifically: ;x j ∈{x1,x2,…,x m}; Among them, x j is the measurement value of the jth sensor on the production equipment corresponding to node i; m represents the total number of sensors on the production equipment, m is a positive integer, R represents a real number set; i represents the node label, i is a positive integer; j represents the sensor number, j is a positive integer, 1≤j≤m; v i Represents the node attribute list of the i-th node; The edges in the equipment operation graph represent the interactions and dependencies between nodes, which are determined in the following way: Create edges based on temporal similarity relationships: Get the time series data list Q for each production equipment c =[Q c (1),Q c (2),…,Q c (T)]; where c represents the production equipment number and is a positive integer; Q c represents the time series data list of the production equipment numbered c; T represents the time step; Q c (T) represents the time series data corresponding to the production equipment numbered c at time step T; Calculate the correlation coefficient: ; Among them, P c,v represents the correlation coefficient between the production equipment numbered c and the production equipment numbered v; cov(Q c ,Q v ) represents the covariance between the time series data list of the production equipment numbered c and the time series data list of the production equipment numbered v; σ(Q c ) represents the standard deviation of the data in the time series data list of the production equipment numbered c; σ(Q v ) represents the standard deviation of the data in the time series data list of the production equipment numbered v; The current time series data of two production equipment c and v are as follows: Real-time current sampling value of device c: Q c =[1.05,1.08,1.03,1.10,1.07,1.04]; Real-time current sampling value of device v: Q v =[1.02,1.06,1.03,1.09,1.05,1.02]; Mean: μ(Q c )=(1.05+1.08+1.03+1.10+1.07+1.04) / 6≈1.062; μ(Q v )=1.045; Standard deviation: σ(Q c )≈0.023; σ(Q v )≈0.025; Covariance: cov(Q c ,Q v )=Σ[(Q c -μ(Qc ))*(Q v -μ(Q v ))] / 6≈0.005583; Correlation coefficient: P c,v =0.005583 / 0.023*0.025≈0.97; Similarity is calculated using Dynamic Time Warping (DTW): ; Among them, DTW(Q c ,Q v ) represents the time series data list Q c With the time series data list Q v The similarity between them; B represents the total number of aligned time steps, B is a positive integer; b represents the aligned time step label, b is a positive integer, 1≤b≤B; Set the correlation coefficient threshold φ; The calculated correlation coefficient is judged according to the correlation coefficient threshold, the edge in the equipment operation graph is determined, and the similarity size is determined as the weight of the edge.

[0029] The correlation coefficient threshold φ is set based on experience and can be the average value of the covariance between the time series data lists of all production equipment in history.

[0030] Among them, no edge is established between production equipment or sensor data with a correlation coefficient less than the threshold φ.

[0031] By analyzing the structural characteristics of the equipment operation diagram, the abnormal state of the production equipment can be identified and an early warning of equipment abnormality can be provided, specifically: The analysis is performed through the graph convolutional network GCN, specifically: According to the constructed equipment operation graph, obtain the adjacency matrix and the initial node attribute matrix; The adjacency matrix A∈R N×N ; Where N represents the number of nodes, and N is a positive integer; The initial node attribute matrix X∈R N×m ; Wherein, each data in the adjacency matrix A indicates whether there is an edge between the production equipment of the corresponding two nodes; Each data in the initial node attribute matrix X represents the device status and sensor data of the corresponding production equipment; The feature update process of the GCN is specifically as follows: Step 1, feature aggregation: use the normalized adjacency matrix to perform weighted aggregation on node features; Step 2, feature transformation: Use a learnable weight matrix to linearly transform the features and learn higher-dimensional feature representations; Step 3, nonlinear activation: introduce nonlinearity through activation function to increase the model's expressiveness; Among them, the formula of the feature update process of the GCN is expressed as: ; Among them, H (l) represents the node attribute matrix of the lth layer; W (l) represents the learnable weight matrix of the lth layer; A* represents the normalized adjacency matrix, which is used to propagate the neighbor information of the node; σ represents the nonlinear activation function; H (l+1) Represents the node attribute matrix of the l+1th layer; Wherein, the nonlinear activation function may be ReLU(σ(x)=max(0,x)); wherein x represents feature update process data; Among them, H (0) =X, X represents the initial node attribute matrix; In the adjacency matrix, 1 indicates that there is an edge between nodes, and 0 indicates that there is no edge between nodes; Obtain the time series data list of production equipment in real time and calculate the abnormal status value of production equipment, specifically: ; Among them, y i represents the abnormal state value of the production equipment of the i-th node; C(v i ) is the clustering coefficient of node i; e represents the total number of neighbor nodes of node i; i' represents a neighbor node of node i; ΔDTW(i,i') represents the difference in similarity between the time series data list of production equipment corresponding to node i and the time series data list of production equipment corresponding to node i'; Set abnormal status thresholds for production equipment; The abnormal state threshold of the production equipment is set according to historical data, and can be set to the average value of the change difference of the similarity between the time series data lists of all abnormal production equipment within a period of time.

[0032] Compare the abnormal status value of the production equipment calculated in real time with the abnormal status threshold of the production equipment, issue an early warning for the production equipment whose value is greater than the abnormal status threshold of the production equipment, and arrange corresponding personnel to conduct inspections; Calculate the clustering coefficient: C(v i )=(2*s) / K; where C(v i ) is the clustering coefficient of node i, which indicates the importance of node i in the equipment operation graph; s indicates the degree of node i; K indicates the total number of edges in the equipment operation graph.

[0033] Specifically: The normalized adjacency matrix A* is obtained by normalizing the node degree; The calculation method of the normalized adjacency matrix A* is specifically as follows: ; Where A* represents the normalized adjacency matrix; D represents the node degree matrix; ; Among them, E represents the identity matrix; A' represents the self-connected adjacency matrix; A represents the adjacency matrix.

[0034] Supports visual display of equipment abnormality warnings, specifically: Mark the real-time node attributes of each node on the equipment operation diagram; Update the weight changes of edges in the equipment operation graph in real time; Highlight abnormal devices; Support playback viewing.

[0035] like Figure 2 As shown, a digital visualization production management system based on MES, the digital visualization production management system includes a distributed network architecture module, a data processing and storage module, a data analysis module, an anomaly detection module and a visualization display module; The distributed network architecture module is used to build a distributed network architecture and collect production equipment status and sensor data; Pre-process the collected production process data; The data analysis module is used to construct an equipment operation diagram based on the preprocessed production process data and analyze the structural characteristics of the equipment operation diagram; The anomaly detection module is used to analyze the adjacency matrix and node attribute matrix of the equipment operation graph through a graph convolution network to calculate the abnormal state value of the production equipment; The visualization display module is used to judge the abnormal state value of the production equipment according to the set abnormal state threshold of the production equipment, trigger an early warning for the equipment exceeding the threshold; and provide a user interaction interface to support visualization display.

[0036] Specifically: Real-time collection of production equipment status and sensor data, including operating status, production efficiency, failure rate, switch status, and temperature, pressure, vibration and current sensor data; Build a distributed network through sensors, PLC and SCADA equipment to achieve distributed data collection; The collected production process data is transmitted to the MES system in real time through a distributed network, providing basic data support for subsequent processing.

[0037] Specifically: Update node attributes and edge weight changes in real time to reflect the dynamic evolution of device status; Summarize abnormal status data, generate early warning reports, and assist managers in making decisions; The visual display provides playback.

[0038] It will be apparent to those skilled in the art that the invention is not limited to the details of the exemplary embodiments described above and that the invention can be implemented in other specific forms without departing from the spirit or essential features of the invention. Therefore, the embodiments should be considered exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description, and it is intended that all variations falling within the meaning and scope of the equivalent elements of the claims be included in the invention. Any reference numeral in a claim should not be considered as limiting the claim to which it relates.

Claims

1. A digital visualization production management method based on MES, characterized by: include: Establish a distributed network architecture to collect production process data; The production process data includes production equipment status and sensor data; Perform data preprocessing on the collected production process data; Based on the pre-processed production process data, the equipment operation diagram is constructed. By analyzing the structural characteristics of the equipment operation diagram, the abnormal state of the production equipment operation is identified and the equipment abnormality warning is provided; Supports visual display of equipment abnormality warnings.

2. According to claim 1, a digital visualization production management method based on MES is characterized in that: Establish a distributed network architecture to collect production process data and ensure that the data can be collected and transmitted to the MES system in real time; the production process data includes production equipment status and sensor data, specifically: Establish a distributed network architecture, install sensors, PLCs and SCADA data acquisition equipment, realize distributed data access of production equipment status and sensor data, and transmit them to the MES system; Integrate a central database in the MES system to store received production equipment status and sensor data; The production equipment status includes operating status, production efficiency, failure rate and equipment switch status; The sensor data includes temperature, pressure, vibration and current.

3. According to claim 2, a digital visualization production management method based on MES is characterized in that: The collected production process data is preprocessed, specifically: Time series data is filled by the previous and next values; Non-time series data are filled in using mean, mode, and median; Data denoising using wavelet transform; Data normalization was performed using Z-score; Set the time step alignment timestamp to align the time series data of different production equipment in chronological order.

4. The digital visualization production management method based on MES according to claim 3 is characterized in that: Based on the pre-processed production process data, the equipment operation diagram model is constructed, specifically: The device operation diagram is constructed in the following manner: The nodes in the equipment operation diagram include production equipment and sensors on the production equipment; the node attributes are composed of production equipment status and sensor data; The node attributes are specifically: ;x j ∈{x1,x2,…,x m }; Among them, x j is the measurement value of the jth sensor on the production equipment corresponding to node i; m represents the total number of sensors on the production equipment, m is a positive integer, R represents a real number set; i represents the node label, i is a positive integer; j represents the sensor number, j is a positive integer, 1≤j≤m; v i Represents the node attribute list of the i-th node; The edges in the equipment operation graph represent the interactions and dependencies between nodes, which are determined in the following way: Create edges based on temporal similarity relationships: Get the time series data list Q for each production equipment c =[Q c (1),Q c (2),…,Q c (T)]; where c represents the production equipment number and is a positive integer; Q c represents the time series data list of the production equipment numbered c; T represents the time step; Q c (T) represents the time series data corresponding to the production equipment numbered c at time step T; The correlation coefficient between production equipment is calculated through the covariance and standard deviation of the time series data between production equipment; Similarity is calculated using Dynamic Time Warping (DTW): Set the correlation coefficient threshold φ; The calculated correlation coefficient is judged according to the correlation coefficient threshold, the edge in the equipment operation graph is determined, and the similarity size is determined as the weight of the edge.

5. The digital visualization production management method based on MES according to claim 4 is characterized in that: By analyzing the structural characteristics of the equipment operation diagram, the abnormal state of the production equipment can be identified and an early warning of equipment abnormality can be provided, specifically: The analysis is performed through the graph convolutional network GCN, specifically: According to the constructed equipment operation graph, obtain the adjacency matrix and the initial node attribute matrix; The adjacency matrix A∈R N×N ; Where N represents the number of nodes, and N is a positive integer; The initial node attribute matrix X∈R N×m ; The feature update process of the GCN is specifically as follows: Step 1, feature aggregation: use the normalized adjacency matrix to perform weighted aggregation on node features; Step 2, feature transformation: Use a learnable weight matrix to linearly transform the features and learn higher-dimensional feature representations; Step 3, nonlinear activation: introduce nonlinearity through activation function to increase the model's expressiveness; Obtain the time series data list of production equipment in real time, and calculate the abnormal state value of production equipment according to the change difference of similarity between the time series data lists of production equipment; Set abnormal status thresholds for production equipment; The real-time calculated abnormal status value of the production equipment is compared with the abnormal status threshold of the production equipment, an early warning is issued for the production equipment whose value is greater than the abnormal status threshold of the production equipment, and corresponding personnel are arranged to conduct inspections.

6. The digital visualization production management method based on MES according to claim 5 is characterized in that: Specifically: The normalized adjacency matrix A* is obtained by normalizing the node degree; The calculation method of the normalized adjacency matrix A* is specifically as follows: ； Where A* represents the normalized adjacency matrix; D represents the node degree matrix; ； Among them, E represents the identity matrix; A' represents the self-connected adjacency matrix; A represents the adjacency matrix.

7. The digital visualization production management method based on MES according to claim 6 is characterized in that: Supports visual display of equipment abnormality warnings, specifically: Mark the real-time node attributes of each node on the equipment operation diagram; Update the weight changes of edges in the equipment operation graph in real time; The abnormal devices are highlighted.

8. A digital visualization production management system based on MES, applying a digital visualization production management method based on MES according to any one of claims 1 to 7, characterized in that: The digital visual production management system includes a distributed network architecture module, a data processing and storage module, a data analysis module, an anomaly detection module and a visual display module; The distributed network architecture module is used to construct a distributed network architecture and collect production process data; the production process data includes production equipment status and sensor data; Pre-process the collected production process data; The data analysis module is used to construct an equipment operation diagram based on the preprocessed production process data and analyze the structural characteristics of the equipment operation diagram; The anomaly detection module is used to analyze the adjacency matrix and node attribute matrix of the equipment operation graph through a graph convolution network to calculate the abnormal state value of the production equipment; The visual display module is used to judge the abnormal state value of the production equipment according to the set abnormal state threshold of the production equipment, and trigger an early warning for the equipment exceeding the threshold; It also provides a user interaction interface and supports visual display.

9. The MES-based digital visual production management system according to claim 8, characterized in that: Specifically: Real-time collection of production equipment status and sensor data, including operating status, production efficiency, failure rate, switch status, and temperature, pressure, vibration and current sensor data; Build a distributed network through sensors, PLC and SCADA equipment to achieve distributed data collection; The collected production process data is transmitted to the MES system in real time through a distributed network, providing basic data support for subsequent processing.

10. The MES-based digital visual production management system according to claim 8, characterized in that: Specifically: Update node attributes and edge weight changes in real time to reflect the dynamic evolution of device status; Summarize abnormal status data, generate early warning reports, and assist managers in making decisions; The visual display provides playback.

Images