Tobacco shred production online monitoring method and system based on edge calculation

Abstract

The invention discloses a cut tobacco production online monitoring method and system based on edge calculation. The method comprises the following steps: receiving real-time data of cut tobacco production metering equipment; performing edge data analysis on the real-time data by using the data processing model, and judging whether the working state of the tobacco shred production metering equipment is abnormal or not and the fault type of the tobacco shred production metering equipment; if the tobacco shred production metering equipment has a fault, giving an alarm corresponding to the fault type; wherein parameters of the data processing model are iterated and optimized regularly or as required. According to the invention, edge data analysis is carried out on the real-time data of the tobacco shred production metering equipment by using the continuously iterated and optimized data processing model in the newly added edge computing equipment, so that the monitoring precision of various abnormal conditions in the tobacco shred production process is effectively improved, and the accuracy of data analysis is ensured; and the response speed of the system is improved, problems can be found and processed in time in the production process, and the user experience is improved.

Description

Technical Field

[0001] The present application relates to the technical field of cigarette production, and more specifically, to an online monitoring method and system for tobacco production based on edge computing. Background Art

[0002] Tobacco processing is a complex process involving many key process steps, including shredding, humidification, drying and flavoring. Precise control of these steps is crucial to ensure the quality of the final product. Traditionally, these process steps rely on the operator's experience and manual adjustments. This dependence makes the production process susceptible to human errors. At the same time, it is difficult to make real-time adjustments when equipment performance fluctuates or the environment changes, thus affecting the consistency and stability of product quality.

[0003] In recent years, although some tobacco companies have introduced automation technologies such as PLC control systems to improve control accuracy and stability, these systems mainly operate based on fixed procedures and parameters and lack in-depth analysis of real-time production data and adaptive learning capabilities. Summary of the invention

[0004] The present application provides an online monitoring method and system for tobacco production based on edge computing, which uses a continuously iterated and optimized data processing model in a newly added edge computing device to perform edge data analysis on the real-time data of tobacco production metering equipment, effectively improving the monitoring accuracy of various abnormal conditions in the tobacco production process and ensuring the accuracy of data analysis; and migrating computing and storage functions from centralized cloud computing to the edge of the network. The edge computing device is directly connected to the tobacco production metering equipment, realizing real-time data collection and real-time processing, greatly shortening the time for data transmission and processing, and improving the response speed of the system. Problems can be discovered and handled in a timely manner during the production process, thereby improving user experience.

[0005] The present application provides an online monitoring method for tobacco production based on edge computing, including:

[0006] Receive real-time data from tobacco production metering equipment;

[0007] Use the data processing model to perform edge data analysis on real-time data to determine whether the working status of tobacco production and metering equipment is abnormal and the type of faults that exist in tobacco production and metering equipment;

[0008] If there is a fault in the tobacco production metering equipment, an alarm corresponding to the type of fault is issued;

[0009] Among them, the parameters of the data processing model are iterated and optimized regularly or as needed.

[0010] Preferably, the method for online monitoring of tobacco production based on edge computing further includes:

[0011] Store the calculation results of edge data analysis;

[0012] Provides the corresponding edge data analysis calculation results based on the query or retrieval request.

[0013] Preferably, the method for online monitoring of tobacco production based on edge computing further includes:

[0014] Synchronize the computation results of real-time data and / or edge data analysis to the cloud server.

[0015] Preferably, the method for online monitoring of tobacco production based on edge computing further includes:

[0016] Update the parameters of the data processing model from the cloud server periodically or as needed.

[0017] Preferably, the edge data analysis includes analyzing whether the working state of the electronic belt scale on the tobacco production line is abnormal and the type of fault that exists, and analyzing whether there is an abnormality in the moisture meter on the tobacco production line.

[0018] The present application also provides an online monitoring system for tobacco production based on edge computing, including an edge computing device and a tobacco production metering device on a tobacco production line;

[0019] The edge computing device includes a main control unit and a communication unit, and the main control unit communicates with the tobacco production metering device through the communication unit;

[0020] The main control unit includes a data receiving module, an edge data analysis module and an output module;

[0021] The data receiving module is used to receive the real-time data of the tobacco production metering equipment;

[0022] The edge data analysis module is used to perform edge data analysis on real-time data using a data processing model to determine whether the working state of the tobacco production metering equipment is abnormal and the type of faults that exist in the tobacco production metering equipment;

[0023] The output module is used to issue an alarm corresponding to the fault type when there is a fault in the tobacco production metering equipment;

[0024] Among them, the parameters of the data processing model are iterated and optimized regularly or as needed.

[0025] Preferably, the edge computing device further comprises a storage unit, the storage unit being used to store the calculation results of the edge data analysis;

[0026] The main control unit also includes a query unit, which is used to provide corresponding edge data analysis calculation results based on a query or retrieval request.

[0027] Preferably, the tobacco production online monitoring system further includes a cloud server;

[0028] The main control unit also includes an edge-cloud collaborative service unit, which is used to synchronize the calculation results of real-time data and / or edge data analysis to the cloud server.

[0029] Preferably, the edge-cloud collaborative service unit is also used to update the parameters of the data processing model from the cloud server regularly or as needed.

[0030] Preferably, the edge data analysis module is used to analyze whether the working state of the electronic belt scale on the tobacco production line is abnormal and the type of fault that exists, and to analyze whether there is an abnormality in the moisture meter on the tobacco production line.

[0031] Other features and advantages of the present application will become apparent from the following detailed description of exemplary embodiments of the present application with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] The accompanying drawings, which are incorporated in and constitute a part of the specification, illustrate embodiments of the application and, together with the description, serve to explain the principles of the application.

[0033] Figure 1 A flow chart of the online monitoring method for tobacco production based on edge computing provided in this application;

[0034] Figure 2 Structural diagram of the online monitoring system for tobacco production based on edge computing provided in this application. DETAILED DESCRIPTION

[0035] Various exemplary embodiments of the present application will now be described in detail with reference to the accompanying drawings. It should be noted that unless otherwise specifically stated, the relative arrangement of components and steps, numerical expressions and numerical values ​​set forth in these embodiments do not limit the scope of the present application.

[0036] The following description of at least one exemplary embodiment is merely illustrative in nature and is in no way intended to limit the present application, its application, or uses.

[0037] Technologies, methods, and equipment known to ordinary technicians in the relevant art may not be discussed in detail, but where appropriate, the technologies, methods, and equipment should be considered part of the specification.

[0038] In all examples shown and discussed herein, any specific values ​​should be interpreted as merely exemplary and not as limiting. Therefore, other examples of the exemplary embodiments may have different values.

[0039] The present application provides an online monitoring method and system for tobacco production based on edge computing, which uses the continuously iterated and optimized data processing model in the newly added edge computing device to perform edge data analysis on the real-time data of the tobacco production metering equipment, effectively improving the monitoring accuracy of various abnormal situations in the tobacco production process and ensuring the accuracy of data analysis; and the computing and storage functions are migrated from centralized cloud computing to the edge of the network, and the edge computing device is directly connected to the tobacco production metering equipment, realizing real-time data collection and real-time processing, greatly shortening the time for data transmission and processing, improving the response speed of the system, and being able to timely discover and handle problems during the production process, improving the user experience. In addition, the edge computing device can synchronize data with the cloud server and update the model through the cloud server, ensuring the stable operation and continuous optimization of the system in different environments, and enhancing the robustness of anomaly detection and the reliability of the system.

[0040] It should be noted that edge computing, as a technology that allows data to be processed near the source of generation, provides a unique solution that can realize real-time, localized processing and analysis of large amounts of data generated during the tobacco production process. By applying edge computing technology to tobacco production lines, data processing processes can be optimized, data transmission delays can be reduced, and the response speed and accuracy to changes in the production process can be improved, thereby more effectively controlling tobacco quality.

[0041] like Figure 1 As shown, the online monitoring method for tobacco production based on edge computing provided by the present application includes:

[0042] S110: Receive real-time data from tobacco production metering equipment.

[0043] Specifically, the edge computing device is connected to the tobacco production metering equipment (such as electronic belt scales, moisture meters, etc.) deployed on the tobacco production line through a high-speed network, and can directly obtain real-time data from the tobacco production metering equipment.

[0044] Real-time data can include quality indicators such as weight, moisture, temperature, and production control parameters such as drum wall temperature, hot air temperature and hot air speed, and moisture content of export materials.

[0045] S120: Use the data processing model to perform edge data analysis on the real-time data to determine whether the working state of the tobacco production metering equipment is abnormal and the type of fault existing in the tobacco production metering equipment.

[0046] Among them, the parameters of the data processing model are iterated and optimized regularly or as needed to improve the accuracy and response speed of detection.

[0047] It can be understood that when edge computing devices receive real-time data, they preprocess, filter, and compress the data before data analysis to remove noise and outliers and reduce data transmission volume and latency.

[0048] As an embodiment, edge data analysis includes real-time abnormal monitoring of electronic belt scales and moisture meters on tobacco production lines, including analyzing whether the working status of the electronic belt scales on the tobacco production line is abnormal (i.e., fault monitoring) and the type of faults that exist (i.e., fault diagnosis), as well as analyzing whether there are abnormalities in the moisture meters on the tobacco production line.

[0049] As an embodiment, analyzing whether the working state of the electronic belt scale is abnormal specifically includes:

[0050] P1: Perform cluster analysis on the real-time data of the electronic belt scale to obtain clusters of different types of data points.

[0051] P2: If the difference between the cluster and the normal cluster is greater than the threshold, the data points in the cluster are abnormal.

[0052] As an embodiment, the fault detection of electronic belt scales is realized by using the DENCLUE clustering model. The DENCLUE clustering algorithm is a density-based clustering algorithm. By dividing the data space into grids and determining the cluster center according to the density of data points, it can automatically identify clusters of different shapes, sizes and densities. DENCLUE can effectively process high-dimensional data and has good robustness to isolated abnormal points.

[0053] Specifically, the DENCLUE clustering algorithm estimates the density of data points and divides the data points into core points, edge points and noise points. In the tobacco production process, normal process data is extracted and clustered using the DENCLUE clustering algorithm to obtain different normal clusters.

[0054] For the newly collected real-time data of the electronic belt scale, the DENCLUE clustering model is used for cluster analysis to obtain clusters of different types of data points. If the difference between a cluster and a normal cluster is less than or equal to the threshold, it means that the current real-time data of the electronic belt scale is normal; otherwise, it means that there is a fault in the current real-time data of the electronic belt scale, such as the tobacco weight error exceeds the threshold or abnormal operation. In this way, the staff can take timely measures to repair the equipment failure and ensure the accuracy and stability of tobacco production.

[0055] When diagnosing the fault of the electronic belt scale, if there is an abnormality in the data points in the cluster, the data points in the cluster are input into the classification model, the data points are classified, and the fault type of the electronic belt scale is obtained.

[0056] As an embodiment, the BTSVM (Binary Tree Support Vector Machine) classification model is used to diagnose the fault of the electronic belt scale. The BTSVM classification algorithm is a binary tree-based support vector machine classification algorithm. It improves the classification effect by gradually dividing the data set into subsets and training the support vector machine classifier on each subset. BTSVM can effectively process large-scale complex data and has good classification performance and generalization ability.

[0057] The trained BTSVM classification model is used to identify multiple fault types. During training, the data of normal sensors (such as electronic belt scales), various on-site environmental data, and original data are generally labeled and used as fault sample features to train the BTSVM classification model and map different types of faults to different categories.

[0058] When the online fault detection of the electronic belt scale finds a fault, the features of the current real-time data of the electronic belt scale are extracted and input into the trained BTSVM classification model for classification and identification to determine the specific type of the fault.

[0059] As an embodiment, analyzing whether the working state of the moisture meter is abnormal specifically includes: predicting key parameters of the tobacco by using a long short-term memory network (LSTM) model based on real-time data of the moisture meter; and outputting the predicted key parameters of the tobacco.

[0060] The LSTM deep learning model is a variant of the recurrent neural network (RNN) that is specifically designed to process sequence data. The LSTM model can remember long-term dependencies and has strong time series modeling capabilities. It has been widely used in natural language processing, speech recognition and other fields.

[0061] In the training stage of the LSTM deep learning model, data such as the instantaneous current of the electronic belt scale, the detection value of the moisture meter and the production operation signal in the corresponding process are collected and labeled as input data. The anomaly detection model is trained based on LSTM. The LSTM model can learn the rules and trends in the tobacco production process and make predictions based on the current data status.

[0062] In the real-time anomaly detection stage, online data (including the current current of the belt electronic scale, the detection value of the moisture meter, and production operation signals) are input into the trained LSTM model, and the output of the model is used to predict whether the status of the moisture meter equipment is abnormal. The LSTM model can predict key parameters such as the humidity and weight of the tobacco, as well as possible abnormal situations. In this way, the staff can adjust the production parameters in time to avoid unnecessary waste and quality problems.

[0063] As another embodiment, the BTSVM classification model can also be used to classify the moisture meter data to determine whether the working state of the moisture meter is abnormal. By analyzing the data collected by the moisture meter, the BTSVM classification model can classify the humidity of the tobacco and determine whether it is within the normal range. If the moisture meter detects an abnormal humidity value, the BTSVM classification model can issue an alarm in time to prompt the staff to adjust the humidity control parameters to ensure the quality and stability of the tobacco production process.

[0064] S130: If there is a fault in the tobacco production metering equipment, an alarm corresponding to the fault type is issued.

[0065] Preferably, the method for online monitoring of tobacco production based on edge computing further includes:

[0066] S140: Storing the calculation results of the edge data analysis.

[0067] S150: Providing corresponding edge data analysis calculation results according to the query or retrieval request.

[0068] The storage of the calculation results of edge data analysis refers to the process of saving the calculation results of anomaly detection in the storage system on the edge device. Its main purpose is to provide fast data query and retrieval so as to monitor the status of tobacco production metering equipment in a timely manner.

[0069] The storage of edge data analysis calculation results specifically includes:

[0070] Storage technology selection: Choose the appropriate storage technology based on actual needs, such as relational database, NoSQL database, or file system. Consider factors such as data structure, access requirements, and performance requirements to make the selection.

[0071] Data model design: Define a reasonable data table structure or document structure to support fast data query and analysis. You can use time series databases, graph databases, or indexes to optimize data access performance.

[0072] Security and reliability: Ensure data security and reliability, including data backup, disaster recovery, and permission management. Use strategies such as data replication, redundant storage, and regular backup to protect data integrity and recoverability.

[0073] Preferably, the method for online monitoring of tobacco production based on edge computing further includes:

[0074] S160: After S120, the calculation results of the real-time data and / or edge data analysis are synchronized to the cloud server.

[0075] Preferably, the method for online monitoring of tobacco production based on edge computing further includes:

[0076] S170: Update the parameters of the data processing model from the cloud server periodically or as needed.

[0077] The interaction between edge data and cloud data refers to the process of data transmission and communication between edge computing devices and cloud servers. Its main purpose is to upload the real-time collected data of tobacco production and metering equipment to the cloud for further processing and analysis, or to obtain model parameters and instructions from the cloud to update the edge computing device, to ensure that the model can adapt to the ever-changing environment and data characteristics of tobacco production and metering equipment, and to improve the accuracy and robustness of anomaly detection.

[0078] Updates to the data processing model within edge computing devices include:

[0079] Offline training and initial model: Offline training is performed on a cloud server to obtain an initial data processing model. Then, training is performed using a large-scale data set to obtain more accurate model parameters and structure.

[0080] Online learning and optimization: The initial model is deployed on the edge computing device, and the model is learned and optimized online using the real-time collected data. By continuously receiving new data samples and updating the model parameters, the model can adapt to the working environment and data characteristics of the tobacco production metering equipment.

[0081] Regular or on-demand updates: Model updates, learning, and optimization can be performed on the cloud server. The edge computing device obtains the latest model parameters or model structure from the cloud server as needed and updates the model on the edge computing device. Model updates can be performed regularly or flexibly adjusted according to specific circumstances.

[0082] Resource consumption and performance optimization: Consider the frequency and resource consumption of model updates to avoid frequent model updates that may cause performance degradation of edge computing devices. Optimize the model update process through reasonable scheduling and resource management strategies to ensure the normal operation of edge devices.

[0083] Based on the above, the present application also provides an online monitoring system for tobacco production based on edge computing, which is used to monitor, record and analyze production data in order to monitor the quality of the tobacco production process in real time. Figure 2 As shown, the tobacco production online monitoring system includes an edge computing device 210 and a tobacco production metering device 220 on the tobacco production line;

[0084] The edge computing device 210 includes a main control unit 2101 and a communication unit 2102 . The main control unit 2101 communicates with the tobacco production metering device 220 through the communication unit 2102 .

[0085] The communication unit 2102 supports WiFi and 4G wireless communications, mainly realizing communication with the cloud server 230 , and supports RS232 and RS485 wired communications, mainly realizing data communication between the edge computing device and the tobacco production metering device 220 .

[0086] As an embodiment, the main control unit 2101 adopts a high-performance GPU and an NVLink card.

[0087] The main control unit 2101 includes a data receiving module 21011 , an edge data analysis module 21012 and an output module 21013 .

[0088] The data receiving module 21011 is used to receive real-time data from tobacco production metering equipment.

[0089] The edge data analysis module 21012 is used to perform edge data analysis on real-time data using a data processing model to determine whether the working state of the tobacco production and metering equipment is abnormal and the type of fault that exists in the tobacco production and metering equipment.

[0090] The output module 21013 is used to issue an alarm corresponding to the fault type when a fault occurs in the tobacco production metering equipment.

[0091] Among them, the parameters of the data processing model are iterated and optimized regularly or as needed.

[0092] Preferably, the edge computing device 210 also includes a storage unit 2103, and the storage unit 2103 is used to store the calculation results of the edge data analysis.

[0093] As an embodiment, the storage unit 2103 supports SD card or SSD hard disk expansion to facilitate storage of historical data on the edge side.

[0094] The main control unit 2101 also includes a query unit 21014, which is used to provide corresponding edge data analysis calculation results based on a query or retrieval request.

[0095] Preferably, the online monitoring system for tobacco production further includes a cloud server 230 .

[0096] The main control unit 2101 also includes an edge-cloud collaborative service unit 21015 , which is used to synchronize the calculation results of real-time data and / or edge data analysis to the cloud server 230 .

[0097] Preferably, the edge-cloud collaborative service unit 21015 is also used to update the parameters of the data processing model from the cloud server 230 periodically or as needed.

[0098] Specifically, the interaction between the edge computing device 210 and the cloud server 230 includes:

[0099] API interface development: Design and develop appropriate API interfaces for communication between edge computing devices and cloud servers. Through the API interface, edge devices can request cloud data, upload collected data, and obtain cloud computing results or model parameters.

[0100] Secure transmission protocol: Use secure transmission protocols for data transmission, such as HTTPS, to ensure the confidentiality and integrity of data during transmission. Encryption and authentication mechanisms are used to prevent data from being tampered with or leaked.

[0101] Data transmission protocol design: Design data transmission protocols to ensure transmission efficiency and security. Data compression, encryption, batch transmission and other methods can be used to improve data transmission efficiency and reduce network bandwidth usage.

[0102] Real-time and fault-tolerant mechanisms: Ensure real-time upload of edge device data and timely delivery of cloud commands. At the same time, consider abnormal situations such as network connection interruption and data loss, and design corresponding fault-tolerant mechanisms to ensure reliable data transmission.

[0103] Preferably, the edge data analysis module 21012 is used to analyze whether the working state of the electronic belt scale on the tobacco production line is abnormal and the type of fault that exists, and to analyze whether there is an abnormality in the moisture meter on the tobacco production line.

[0104] Preferably, the data processing model includes a clustering model, which is used to perform cluster analysis on the real-time data of the electronic belt scale to obtain clusters of different types of data points. When the difference between the cluster and the normal cluster is greater than a threshold, it is determined that the data points in the cluster are abnormal.

[0105] Preferably, the data processing model further includes a classification model, and the classification model is used to classify the data points in the cluster when an abnormality exists in the data points in the cluster, so as to obtain the fault type of the electronic belt scale.

[0106] Preferably, the data processing model also includes a long short-term memory network model, which is used to predict key parameters of the tobacco according to the real-time data of the moisture meter and output them.

[0107] like Figure 2 As shown, the edge computing device 210 also includes a positioning unit 2104 and a power supply unit 2105. The positioning unit 2104 mainly realizes the location positioning of the tobacco production metering equipment, and supports GPS and Beidou positioning. The power supply unit 2105 provides power for each module and unit of the system, and realizes functions such as charging and power detection of the backup battery.

[0108] This application utilizes the combination of edge computing technology and multiple algorithm models to achieve real-time data collection, analysis and processing in the tobacco production process, thereby improving the detection accuracy and response speed, and ensuring the quality stability and safety of tobacco production.

[0109] Although some specific embodiments of the present application have been described in detail by way of example, it should be understood by those skilled in the art that the above examples are only for illustration, not for limiting the scope of the present application. It should be understood by those skilled in the art that the above embodiments may be modified without departing from the scope and spirit of the present application. The scope of the present application is defined by the appended claims.

Claims

1. A method for online monitoring of tobacco production based on edge computing, characterized in that: include: Receive real-time data from tobacco production metering equipment; Performing edge data analysis on the real-time data using a data processing model to determine whether the working state of the tobacco production and metering equipment is abnormal and the type of fault existing in the tobacco production and metering equipment; If the tobacco production metering equipment has a fault, an alarm corresponding to the fault type is issued; The parameters of the data processing model are iterated and optimized periodically or as needed.

2. The method for online monitoring of tobacco production based on edge computing according to claim 1, characterized in that: Also includes: Store the calculation results of edge data analysis; Provides the corresponding edge data analysis calculation results based on the query or retrieval request.

3. The method for online monitoring of tobacco production based on edge computing according to claim 1 or 2, characterized in that: Also includes: The calculation results of the real-time data and / or edge data analysis are synchronized to the cloud server.

4. The method for online monitoring of tobacco production based on edge computing according to claim 3 is characterized in that: Also includes: The parameters of the data processing model are updated from the cloud server periodically or as needed.

5. The method for online monitoring of tobacco production based on edge computing according to claim 1, characterized in that: The edge data analysis includes analyzing whether the working state of the electronic belt scale on the tobacco production line is abnormal and the type of fault that exists, and analyzing whether there is an abnormality in the moisture meter on the tobacco production line.

6. An online monitoring system for tobacco production based on edge computing, characterized in that: Including edge computing equipment and tobacco production metering equipment on tobacco production lines; The edge computing device comprises a main control unit and a communication unit, and the main control unit communicates with the tobacco production metering device through the communication unit; The main control unit includes a data receiving module, an edge data analysis module and an output module; The data receiving module is used to receive real-time data of tobacco production and metering equipment; The edge data analysis module is used to perform edge data analysis on the real-time data using a data processing model to determine whether the working state of the tobacco production and metering equipment is abnormal and the type of fault existing in the tobacco production and metering equipment; The output module is used to issue an alarm corresponding to the fault type when the tobacco production metering equipment fails; The parameters of the data processing model are iterated and optimized periodically or as needed.

7. The online monitoring system for tobacco production based on edge computing according to claim 6 is characterized in that: The edge computing device further includes a storage unit, which is used to store the calculation results of the edge data analysis; The main control unit also includes a query unit, which is used to provide a calculation result of the corresponding edge data analysis according to a query or retrieval request.

8. The online monitoring system for tobacco production based on edge computing according to claim 6 or 7, characterized in that: The online monitoring system for tobacco production also includes a cloud server; The main control unit also includes an edge-cloud collaborative service unit, which is used to synchronize the calculation results of the real-time data and / or edge data analysis to the cloud server.

9. The online monitoring system for tobacco production based on edge computing according to claim 8 is characterized in that: The edge-cloud collaborative service unit is also used to update the parameters of the data processing model from the cloud server regularly or as needed.

10. The online monitoring system for tobacco production based on edge computing according to claim 6, characterized in that: The edge data analysis module is used to analyze whether the working state of the electronic belt scale on the tobacco production line is abnormal and the type of fault that exists, and to analyze whether there is an abnormality in the moisture meter on the tobacco production line.

Images