Meat product production line on-line monitoring system and monitoring method based on digital twinning

Abstract

The invention discloses a meat product production line online monitoring system and monitoring method based on digital twinning, and belongs to the technical field of meat product production line management, the system comprises a real-time data acquisition unit and a digital twinning model unit, and the real-time data acquisition unit is connected with the digital twinning model unit through a virtual-real synchronization unit; the real-time data acquisition unit is used for acquiring real-time data of a meat product production line, the digital twin model unit comprises a multi-source heterogeneous data analysis module, the multi-source heterogeneous data analysis module is connected with a production line data module and a production line description module, and the production line description module is connected with an early warning module and a tracing module. The early warning module, the tracing module and the production line data module are all connected with the database. By adopting the online monitoring system and monitoring method for the meat product production line based on the digital twinning, monitoring and prediction of the meat product production process are realized, the management efficiency is improved, the loss is effectively reduced, and effective support is provided for meat product production management.

Description

Technical Field

[0001] The present invention relates to the technical field of meat product production line management, and specifically refers to an online monitoring system and monitoring method for meat product production lines based on digital twins. Background Art

[0002] With the progress of technology and the improvement of industrial automation levels, the management of meat product production lines has become increasingly important. Traditional meat product production management methods usually rely on manual monitoring and manual recording, which have problems such as untimely data updates, inaccurate information transmission, and inability to give early warnings of equipment failures in a timely manner. These deficiencies not only affect the smoothness of the meat product production process but may also lead to fluctuations in product quality and an increase in potential safety hazards. Against this background, we need a more efficient and advanced meat product production line management method. Summary of the Invention

[0003] The purpose of the present invention is to provide an online monitoring system and monitoring method for meat product production lines based on digital twins to solve the above technical problems.

[0004] To achieve the above purpose, the present invention provides an online monitoring system for meat product production lines based on digital twins, including a real-time data acquisition unit and a digital twin model unit. The real-time data acquisition unit is connected to the digital twin model unit through a virtual-real synchronization unit;

[0005] The real-time data acquisition unit includes an equipment data acquisition module set in meat product production equipment, a production environment data acquisition module set in the processing workshop, and a raw material data acquisition module set in the meat product raw material storage tank. The equipment data acquisition module, the production environment data acquisition module, and the raw material data acquisition module are all connected to the virtual-real synchronization unit;

[0006] The digital twin model unit includes a multi-source heterogeneous data parsing module for integrating data from different data sources and unifying the data format. The multi-source heterogeneous data parsing module is connected to a production line data module and a production line description module. The production line description module is connected to an early warning module and a traceability module. The early warning module, the traceability module, and the production line data module are all connected to a database.

[0007] Preferably, the equipment data acquisition module includes a temperature sensor, a humidity sensor, a vibration sensor, a pressure sensor, a current sensor, a voltage sensor, and a scanner, which are used to collect the operating parameters of meat product production equipment in real time and input the information identification tags of products;

[0008] The production environment data acquisition module includes a temperature sensor, a humidity sensor, and a gas sensor, which are used to collect the environmental temperature, humidity, and gas composition of the production workshop in real time;

[0009] The raw material data acquisition module includes a temperature sensor, a humidity sensor, a PH sensor, an optical sensor, an olfactory sensor, and an infrared sensor, which are used to collect the temperature, humidity, acidity, freshness, and composition of meat product raw materials in real time.

[0010] Preferably, the production line data module includes a sensor data sub-module, an equipment status record sub-module, and an equipment maintenance record sub-module;

[0011] The sensor data sub-module, the equipment status record sub-module, and the equipment maintenance record sub-module are all connected to the multi-source heterogeneous data parsing module;

[0012] The sensor data sub-module is used to receive the operating parameters of meat product production equipment; the environmental temperature, humidity, and gas composition in the production workshop; the temperature, humidity, acidity, freshness, and composition of meat product raw materials;

[0013] The equipment status record sub-module is used to record the startup time, shutdown time, failure time, processing time, and the number of processed products of the production equipment;

[0014] The equipment maintenance record sub-module is used to record the maintenance history and maintenance plan of the production equipment.

[0015] Preferably, the production line description module includes an equipment monitoring sub-module and a product information sub-module;

[0016] The equipment monitoring sub-module and the product information sub-module are both connected to the multi-source heterogeneous data parsing module;

[0017] The equipment monitoring sub-module is respectively connected to the sensor data sub-module, the equipment status record sub-module, and the equipment maintenance record sub-module, and is used to monitor the status information, operating parameters, and maintenance time of the production equipment;

[0018] The product information sub-module is connected to the sensor data sub-module, and is used to monitor the temperature, humidity, acidity, freshness, composition of meat product raw materials, and the basic information of the product. The basic information includes product type, batch number, production date, and expiration date.

[0019] Preferably, the database includes a time series database, a file storage database, a relational database, and a blockchain.

[0020] A monitoring method for an online monitoring system of a meat product production line based on digital twin as described above, the specific steps are as follows:

[0021] Step S1: Connect the virtual-real synchronization unit to the real-time data acquisition unit and the digital twin model unit respectively;

[0022] Step S2: Collect the data of the meat product production line in real time and transmit the collected data to the multi-source heterogeneous data parsing module;

[0023] During the production process, the equipment data acquisition module collects the parameter values ​​of temperature, humidity, vibration, pressure, voltage, and current of meat production equipment in real time through sensors and scanners and enters the information identification tags; the production environment data acquisition module collects the ambient temperature, humidity, and gas composition of the production workshop in real time through sensors; the raw material data acquisition module collects the temperature, humidity, pH, freshness, and composition of meat raw materials in real time through sensors; all collected data are transmitted to the multi-source heterogeneous data analysis module and converted into digital data format;

[0024] Step S3: The multi-source heterogeneous data analysis module analyzes the received raw data to form a unified digital data format and transmits it to the production line data module and the production line description module, displays and updates the data information of the production line equipment, environment, raw materials, and products in real time, and stores all data in the database to provide data support for the early warning module and the tracing module;

[0025] Step S4: The early warning module preprocesses, extracts features and analyzes the data, predicts and analyzes the real-time data through a deep learning neural network, predicts anomalies and risks in advance and issues early warnings;

[0026] The traceability module queries relevant data information from the database according to user needs, and summarizes and displays the queried information so that users can view it quickly and intuitively.

[0027] Preferably, the equipment data acquisition module, the production environment data acquisition module and the raw material data acquisition module collect data in real time through sensors and scanners, and pass the data to the multi-source heterogeneous data analysis module. The multi-source heterogeneous data analysis module parses the data and forms a unified digital format, and then passes it to the equipment detection submodule and the product information submodule. The equipment and product information are updated through the sensor values ​​and product information.

[0028] Preferably, the early warning module performs prediction and early warning based on the real-time updated data in the production line data module and the production line description module, and performs prediction and early warning through a deep learning neural network method. The specific process is as follows:

[0029] Step S41: data collection, obtaining real-time updated monitoring data of the meat production line, including temperature, humidity, vibration, pressure, voltage, current, pH, freshness, ambient gas composition and raw material composition;

[0030] Step S42: Data preprocessing, performing data preprocessing on the monitoring data, supplementing missing values ​​in the data, removing abnormal values ​​in the data, and performing noise reduction, feature extraction and normalization on the data;

[0031] Step S43: Divide the processed data into a training set and a test set;

[0032] Step S44: Construct a deep learning neural network model and train and test it using the training set and the test set;

[0033] Step S45: Input the real-time collected data into the trained deep learning neural network model, formulate an early warning strategy based on the prediction results. When the prediction results show that the equipment, environment or raw materials are abnormal, corresponding early warning signals will be sent to prompt the staff to take measures.

[0034] Preferably, in Step S44, the specific process of constructing the deep learning neural network model is as follows:

[0035] First, input the preprocessed data into the convolutional neural network to extract the data time series features. The one-dimensional convolution formula is shown as Equation (1) below:

[0036]

[0037] In the formula, y t is the output feature data, w t is the convolution kernel, x t-k+1 is the input data, b is the bias, and k is the data length.

[0038] Then, introduce the attention mechanism to dynamically weight the data time series features extracted by the convolutional neural network, which is beneficial to solving the long-term dependence problem and improving the performance and robustness of the model; the mathematical expression of the attention mechanism is shown as Equation (2) below:

[0039]

[0040] In the formula, Q, K, and V are the query value, calculation value, and weight respectively; f(Q, K τ ) is the scoring function used to calculate the correlation between different positions; d k is the feature vector dimension; the Softmax function is the data normalization operation.

[0041] Then, input the features dynamically weighted by the attention mechanism into the long short-term memory neural network; the mathematical expressions of the long short-term memory neural network are shown as Equations (3-7) below:

[0042] f t = σ(W f ·[h t-1 , x t +b f )

[0043]

[0044]

[0045] o t = σ(W o ·[h t-1 , x t +b o )

[0046]

[0047] In the formula, f t is the genetic factor; W f , b f are the parameters to be trained in the network; σ is the Sigmoid function; is the content to be determined for update; C t is the updated cell state; W i , b i , W c , b c , W o , b o are all parameters to be trained in the network; tanh is the hyperbolic tangent function; is the element-wise multiplication of matrices; h t is the output at the current moment; o t is the output result.

[0048] Finally, the hyperparameters in the deep learning neural network model are optimized through iterative learning, and accurate prediction results of production line equipment parameters, environmental parameters, and raw material parameters are obtained.

[0049] Therefore, the present invention adopts the above-mentioned online monitoring system and monitoring method for a meat product production line based on digital twin, and has the following advantages:

[0050] (1) The real-time monitoring and prediction method for a meat product production line based on digital twin technology, big data analysis technology, Internet of Things technology, and deep learning algorithm is beneficial to solving the technical problems of insufficient timeliness and accuracy of information in the meat product production process, ensuring product quality and safety, optimizing production efficiency, and promoting the intelligent development of the meat product industry;

[0051] (2) Using Internet of Things technology to collect various data in the meat product production line in real time, including information such as temperature, humidity, acidity, freshness, etc., and using digital twin technology to build a virtual meat product production line model, making the monitoring of the production line more accurate;

[0052] (3) Using digital twin technology to build a virtual meat product production line model can better simulate and optimize the production process, providing new ideas and methods for the management of the meat product production line;

[0053] (4) Organize and analyze the collected data, which can comprehensively understand the situation in the meat product production process, improve corresponding early warnings for equipment, environment and raw materials, and effectively improve production efficiency and production safety. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the drawings in the following description are some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0055] Figure 1 It is a block diagram of an online monitoring system for a meat product production line based on digital twin according to the present invention;

[0056] Figure 2 It is a layout diagram of data collection for a meat product production line according to the present invention;

[0057] Figure 3 It is an architecture diagram of an online monitoring system for a meat product production line based on digital twin according to the present invention;

[0058] Figure 4 It is a monitoring diagram of the meat product production environment state in the digital twin model according to the present invention;

[0059] Figure 5 It is a monitoring diagram of the meat product production equipment state in the digital twin model according to the present invention;

[0060] Figure 6 It is a monitoring diagram of the meat product production raw material state in the digital twin model according to the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0061] The following will further elaborate on the present invention with reference to the drawings.

[0062] In the description of the present invention, it should be noted that the orientation or positional relationship indicated by terms such as "upper", "lower", "inner", "outer", etc. is based on the orientation or positional relationship shown in the drawings, or the orientation or positional relationship in which the inventive product is customarily placed during use. It is only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and thus should not be construed as a limitation to the present invention. In the description of the present invention, it should also be noted that unless otherwise clearly specified and defined, the terms "set", "installed", "connected" should be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be directly connected, or indirectly connected through an intermediate medium, and it can be the communication inside two elements. For those of ordinary skill in the art, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0063] The following will describe the embodiments of the present invention in detail with reference to the drawings.

[0064] As Figure 1 shown, an online monitoring system for a meat product production line based on digital twin includes a real-time data acquisition unit and a digital twin model unit. The real-time data acquisition unit is connected to the digital twin model unit through a virtual-real synchronization unit. The virtual-real synchronization unit can be a PLC, an industrial computer, or a single-chip microcomputer, and is used to connect the physical space and the digital space.

[0065] As Figures 2-3 shown, Figure 2 is the data acquisition layout diagram of the meat product production line of the present invention, Figure 3 is the architecture diagram of an online monitoring system for a meat product production line based on digital twin of the present invention. The real-time data acquisition unit is used to collect data in the production line in real time. The real-time data acquisition unit includes an equipment data acquisition module arranged in the meat product production equipment, a production environment data acquisition module arranged in the processing workshop, and a raw material data acquisition module arranged in the meat product raw material storage tank. The equipment data acquisition module, the production environment data acquisition module, and the raw material data acquisition module are all connected to the virtual-real synchronization unit. The equipment data acquisition module includes a temperature sensor, a humidity sensor, a vibration sensor, a pressure sensor, a current sensor, a voltage sensor, and a scanner, and is used to collect the operation parameters of the meat product production equipment in real time and input the information identification label of the product; the production environment data acquisition module includes a temperature sensor, a humidity sensor, and a gas sensor, and is used to collect the environmental temperature, humidity, and gas composition of the production workshop in real time; the raw material data acquisition module includes a temperature sensor, a humidity sensor, a PH sensor, an optical sensor, an olfactory sensor, and an infrared sensor, and is used to collect the temperature, humidity, acidity, freshness, and composition of the meat product raw materials in real time.

[0066] The digital twin model unit includes a multi-source heterogeneous data parsing module for unifying data formats. The multi-source heterogeneous data parsing module is respectively connected to the production line data module and the production line description module. The production line description module is connected to an early warning module and a traceability module. The early warning module, the traceability module, and the production line data module are all connected to the database. The production line data module includes a sensor data sub-module, an equipment status record sub-module, and an equipment maintenance record sub-module. The sensor data sub-module is used to receive the operating parameters of meat product production equipment; the environmental temperature, humidity, and gas composition of the production workshop; the temperature, humidity, pH value, freshness, and composition of meat product raw materials. The equipment status record sub-module is used to record the startup time, shutdown time, failure time, processing time, and the number of processed products of the production equipment. The equipment maintenance record sub-module is used to record the maintenance history and maintenance plan of the production equipment. The production line description module includes an equipment monitoring sub-module and a product information sub-module. The equipment monitoring sub-module is respectively connected to the sensor data sub-module, the equipment status record sub-module, and the equipment maintenance record sub-module, and is used to monitor the status information, operating parameters, and maintenance time of the production equipment. The product information sub-module is connected to the sensor data sub-module and is used to monitor the temperature, humidity, pH value, freshness, composition of meat product raw materials, and the basic information of the product. The basic information includes product type, batch number, production date, and expiration date. The database includes a time series database, a file storage database, a relational database, and a blockchain.

[0067] A monitoring method for an online monitoring system of a meat product production line based on digital twin as described above is as follows:

[0068] Step S1: Connect the virtual-real synchronization unit to the real-time data acquisition unit and the digital twin model unit respectively.

[0069] Step S2: Real-time collect the data of the meat product production line and transmit the collected data to the multi-source heterogeneous data parsing module.

[0070] During the production process, the equipment data acquisition module uses sensors and scanners to real-time collect the parameter values of the temperature, humidity, vibration, pressure, voltage, and current of the meat product production equipment and enter them into the information identification label. The production environment data acquisition module uses sensors to real-time collect the environmental temperature, humidity, and gas composition of the production workshop. The raw material data acquisition module uses sensors to real-time collect the temperature, humidity, pH value, freshness, and composition of meat product raw materials. All the collected data is transmitted to the multi-source heterogeneous data parsing module and converted into a digital data format.

[0071] Step S3: The multi-source heterogeneous data analysis module analyzes the received raw data to form a unified digital data format and transmits it to the production line data module and the production line description module, displays and updates the data information of the production line equipment, environment, raw materials, and products in real time, and stores all data in the database to provide data support for the early warning module and the tracing module;

[0072] Step S4: The early warning module preprocesses, extracts features and analyzes the data, predicts and analyzes the real-time data through a deep learning neural network, predicts anomalies and risks in advance and issues early warnings;

[0073] The traceability module queries relevant data information from the database according to user needs, and summarizes and displays the queried information so that users can view it quickly and intuitively.

[0074] The prediction and warning information is presented to users in a visual way, allowing users to intuitively understand potential problems that may occur in the production line and make timely adjustments and interventions.

[0075] The equipment data acquisition module, production environment data acquisition module and raw material data acquisition module collect data in real time through sensors and scanners, and pass it to the multi-source heterogeneous data analysis module. The multi-source heterogeneous data analysis module parses the data and forms a unified digital format, which is then passed to the equipment detection submodule and the product information submodule. The equipment and product information are updated through the sensor values ​​and product information.

[0076] In this embodiment, the data transmission format is Json format, the communication protocol is MTQQ, and the digital twin space is updated according to the data collected in real time in the production line to maintain the accuracy and real-time nature of the digital model information.

[0077] The early warning module makes predictions and early warnings based on the real-time updated data in the production line data module and the production line description module. It uses deep learning neural network methods to make predictions and early warnings. The specific process is as follows:

[0078] Step S41: data collection, obtaining real-time updated monitoring data of the meat production line, including temperature, humidity, vibration, pressure, voltage, current, pH, freshness, ambient gas composition and raw material composition;

[0079] Step S42: Data preprocessing, performing data preprocessing on the monitoring data, supplementing missing values ​​in the data, removing abnormal values ​​in the data, and performing noise reduction, feature extraction and normalization on the data;

[0080] Step S43: Divide the processed data into a training set and a test set;

[0081] Step S44: Construct a deep learning neural network model and train and test it using a training set and a test set;

[0082] The specific process of constructing the deep learning neural network model is as follows:

[0083] First, input the preprocessed data into the convolutional neural network to extract the data time series features. The one-dimensional convolution formula is shown in Equation (1) below:

[0084]

[0085] In the formula, y t is the output feature data, w t is the convolution kernel, x t-k+1 is the input data, b is the bias, and k is the data length.

[0086] Then, introduce the attention mechanism to dynamically weight the data time series features extracted by the convolutional neural network, which is beneficial to solving the long-term dependence problem and improving the performance and robustness of the model. The mathematical expression of the attention mechanism is shown in Equation (2) below:

[0087]

[0088] In the formula, Q, K, and V are the query value, calculation value, and weight respectively; f(Q, K τ ) is the scoring function used to calculate the correlation between different positions; d k is the feature vector dimension; the Softmax function is the data normalization operation.

[0089] Then, input the features dynamically weighted by the attention mechanism into the long short-term memory neural network. The mathematical expressions of the long short-term memory neural network are shown in Equations (3-7) below:

[0090] f t =σ(W f ·[h t-1 ,x t +b f )(3)

[0091]

[0092]

[0093] o t =σ(W o ·[h t-1 ,x t +b o )(6)

[0094]

[0095] where f t is a genetic factor; W f , b f are the parameters to be trained in the network; σ is the Sigmoid function; is the content to be determined for update; C t is the updated cell state; W i , b i , W c , b c , W o , b o are all the parameters to be trained in the network; tanh is the hyperbolic tangent function; is the element-wise multiplication of matrices; h t is the output at the current moment; o t is the output result.

[0096] Finally, the hyperparameters in the deep learning neural network model are optimized through iterative learning, and accurate prediction results of the production line equipment parameters, environmental parameters, and raw material parameters are obtained.

[0097] Meanwhile, the deep learning neural network model is evaluated, and the evaluation process is as follows:

[0098] First, the constructed deep learning neural network model is used to predict the sample data in the test set to obtain the prediction results;

[0099] Then, by comparing the prediction results with the actual results, and calculating the evaluation metrics of the deep learning neural network model, including the root mean square error (RMSE) and the coefficient of determination (R 2 ).

[0100] RMSE is the square root of the mean of the squared differences between the predicted values and the actual values, which can quantify the error range of the model prediction. R2 is a statistical metric used to measure the goodness of fit of a regression model. The R2 value ranges from 0 to 1, and the higher the value, the better the model fitting effect. The calculation formulas for the two evaluation metrics are as follows:

[0101]

[0102]

[0103] where r i ' is the predicted value, and r i is the actual value.

[0104] Then, the evaluation results are interpreted to explain the performance of the deep learning neural network model in the prediction and early warning process, and to evaluate the advantages and disadvantages of the model.

[0105] Finally, the hyperparameters of the deep learning neural network model are adjusted to improve the performance of the model.

[0106] Step S45: Input the real-time collected data into the trained deep learning neural network model, formulate an early warning strategy according to the prediction result. When the prediction result shows that there are abnormal situations in the equipment, environment or raw materials, corresponding early warning signals will be sent out to prompt the staff to take measures.

[0107] As Figure 4 shown, through the real-time monitoring of the digital twin of the meat product production line, the environmental state of the production line is monitored, the production situation of the production line is displayed in real time, and various data are obtained, including temperature, humidity and gas composition. As Figure 5 shown, through the real-time monitoring of the digital twin of the meat product production line, the equipment state of the production line is monitored, and the equipment state and various parameter values are displayed in real time, including humidity, humidity, vibration, voltage, current and pressure. As Figure 6 shown, through the real-time monitoring of the digital twin of the meat product production line, the state of the meat product raw materials in the storage tank is monitored, and the temperature, humidity, pH value, freshness and raw material composition of the raw materials are displayed in real time. Through the real-time monitoring system, the production line environment, equipment and raw materials are continuously monitored, abnormal situations are detected in time and early warnings are given to prompt relevant personnel to intervene and regulate. At the same time, relevant data are recorded to provide more comprehensive data support for early warnings.

[0108] Utilize the Internet of Things technology to collect various data in the meat product production line in real time, and build a virtual meat product production line model through the digital twin technology to realize the real-time monitoring and early warning of the meat product production line. At the same time, combined with big data analysis technology and deep learning technology, analyze and predict historical data and real-time data. At the same time, trace the product information to provide timely early warnings and adjustment suggestions for problems in the meat product production process.

[0109] The above are only the preferred embodiments of the present invention. It should be noted that for those skilled in the art of this technology, without departing from the principle of the present invention, several improvements and refinements can still be made, and these improvements and refinements all belong to the protection scope of the technology of the present invention.

Claims

1. An online monitoring system for a meat product production line based on digital twin, characterized in that: It includes a real-time data acquisition unit and a digital twin model unit, and the real-time data acquisition unit is connected to the digital twin model unit through a virtual-real synchronization unit; The real-time data acquisition unit includes a device data acquisition module, a production environment data acquisition module, and a raw material data acquisition module; The digital twin model unit includes a multi-source heterogeneous data parsing module for integrating data from different data sources and unifying the data format. The multi-source heterogeneous data parsing module is connected to a production line data module and a production line description module. The production line description module is connected to an early warning module and a traceability module. The early warning module, the traceability module, and the production line data module are all connected to a database.

2. The on-line monitoring system for a meat product production line based on digital twin according to claim 1, wherein: The device data acquisition module is set in the meat product production equipment and is used to collect the operation parameters of the meat product production equipment in real time and input the information identification label of the product; The production environment data acquisition module is set in the processing workshop and includes sensors for collecting the environmental temperature, humidity, and gas composition of the production workshop in real time; The raw material data acquisition module is set in the meat product raw material storage tank and is used to collect the temperature, humidity, pH value, freshness, and composition of the meat product raw materials in real time.

3. An online monitoring system for a meat product production line based on digital twin according to claim 1 or 2, characterized in that: The device data acquisition module includes a temperature sensor, a humidity sensor, a vibration sensor, a pressure sensor, an electric current sensor, a voltage sensor, and a scanner The production environment data acquisition module includes a temperature sensor, a humidity sensor, and a gas sensor, The raw material data acquisition module includes a temperature sensor, a humidity sensor, a pH sensor, an optical sensor, an olfactory sensor, and an infrared sensor.

4. An online monitoring system for a meat product production line based on digital twin according to claim 1 or 2, characterized in that: The production line data module includes a sensor data sub-module, a device status record sub-module, and a device maintenance record sub-module; The sensor data sub-module, the device status record sub-module, and the device maintenance record sub-module are all connected to the multi-source heterogeneous data parsing module; The sensor data sub-module is used to receive the operation parameters of the meat product production equipment; the environmental temperature, humidity, and gas composition of the production workshop; the temperature, humidity, pH value, freshness, and composition of the meat product raw materials; The device status record sub-module is used to record the startup time, shutdown time, failure time, processing time, and the number of processed products of the production equipment; The device maintenance record sub-module is used to record the maintenance history and maintenance plan of the production equipment.

5. The on-line monitoring system of a meat product production line based on digital twin according to claim 4, characterized in that: The production line description module includes a device monitoring sub-module and a product information sub-module; The device monitoring sub-module and the product information sub-module are both connected to the multi-source heterogeneous data parsing module; The device monitoring sub-module is respectively connected to the sensor data sub-module, the device status record sub-module, and the device maintenance record sub-module, and is used to monitor the status information, operation parameters, and maintenance time of the production equipment; The product information sub-module is connected to the sensor data sub-module and is used to monitor the temperature, humidity, pH value, freshness, composition of the meat product raw materials, and the basic information of the product. The basic information includes product type, batch number, production date, and expiration date.

6. The on-line monitoring system for a meat product production line based on digital twin according to claim 1, wherein: The database includes a time series database, a file storage database, a relational database, and a blockchain.

7. A monitoring method for the online monitoring system of a meat product production line based on digital twin according to any one of claims 1-6, characterized in that, The specific steps are as follows: Step S1: Connect the virtual-real synchronization unit to the real-time data acquisition unit and the digital twin model unit respectively; Step S2: Collecting data from the meat product production line in real time and transmitting the collected data to a multi-source heterogeneous data analysis module; All collected data are transferred to the multi-source heterogeneous data analysis module and converted into digital data format; Step S3: The multi-source heterogeneous data analysis module analyzes the received raw data to form a unified digital data format and transmits it to the production line data module and the production line description module, displays and updates the data information of the production line equipment, environment, raw materials, and products in real time, and stores all data in the database to provide data support for the early warning module and the tracing module; Step S4: The early warning module preprocesses, extracts features and analyzes the data, predicts and analyzes the real-time data through a deep learning neural network, predicts anomalies and risks in advance and issues early warnings; The traceability module queries relevant data information from the database according to user needs, and summarizes and displays the queried information so that users can view it quickly and intuitively.

8. A monitoring method according to claim 7, characterized in that: The equipment data acquisition module, production environment data acquisition module and raw material data acquisition module collect data in real time through sensors and scanners: The equipment data acquisition module collects the parameter values ​​of temperature, humidity, vibration, pressure, voltage and current of meat production equipment in real time through sensors and scanners and enters the information identification tags; The production environment data acquisition module collects the ambient temperature, humidity and gas composition of the production workshop in real time through sensors; The raw material data collection module collects the temperature, humidity, pH, freshness and composition of meat raw materials in real time through sensors; The collected data is passed to the multi-source heterogeneous data analysis module, which parses the data and forms a unified digital format, and then passes it to the equipment detection submodule and the product information submodule to update the equipment and product information through the sensor values ​​and product information.

9. A monitoring method according to claim 7 or 8, characterized in that: The warning module predicts and warns based on the real-time updated data in the production line data module and the production line description module, and predicts and warns through a deep learning neural network method. The specific process is as follows: Step S41: data collection, obtaining real-time updated monitoring data of the meat production line, including temperature, humidity, vibration, pressure, voltage, current, pH, freshness, ambient gas composition and raw material composition; Step S42: Data preprocessing, performing data preprocessing on the monitoring data, supplementing missing values ​​in the data, removing abnormal values ​​in the data, and performing noise reduction, feature extraction and normalization on the data; Step S43: Divide the processed data into a training set and a test set; Step S44: construct a deep learning neural network model, and perform training and testing using a training set and a test set; Step S45: Input the real-time collected data into the trained deep learning neural network model, and formulate an early warning strategy based on the prediction results. When the prediction results show that the equipment, environment or raw materials are abnormal, a corresponding early warning signal will be issued to prompt the staff to take measures.

10. A monitoring method according to claim 9, characterized in that: In step S44, the specific process of constructing a deep learning neural network model is as follows: First, input the preprocessed data into a convolutional neural network to extract the data time series features. The one-dimensional convolution formula is shown in Equation (1) below: where y t is the output feature data, w t is the convolution kernel, x t-k+1 is the input data, b is the bias, and k is the data length. Then, introduce an attention mechanism to dynamically weight the data time series features extracted by the convolutional neural network, which helps to solve the long-term dependence problem and improve the performance and robustness of the model. The mathematical expression of the attention mechanism is shown in Equation (2) below: Wherein, Q, K, and V are the query value, calculation value, and weight respectively; f(Q, K τ ) is a scoring function used to calculate the correlation between different positions; d k is the dimension of the feature vector; the Softmax function is a data normalization operation. Then, input the features dynamically weighted by the attention mechanism into a long short-term memory neural network. The mathematical expressions of the long short-term memory neural network are shown in Equations (3-7) below: f t = σ(W f · [h t-1 , x t + b f ) o t = σ(W o · [h t-1 , x t + b o ) where f t is a genetic factor; W f , b f are parameters to be trained in the network; σ is the Sigmoid function; is the content to be updated; C t is the updated cell state; W i , b i , W c , b c , W o , b o are all parameters to be trained in the network; tanh is the hyperbolic tangent function; is the element-wise multiplication of matrices; h t is the output at the current time; o t is the output result. Finally, optimize the hyperparameters in the deep learning neural network model through iterative learning and obtain accurate prediction results of the production line equipment parameters, environmental parameters, and raw material parameters.

Images