Pet food production line intelligent scheduling system and method based on Internet of Things

Abstract

The invention discloses a pet food production line intelligent scheduling system and method based on the Internet of Things, and relates to the technical field of production line scheduling. Internet of Things nodes corresponding to the positions of a plurality of pet food production lines are deployed, and convergence channels corresponding to different Internet of Things nodes are established for collecting production line data of the corresponding pet food production lines; the method comprises the steps of establishing a data slicing window of each pet food production line, dividing respective production line data into a plurality of sub-stage data, generating a scheduling atlas database based on historical production line data, establishing respective real-time scheduling atlases based on the sub-stage data of each pet food production line, and comparing and traversing the real-time scheduling atlases in the scheduling atlas database. And obtaining a plurality of stage scheduling maps of each pet food production line, integrating all stage scheduling maps of each pet food production line, obtaining a global production line scheduling map of the corresponding pet food production line, and executing production line scheduling of the corresponding pet food production line.

Description

Technical Field

[0001] This invention relates to the field of production line scheduling technology, specifically to an intelligent scheduling system and method for pet food production lines based on the Internet of Things. Background Technology

[0002] Pet food production is upgrading from standardization to flexible customization and intelligent manufacturing. However, existing production line scheduling methods in pet food production have the following limitations: Sensor data from each process is collected and transmitted independently, without forming a unified view of the production process in terms of timing and content. The collected data stream lacks a direct connection with specific production stages, making it difficult to use for detailed management at the specific process level. At the same time, scheduling relies heavily on fixed formulas and experience, failing to fully utilize historical high-quality production data for real-time matching and optimization.

[0003] Therefore, there is an urgent need for a system that can effectively integrate production line data, realize phased analysis of the production process, and perform scheduling based on historical patterns, so as to improve overall production efficiency and quality control. Summary of the Invention

[0004] The purpose of this invention is to provide an intelligent scheduling system and method for pet food production lines based on the Internet of Things (IoT) to address the shortcomings in the prior art.

[0005] To achieve the above objectives, the present invention provides the following technical solution: an intelligent scheduling system for pet food production lines based on the Internet of Things, the system comprising: The Internet of Things (IoT) module is used to deploy IoT nodes at the locations of several pet food production lines, establish aggregation channels for different IoT nodes, and collect production line data from several pet food production lines through the aggregation channels. The data processing module is used to create a data slice window for each pet food production line, which divides the production line data into several sub-stage data. The scheduling graph module generates a scheduling graph library based on historical production line data. Based on the sub-stage data corresponding to each pet food production line, it establishes its own real-time scheduling graph and compares and traverses the real-time scheduling graph in the scheduling graph library to obtain several stage scheduling graphs corresponding to each pet food production line. The production line scheduling module integrates all stage scheduling maps of each pet food production line to obtain the global production line scheduling map of the corresponding pet food production line, and performs production line scheduling based on the global production line scheduling map.

[0006] Furthermore, the process of deploying IoT nodes corresponding to the locations of several pet food production lines includes: Several deployment nodes are set up on each pet food production line. Several types of sensors are installed and debugged at each deployment node, and the sensors at each deployment node are locally networked. Set up a data sentinel for each local network, establish a message channel between several local networks corresponding to each production line, and generate a defense event corresponding to the abnormal event when any local network detects an abnormal event. Data sentinels carry defense events and abnormal events and broadcast them in the message channel to data sentinels in other local networks. The local networks that receive the broadcast determine whether an abnormal event has occurred and, based on the determination result, choose to convert the deployed node into an IoT node.

[0007] Furthermore, the process of establishing aggregation channels corresponding to different IoT nodes, and collecting production line data from several pet food production lines through these aggregation channels, includes: Set up an upload point and a backup upload point at each IoT node; Set up a cloud integration point to connect each IoT node, and switch between the upload point and the backup upload point by setting a switching line at the IoT node; Each IoT node collects point-to-point production line data of its respective pet food production line at its current location, and connects to the cloud integration point by selecting its own upload point or backup upload point to build its own aggregation channel; Based on the production timeline, several convergence channels corresponding to the same pet food production line are arranged, and the point production line data obtained from each convergence channel is integrated to collect the final production line data for each pet food production line.

[0008] Furthermore, the process of establishing a data slicing window corresponding to each pet food production line, and dividing the production line data into several sub-stage data using the data slicing window, includes: Establish corresponding slice windows based on the production sequence of each pet food production line; Based on the time required for different processes in the production of pet food, a corresponding window traversal time is set for each slice window, thereby converting each slice window into a corresponding data slice window. Data slice windows with the same window traversal time are grouped into a set of windows, and then several sets of windows are generated. Corresponding resource overhead is allocated to each set of windows. The pet food production line is divided into several production stages based on the data slice windows. The production line data is sliced ​​in the corresponding production stage through the data slice windows to obtain the sub-stage data of each production stage in the entire production process.

[0009] Furthermore, the process of allocating corresponding resource overhead to each set of windows includes: Set the number of concurrent resource requests for each window collection; The number of concurrent resource requests corresponds to the resource requests from the data slice window to the preset cloud server. A mirror request channel is established for each data slice window that initiates a resource request. The mirror request channel includes several mirror node groups. A data communication path is established between several mirror node groups. The mirror node group includes a probe node and an execution node. The probe node generates a fake data stream and transmits it to other mirror node groups on the data communication path. It also determines whether the fake data stream has been altered during transmission. Based on the determination result, it selects whether the corresponding execution node at each transmission endpoint performs data restoration of the fake data stream. The data restoration yields the corresponding part of the resource request data. At the corresponding endpoint location of each data slice window, the mirror node group integrates the resource request data of the data slice window across all data communication paths, constructs the final resource request file received by the cloud server, and the cloud server allocates resource overhead to the corresponding window set.

[0010] Furthermore, the process of generating a scheduling map library based on historical production line data includes: Historical production line data includes historical scheduling modal data corresponding to several categories of production line scheduling processes. Data modeling is performed on each historical scheduling modal data to obtain its corresponding production line scheduling map. The production line scheduling map is used to characterize the production line scheduling features corresponding to a category of production line scheduling process. Each production line scheduling map is divided into several corresponding scheduling sub-maps based on different scheduling stages of the corresponding production line scheduling process, and all scheduling sub-maps are stored in a preset database, which is then converted into a scheduling map library.

[0011] Furthermore, based on the sub-stage data corresponding to each pet food production line, a real-time scheduling graph is established for each line. This real-time scheduling graph is then compared and traversed in a scheduling graph database to obtain several stage scheduling graphs corresponding to each pet food production line. The process includes: Establish a corresponding production timeline coordinate axis for each pet food production line; Based on the various production stages of the pet food production line, the sub-stage data of each production stage are mapped to the corresponding coordinate positions on the production timeline coordinate axis based on their respective production timelines, and a scheduling layer is bound at each coordinate position. The scheduling layer is used to store the real-time scheduling map corresponding to the sub-stage data. An exchange channel is established for each scheduling layer to interact with the scheduling map library and transmit the real-time scheduling map within the exchange channel. When the scheduling graph library receives a real-time scheduling graph, it sequentially traverses the real-time scheduling graph with several scheduling sub-graphs corresponding to the production line scheduling graph stored in its own database, compares them to obtain the scheduling sub-graph that matches each real-time scheduling graph, and uses the several matching scheduling sub-graphs as several stage scheduling graphs corresponding to the pet food production line.

[0012] Furthermore, the process of integrating all stage scheduling maps of each pet food production line to obtain the global production line scheduling map for the corresponding pet food production line, and then performing production line scheduling based on the global production line scheduling map, includes: Data connection points are set in the stage scheduling graphs of every two adjacent stages in the same pet food production line. All stage scheduling graphs are connected through the data connection points, thereby generating the global production line scheduling graph of the corresponding pet food production line. Based on the global production line scheduling map of each pet food production line, generate their respective scheduling tasks, analyze the task correlation between each scheduling task, set the scheduling strategy for each scheduling task, and execute the final production line scheduling of the corresponding pet food production line based on the scheduling strategy.

[0013] This invention also provides an intelligent scheduling method for an IoT-based intelligent scheduling system for pet food production lines, comprising the following steps: Step S1: Deploy IoT nodes corresponding to the locations of several pet food production lines, establish aggregation channels for different IoT nodes, and collect production line data corresponding to several pet food production lines through the aggregation channels. Step S2: Used to create a data slice window corresponding to each pet food production line, which divides the production line data into several sub-stage data. Step S3: Generate a scheduling map library based on historical production line data. Based on the sub-stage data corresponding to each pet food production line, establish their own real-time scheduling map. Compare and traverse the real-time scheduling map in the scheduling map library to obtain several stage scheduling maps corresponding to each pet food production line. Step S4: Integrate all stage scheduling maps of each pet food production line to obtain the global production line scheduling map of the corresponding pet food production line, and execute production line scheduling based on the global production line scheduling map.

[0014] In the above technical solution, the technical effects and advantages provided by the present invention are as follows: The present invention realizes the unified aggregation and time sequence alignment of production line data through the Internet of Things module; the data processing module transforms the original production line data into sub-stage data associated with specific production processes through the data slicing window, providing a data foundation for subsequent detailed management; the scheduling map module realizes dynamic optimization of scheduling based on historical high-quality production modes by constructing a historical scheduling map library and intelligently matching it with the map generated by real-time production data. Attached Figure Description

[0015] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this invention. For those skilled in the art, other drawings can be obtained based on these drawings.

[0016] Figure 1 This is a system block diagram of the present invention.

[0017] Figure 2 This is a flowchart of the method of the present invention. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0019] Please see Figure 1 As shown, the intelligent scheduling system for a pet food production line based on the Internet of Things includes: The Internet of Things (IoT) module is used to deploy IoT nodes at the locations of several pet food production lines, establish aggregation channels for different IoT nodes, and collect production line data from several pet food production lines through the aggregation channels. The data processing module is used to create a data slice window for each pet food production line, which divides the production line data into several sub-stage data. The scheduling graph module generates a scheduling graph library based on historical production line data. Based on the sub-stage data corresponding to each pet food production line, it establishes its own real-time scheduling graph and compares and traverses the real-time scheduling graph in the scheduling graph library to obtain several stage scheduling graphs corresponding to each pet food production line. The production line scheduling module integrates all stage scheduling maps of each pet food production line to obtain the global production line scheduling map of the corresponding pet food production line, and performs production line scheduling based on the global production line scheduling map.

[0020] It should be further explained that, in the specific implementation process, the deployment of IoT nodes corresponding to the locations of several pet food production lines includes: On each pet food production line, several deployment nodes are set up and labeled. Let the label be i, then i = 1, 2, 3, ..., n, where n is a natural number greater than 0. Several types of sensors are installed and debugged at each deployment node. Several sensors at each deployment node are locally networked. The sensors specifically include RFID readers, weight sensors, temperature and humidity sensors, vibration sensors and pressure sensors. Among them, RFID readers are used to automatically identify and store information such as batch, variety, supplier, and shelf life of the raw materials corresponding to the pet food raw material bags or containers. The purpose is to achieve accurate traceability and prevent incorrect feeding in the future, serving as the identity traceability information for each pet food item. Weight sensors are integrated into corresponding silos and batching scales on the pet food production line to monitor the weight of pet food in real time; temperature and humidity sensors are used to detect the environmental conditions of the pet food and determine whether there are moldy or spoiled environmental conditions. Vibration sensors and pressure sensors are used to monitor the vibration parameters and pressure information of the corresponding production line equipment in the pet food production line. The data is combined as production line operation and maintenance reference data and directly uploaded to the production line operation and maintenance personnel in real time. The production line operation and maintenance personnel decide whether to carry out maintenance on the corresponding production line location based on the preset vibration parameter table and pressure parameter table. The vibration parameter table is used to record the vibration parameters corresponding to each production line position under normal working conditions. Vibration sensors are installed at the main drive and screw bearing positions of the corresponding extruder on the production line for predictive maintenance and to identify serious faults such as screw wear and bearing failure. The pressure parameter table is used to record the pressure information corresponding to each production line position under normal working conditions. Among them, the temperature and humidity sensor is also used to monitor the material temperature and moisture of the raw materials related to pet food in real time, and the pressure sensor is installed at the corresponding puffing cavity and die head of the production line to monitor the extrusion pressure, which directly affects the density and puffing degree of pet food products. Set up a data sentinel for each local network, and use the network IP of the local network as the identity authentication information of each data sentinel. Establish a message channel between several local networks corresponding to each production line. When any local network detects an abnormal event, a defense event corresponding to the abnormal event is generated. The abnormal events include connection failures between single or multiple IoT sensors and edge gateways or controllers, such as node signal loss, persistently unstable node signals, and abnormal node data (freezing or jumping). The corresponding defense events include the gateway initiating protocol reconnection at the deployment node corresponding to the node signal loss and sending diagnostic instructions to indicate the location of the abnormality and the corresponding abnormal sensor ID; adjusting the rate of deployment nodes with unstable signals to reduce the data transmission rate at the corresponding deployment nodes, or performing channel switching and data interpolation on the corresponding deployment nodes; and marking the sensors that have frozen or jumped as abnormal and arranging relevant management personnel to conduct on-site calibration and inspection.

[0021] Data sentinels carry defense events and abnormal events and broadcast them in the message channel to data sentinels in other local networks. The local networks that receive the broadcast determine whether they have experienced abnormal events and, based on the determination result, choose to convert the deployed nodes into IoT nodes. If an abnormal event is determined to have occurred, then it is further determined whether the abnormal event that occurred is consistent with the abnormal event broadcast. If they match, the defense event corresponding to the broadcast exception event is directly called as the target defense event corresponding to the current local network, and the target defense event is executed to handle the exception event. If there is a discrepancy, the abnormal event is analyzed to generate a corresponding new defense event. The data sentinel at the current local network location carries the new abnormal event and defense event to the message channel, merges the original broadcast content with the new abnormal event and defense event to form the latest broadcast message list, and continues to broadcast the broadcast message list to other local network locations. If no abnormal event is determined to have occurred, no action will be taken.

[0022] It should be further explained that, in the specific implementation process, the process of establishing aggregation channels corresponding to different IoT nodes, and collecting production line data from several pet food production lines through these aggregation channels, includes: Set up an upload point and a backup upload point at each IoT node; The upload point is used to upload all relevant data collected by the sensors at the corresponding IoT node. When the upload point fails, the backup upload point at the same IoT node will continue to take over the upload of the relevant data. Set up a cloud integration point to connect each IoT node, and switch between the upload point and the backup upload point by setting a switching line at the IoT node; In the initial state, the switching line is connected to the cloud integration point and the upload point corresponding to each IoT node. When the upload point fails, the switching line is switched to connect to the backup upload point. Each IoT node collects point-to-point production line data of its respective pet food production line at its current location, and connects to the cloud integration point by selecting its own upload point or backup upload point to build its own aggregation channel; Based on the production timeline, several convergence channels corresponding to the same pet food production line are arranged, and the point production line data obtained from each convergence channel is integrated to collect the final production line data for each pet food production line.

[0023] It should be further explained that, in the specific implementation process, the process of establishing a data slice window corresponding to each pet food production line, and dividing the production line data into several sub-stage data by the data slice window, includes: Establish corresponding slice windows based on the production sequence of each pet food production line; Each slice window is used to collect production line related data when the pet food production line performs a certain type of work at the corresponding production line position under the corresponding production sequence. The timestamp of each slice window is serialized and used as its own window label. The window label is used to distinguish different slice windows. Based on the time required for different processes in the production of pet food, a corresponding window traversal time is set for each slice window, thereby converting each slice window into a corresponding data slice window; wherein, the window traversal time of each slice window is consistent with the time required for the corresponding process in the production process. Data slice windows with the same window traversal time are grouped into a set of windows, and then several sets of windows are generated. The corresponding resource overhead is allocated to each set of windows. After the window set obtains the corresponding resource overhead, the corresponding resource overhead is allocated to each data slice window under the window set on an average basis. The pet food production line is divided into several production stages based on the data slicing window. The production line data is sliced ​​in the corresponding production stage through the data slicing window to obtain the sub-stage data of each production stage in the entire production process. Each sub-stage data is used to represent all production line related data of the pet food production line in a production stage. All production line related data corresponding to different production stages of the same pet food production line represent the production line related data of the corresponding pet food production line in the entire production process. The specific details regarding resource allocation for window collections are as follows: Set the number of concurrent resource requests for each window collection; The number of concurrent resource requests is used to characterize the maximum number of data slice windows in the window set that initiates resource requests at the same time. By using concurrent operations, several data slice windows can make synchronous resource requests, thereby accelerating the efficiency of resource acquisition. The number of concurrent resource requests corresponds to the resource requests from the data slice window to the preset cloud server. A mirror request channel is established for each data slice window that initiates a resource request. The mirror request channel includes several mirror node groups. A data communication path is established between several mirror node groups. The mirror node group includes probe nodes and execution nodes. The probe nodes generate fake data streams and transmit them to other mirror node groups on the data communication path. Determine whether the forged data stream has been altered during transmission. Based on the determination result, select the corresponding execution node at each transmission endpoint to perform data restoration of the forged data stream. The data restoration yields the corresponding part of the resource request data. The forged data stream is obtained by processing the request file of the corresponding resource request, copying the request file to obtain an initial request file and a copied request file, and then folding the copied request file based on a preset data protection structure to construct the corresponding forged data stream. When a forged data stream is tampered with by external parties during transmission, the IP address corresponding to the party that performed the external tampering is obtained, and the execution object corresponding to the IP address is added to the blacklist. The data communication path after the external tampering is processed is marked as a safe path, and the initial request file is transmitted again on the safe path to ensure data security. When the forged data stream is not tampered with by external forces during transmission, the mirror node group that receives the forged data stream restores the forged data stream based on the corresponding data protection structure to obtain the original corresponding resource request file. The restoration operation for the forged data stream is determined by whether the forged data stream itself is secure, thereby reducing the transmission overhead workload and speeding up the transmission efficiency. At the corresponding endpoint location of each data slice window, the mirror node group integrates the resource request data of the data slice window across all data communication paths, constructs the final resource request file received by the cloud server, and the cloud server allocates resource overhead to the corresponding window set.

[0024] It should be noted that grouping several data slice windows based on the window traversal time enables batch allocation of resource overhead, thereby accelerating the allocation efficiency of resource overhead.

[0025] It should be further explained that, in the specific implementation process, the process of generating a scheduling map library based on historical production line data, establishing a real-time scheduling map based on the sub-stage data corresponding to each pet food production line, and comparing and traversing the real-time scheduling map in the scheduling map library to obtain the scheduling map for several stages corresponding to each pet food production line includes: Historical production line data includes historical scheduling modal data corresponding to several categories of production line scheduling processes. Data modeling is performed on each historical scheduling modal data to obtain its corresponding production line scheduling map. The production line scheduling map is used to characterize the production line scheduling features corresponding to a category of production line scheduling process. Each production line scheduling map is divided into several corresponding scheduling sub-maps based on different scheduling stages of the corresponding production line scheduling process, and all scheduling sub-maps are stored in a preset database, which is then converted into a scheduling map library.

[0026] Establish a corresponding production timeline coordinate axis for each pet food production line; Based on the various production stages of the pet food production line, the sub-stage data of each production stage are mapped to the corresponding coordinate positions on the production timeline coordinate axis based on their respective production timelines, and a scheduling layer is bound at each coordinate position. The scheduling layer is used to store the real-time scheduling map corresponding to the sub-stage data. An exchange channel is established for each scheduling layer to interact with the scheduling map library and transmit the real-time scheduling map within the exchange channel. When the scheduling graph library receives a real-time scheduling graph, it sequentially traverses the real-time scheduling graph with several scheduling sub-graphs corresponding to the production line scheduling graph stored in its own database, compares them to obtain the scheduling sub-graph that matches each real-time scheduling graph, and uses the several matching scheduling sub-graphs as several stage scheduling graphs corresponding to the pet food production line.

[0027] For example, for chicken-based dry pet food products, the corresponding pet food production line includes five production stages: raw material pretreatment, mixing and stirring, high-temperature extrusion, drying and dehydration, and cooling and packaging. By matching and comparing the historical data corresponding to these five production stages with the real-time data of the current production line, dynamic optimization of the entire production process at each sub-stage can be achieved. Based on historical data from all pet food production lines across all production stages, various categories of historical scheduling modal data were obtained. Specifically, these include Modal A (related to production line operation), which describes the standard production formula for chicken-based pet dry food products and the corresponding parameters for stable, uninterrupted operation of equipment at full load on the production line; Modal B (related to raw material switching transition), which describes the cleaning, parameter adjustment, and capacity ramp-up process when switching pet dry food products from chicken to fish formula; and Modal C (minor equipment degradation), which describes the compensatory adjustment mode of temperature and pressure parameters of the extruder screw due to slight wear, while maintaining the same product quality.

[0028] For each type of modal data, extract the corresponding key scheduling features (such as the power time-series curves of each device, the inventory change curves of materials in each buffer bin, and the signal instruction sequence of stage transition) and construct the corresponding production line scheduling map. This map is a structured data model where nodes represent equipment or processes, edges represent material flow, energy flow, or control logic, and time parameters and key performance indicators are attached to the nodes and edges. Each complete production line scheduling map is broken down into smaller scheduling sub-maps according to its natural stages. Taking mode A (related to production line operation) as an example, its complete scheduling map is broken down into pre-processing sub-map, mixing sub-map, extrusion sub-map, drying sub-map, and cooling and packaging sub-map. Each sub-map focuses on the ideal scheduling state characteristics of a stage. All sub-maps of all historical modes are stored in the scheduling map library as reference data for other pet food production lines in the future.

[0029] Select a pet food production line currently producing a batch of pet food, create a corresponding production timeline coordinate axis, and set the coordinate zero point (T0) to the start time of raw material feeding for this batch. Receive sub-stage data from the current production batch in real time. For example, at T0+25 minutes, the production line is in the high-temperature extrusion stage. The latest sub-stage data (including the extruder's temperature, pressure, speed, and outlet density monitoring values ​​within the past minute) is mapped to the corresponding coordinate position on the production timeline coordinate axis based on its timestamp (T0+25min). A scheduling layer is dynamically created at this coordinate position. This scheduling layer uses the bound sub-stage data to quickly generate a real-time scheduling graph describing the current instantaneous (T0+25min) extrusion stage scheduling status. This graph is used to depict the instantaneous relationship and status of the current extruder and its upstream and downstream buffer units. Simultaneously, this real-time scheduling graph is sent as a query request to the scheduling graph library in real time. First, based on metadata (such as production line ID and current stage), all candidate scheduling subgraphs related to puffing in the scheduling graph library are selected. Then, using a graph similarity algorithm, the similarity between the real-time graph and each candidate subgraph in terms of topology and node / edge attributes is calculated. If the current real-time graph has a similarity of 92% with the puffing subgraph of mode A (related to production line operation) in the scheduling graph library and a similarity of 78% with the puffing subgraph of mode C (slight equipment degradation), then the mode A-puffing subgraph with the highest similarity (92%) is taken as the best matching result. The output is a stage scheduling graph corresponding to the current batch of pet food production line at the current moment. It is not a fixed instruction, but a scheduling suggestion template containing information such as the range of equipment target parameters, ideal material flow rate, and expected energy consumption range.

[0030] It should be further explained that, in the specific implementation process, the process of integrating all stage scheduling maps of each pet food production line to obtain the global production line scheduling map for the corresponding pet food production line, and executing production line scheduling based on the global production line scheduling map, includes: Data connection points are set in the stage scheduling graphs of every two adjacent stages in the same pet food production line. All stage scheduling graphs are connected through the data connection points, thereby generating the global production line scheduling graph of the corresponding pet food production line. The data connection point is a pre-configured data processing and logic judgment node, which is embedded with constraint rules and state transition conditions defined according to production process knowledge. Preferably, each data connection point is used to: receive the final state data packet output by its upstream stage scheduling map. The state data packet at least includes key indicators of material status when the stage is completed, equipment operation endpoint parameters and stage completion event signals. The received status data is verified for compliance based on preset process constraint rules; When the verification is successful, the initialization instructions and trigger signals of the downstream stage scheduling graph are generated according to the embedded state transition logic, and the necessary state parameters are passed to the downstream graph, thereby driving the production process to smoothly transition from one stage to the next. In the above way, multiple independent stage scheduling maps that represent the production stages such as raw material pretreatment, mixing and stirring, high-temperature puffing, drying and dehydration and cooling and packaging are linked together into a complete global production line scheduling map with temporal logic and state dependency relationship. This global map constitutes a complete digital process model for the production line to execute a single production batch task. Each pet food production line generates its own scheduling task based on the global production line scheduling map. Specifically, the global production line scheduling map is analyzed, and the elements such as equipment operation instructions, parameter setpoints, quality inspection trigger conditions, and material flow control are decomposed and instantiated into scheduling tasks with clear execution objects, target values, and time windows. For example, specific tasks such as starting the mixer at time T1 and running it at speed R1, opening the extruder feed valve when the material temperature reaches T2, and maintaining the temperature of the first section of the drying zone within the set value S±Δ are generated. Analyze the degree of task correlation between various scheduling tasks, set the scheduling strategy for each scheduling task, and execute the final production line scheduling of the corresponding pet food production line based on the scheduling strategy; For tasks with strong time-series dependencies, a sequential triggering strategy is adopted to ensure that the successor task can only start after the predecessor task is completed. For tasks competing for shared resources (such as a common steam source or total power load), a priority-based resource allocation strategy or a time-sharing reuse strategy is adopted. For tasks with coupled process parameters (such as drying temperature and wind speed), a collaborative control strategy is adopted to adjust their setpoints in conjunction with the coupling model. For tasks that can be executed independently and have no resource conflicts, a parallel execution strategy is adopted to improve efficiency.

[0031] Please see Figure 2 As shown, the present invention also provides an intelligent scheduling method for an IoT-based intelligent scheduling system for pet food production lines, comprising the following steps: Step S1: Deploy IoT nodes corresponding to the locations of several pet food production lines, establish aggregation channels for different IoT nodes, and collect production line data corresponding to several pet food production lines through the aggregation channels. Step S2: Used to create a data slice window corresponding to each pet food production line, which divides the production line data into several sub-stage data. Step S3: Generate a scheduling map library based on historical production line data. Based on the sub-stage data corresponding to each pet food production line, establish their own real-time scheduling map. Compare and traverse the real-time scheduling map in the scheduling map library to obtain several stage scheduling maps corresponding to each pet food production line. Step S4: Integrate all stage scheduling maps of each pet food production line to obtain the global production line scheduling map of the corresponding pet food production line, and execute production line scheduling based on the global production line scheduling map.

[0032] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. The pet food production line intelligent scheduling system based on the Internet of Things, characterized in that, The system comprises: An Internet of Things module is configured to deploy a plurality of Internet of Things nodes at positions of pet food production lines, establish convergence channels corresponding to different Internet of Things nodes, and collect production line data corresponding to the plurality of pet food production lines by the convergence channels; A data processing module is configured to establish data slice windows corresponding to each pet food production line, and divide the respective production line data into a plurality of sub-stage data by the data slice windows; A scheduling atlas module is configured to generate a scheduling atlas library based on historical production line data, establish real-time scheduling atlases of each pet food production line based on the sub-stage data corresponding to each pet food production line, and compare and traverse the real-time scheduling atlases in the scheduling atlas library to obtain a plurality of stage scheduling atlases corresponding to each pet food production line; A production line scheduling module is configured to integrate all the stage scheduling atlases of each pet food production line to obtain a global production line scheduling atlas of the corresponding pet food production line, and perform production line scheduling based on the global production line scheduling atlas. 2.The IoT-based intelligent scheduling system for pet food production lines according to claim 1, characterized in that, The process of deploying a plurality of Internet of Things nodes at positions of pet food production lines comprises: A plurality of deployment nodes are respectively arranged on each pet food production line, a plurality of types of sensors are installed and debugged at each deployment node, and local networking is performed between the plurality of sensors at each deployment node; A data sentinel is set for each local network, a message channel is established between a plurality of local networks corresponding to each production line, and when any local network detects an abnormal event, a defense event corresponding to the abnormal event is generated; The data sentinels carry the defense event and the abnormal event and broadcast them in the message channel to the data sentinels at other local networks, the broadcasted local networks determine whether an abnormal event occurs in themselves, and based on the determination result, the deployment nodes are converted into Internet of Things nodes. 3.The IoT-based intelligent scheduling system for pet food production lines according to claim 2, characterized in that, The process of establishing convergence channels corresponding to different Internet of Things nodes and collecting production line data corresponding to the plurality of pet food production lines by the convergence channels comprises: An upload point and a backup upload point are set at each Internet of Things node; A cloud integration point is set, which is used to connect each Internet of Things node and switch between the upload point and the backup upload point by setting a switching line at the Internet of Things node; Each Internet of Things node collects point position production line data of the pet food production line at the current production line position, and connects to the cloud integration point by selecting the upload point or the backup upload point to build a respective convergence channel; Based on the production time sequence, a plurality of convergence channels corresponding to the same pet food production line are arranged, the point position production line data obtained by each convergence channel is integrated, and the final production line data of each pet food production line is collected. 4.The IoT-based intelligent scheduling system for pet food production lines according to claim 3, characterized in that, The process of establishing data slice windows corresponding to each pet food production line and dividing the respective production line data into a plurality of sub-stage data by the data slice windows comprises: A slice window corresponding to each pet food production line is established based on the production time sequence; A window traversal time corresponding to each slice window is associated and set based on the time length of different processes in the production process of pet food, and each slice window is converted into a corresponding data slice window. The data slice windows under the same window traversal duration are classified into a group of window collections, and then a plurality of groups of window collections are generated, each group of window collections is allocated a corresponding resource overhead, and the pet food production line is divided into a plurality of production stages based on the data slice windows, the production line data is sliced in the corresponding production stage through the data slice windows, and the respective sub-stage data of each production stage of the entire production process is obtained. 5.The IoT-based intelligent scheduling system for pet food production lines according to claim 4, characterized in that, The process of allocating a corresponding resource overhead to each group of window collections includes: setting the number of concurrent resource requests corresponding to each window collection; performing resource requests of the data slice windows corresponding to the number of concurrent resource requests to the preset cloud server, establishing a mirror request channel for each data slice window initiating a resource request, the mirror request channel including a plurality of mirror node groups; a data intercommunication path is established between the plurality of mirror node groups, the mirror node group includes a detection node and an execution node, the detection node generates a fake data stream and transmits it to other mirror node groups on the data intercommunication path, and judges whether the fake data stream is changed in the transmission process, and selects whether to perform data restoration of the fake data stream at each transmission endpoint corresponding execution node based on the judgment result, and the data restoration obtains the corresponding part of the resource request data; at the mirror node group at the endpoint position of each data slice window, integrate the resource request data of the data slice window on all data intercommunication paths, and construct a resource request file received by the cloud server, and allocate resource overhead to the corresponding window collection by the cloud server. 6.The IoT-based intelligent scheduling system for pet food production lines according to claim 5, characterized in that, The process of generating a scheduling atlas library based on historical production line data includes: The historical production line data includes historical scheduling modal data corresponding to production line scheduling processes of several categories, data modeling is performed on each historical scheduling modal data to obtain respective corresponding production line scheduling atlases, and the production line scheduling atlas is used to represent the production line scheduling characteristics corresponding to a category of production line scheduling process; each production line scheduling atlas is split into respective corresponding plurality of scheduling sub-atlases based on different scheduling stages of the respective corresponding production line scheduling process, and all scheduling sub-atlases are stored in a preset database, and the database is converted into a scheduling atlas library. 7.The IoT-based intelligent scheduling system for pet food production lines according to claim 6, characterized in that, The process of establishing respective real-time scheduling atlases based on the sub-stage data corresponding to each pet food production line, and comparing and traversing the real-time scheduling atlases in the scheduling atlas library to obtain a plurality of stage scheduling atlases corresponding to each pet food production line includes: establishing a corresponding production time sequence coordinate axis for each pet food production line; based on the respective production stages corresponding to the pet food production line, mapping the sub-stage data of each production stage to the corresponding coordinate position on the production time sequence coordinate axis based on the respective production time sequence, and binding a scheduling graph layer at each coordinate position; the scheduling graph layer is used to store the real-time scheduling atlas established by the sub-stage data, and an exchange channel for data interaction with the scheduling atlas library is established for each scheduling graph layer, and the real-time scheduling atlas is transmitted in the exchange channel; When the real-time scheduling graph is received by the scheduling graph library, the real-time scheduling graph and the corresponding several scheduling sub-graphs of the production line scheduling graph stored by itself are sequentially traversed, and the scheduling sub-graphs matched with each real-time scheduling graph are obtained, and the matched several scheduling sub-graphs are used as the several stage scheduling graphs corresponding to the pet food production line. 8.The IoT-based pet food production line intelligent scheduling system according to claim 7, characterized in that, The global production line scheduling graph corresponding to the pet food production line is obtained by integrating all the stage scheduling graphs of each pet food production line, and the process of executing the production line scheduling based on the global production line scheduling graph comprises: Data connection points are set between the stage scheduling graphs adjacent to each other in the order of each two stages corresponding to the same pet food production line, all the stage scheduling graphs are connected through the data connection points, and then the global production line scheduling graph corresponding to the pet food production line is generated; Based on the global production line scheduling graph of each pet food production line, the scheduling tasks of each pet food production line are generated, the task correlation degree between each scheduling task is analyzed, the scheduling strategy corresponding to each scheduling task is set, and the final production line scheduling of the corresponding pet food production line is executed based on the scheduling strategy.

9. The intelligent scheduling method of the pet food production line intelligent scheduling system based on the Internet of Things, used for realizing the pet food production line intelligent scheduling system of any one of claims 1 to 8, characterized in that, The steps comprise: Step S1: The Internet of Things nodes corresponding to the positions of the several pet food production lines are deployed, the aggregation channels corresponding to the different Internet of Things nodes are established, and the production line data corresponding to the several pet food production lines is collected by the aggregation channels; Step S2: The data slice window corresponding to each pet food production line is established, and the production line data of each pet food production line is divided into several sub-stage data by the data slice window; Step S3: The scheduling graph library is generated based on the historical production line data, the real-time scheduling graph of each pet food production line is established based on the sub-stage data corresponding to each pet food production line, and the real-time scheduling graph is compared and traversed in the scheduling graph library to obtain the several stage scheduling graphs corresponding to each pet food production line; Step S4: The global production line scheduling graph corresponding to the pet food production line is obtained by integrating all the stage scheduling graphs of each pet food production line, and the production line scheduling is executed based on the global production line scheduling graph.

Images