Real-time control method of full-automatic production line based on Internet of Things
By deploying IoT nodes and edge computing modules on a fully automatic production line, and combining dynamic time prediction models for hierarchical interception control, the problem of low interception efficiency of defective products in the existing technology is solved, and precise interception of defective products and synchronous optimization of production efficiency is achieved.
Patent Information
- Application Number
- CN202510475798.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-04-16
- Publication Date
- 2025-08-19
AI Technical Summary
The existing real-time control methods of fully automatic production lines have high data heterogeneity and significant response delay, which leads to low interception efficiency of defective products and mostly rely on a single shutdown interception strategy, which can easily cause fluctuations in production line operation efficiency and intensify equipment losses.
By deploying IoT nodes to collect product images, locations and electronic tag data in real time, generate standardized data flows, use edge computing modules to identify defects, and combine dynamic time prediction models for hierarchical interception control, including first-level interception, two-level interception and three-level interception to avoid full-line downtime.
The precise interception of defective products and the synchronous optimization of production efficiency has been achieved, the product pass rate has been improved, the production costs and resource waste have been reduced, and the production line operation fluctuations have been reduced.
Smart Images
Figure CN120508053A_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the technical field of fully automatic production lines, and in particular relates to a real-time control method for a fully automatic production line based on the Internet of Things. Background Art
[0002] In the field of industrial automation, the application of real-time control technology for fully automated production lines has become a key means of improving manufacturing efficiency and quality. Currently, developed countries, represented by Germany and Japan, have widely adopted technologies such as the Internet of Things and edge computing to achieve dynamic collaborative control of production lines, improving production line efficiency and stability, especially in the detection of defective products. Traditional real-time control of fully automated production lines generally uses visual technology to inspect products on the production conveyor line, but this detection method has certain limitations in practical applications.
[0003] Because it usually collects production line data by deploying sensor networks and combines preset rules to achieve anomaly detection and coordinated control of production line equipment, it inevitably leads to problems such as high data heterogeneity and significant response delay, which in turn leads to low efficiency in intercepting defective products. At the same time, it relies too much on a single shutdown interception strategy and lacks hierarchical collaborative interception control logic, which can easily cause fluctuations in production line operation efficiency and increased equipment loss, thereby greatly restricting the production line.
[0004] Therefore, it is urgent to propose a real-time control method for a fully automatic production line based on the Internet of Things. Summary of the Invention
[0005] The present invention provides a real-time control method for a fully automatic production line based on the Internet of Things, which can solve the technical problem in the prior art that it is difficult to achieve precise interception of defective products on the production line and simultaneous optimization of production efficiency.
[0006] In order to achieve the above object, the present invention is implemented by the following technical solutions:
[0007] The present application provides a real-time control method for a fully automatic production line based on the Internet of Things, which specifically includes the following steps:
[0008] S1. Deploy key nodes; deploy interception areas at each workstation of the production line, and configure IoT nodes on the side of the upstream production line near each interception area; the IoT nodes collect product image data, location information and electronic tag data in real time, and generate standardized data streams in a unified format through data fusion technology; S2. Perform defect identification; analyze the standardized data stream to identify defective products; S3. Calculate the remaining time; build a dynamic time prediction model to calculate the remaining time for the defective product to reach the target interception area based on the current location of the defective product, the real-time speed of the production line and the distance information of the target interception area; S4. Execute the control strategy; compare the remaining time with the preset time threshold, and send the comparison result to the production line on-site control end to execute hierarchical interception control; Among them, the time threshold includes a first time threshold and a second time threshold, and the hierarchical interception control includes: first-level interception; when the remaining time is greater than the first time threshold, the production line on-site control end reduces the production line transmission speed at the current workstation to 70% of the original speed and triggers a warning signal; second-level interception; when the remaining time is between the first time threshold and the second time threshold, the production line on-site control end jointly adjusts the production line transmission speed between adjacent workstations to 50% of the original speed, and starts the physical isolation device at the target interception area; third-level interception; when the remaining time is less than the second time threshold, the production line on-site control end triggers an emergency shutdown of the entire production line and pushes an alarm signal to the production management terminal; S5, forming feedback optimization; the production management terminal records the interception data for analysis, and optimizes the production line operation parameters.
[0009] It should be noted that the above technical solution effectively solves the lag problem of the existing control method and realizes the precise interception of defective products and the simultaneous optimization of production efficiency. Specifically, by deploying IoT nodes at each workstation on the production line, real-time collection of product images, positions and electronic tag data, and generating standardized data streams, defective products can be discovered in a timely and accurate manner, avoiding the problem of defective products entering the next process due to detection lag in traditional methods. This not only improves the product qualification rate, but also reduces subsequent production problems caused by defective products and reduces production costs. At the same time, this solution introduces a dynamic time prediction model, which can accurately calculate the remaining time based on the current position of the defective product, the production line speed and the distance to the target interception area, providing an accurate time basis for hierarchical interception control, and enabling the production line to take corresponding interception measures in time before the defective product reaches the target interception area, avoiding the problem of interception failure or reduced production efficiency due to insufficient time, thereby realizing the precise interception of defective products on the production line and the simultaneous optimization of production efficiency.
[0010] In a further technical solution, in S1, the Internet of Things node includes: a perception module and a data transmission module that are signal-connected to each other, wherein the perception module is used to collect product data, and an RFID reading and writing unit, a camera unit and a data processing unit are integrated therein, the RFID reading and writing unit is used to read the product electronic tag information and location information, the camera unit is used to collect product image information, the data processing unit is used to process the product electronic tag information, location information and product image information through data fusion technology to form a standardized data stream, and the data transmission module is used to transmit the standardized data stream over the network.
[0011] Specifically, the data fusion technology performs spatiotemporal alignment on the product electronic tag information, location information, and product image information to generate a standardized data stream containing product ID, coordinate matrix, and timestamp.
[0012] In a further technical solution, in S2, an edge computing module is used to analyze the standardized data stream and identify defective products. The edge computing module is signal-connected to the data transmission module, and a defect recognition model is pre-built inside the edge computing module. The standardized data stream is multimodally analyzed through the defect recognition model to identify defective products.
[0013] Furthermore, the defect recognition model is constructed based on a convolutional neural network structure;
[0014] The convolutional neural network structure includes an extraction layer, an intermediate layer, and an output layer. The extraction layer is used to extract mutation data features from the standardized data stream. The intermediate layer is used to parse and analyze the mutation data features and perform similarity comparison analysis with a pre-stored product defect database to determine the defect type. The output layer generates a defect determination result based on the defect type.
[0015] As a more specific technical solution, in S3, a dynamic time prediction model is built based on the production line site layout diagram and digital twin model;
[0016] The target interception area is the downstream interception area closest to the current location of the defective product.
[0017] In a further solution, the interception area includes guard plates arranged on both sides of the production line near each workstation, and the physical isolation device includes a cache component and an interception component respectively arranged on the two guard plates. When the physical isolation device is activated, the defective product is pushed into the cache component through the interception component.
[0018] In a further embodiment, the buffer assembly includes a buffer bin, which is located on a side of one of the guard plates away from the production line, and has a side opening that passes through the guard plate and is flush with the surface of the production line.
[0019] The interception assembly includes an interception push plate, which is located on the side of the other guard plate close to the production line and coincides with the side opening position of the buffer bin, and the interception push plate and the guard plate are connected via a pneumatic actuator.
[0020] Compared with the prior art, the present invention has the following advantages and beneficial effects:
[0021] 1. This invention deploys IoT nodes that integrate RFID reading and writing units, camera units, and data processing units to achieve real-time collection and localized processing of multi-source heterogeneous data (images, locations, electronic tags). It also uses data fusion technology to align multimodal data in time and space, generating a standardized data stream containing product IDs, coordinate matrices, and timestamps. This solves the spatiotemporal asynchrony and data silos inherent in traditional single-sensor detection. By leveraging the convolutional neural network defect recognition model pre-installed in the edge computing module and extracting mutation data features in layers, comparing pre-stored defect patterns, and generating defect determination results, this method achieves real-time multimodal analysis of defective products, avoiding the problems of high data heterogeneity and response delays. It also provides high timeliness for subsequent dynamic prediction and interception control, significantly improving the success rate of intercepting defective products on the production line.
[0022] 2. The present invention uses a dynamic time prediction model, combined with the digital twin layout and real-time operating parameters of the production line, to construct an accurate spatiotemporal mapping of the movement trajectory of defective products. By calculating the remaining time for the defective products to reach the nearest downstream interception area and combining it with a graded threshold strategy, it achieves dynamic priority division of interception control of defective products while avoiding shutdown of the entire line. That is, through the graded interception strategy combined with local speed reduction and physical isolation, it reduces the fluctuation of production line operating efficiency and the aggravation of equipment loss, avoids production line fluctuations caused by a single shutdown interception strategy, reduces the waste of production line resources caused by shutdown losses, and achieves coordinated optimization of interception efficiency, equipment life and operation stability. BRIEF DESCRIPTION OF THE DRAWINGS
[0023] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.
[0024] Figure 1 A schematic diagram of the steps of a real-time control method for a fully automatic production line based on the Internet of Things provided by an embodiment of the present invention;
[0025] Figure 2A schematic diagram of a flow chart of a real-time control method for a fully automatic production line based on the Internet of Things provided by an embodiment of the present invention;
[0026] Figure 3 This is a schematic diagram of the top structure of the production line provided by the present invention.
[0027] Icons: 1. Workstation; 2. IoT node; 30. Guard plate; 31. Cache bin; 32. Interceptor push plate; 33. Pneumatic actuator. DETAILED DESCRIPTION
[0028] The embodiments of the present application are described in detail below with reference to the accompanying drawings.
[0029] In the description of this application, it should be understood that the terms "center", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside", etc., indicating the orientation or position relationship, are based on the orientation or position relationship shown in the accompanying drawings, and are only for the convenience of describing this application and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation on this application.
[0030] The terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of the technical features being referred to. Thus, a feature specified as "first" or "second" may explicitly or implicitly include one or more of such features. Throughout this application, unless otherwise specified, "plurality" means two or more.
[0031] In the description of this application, it should be noted that, unless otherwise expressly specified or limited, the terms "installation," "connection," and "connection" should be understood in a broad sense. For example, they can refer to welding, bolting, or riveting; fixed connection, detachable connection, or integral connection; mechanical connection or electrical connection; direct connection or indirect connection through an intermediate medium; or internal communication between two components. Those skilled in the art will understand the specific meanings of the above terms in this application based on the specific circumstances.
[0032] Example:
[0033] Please refer to Figures 1 to 2 This embodiment provides a real-time control method for a fully automatic production line based on the Internet of Things, which specifically includes the following steps:
[0034] S1. Deploy key nodes; deploy interception areas at each workstation 1 of the production line, and configure IoT nodes 2 on the side of the upstream production line near each interception area; the IoT nodes 2 collect image data, location information and electronic tag data of the product in real time, and generate standardized data streams in a unified format through data fusion technology; S2. Perform defect identification; analyze the standardized data stream to identify defective products; S3. Calculate the remaining time; build a dynamic time prediction model to calculate the remaining time for the defective product to reach the target interception area based on the current location of the defective product, the real-time speed of the production line and the distance information of the target interception area; S4. Execute the control strategy; compare the remaining time with the preset time threshold, and send the comparison result to the production line on-site control end to execute hierarchical interception control; Among them, the time threshold includes a first time threshold and a second time threshold, and the hierarchical interception control includes: first-level interception; when the remaining time is greater than the first time threshold, the production line on-site control end reduces the production line transmission speed at the current workstation 1 to 70% of the original speed, and triggers a warning signal; second-level interception; when the remaining time is between the first time threshold and the second time threshold, the production line on-site control end jointly adjusts the production line transmission speed between adjacent workstations 1 to 50% of the original speed, and starts the physical isolation device at the target interception area; third-level interception; when the remaining time is less than the second time threshold, the production line on-site control end triggers an emergency shutdown of the entire production line and pushes an alarm signal to the production management terminal; S5, forming feedback optimization; the production management terminal records the interception data for analysis, and optimizes the production line operation parameters.
[0035] It should be understood that the first time threshold is the maximum critical point of the remaining time for the defective product to reach the target interception area, while the second time threshold is the minimum critical point of the remaining time for the defective product to reach the target interception area. Both are theoretical requirements and are calculated based on the specific coordinates of the defective product and the average transmission speed of the production line.
[0036] That is, when the calculated remaining time is greater than the first time threshold, it indicates that the defective products still have sufficient time to reach the interception area. In this case, first-level interception is prioritized over destructive intervention. This involves reducing the conveyor speed to 70% of the original speed and triggering a warning signal, thereby slowing the flow of defective products while minimizing the impact on production line efficiency.
[0037] When the calculated remaining time is between the first time threshold and the second time threshold, it indicates that the defective product has approached the interception area and time is urgent. It is necessary to start collaborative control by linking the adjacent workstation 1 to reduce the speed to 50% and start the physical isolation device to limit the spread of the defective product and prepare for interception; when the calculated remaining time is less than the second time threshold, it indicates that it is not enough to stop the flow of defective products through conventional interception means, and the third-level interception must be triggered, that is, through emergency shutdown of the entire line and alarm push, the defective products are completely blocked from entering the downstream.
[0038] It should be noted that the above technical solution effectively solves the hysteresis problem of the existing control method and realizes the precise interception of defective products and the simultaneous optimization of production efficiency. Specifically, by deploying the Internet of Things node 2 at each workstation 1 of the production line, the product image, location and electronic tag data are collected in real time, and a standardized data stream is generated, so that defective products can be discovered in a timely and accurate manner, avoiding the problem of defective products entering the next process due to detection lag in the traditional method. This not only improves the product qualification rate, but also reduces subsequent production problems caused by defective products and reduces production costs. At the same time, by calculating the remaining time for the defective product to reach the target interception area in real time and performing time threshold comparison, the hierarchical interception control is triggered, so that it can accurately match the interception strategy with the flow state of the defective product, avoid production line fluctuations caused by a single shutdown interception strategy, reduce the waste of production line resources caused by shutdown losses, and realize the coordinated optimization of interception efficiency, equipment life and operation stability.
[0039] As a preferred embodiment, in S1 above, the Internet of Things node 2 includes: a perception module and a data transmission module that are signal-connected to each other, wherein the perception module is used to collect product data, and an RFID reading and writing unit, a camera unit and a data processing unit are integrated therein, the RFID reading and writing unit is used to read the product electronic tag information and location information, the camera unit is used to collect product image information, the data processing unit is used to process the product electronic tag information, location information and product image information through data fusion technology to form a standardized data stream, and the data transmission module is used to transmit the standardized data stream over the network.
[0040] The present invention realizes the real-time collection and localization processing of multi-source heterogeneous data (images, locations, electronic tags) by deploying IoT nodes 2 to integrate RFID reading and writing units, camera units and data processing units. It also uses data fusion technology to align multimodal data in time and space, generating a standardized data stream containing product ID, coordinate matrix and timestamp, thus solving the problems of time and space asynchrony and data islands in traditional single-sensor detection.
[0041] Based on the above embodiment, a standardized data stream in a unified format is generated through data fusion technology to solve the asynchronous and heterogeneous problems of various types of data, thereby providing highly consistent input for defect identification.
[0042] Specifically, the data fusion technology performs spatiotemporal alignment on the product electronic tag information, location information, and product image information to generate a standardized data stream containing product ID, coordinate matrix, and timestamp.
[0043] Furthermore, the spatiotemporal alignment algorithm is used to resolve the inconsistencies in timestamps, coordinate systems, and physical dimensions of various data. The expression is as follows:
[0044] ;
[0045] In the above formula, is the position coordinate of the product on the production line, is the timestamp after time-space synchronization, are the image pixel coordinates in the image data, is the original timestamp; The spatiotemporal calibration matrix is obtained through offline calibration. It is an environmental compensation vector used to correct environmental interference factors of the production line, including signal delay and mechanical vibration. The specific correction measure is to use data pre-reading technology to perform timestamp interpolation compensation on continuous data.
[0046] As a preferred implementation method, in S2 above, an edge computing module is used to analyze the standardized data stream and identify defective products. The edge computing module is signal-connected to the data transmission module, and a defect recognition model is pre-built inside the edge computing module. The defect recognition model is used to perform multimodal analysis on the standardized data stream and identify defective products.
[0047] Furthermore, the defect recognition model is constructed based on a convolutional neural network structure;
[0048] The convolutional neural network structure includes an extraction layer, an intermediate layer, and an output layer. The extraction layer is used to extract mutation data features from the standardized data stream. The intermediate layer is used to parse and analyze the mutation data features and perform similarity comparison analysis with a pre-stored product defect database to determine the defect type. The output layer generates a defect determination result based on the defect type.
[0049] In the above embodiment, the edge computing module is combined with the defect recognition model to build an efficient closed-loop control logic for real-time processing of multimodal data, accurate defect recognition, and dynamic hierarchical interception, thereby achieving efficient and accurate defect recognition of products. The edge computing module uses the defect recognition model to perform multimodal analysis on the standardized data stream to process the data stream in real time, which can quickly respond to changes in product status on the production line and accurately identify defective products in a timely manner, greatly reducing subsequent production problems caused by defective products entering the next process.
[0050] Specifically, the defect recognition model is deployed in the edge computing module and consists of a hierarchical processing architecture consisting of an extraction layer, an intermediate layer, and an output layer. The extraction layer uses a multi-dimensional convolution kernel to simultaneously analyze the image pixel matrix, position coordinate sequence, and electronic tag code in the standardized data stream, extracting multimodal features such as image grayscale mutations, position trajectory, acceleration, and tag parameter deviation to generate a primary feature map.
[0051] The middle layer compresses feature dimensions through a maximum pooling operation and introduces a channel attention mechanism to dynamically assign feature weights to image, position, and label data. The weighted feature vectors are then matched against standard defect types in a pre-stored defect database for similarity and mapped to the defect type probability space through a fully connected layer.
[0052] The output layer uses the Softmax function to normalize the probability distribution and combines it with the preset confidence scoring criteria to determine the defect type, generating a result signal that includes the defect category, spatiotemporal coordinates, and confidence score. This enables efficient and accurate defect identification of products, improves the product qualification rate, and triggers subsequent hierarchical interception control instructions, including local speed reduction, physical isolation, or emergency shutdown, ensuring the effective interception of defective products while minimizing the impact on production efficiency.
[0053] As a preferred implementation, in the above S3, the dynamic time prediction model is built based on the production line site layout diagram and the digital twin model;
[0054] It should be understood that the calculation of the remaining time is based on the current position of the defective product, the real-time speed of the production line, and the distance information of the target interception area. It is calculated to calculate the remaining time for the defective product to reach the target interception area. It should be further explained that during the actual operation of the production line, unexpected situations are inevitable. For example, the conveyor belt may suddenly fluctuate due to mechanical failure or external disturbance, which may affect the accuracy of the calculated remaining time and adversely affect the subsequent interception of defective products.
[0055] In view of this, this solution introduces a dynamic time prediction model, which constructs a digital twin topology map based on the physical layout of the production line, obtains the current position coordinates, production line transmission speed and acceleration information of the defective product in real time, and obtains the path distance of the target interception area through the constructed digital twin topology map, and dynamically calculates the remaining time through the kinematic equation. Specifically, it cleverly incorporates the sudden fluctuation of the production line speed into the calculation scope, and calculates the accurate remaining time through the time integral function. It significantly reduces the error in the remaining time calculation caused by speed mutation or environmental interference, and can accurately calculate the remaining time for the defective product to reach the target interception area, providing an accurate time basis for hierarchical interception control.
[0056] For the dynamic time prediction model, its expression is as follows:
[0057] ,
[0058] In the above formula, To calculate the exact remaining time, The final time when the defective product is expected to arrive at the target interception area, To identify the current moment of defective products, is the path distance of the defective product on the production line, For the production line Real-time speed at all times, For the production line The acceleration of the momentary fluctuations, For the current moment arrive The time interval between moments, For the integration process Small increments of time, is the Kalman filter, and Correction of measurement noise.
[0059] The target interception area is the downstream interception area closest to the current location of the defective product.
[0060] Based on the above embodiments, a more preferred implementation method of the interception area is proposed here. It should be understood that the interception area includes guard plates 30 respectively arranged on both sides of the production line near each workstation 1, and the physical isolation device includes a cache component and an interception component respectively arranged on the two guard plates 30. When the physical isolation device is started, the defective product is pushed into the cache component through the interception component.
[0061] For further solutions, please refer to Figure 3The cache component includes a cache bin 31, which is located on a side of one of the guard plates 30 away from the production line, and the cache bin 31 has a side opening and passes through the guard plate 30 and is flush with the surface of the production line;
[0062] The interception assembly includes an interception push plate 32, which is located on the side of the other guard plate 30 close to the production line and coincides with the side opening position of the buffer bin 31, and the interception push plate 32 and the guard plate 30 are connected through a pneumatic actuator 33.
[0063] Regarding pneumatic actuator 33, it should be understood that it is widely used in industrial automated production lines, enabling fast, efficient, and precise mechanical motion control. In the application scenario you mentioned, pneumatic actuator 33 is primarily used to push intercepting push plate 32 toward buffer bin 31. To achieve this function, various actuator types can be used, such as robotic arms, cylinders, and hydraulic push rods. In actual implementation, the choice can be tailored to the specific situation. In this solution, pneumatic actuator 33 uses a cylinder.
[0064] It should be noted that when intercepting defective products on the production line, the traditional method is mostly to deploy robotic arms on both sides of the production line or to block them through interception devices. However, in actual application, the traditional method is often inconvenient. Since the production line is always in motion during operation, if a direct interception method is adopted, it will cause congestion to other products on the production line, and in serious cases, products will collide with each other, causing chaos in the entire production line. In view of this, the present embodiment forms a dynamic directional disengagement channel by symmetrically setting guard plates 30 on both sides of the workstation 1, integrating a buffer bin 31 on one side and deploying an interception push plate 32 on the other side. When the defective product reaches the target interception area, the cylinder responds in time and pushes the interception push plate 32 out horizontally, thereby accurately pushing the defective product across the conveyor belt into the opposite side buffer bin 31 to avoid the interference of the interception action on non-defective products, thereby improving the interception efficiency of defective products.
[0065] In a further preferred embodiment, an oblique guide slope is designed on the inner wall of the buffer bin 31 near its opening to ensure that defective products can quickly and directionally slide into the bin for storage under the action of gravity and thrust, eliminating the risk of secondary collision caused by product retention at the edge after traditional interception.
[0066] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any modifications or substitutions that can be easily conceived by a person skilled in the art within the technical scope of the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be based on the scope of protection of the claims.
Claims
1. A real-time control method for a fully automatic production line based on the Internet of Things, characterized in that: The method specifically comprises the following steps: S1. Deploy key nodes; deploy interception areas at each workstation (1) of the production line, and configure IoT nodes (2) on the side of the upstream production line near each interception area. The Internet of Things node (2) collects image data, location information and electronic tag data of the product in real time, and generates a standardized data stream in a unified format through data fusion technology; S2. Perform defect identification: Analyze standardized data streams to identify defective products; S3. Calculate the remaining time: Build a dynamic time prediction model to calculate the remaining time for the defective product to reach the target interception area based on the current location of the defective product, the real-time speed of the production line, and the distance information of the target interception area; S4, executing the control strategy; comparing the remaining time with the preset time threshold, and sending the comparison result to the production line on-site control terminal to execute hierarchical interception control; Wherein, the time threshold includes a first time threshold and a second time threshold; The hierarchical interception control includes: Level 1 interception: When the remaining time is greater than the first time threshold, the production line on-site control terminal reduces the production line transmission speed at the current workstation (1) to 70% of the original speed and triggers a warning signal; Secondary interception: When the remaining time is between the first time threshold and the second time threshold, the production line field control terminal will adjust the production line transmission speed between adjacent workstations (1) to 50% of the original speed, and start the physical isolation device at the target interception area; Level 3 interception: When the remaining time is less than the second time threshold, the production line on-site control terminal triggers an emergency shutdown of the entire production line and sends an alarm signal to the production management terminal; S5. Feedback optimization is formed; the production management terminal records the intercepted data for analysis and optimizes the production line operation parameters.
2. The real-time control method for a fully automatic production line based on the Internet of Things according to claim 1 is characterized in that: In S1, the Internet of Things node (2) includes: a perception module and a data transmission module that are mutually signal-connected, wherein the perception module is used to collect product data, and internally integrates an RFID reading and writing unit, a camera unit, and a data processing unit, the RFID reading and writing unit is used to read product electronic tag information and location information, the camera unit is used to collect product image information, the data processing unit is used to process product electronic tag information, location information, and product image information through data fusion technology to form a standardized data stream, and the data transmission module is used to transmit the standardized data stream over a network.
3. The real-time control method for a fully automatic production line based on the Internet of Things according to claim 2 is characterized in that: The data fusion technology performs spatiotemporal alignment on the product electronic tag information, location information, and product image information to generate a standardized data stream containing product ID, coordinate matrix, and timestamp.
4. The real-time control method for a fully automatic production line based on the Internet of Things according to claim 1, characterized in that: In S2, an edge computing module is used to analyze the standardized data stream and identify defective products. The edge computing module is signal-connected to the data transmission module, and a defect recognition model is pre-built inside the edge computing module. The standardized data stream is multimodally analyzed through the defect recognition model to identify defective products.
5. The real-time control method for a fully automatic production line based on the Internet of Things according to claim 4 is characterized in that: The defect recognition model is constructed based on a convolutional neural network structure; The convolutional neural network structure includes an extraction layer, an intermediate layer, and an output layer. The extraction layer is used to extract mutation data features from the standardized data stream. The intermediate layer is used to parse and analyze the mutation data features and perform similarity comparison analysis with a pre-stored product defect database to determine the defect type. The output layer generates a defect determination result based on the defect type.
6. The real-time control method for a fully automatic production line based on the Internet of Things according to claim 1 is characterized in that: In S3, the dynamic time prediction model is built based on the production line site layout diagram and digital twin model; The target interception area is the downstream interception area closest to the current location of the defective product.
7. The real-time control method for a fully automatic production line based on the Internet of Things according to claim 1 is characterized in that: The interception area includes guard plates (30) respectively arranged on both sides of the production line near each workstation (1), and the physical isolation device includes a cache component and an interception component respectively arranged on the two guard plates (30). When the physical isolation device is activated, the defective product is pushed into the cache component through the interception component.
8. The real-time control method for a fully automatic production line based on the Internet of Things according to claim 7 is characterized in that: The cache component comprises a cache bin (31), the cache bin (31) being located on a side of one of the guard plates (30) away from the production line, and the cache bin (31) is open on a side and passes through the guard plate (30) to be flush with the surface of the production line; The interception assembly includes an interception push plate (32), which is located on the side of the other guard plate (30) close to the production line and coincides with the side opening position of the buffer bin (31), and the interception push plate (32) and the guard plate (30) are connected via a pneumatic actuator (33).