Edge processing apparatus, digital factory, and equipment control method

Abstract

The invention relates to the technical field of hydrogen storage tank production, and discloses an edge processing device, a digital factory and an equipment control method, the edge processing device comprises a communication unit, an information processing unit, a visual detection unit, a strategy unit, a control signal unit and a visualization unit; wherein the communication unit is used for information interaction; the information processing unit is used for processing product information; the visual detection unit is used for performing defect detection on the product; the strategy unit configures a production control strategy generated based on the product information and the defect detection result; the control signal unit is used for generating a control signal; and the visualization unit is used for processing the product information and the production control strategy into a visual format and sending the visual format to the cloud module. The intelligent level of the hydrogen storage tank production line is remarkably improved, and a powerful guarantee is provided for efficient and high-quality production of hydrogen storage tanks.

Description

Technical Field

[0001] The present invention relates to the technical field of hydrogen storage tank production, and in particular to an edge processing device, a digital factory and an equipment control method. Background Art

[0002] As a key component of the hydrogen energy industry chain, hydrogen storage tanks not only affect the transportation and storage efficiency of hydrogen, but are also an important link in ensuring the safe use of hydrogen energy. The production of hydrogen storage tanks involves multiple complex process flows, including but not limited to extrusion molding, injection molding, welding, etc. These process flows require not only a high degree of precision and stability, but also efficient automated production and a strict quality control system. Traditional hydrogen storage tank production lines mostly use centralized control systems. Although this system can achieve a certain degree of automated production, it has obvious deficiencies in flexibility, response speed, data processing capabilities, etc.

[0003] The data processing capabilities of traditional edge computing devices are relatively limited, and it is difficult to meet the real-time processing requirements of massive data generated in complex process flows. In particular, in the production process of hydrogen storage tanks, a large amount of unstructured data such as sensor data and image data is involved, which puts higher requirements on the computing power and storage capacity of edge computing devices. The quality of hydrogen storage tanks is directly related to the safe use of hydrogen energy, so the quality control requirements in the production process are extremely high. Traditional quality inspection methods mostly rely on manual visual inspection or simple machine vision inspection, with low inspection accuracy and efficiency, and are prone to missed inspections or false inspections. Especially in key processes such as welding, how to achieve high-precision and high-efficiency quality inspection is a major problem. Existing hydrogen storage tank production lines mostly adopt a fixed production mode, and once the production plan is formulated, it is difficult to adjust. Faced with market changes or emergencies, the production line often cannot respond quickly, resulting in low production efficiency. In addition, in the production process of hydrogen storage tanks of different batches and models, how to achieve dynamic scheduling and optimize resource allocation is also a major challenge. At present, most of the equipment in each link of the hydrogen storage tank production line operates independently, lacking an effective coordination mechanism. For example, the lack of real-time data sharing and linkage control between process flows such as extrusion molding, injection molding, and welding has affected overall production efficiency and product quality. How to achieve efficient collaboration between the various process flows and improve the intelligence level of the entire production line is an issue that needs to be addressed urgently.

[0004] For example, a Chinese patent application with publication number CN111580471A discloses a production line control system and method, wherein the production line control system includes a shooting device, a control module, a mobile device and a scanning device. The shooting device generates a first image, the control module is connected to the shooting device in communication, and the control module analyzes the first image and generates positioning information. The mobile device is connected to the control module in communication, and the scanning device is connected to the control module in communication. The scanning device is arranged on the mobile device and generates scanning information. The scanning device and the shooting device are spaced apart from each other in a first direction, the scanning device has the ability to move in a second direction through the mobile device, and the first direction is not parallel to the second direction. The control module controls the mobile device according to the positioning information, so that the mobile device moves the scanning device to a specific position, and the control module controls the scanning device to generate scanning information at a specific position. This patent has the problem raised by this background technology: how to achieve high-precision and high-efficiency quality inspection is a major problem.

[0005] The information disclosed in this background technology section is only intended to enhance the understanding of the overall background of the invention and should not be regarded as an acknowledgement or any form of suggestion that the information constitutes the prior art already known to ordinary technicians in this field. Summary of the invention

[0006] The technical problem to be solved by the present invention is to overcome the defects of the prior art, provide an edge processing device, a digital factory and an equipment control method, improve the intelligence level of the hydrogen storage tank production line, and provide guarantee for the efficient and high-quality production of hydrogen storage tanks.

[0007] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0008] In a first aspect, the present invention provides an edge processing device, including a communication unit, an information processing unit, a visual detection unit, a strategy unit, a control signal unit, and a visualization unit; wherein:

[0009] The communication unit is used for information exchange between the edge processing device and the field module and the cloud module;

[0010] The information processing unit is used to process the product information collected from the field module and the cloud module, and transmit the processed product information to the cloud module;

[0011] The visual inspection unit is used to process product images collected by the on-site module and perform defect detection on the product;

[0012] The strategy unit configures a production control strategy generated based on the product information and defect detection results;

[0013] The control signal unit generates a control signal based on the production control strategy and sends the control signal to the field module;

[0014] The visualization unit is used to process the product information and production control strategy into a visualization format and send it to the cloud module.

[0015] As a preferred solution of the edge processing device of the present invention, wherein: the product information includes product production information and product inventory information;

[0016] The product production information includes the production information of the extruded tube, the production information of the end cap, and the production information of the welded liner; wherein the production information of the extruded tube includes the number, specification parameters, and extrusion production time of the extruded tube; the production information of the end cap includes the number, specification parameters, and injection production time of the end cap; the production information of the welded liner includes the number of the welded liner, the number and specification parameters of the extruded tube involved in welding, the number of the end cap involved in welding, and the welding production time;

[0017] The product inventory information includes inventory information of extruded tubes, inventory information of end caps, and inventory information of welded liners; wherein, the inventory information of extruded tubes includes the inventory quantity of extruded tubes, the total number of storage locations of the extruded tube cache vertical warehouse, and the storage location number and storage location status of each storage location; the storage location status of the extruded tube cache vertical warehouse includes free and occupied, if the storage location status is occupied, the storage location status also includes the number and storage time of the stored extruded tubes; the inventory information of end caps includes the inventory quantity of end caps, the total number of storage locations of the end cap cache vertical warehouse, and the storage location number and storage location status of each storage location; the storage location status of the end cap cache vertical warehouse includes free and occupied, if the storage location status is occupied, the storage location status also includes the number and storage time of the stored end caps; the inventory information of welded liners includes the inventory quantity of welded liners, the total number of storage locations of the welded liners cache vertical warehouse, and the storage location number and storage location status of each storage location; the storage location status of the welded liners cache vertical warehouse includes free and occupied, if the storage location status is occupied, the storage location status also includes the number and storage time of the stored welded liners.

[0018] As a preferred solution of the edge processing device of the present invention, the method in which the information processing unit processes the product information is as follows:

[0019] Receive and record the production information of the extruded tube; if the extruded tube is stored in the extruded tube buffer warehouse, add the number and storage time of the extruded tube to the storage location status of the corresponding storage location, and update the inventory information of the extruded tube; when any extruded tube is out of the warehouse, reset the corresponding storage location status to free, and update the inventory information of the extruded tube;

[0020] Receive and record the production information of the head; if the head is stored in the head cache, add the head number and storage time to the storage location status of the corresponding storage location, and update the inventory information of the head; when any head is shipped out, reset the corresponding storage location status to free, and update the inventory information of the head;

[0021] If the extruded tube and the head are transported to the welding equipment for welding, the number and specification parameters of the extruded tube involved in the welding, the number of the head involved in the welding, and the welding production time are added to the production information of the welding liner; if the welding liner is stored in the welding liner buffer warehouse, the number and warehousing time of the welding liner are added to the storage location status of the corresponding storage location, and the inventory information of the welding liner is updated; when any welding liner is shipped out, the corresponding storage location status is reset to idle, and the inventory information of the welding liner is updated.

[0022] As a preferred solution of the edge processing device of the present invention, wherein: the product images include a CCD image of the extruded tube, a CCD image of the end cap, and an ultrasonic image of the welded liner;

[0023] The visual inspection unit includes an image processing subunit, a first inspection model, a second inspection model, a third inspection model, and a defect prompting subunit; wherein:

[0024] The image processing subunit is used to pre-process the CCD image of the extruded tube, the CCD image of the end cap, and the ultrasonic image of the welded liner, and transmit them to the first detection model, the second detection model, and the third detection model respectively;

[0025] The first detection model is a convolutional neural network model trained based on a CCD image set of an extruded tube, and is used to process the CCD image of the extruded tube and perform defect detection on the extruded tube;

[0026] The second detection model is a convolutional neural network model trained based on the CCD image set of the head, which is used to process the CCD image of the head and perform defect detection on the head;

[0027] The third detection model is a convolutional neural network model trained based on an ultrasonic image set of a welded liner, which is used to process the ultrasonic image of the welded liner and perform defect detection on the welded liner;

[0028] The defect prompt subunit is used to send the defect type and the number of the corresponding product to the strategy unit when any extruded tube, end cap or welded liner is detected to have a defect.

[0029] As a preferred solution of the edge processing device of the present invention, wherein: the strategy unit includes a query subunit, a calculation subunit, a data subunit, a control subunit, and a strategy subunit; the production control strategy includes a first control strategy, a second control strategy, and a third control strategy;

[0030] The strategy unit generates a second control strategy based on the defect detection result, which is as follows:

[0031] Record the defect type, corresponding product number, and production time of the defective product and send them to the cloud module; move the defective product to the manual maintenance station; suspend the production equipment of the defective product and perform working status detection and reset;

[0032] When the extruded tube, the end cap or the welded liner is shipped out of the warehouse, the strategy unit generates a third control strategy based on the product information, which is as follows:

[0033] The query subunit obtains the corresponding inventory information; the strategy subunit counts the storage location status of each storage location in the corresponding inventory information and sorts them by the storage time, and then carries out the shipment in order from the earliest to the latest storage time.

[0034] As a preferred solution of the edge processing device of the present invention, the strategy unit is configured with a genetic algorithm, and at the beginning of each control cycle, a first control strategy is generated based on the product information, which is as follows:

[0035] The query subunit queries product inventory information from the cloud module;

[0036] The data subunit generates the primary population based on the input population size, crossover probability, and mutation probability, and executes the population iteration instruction; the chromosome of any individual in the primary population includes the welding rate, injection rate, and extrusion rate in the current regulation cycle;

[0037] The welding rate indicates the number of welded inner shells produced by the welding equipment per unit time; the injection molding rate indicates the number of heads produced by the injection molding equipment per unit time; the extrusion rate indicates the number of extruded tubes produced by the extrusion equipment per unit time;

[0038] The calculation subunit calculates the fitness of each individual based on the product inventory information;

[0039] The control subunit generates the population iteration instruction based on the input maximum number of iterations; the population iteration instruction is specifically as follows: at each iteration, the fitness of each individual is counted, and the selection operation, crossover operation, and mutation operation are performed in sequence to generate the next generation population; the population iteration is repeated until the maximum number of iterations is reached;

[0040] The strategy subunit extracts the final welding rate, injection rate, and extrusion rate from the individual with the highest fitness in the last iteration.

[0041] As a preferred solution of the edge processing device of the present invention, the calculation formula of the fitness is as follows:

[0042] H=w1·S-w2·(a1+a2+a3)-w3·F;

[0043] Among them, H represents the fitness of any individual; S represents the expected number of welded liners to be produced, which is calculated by multiplying the welding rate by the length of the control cycle; a1 represents the inventory adjustment of welded liners; a2 represents the inventory adjustment of heads; a3 represents the inventory adjustment of extruded tubes; w1, w2, and w3 are all weight coefficients; and F represents the penalty term.

[0044] As a preferred solution of the edge processing device of the present invention, the calculation formula of the penalty term F is as follows:

[0045] F = r1 + r2 + r3;

[0046] Among them, r1 represents the penalty factor of the welding liner; if the welding liner cache vertical warehouse is overloaded or overspent, r1 is assigned to 1, otherwise, r1 is assigned to 0; r2 represents the penalty factor of the head; if the head cache vertical warehouse is overloaded or overspent, r2 is assigned to 1, otherwise, r2 is assigned to 0; r3 represents the penalty factor of the extruded tube; if the extruded tube cache vertical warehouse is overloaded or overspent, r3 is assigned to 1, otherwise, r3 is assigned to 0;

[0047] The method for judging whether any warehouse among the welding liner cache warehouse, the head cache warehouse, and the extruded tube cache warehouse is overrun or overloaded is as follows: if the inventory quantity of the product corresponding to the warehouse at the beginning of the current control cycle plus the expected inventory adjustment is greater than the total number of warehouse locations, the warehouse is overloaded; if the inventory quantity of the product corresponding to the warehouse at the beginning of the current control cycle plus the expected inventory adjustment is less than 0, the warehouse is overrun.

[0048] In a second aspect, the present invention provides a digital factory, including an edge processing device, a field module, and a cloud module; wherein:

[0049] The cloud module includes a data acquisition device and a control device; wherein the data acquisition device is used to collect product information and transmit the product information to the edge computing module; the edge computing module is used to process the product information and generate a control signal based on the production control strategy; and send the control signal to the control device; the control device controls the operation of the equipment on the hydrogen storage tank production line based on the control signal;

[0050] The cloud module includes a database and a visualization device; wherein the database is used to store product information processed by the edge computing module; the visualization device is used to visually display and store product information processed by the edge computing module and production control strategies.

[0051] In a third aspect, the present invention provides a device control method, comprising the following steps:

[0052] S1: Collect product information and product pictures, and process the product information;

[0053] S2: Performing defect detection on the product based on the product image;

[0054] S3: generating a production control strategy based on the product information and defect detection results;

[0055] S4: generating a control signal based on the production control strategy;

[0056] S5: Control the operation of equipment on the hydrogen storage tank production line based on the control signal.

[0057] Compared with the prior art, the beneficial effects achieved by the present invention are as follows:

[0058] The edge processing device of the present invention is equipped with a high-performance information processing unit, which can process a large amount of data from the field module and the cloud module in real time, including product production information and inventory information. This not only improves the speed of data processing, but also ensures the accuracy of data processing.

[0059] Through the detection model based on convolutional neural network, various defects of extruded tubes, heads and welded liners, such as deformation, bubbles, voids, cracks, incomplete penetration, slag inclusion, etc., can be detected efficiently and accurately, significantly improving the quality control level of products. When a defect is detected, the defect prompt subunit will send the defect type and the corresponding product number to the strategy unit, automatically record and process the defective products, reduce manual intervention, and improve production efficiency.

[0060] The strategy unit is equipped with a genetic algorithm, which can generate the optimal production control strategy based on product information and inventory information. By dynamically adjusting the welding rate, injection rate and extrusion rate, the production plan can be flexibly adjusted to cope with market changes and emergencies. The strategy unit can monitor inventory information in real time and generate a reasonable outbound strategy based on the inventory status to avoid inventory backlogs or shortages and improve the efficiency of inventory management.

[0061] The edge processing device of the present invention includes a communication unit, an information processing unit, a visual detection unit, a strategy unit, a control signal unit and a visualization unit. The modules achieve efficient collaboration to ensure the smooth operation of the entire production process. Through the communication unit and the information processing unit, real-time data sharing and linkage control between the various process flows are achieved to improve the overall production efficiency and product quality. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. 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 creative labor. Among them:

[0063] Figure 1 A schematic diagram of the structure of an edge processing device provided by the present invention;

[0064] Figure 2 A structural schematic diagram of a digital factory provided by the present invention;

[0065] Figure 3 The present invention provides a flow chart of a device control method. DETAILED DESCRIPTION

[0066] The technical solution of the present invention is described in detail below through the accompanying drawings and specific embodiments. It should be understood that the embodiments of the present invention and the specific features in the embodiments are detailed descriptions of the technical solution of the present invention, rather than limitations on the technical solution of the present invention. The embodiments of the present invention and the technical features in the embodiments may be combined with each other unless there is a conflict.

[0067] Example 1

[0068] This embodiment introduces an edge processing device, referring to Figure 1 , the edge processing device includes a communication unit, an information processing unit, a visual detection unit, a strategy unit, a control signal unit, and a visualization unit; wherein:

[0069] The communication unit is used for information exchange between the edge processing device and the field module and the cloud module;

[0070] The communication unit is equipped with a network interface card, which supports wired (such as Ethernet) and wireless (such as WiFi, Bluetooth, Zigbee) connections between the field module and the cloud module; it is also equipped with a switch to manage and forward data packets to achieve communication between different networks. At the same time, the communication unit supports a variety of industrial communication protocols, such as OPC UA, Modbus, EtherCAT, and uses encryption technology to ensure the security of data transmission; thus, it can receive product information from the field module, upload the product information to the cloud module, and send control signals to the field module.

[0071] The information processing unit is used to process the product information collected from the field module and the cloud module, and transmit the processed product information to the cloud module;

[0072] The product information includes product production information and product inventory information;

[0073] The product production information includes the production information of the extruded tube, the production information of the end cap, and the production information of the welded liner; wherein the production information of the extruded tube includes the number, specification parameters, and extrusion production time of the extruded tube; the production information of the end cap includes the number, specification parameters, and injection production time of the end cap; the production information of the welded liner includes the number of the welded liner, the number and specification parameters of the extruded tube involved in welding, the number of the end cap involved in welding, and the welding production time;

[0074] The product inventory information includes inventory information of extruded tubes, inventory information of end caps, and inventory information of welded liners; wherein, the inventory information of extruded tubes includes the inventory quantity of extruded tubes, the total number of storage locations of the extruded tube cache vertical warehouse, and the storage location number and storage location status of each storage location; the storage location status of the extruded tube cache vertical warehouse includes free and occupied, if the storage location status is occupied, the storage location status also includes the number and storage time of the stored extruded tubes; the inventory information of end caps includes the inventory quantity of end caps, the total number of storage locations of the end cap cache vertical warehouse, and the storage location number and storage location status of each storage location; the storage location status of the end cap cache vertical warehouse includes free and occupied, if the storage location status is occupied, the storage location status also includes the number and storage time of the stored end caps; the inventory information of welded liners includes the inventory quantity of welded liners, the total number of storage locations of the welded liners cache vertical warehouse, and the storage location number and storage location status of each storage location; the storage location status of the welded liners cache vertical warehouse includes free and occupied, if the storage location status is occupied, the storage location status also includes the number and storage time of the stored welded liners;

[0075] The method for the information processing unit to process the product information is as follows:

[0076] Receive and record the production information of the extruded tube; if the extruded tube is stored in the extruded tube buffer warehouse, add the number and storage time of the extruded tube to the storage location status of the corresponding storage location, and update the inventory information of the extruded tube; when any extruded tube is out of the warehouse, reset the corresponding storage location status to free, and update the inventory information of the extruded tube;

[0077] Receive and record the production information of the head; if the head is stored in the head cache, add the head number and storage time to the storage location status of the corresponding storage location, and update the inventory information of the head; when any head is shipped out, reset the corresponding storage location status to free, and update the inventory information of the head;

[0078] If the extruded tube and the head are transported to the welding equipment for welding, the number and specification parameters of the extruded tube involved in the welding, the number of the head involved in the welding, and the welding production time are added to the production information of the welding liner; if the welding liner is stored in the welding liner buffer warehouse, the number and warehousing time of the welding liner are added to the storage location status of the corresponding storage location, and the inventory information of the welding liner is updated; when any welding liner is shipped out, the corresponding storage location status is reset to idle, and the inventory information of the welding liner is updated.

[0079] The visual inspection unit is used to process product images collected by the on-site module and perform defect detection on the product;

[0080] The product images include CCD images of extruded tubes, CCD images of end caps, and ultrasonic images of welded liner;

[0081] The visual inspection unit includes an image processing subunit, a first inspection model, a second inspection model, a third inspection model, and a defect prompting subunit; wherein:

[0082] The image processing subunit is used to preprocess the CCD image of the extruded tube, the CCD image of the head, and the ultrasonic image of the welded liner, and transmit them to the first detection model, the second detection model, and the third detection model respectively; the preprocessing includes image denoising, contrast enhancement, image cropping, etc.; remove noise, enhance image quality, ensure the accuracy of subsequent processing, and adjust the image to a uniform size and format for input into the detection model.

[0083] The first detection model is a convolutional neural network model trained based on the CCD image set of the extruded tube. It is used to process the CCD image of the extruded tube and perform defect detection on the extruded tube. It can effectively identify various defects in the extruded tube, such as deformation, bubbles, and voids.

[0084] The second detection model is a convolutional neural network model trained based on the CCD image set of the head. It is used to process the CCD image of the head and perform defect detection on the head. It can effectively identify various defects in the head, such as deformation, bubbles, and voids.

[0085] The third detection model is a convolutional neural network model trained based on the ultrasonic image set of the welded liner. It is used to process the ultrasonic image of the welded liner and perform defect detection on the welded liner. It can effectively identify various defects in the welded liner, such as cracks, incomplete penetration, slag inclusions, etc.

[0086] The defect prompt subunit is used to send the defect type and the corresponding product number to the strategy unit when any defect is detected in any extruded tube, end cap or welded liner. With this configuration, the visual inspection unit can efficiently and accurately complete the defect detection task of extruded tubes, end caps and welded liner, thereby improving the quality control level of the entire production line.

[0087] The strategy unit configures a production control strategy generated based on the product information and defect detection results;

[0088] The strategy unit includes a query subunit, a calculation subunit, a data subunit, a control subunit, and a strategy subunit; the production control strategy includes a first control strategy, a second control strategy, and a third control strategy;

[0089] The strategy unit is configured with a genetic algorithm. At the beginning of each regulation cycle, a first control strategy is generated based on the product information, which is as follows:

[0090] The query subunit queries product inventory information from the cloud module;

[0091] The data subunit generates the primary population based on the input population size, crossover probability, and mutation probability, and executes the population iteration instruction; the chromosome of any individual in the primary population includes the welding rate, injection rate, and extrusion rate in the current regulation cycle;

[0092] The welding rate indicates the number of welding liners produced by the welding equipment per unit time; the injection molding rate indicates the number of heads produced by the injection molding equipment per unit time; the extrusion rate indicates the number of extruded tubes produced by the extrusion equipment per unit time; the data subunit is respectively configured with upper and lower thresholds of the welding rate, injection molding rate, and extrusion rate, and the welding rate, injection molding rate, and extrusion rate are randomly generated in the corresponding upper and lower thresholds to form a primary population;

[0093] The calculation subunit calculates the fitness of each individual based on the product inventory information; the formula is as follows:

[0094] H=w1·S-w2·(a1+a2+a3)-w3·F;

[0095] Among them, H represents the fitness of any individual; S represents the expected number of welded liner production, which is calculated by multiplying the welding rate by the length of the control cycle; a1 represents the inventory adjustment of the welded liner; a2 represents the inventory adjustment of the head; a3 represents the inventory adjustment of the extruded tube;

[0096] The calculation method of a1 is as follows: Based on the ultrasonic inspection rate of the welded inner liner, the number of inner liner D expected to complete ultrasonic inspection within the current control cycle is calculated. The difference between D and the expected number of welded inner liner production S is a1:

[0097] The calculation method of a2 is as follows: the expected number of heads produced Z is calculated by multiplying the injection rate by the length of the control cycle. Since two heads are consumed to weld one welding liner, the difference between Z and twice S is a2;

[0098] The calculation method of a3 is as follows: the expected number of extruded tubes produced P is calculated by multiplying the extrusion rate by the length of the control cycle. Since one extruded tube is consumed to weld one welding liner, the difference between Z and S is a3;

[0099] w1, w2, and w3 are all weight coefficients, which are set by technicians in this field based on actual needs; F represents the penalty term, and the calculation formula is as follows:

[0100] F = r1 + r2 + r3;

[0101] Among them, r1 represents the penalty factor of the welding liner; if the welding liner cache vertical warehouse is overloaded or overspent, r1 is assigned to 1, otherwise, r1 is assigned to 0; r2 represents the penalty factor of the head; if the head cache vertical warehouse is overloaded or overspent, r2 is assigned to 1, otherwise, r2 is assigned to 0; r3 represents the penalty factor of the extruded tube; if the extruded tube cache vertical warehouse is overloaded or overspent, r3 is assigned to 1, otherwise, r3 is assigned to 0;

[0102] The method for judging whether any warehouse among the welding liner cache warehouse, the head cache warehouse, and the extruded tube cache warehouse is overrun or overloaded is as follows: if the inventory quantity of the product corresponding to the warehouse at the beginning of the current control cycle plus the expected inventory adjustment is greater than the total number of warehouse locations, the warehouse is overloaded; if the inventory quantity of the product corresponding to the warehouse at the beginning of the current control cycle plus the expected inventory adjustment is less than 0, the warehouse is overrun.

[0103] The control subunit generates the population iteration instruction based on the input maximum number of iterations; the population iteration instruction is specifically as follows: at each iteration, the fitness of each individual is counted, and the selection operation, crossover operation, and mutation operation are performed in sequence to generate the next generation population; the population iteration is repeated until the maximum number of iterations is reached;

[0104] The strategy subunit extracts the final welding rate, injection rate, and extrusion rate from the individual with the highest fitness in the last iteration.

[0105] The strategy unit generates a second control strategy based on the defect detection result, which is as follows:

[0106] The defect type, corresponding product number, and production time of the defective product are recorded and sent to the cloud module; the defective product is transported to the manual maintenance station; the production equipment of the defective product is suspended and the working status is detected and reset.

[0107] When the extruded tube, the end cap or the welded liner is shipped out of the warehouse, the strategy unit generates a third control strategy based on the product information, which is as follows:

[0108] The query subunit obtains the corresponding inventory information; the strategy subunit counts the storage location status of each storage location in the corresponding inventory information and sorts them by the storage time, and then carries out the shipment in order from the earliest to the latest storage time.

[0109] The control signal unit generates a control signal based on the production control strategy and sends the control signal to the field module; the control signal is sent to the field module through the communication unit, and the field module performs specific production activity control based on the control signal.

[0110] The visualization unit is used to process the product information and production control strategy into a visualization format and send it to the cloud module;

[0111] The visualization unit is equipped with a high-performance processor and large-capacity memory, which can quickly process data and generate graphical interfaces to ensure smooth operation when processing large amounts of data. At the same time, the visualization unit is also equipped with data processing engines such as Apache Flink and Spark Streaming for real-time data processing; it is also equipped with chart libraries such as D3.js and ECharts for drawing dynamic charts.

[0112] Based on the above configuration, the visualization unit converts the processed product information and control strategies into charts, dashboards and other forms, which are convenient for relevant managers to understand intuitively; when an abnormal situation is detected, the management personnel are reminded to conduct abnormal investigation and processing by changing the color of the chart, adding warning information, etc. In addition, the visualization unit provides an entry for viewing historical data to help analyze long-term trends. Through such a configuration, the visualization unit can effectively generate control strategies for product information and present relevant information in a visual form, helping managers to better monitor and manage the entire production process.

[0113] Example 2

[0114] This embodiment is the second embodiment of the present invention; based on the same inventive concept as embodiment 1, this embodiment introduces a digital factory, referring to Figure 2 , the digital factory includes edge processing devices, field modules, and cloud modules; among them:

[0115] The cloud module includes a data acquisition device and a control device; wherein the data acquisition device is used to collect product information and transmit the product information to the edge computing module; the edge computing module is used to process the product information and generate a control signal based on the production control strategy; and send the control signal to the control device; the control device controls the operation of the equipment on the hydrogen storage tank production line based on the control signal;

[0116] The cloud module includes a database and a visualization device; wherein the database is used to store product information processed by the edge computing module; the visualization device is used to visually display and store product information processed by the edge computing module and production control strategies.

[0117] Example 3

[0118] Based on the same inventive concept as other embodiments, this embodiment introduces a device control method, the steps of which refer to Figure 3 , as follows:

[0119] S1: Collect product information and product pictures, and process the product information;

[0120] S2: Performing defect detection on the product based on the product image;

[0121] S3: generating a production control strategy based on the product information and defect detection results;

[0122] S4: generating a control signal based on the production control strategy;

[0123] S5: Control the operation of equipment on the hydrogen storage tank production line based on the control signal.

[0124] The embodiments of the present invention are described above in conjunction with the accompanying drawings, but the present invention is not limited to the above-mentioned specific implementation methods. The above-mentioned specific implementation methods are merely illustrative and not restrictive. Under the guidance of the present invention, ordinary technicians in this field can also make many forms without departing from the purpose and scope of protection of the present invention, which are all within the protection of the present invention.

Claims

1. An edge processing device, characterized in that: It includes a communication unit, an information processing unit, a visual detection unit, a strategy unit, a control signal unit, and a visualization unit; wherein: The communication unit is used for information exchange between the edge processing device and the field module and the cloud module; The information processing unit is used to process the product information collected from the field module and the cloud module, and transmit the processed product information to the cloud module; The visual inspection unit is used to process product images collected by the on-site module and perform defect detection on the product; The strategy unit configures a production control strategy generated based on the product information and defect detection results; The control signal unit generates a control signal based on the production control strategy and sends the control signal to the field module; The visualization unit is used to process the product information and production control strategy into a visualization format and send it to the cloud module.

2. The edge processing device according to claim 1, characterized in that: The product information includes product production information and product inventory information; The product production information includes the production information of the extruded tube, the production information of the end cap, and the production information of the welded liner; wherein the production information of the extruded tube includes the number, specification parameters, and extrusion production time of the extruded tube; the production information of the end cap includes the number, specification parameters, and injection production time of the end cap; the production information of the welded liner includes the number of the welded liner, the number and specification parameters of the extruded tube involved in welding, the number of the end cap involved in welding, and the welding production time; The product inventory information includes inventory information of extruded tubes, inventory information of end caps, and inventory information of welded liners; wherein, the inventory information of extruded tubes includes the inventory quantity of extruded tubes, the total number of storage locations of the extruded tube cache vertical warehouse, and the storage location number and storage location status of each storage location; the storage location status of the extruded tube cache vertical warehouse includes free and occupied, if the storage location status is occupied, the storage location status also includes the number and storage time of the stored extruded tubes; the inventory information of end caps includes the inventory quantity of end caps, the total number of storage locations of the end cap cache vertical warehouse, and the storage location number and storage location status of each storage location; the storage location status of the end cap cache vertical warehouse includes free and occupied, if the storage location status is occupied, the storage location status also includes the number and storage time of the stored end caps; the inventory information of welded liners includes the inventory quantity of welded liners, the total number of storage locations of the welded liners cache vertical warehouse, and the storage location number and storage location status of each storage location; the storage location status of the welded liners cache vertical warehouse includes free and occupied, if the storage location status is occupied, the storage location status also includes the number and storage time of the stored welded liners.

3. The edge processing device according to claim 1, wherein: The method for the information processing unit to process the product information is as follows: Receive and record the production information of the extruded tube; if the extruded tube is stored in the extruded tube buffer warehouse, add the number and storage time of the extruded tube to the storage location status of the corresponding storage location, and update the inventory information of the extruded tube; when any extruded tube is out of the warehouse, reset the corresponding storage location status to free, and update the inventory information of the extruded tube; Receive and record the production information of the head; if the head is stored in the head cache, add the head number and storage time to the storage location status of the corresponding storage location, and update the inventory information of the head; when any head is shipped out, reset the corresponding storage location status to free, and update the inventory information of the head; If the extruded tube and the head are transported to the welding equipment for welding, the number and specification parameters of the extruded tube involved in the welding, the number of the head involved in the welding, and the welding production time are added to the production information of the welding liner; if the welding liner is stored in the welding liner buffer warehouse, the number and warehousing time of the welding liner are added to the storage location status of the corresponding storage location, and the inventory information of the welding liner is updated; when any welding liner is shipped out, the corresponding storage location status is reset to idle, and the inventory information of the welding liner is updated.

4. The edge processing device according to claim 1, wherein: The product images include CCD images of extruded tubes, CCD images of end caps, and ultrasonic images of welded liner; The visual inspection unit includes an image processing subunit, a first inspection model, a second inspection model, a third inspection model, and a defect prompting subunit; wherein: The image processing subunit is used to pre-process the CCD image of the extruded tube, the CCD image of the end cap, and the ultrasonic image of the welded liner, and transmit them to the first detection model, the second detection model, and the third detection model respectively; The first detection model is a convolutional neural network model trained based on a CCD image set of an extruded tube, and is used to process the CCD image of the extruded tube and perform defect detection on the extruded tube; The second detection model is a convolutional neural network model trained based on the CCD image set of the head, which is used to process the CCD image of the head and perform defect detection on the head; The third detection model is a convolutional neural network model trained based on an ultrasonic image set of a welded liner, which is used to process the ultrasonic image of the welded liner and perform defect detection on the welded liner; The defect prompt subunit is used to send the defect type and the number of the corresponding product to the strategy unit when any extruded tube, end cap or welded liner is detected to have a defect.

5. The edge processing device according to claim 1, wherein: The strategy unit includes a query subunit, a calculation subunit, a data subunit, a control subunit, and a strategy subunit; the production control strategy includes a first control strategy, a second control strategy, and a third control strategy; The strategy unit generates a second control strategy based on the defect detection result, which is as follows: Record the defect type, corresponding product number, and production time of the defective product and send them to the cloud module; move the defective product to the manual maintenance station; suspend the production equipment of the defective product and perform working status detection and reset; When the extruded tube, the end cap or the welded liner is shipped out of the warehouse, the strategy unit generates a third control strategy based on the product information, which is as follows: The query subunit obtains the corresponding inventory information; the strategy subunit counts the storage location status of each storage location in the corresponding inventory information and sorts them by the storage time, and then carries out the shipment in order from the earliest to the latest storage time.

6. The edge processing device according to claim 5, characterized in that: The strategy unit is configured with a genetic algorithm. At the beginning of each regulation cycle, a first control strategy is generated based on the product information, which is as follows: The query subunit queries product inventory information from the cloud module; The data subunit generates the primary population based on the input population size, crossover probability, and mutation probability, and executes the population iteration instruction; the chromosome of any individual in the primary population includes the welding rate, injection rate, and extrusion rate in the current regulation cycle; The welding rate indicates the number of welded inner shells produced by the welding equipment per unit time; the injection molding rate indicates the number of heads produced by the injection molding equipment per unit time; the extrusion rate indicates the number of extruded tubes produced by the extrusion equipment per unit time; The calculation subunit calculates the fitness of each individual based on the product inventory information; The control subunit generates the population iteration instruction based on the input maximum number of iterations; the population iteration instruction is specifically as follows: at each iteration, the fitness of each individual is counted, and the selection operation, crossover operation, and mutation operation are performed in sequence to generate the next generation population; the population iteration is repeated until the maximum number of iterations is reached; The strategy subunit extracts the final welding rate, injection rate, and extrusion rate from the individual with the highest fitness in the last iteration.

7. The edge processing device according to claim 6, characterized in that: The calculation formula of the fitness is as follows: H=w1·S-w2·(a1+a2+a3)-w3·F; Among them, H represents the fitness of any individual; S represents the expected number of welded liners to be produced, which is calculated by multiplying the welding rate by the length of the control cycle; a1 represents the inventory adjustment of welded liners; a2 represents the inventory adjustment of heads; a3 represents the inventory adjustment of extruded tubes; w1, w2, and w3 are all weight coefficients; and F represents the penalty term.

8. The edge processing device according to claim 7, characterized in that: The calculation formula of the penalty term F is as follows: F＝r1+r2+r3； Among them, r1 represents the penalty factor of the welding liner; if the welding liner cache vertical warehouse is overloaded or overspent, r1 is assigned to 1, otherwise, r1 is assigned to 0; r2 represents the penalty factor of the head; if the head cache vertical warehouse is overloaded or overspent, r2 is assigned to 1, otherwise, r2 is assigned to 0; r3 represents the penalty factor of the extruded tube; if the extruded tube cache vertical warehouse is overloaded or overspent, r3 is assigned to 1, otherwise, r3 is assigned to 0; The method for judging whether any warehouse among the welding liner cache warehouse, the head cache warehouse, and the extruded tube cache warehouse is overrun or overloaded is as follows: if the inventory quantity of the product corresponding to the warehouse at the beginning of the current control cycle plus the expected inventory adjustment is greater than the total number of warehouse locations, the warehouse is overloaded; if the inventory quantity of the product corresponding to the warehouse at the beginning of the current control cycle plus the expected inventory adjustment is less than 0, the warehouse is overrun.

9. A digital factory, characterized in that: The method comprises an edge processing device, a field module, and a cloud module as claimed in any one of claims 1 to 8; wherein: The cloud module includes a data acquisition device and a control device; wherein the data acquisition device is used to collect product information and transmit the product information to the edge computing module; the edge computing module is used to process the product information and generate a control signal based on the production control strategy; and send the control signal to the control device; the control device controls the operation of the equipment on the hydrogen storage tank production line based on the control signal; The cloud module includes a database and a visualization device; wherein the database is used to store product information processed by the edge computing module; the visualization device is used to visually display and store product information processed by the edge computing module and production control strategies.

10. A device control method, implemented based on the digital factory according to claim 9, characterized in that: The following steps are involved: S1: Collect product information and product pictures, and process the product information; S2: Performing defect detection on the product based on the product image; S3: generating a production control strategy based on the product information and defect detection results; S4: generating a control signal based on the production control strategy; S5: Control the operation of equipment on the hydrogen storage tank production line based on the control signal.

Images