Packaging equipment automatic rejection control system based on industrial ethernet

Abstract

The application discloses a packaging equipment automatic rejection control system based on an industrial Ethernet, relates to the technical field of rejection control, and aims to improve the rejection efficiency of a defective product and reduce energy consumption while avoiding secondary damage. The application discloses a packaging equipment automatic rejection control system based on an industrial Ethernet, relates to the technical field of rejection control, and aims to improve the rejection efficiency of a defective product and reduce energy consumption while avoiding secondary damage. The application discloses a packaging equipment automatic rejection control system based on an industrial Ethernet, relates to the technical field of rejection control, and aims to improve the rejection efficiency of a defective product and reduce energy consumption while avoiding secondary damage. The application discloses a packaging equipment automatic rejection control system based on an industrial Ethernet, relates to the technical field of rejection control, and aims to improve the rejection efficiency of a defective product and reduce energy consumption while avoiding secondary damage. The application discloses a packaging equipment automatic rejection control system based on an industrial Ethernet, relates to the technical field of rejection control, and aims to improve the rejection efficiency of a defective product and reduce energy consumption while avoiding secondary damage.

Description

Technical Field

[0001] This invention relates to the field of rejection control technology, specifically an automatic rejection control system for packaging equipment based on industrial Ethernet. Background Technology

[0002] In modern packaging production lines, product quality inspection and defective product rejection are key links to ensure product quality and production efficiency. Automatic rejection control of packaging equipment can leverage the high-speed and stable data transmission characteristics of industrial Ethernet to achieve real-time monitoring of the operating status of packaging equipment, accurate detection and automatic rejection of defective products, thus providing reliable quality assurance for packaging production.

[0003] In the existing technology, there are various methods for detecting defective products, especially image recognition. However, existing technologies often only determine whether there are defects without further obtaining the defective area. Moreover, the existing rejection methods often cause secondary damage to the product. If, based on obtaining the defective area, an appropriate clamping force is selected to clamp the parts outside the defective area to achieve rejection, the above problems can be solved. Therefore, this invention provides an automatic rejection control system for packaging equipment based on industrial Ethernet. Summary of the Invention

[0004] The purpose of this invention is to provide an automatic rejection control system for packaging equipment based on industrial Ethernet.

[0005] The objective of this invention can be achieved through the following technical solution: an automatic rejection control system for packaging equipment based on industrial Ethernet, comprising the following modules:

[0006] The data acquisition module is used to acquire equipment parameters and operating parameters of different components of the packaging equipment, and to perform protocol conversion on the acquired operating parameters;

[0007] The data simulation module is used to build corresponding digital twin models based on the equipment parameters and operating parameters of different components, and to simulate the built digital twin models.

[0008] The image recognition module is used to acquire defective image data of rejected products and build corresponding image recognition models, acquire real-time image data of non-rejected products, and combine the image recognition model to obtain defective products and their defective areas.

[0009] The clamping control module is used to acquire the first data set of different rejected products, use the digital twin model to obtain the corresponding first clamping force, and construct the clamping control model based on the different first data sets and their first clamping forces.

[0010] The product rejection module is used to obtain a second data set based on defective products and their defective areas, combine it with the clamping control model to obtain a second clamping force, and use the second clamping force to reject defective products.

[0011] Furthermore, the process of acquiring equipment parameters and operating parameters of different components of the packaging equipment, and then performing protocol conversion on the acquired operating parameters, includes:

[0012] The packaging equipment comprises the following different components: a material supply device, a filling device, a sealing device, a rejection detection device, a product conveying device, and a drive device;

[0013] The equipment parameters include basic parameters, communication parameters, structural parameters, and operating parameters. The operating parameters include temperature parameters, humidity parameters, pressure parameters, vibration parameters, and electrical parameters.

[0014] Set corresponding protocol adapters for different components, which can automatically identify the protocol characteristics of different components and dynamically convert the running parameters of different components into standardized JSON format.

[0015] Furthermore, based on the equipment parameters and operating parameters of different components, corresponding digital twin models are constructed. The process of simulating the constructed digital twin models includes:

[0016] Using 3D modeling technology, a physical model of the packaging equipment is constructed based on the equipment parameters of different components. Using digital twin technology, a digital twin model of the packaging equipment is constructed based on the operating parameters of different components and their physical models.

[0017] The rejection detection device is simulated in a digital twin model using simulation software. The rejection detection device works by gripping a robotic arm. The simulation can obtain the gripping force generated by the gripping robotic arm during its operation.

[0018] Furthermore, the process of acquiring defective image data of the rejected products and constructing corresponding image recognition models includes:

[0019] The rejected products refer to packaged products that have been rejected by the rejection detection device due to product defects. The product defects include the following types of defects: sealing defects, appearance damage, printing defects, surface contamination, and labeling errors. Different rejected products are linked to their defect types.

[0020] Acquire defect image data of rejected products with different defect types, including multi-angle images of rejected products and corresponding defect areas. Generate an image recognition set based on the defect image data of rejected products with different defect types and their defect areas, and divide it into an image training set and an image test set.

[0021] Construct a first convolutional neural network, using the defective image data of different eliminated products in the image training set as the input data of the first convolutional neural network, and using the corresponding defect type and defect region in the image training set as the output data of the first convolutional neural network;

[0022] The first convolutional neural network is trained using an image training set to obtain an initial first convolutional neural network. The initial first convolutional neural network is then validated using an image test set. The initial first convolutional neural network whose output is less than or equal to a preset image test error threshold is used as the image recognition model.

[0023] Furthermore, the process of acquiring real-time image data of products that have not been removed, and combining this data with an image recognition model to identify defective products and their defective areas, includes:

[0024] The unrejected product refers to the packaged product in the current packaging equipment that has not yet passed the rejection detection device. The multi-angle image of a single unrejected product is used as the corresponding real-time image data and input into the image recognition model to determine whether there is a product defect.

[0025] If it exists, the single product that was not removed will be treated as a defective product, and the defect type and defect area of ​​the defective product will be output. If it does not exist, no other operations will be performed on it.

[0026] Furthermore, the process of obtaining the first set of data for different excluded products and using a digital twin model to obtain the corresponding first clamping force includes:

[0027] Obtain the first set of data for each rejected product at the time of its rejection, including the gripping area, conveying speed, product weight, product specifications, and packaging material of the rejected product. The gripping area refers to the area of ​​the rejected product other than its defective area.

[0028] In the digital twin model, the clamping force of the gripping robot arm when it can grip a single rejected product is obtained. The gripping position of the gripping robot arm in its gripping area is continuously adjusted to obtain different clamping forces. The minimum clamping force is taken as the first clamping force, the corresponding clamping position is taken as the first clamping position, and it is bound to the first data set of the single rejected product.

[0029] Furthermore, the process of constructing a clamping control model based on different first data sets and their first clamping forces includes:

[0030] A clamping control set is generated based on the first data set of different eliminated products and their corresponding first clamping force and first clamping position, and it is divided into a clamping training set and a clamping test set.

[0031] Construct a second convolutional neural network, using the first dataset of different removed products in the clamping training set as the input data of the second convolutional neural network, and using the first clamping force and the first clamping position corresponding to the clamping training set as the output data of the second convolutional neural network;

[0032] The second convolutional neural network is trained using a clamping training set to obtain an initial second convolutional neural network. The initial second convolutional neural network is then validated using a clamping test set. The initial second convolutional neural network whose output is less than or equal to a preset clamping test error threshold is used as the clamping control model.

[0033] Furthermore, a second dataset is obtained based on the defective products and their defective areas. A second clamping force is then obtained using the clamping control model. The process of removing defective products using this second clamping force includes:

[0034] When a defective product is identified, a second set of data for the defective product is obtained, including its clamping area, conveying speed, product weight, product specifications, and packaging material.

[0035] The second data set is input into the clamping control model to obtain the corresponding second clamping force and second clamping position. The clamping robot arm is controlled to clamp the defective product at the second clamping position with the second clamping force to remove the defective product.

[0036] Compared with the prior art, the beneficial effects of the present invention are:

[0037] This invention enables timely protocol conversion of acquired operating parameters by setting protocol converters in different components of the packaging equipment. It also allows for the construction of a digital twin model of the packaging equipment and simulation of the rejection detection device to obtain its clamping force during operation.

[0038] By constructing image recognition and clamping control models, it is possible to promptly determine whether the packaged product has defects and obtain the defective area. Based on this, the clamping control model, combined with other parameters of the defective product, can quickly obtain the optimal clamping position and optimal clamping force, which can significantly improve the removal efficiency of defective products, reduce energy consumption while avoiding secondary damage. Attached Figure Description

[0039] Figure 1 This is a schematic diagram of the modules of the present invention. Detailed Implementation

[0040] like Figure 1 As shown, an automatic rejection control system for packaging equipment based on industrial Ethernet includes the following modules:

[0041] The data acquisition module is used to acquire equipment parameters and operating parameters of different components of the packaging equipment, and to perform protocol conversion on the acquired operating parameters;

[0042] The data simulation module is used to build corresponding digital twin models based on the equipment parameters and operating parameters of different components, and to simulate the built digital twin models.

[0043] The image recognition module is used to acquire defective image data of rejected products and build corresponding image recognition models, acquire real-time image data of non-rejected products, and combine the image recognition model to obtain defective products and their defective areas.

[0044] The clamping control module is used to acquire the first data set of different rejected products, use the digital twin model to obtain the corresponding first clamping force, and construct the clamping control model based on the different first data sets and their first clamping forces.

[0045] The product rejection module is used to obtain a second data set based on defective products and their defective areas, combine it with the clamping control model to obtain a second clamping force, and use the second clamping force to reject defective products.

[0046] It should be further explained that, in the specific implementation process, the process of obtaining equipment parameters and operating parameters of different components of the packaging equipment, and then performing protocol conversion on the obtained operating parameters, includes:

[0047] A complete packaging equipment typically includes the following different components: material supply device, filling device, sealing device, rejection detection device, product conveying device, and drive device.

[0048] The equipment parameters refer to the inherent relevant data of different components of the packaging equipment, which are used to identify and manage the corresponding components, including basic parameters, communication parameters, structural parameters, and operating parameters;

[0049] The basic parameters include device model and device serial number; the communication parameters include IP address, MAC address, communication protocol, and port information; the structural parameters include size, weight, and material; and the operating parameters include operating temperature range, power requirements, and power consumption.

[0050] The operating parameters refer to the relevant data on the dynamic changes of different components of the packaging equipment during its operation, used to monitor the operating status and environmental conditions of the corresponding components, including temperature parameters, humidity parameters, pressure parameters, vibration parameters, and electrical parameters;

[0051] Corresponding protocol adapters are set up on different components. The hardware consists of an edge gateway and an FPGA, and the software consists of a protocol feature library and a lightweight CNN model. It can automatically identify the protocol features of different components, such as Modbus function codes and ROS Topic structures, and dynamically convert the running parameters of different components into standardized JSON format.

[0052] It should be further explained that, in the specific implementation process, the process of constructing corresponding digital twin models based on the equipment parameters and operating parameters of different components, and simulating the constructed digital twin models, includes:

[0053] Using 3D modeling technology, a physical model of the packaging equipment is constructed based on the equipment parameters of different components. Using digital twin technology, a digital twin model of the packaging equipment is constructed based on the operating parameters of different components and their physical models.

[0054] Since the technical solution of this invention is for rejection control of packaging equipment, only the rejection detection device is simulated. The rejection detection device mainly works by gripping a robotic arm. The rejection detection device is simulated in a digital twin model using simulation software. Through simulation, the gripping force generated by the gripping robotic arm during its operation can be obtained.

[0055] It should be further explained that, in the specific implementation process, the process of acquiring defective image data of the rejected products and constructing the corresponding image recognition model includes:

[0056] The rejected products refer to packaged products that have been rejected by the rejection detection device due to product defects. The product defects include the following types of defects, such as sealing defects, appearance damage, printing defects, surface contamination, and labeling errors. Different rejected products are bound to their defect types.

[0057] Acquire defect image data of rejected products with different defect types, including multi-angle images of rejected products and local areas of corresponding product defects, which are denoted as defect areas and highlighted. Generate corresponding image recognition sets based on the defect image data of rejected products with different defect types and their defect areas, and divide them into image training sets and image test sets.

[0058] Construct a first convolutional neural network, using the defective image data of different eliminated products in the image training set as the input data of the first convolutional neural network, and using the corresponding defect type and defect region in the image training set as the output data of the first convolutional neural network;

[0059] The first convolutional neural network is trained using an image training set to obtain an initial first convolutional neural network. The initial first convolutional neural network is then validated using an image test set. The initial first convolutional neural network whose output is less than or equal to a preset image test error threshold is used as the image recognition model.

[0060] It should be further explained that, in the specific implementation process, the process of acquiring real-time image data of products that have not been removed, and combining this data with an image recognition model to identify defective products and their defective areas, includes:

[0061] The unrejected products refer to the packaged products in the current packaging equipment that have not yet passed the rejection detection device. Multi-angle images of each unrejected product are used as corresponding real-time image data. The real-time image data of a single unrejected product is input into the image recognition model to determine whether it has product defects.

[0062] If a defective product exists, it is treated as a defective product, and its defect type and defect area are output and highlighted. If a defective product does not exist, no other operation is performed on it. The same method is used to obtain all defective products and their corresponding defect types and defect areas from all the non-rejected products.

[0063] It should be further explained that, in the specific implementation process, the process of obtaining the first set of data for different eliminated products and using the digital twin model to obtain the corresponding first clamping force includes:

[0064] Obtain the first set of data for each rejected product at the time of its rejection, including the gripping area, conveying speed, product weight, product specifications, and packaging material of the rejected product;

[0065] The clamping area refers to the area of ​​the rejected product other than its defective area; the conveying speed refers to the moving speed of the rejected product on the conveyor belt; the product weight refers to the total weight of the rejected product; the product specifications refer to the length, width, and height of the rejected product; and the packaging material refers to the material of the outer packaging of the rejected product.

[0066] In the digital twin model, the clamping force of the gripping robot arm when it can hold a single rejected product is obtained, and the clamping position of the gripping robot arm in the gripping area is continuously adjusted to obtain different clamping forces.

[0067] The smallest clamping force is taken as the first clamping force, and its corresponding clamping position is taken as the first clamping position. It is then bound to the first data set of the single rejected product. The same method is used to obtain the first data set of different rejected products and their corresponding first clamping force and first clamping position.

[0068] It should be further explained that, in the specific implementation process, the process of constructing a clamping control model based on different first data sets and their first clamping forces includes:

[0069] A clamping control set is generated based on the first data set of different eliminated products and their corresponding first clamping force and first clamping position, and it is divided into a clamping training set and a clamping test set.

[0070] Construct a second convolutional neural network, using the first dataset of different removed products in the clamping training set as the input data of the second convolutional neural network, and using the first clamping force and the first clamping position corresponding to the clamping training set as the output data of the second convolutional neural network;

[0071] The second convolutional neural network is trained using a clamping training set to obtain an initial second convolutional neural network. The initial second convolutional neural network is then validated using a clamping test set. The initial second convolutional neural network whose output is less than or equal to a preset clamping test error threshold is used as the clamping control model.

[0072] It should be further explained that, in the specific implementation process, the process of obtaining a second data set based on the defective products and their defective areas, obtaining a second clamping force in conjunction with the clamping control model, and using the second clamping force to remove the defective products includes:

[0073] When it is determined that there is a defective product, a second set of data for the defective product is obtained, including its clamping area, conveying speed, product weight, product specifications, and packaging material. The clamping area refers to the area of ​​the defective product other than its defective area.

[0074] The acquired second set of data is input into the clamping control model to obtain the corresponding second clamping force and second clamping position. The clamping robot arm is controlled to clamp the defective product at the second clamping position with the second clamping force to remove the defective product. The same method is used to remove each defective product separately.

[0075] The above embodiments are only used to illustrate the technical methods of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical methods of the present invention without departing from the spirit and scope of the technical methods of the present invention.

Claims

1. An automatic rejection control system for packaging equipment based on industrial Ethernet, characterized in that, Includes the following modules: The data acquisition module is used to acquire equipment parameters and operating parameters of different components of the packaging equipment, and to perform protocol conversion on the acquired operating parameters; The data simulation module is used to build corresponding digital twin models based on the equipment parameters and operating parameters of different components, and to simulate the built digital twin models. The image recognition module is used to acquire defective image data of rejected products and build corresponding image recognition models, acquire real-time image data of non-rejected products, and combine the image recognition model to obtain defective products and their defective areas. The clamping control module is used to acquire the first data set of different rejected products, use the digital twin model to obtain the corresponding first clamping force, and construct the clamping control model based on the different first data sets and their first clamping forces. The product rejection module is used to obtain a second data set based on defective products and their defective areas, combine it with the clamping control model to obtain a second clamping force, and use the second clamping force to reject defective products. The process of obtaining the first data set and its first clamping force includes: Obtain the first set of data for each rejected product at the time of its rejection, including the gripping area, conveying speed, product weight, product specifications, and packaging material of the rejected product. The gripping area refers to the area of ​​the rejected product other than its defective area. In the digital twin model, the clamping force of the gripping robot arm when it can grip a single rejected product is obtained. The gripping position of the gripping robot arm in its gripping area is continuously adjusted to obtain different clamping forces. The minimum clamping force is taken as the first clamping force, the corresponding clamping position is taken as the first clamping position, and it is bound to the first data set of the single rejected product. The process of acquiring a second data set and its second clamping force, and using the second clamping force to remove defective products, includes: When a defective product is identified, a second set of data for the defective product is obtained, including its clamping area, conveying speed, product weight, product specifications, and packaging material. The second data set is input into the clamping control model to obtain the corresponding second clamping force and second clamping position. The clamping robot arm is controlled to clamp the defective product at the second clamping position with the second clamping force to remove the defective product.

2. The automatic rejection control system for packaging equipment based on industrial Ethernet according to claim 1, characterized in that, The process of acquiring device parameters and operating parameters from different components, and then performing protocol conversion on the operating parameters, includes: The packaging equipment comprises the following different components: a material supply device, a filling device, a sealing device, a rejection detection device, a product conveying device, and a drive device; The equipment parameters include basic parameters, communication parameters, structural parameters, and operating parameters. The operating parameters include temperature parameters, humidity parameters, pressure parameters, vibration parameters, and electrical parameters. Set corresponding protocol adapters for different components, which can automatically identify the protocol characteristics of different components and dynamically convert the running parameters of different components into standardized JSON format.

3. The automatic rejection control system for packaging equipment based on industrial Ethernet according to claim 2, characterized in that, The process of constructing a digital twin model based on equipment parameters and operating parameters and then simulating it includes: Using 3D modeling technology, a physical model of the packaging equipment is constructed based on the equipment parameters of different components. Using digital twin technology, a digital twin model of the packaging equipment is constructed based on the operating parameters of different components and their physical models. The rejection detection device is simulated in a digital twin model using simulation software. The rejection detection device works by gripping a robotic arm. The simulation can obtain the gripping force generated by the gripping robotic arm during its operation.

4. The automatic rejection control system for packaging equipment based on industrial Ethernet according to claim 3, characterized in that, The process of acquiring flawed image data and building an image recognition model includes: The rejected products refer to packaged products that have been rejected by the rejection detection device due to product defects. The product defects include the following types of defects: sealing defects, appearance damage, printing defects, surface contamination, and labeling errors. Different rejected products are linked to their defect types. Acquire defect image data of rejected products with different defect types, including multi-angle images of rejected products and corresponding defect areas. Generate an image recognition set based on the defect image data of rejected products with different defect types and their defect areas, and divide it into an image training set and an image test set. Construct a first convolutional neural network, using defective image data of different removed products in the image training set as input data for the first convolutional neural network, and using the corresponding defect types and defect regions in the image training set as output data for the first convolutional neural network. The first convolutional neural network is trained using an image training set to obtain an initial first convolutional neural network. The initial first convolutional neural network is then validated using an image test set. The initial first convolutional neural network whose output is less than or equal to a preset image test error threshold is used as the image recognition model.

5. The automatic rejection control system for packaging equipment based on industrial Ethernet according to claim 4, characterized in that, The process of acquiring real-time image data and identifying defective products and their defective areas includes: The unrejected product refers to the packaged product in the current packaging equipment that has not yet passed the rejection detection device. The multi-angle image of a single unrejected product is used as the corresponding real-time image data and input into the image recognition model to determine whether there is a product defect. If it exists, the single product that was not removed will be treated as a defective product, and the defect type and defect area of ​​the defective product will be output. If it does not exist, no other operations will be performed on it.

6. The automatic rejection control system for packaging equipment based on industrial Ethernet according to claim 5, characterized in that, The process of constructing the clamping control model includes: A clamping control set is generated based on the first data set of different eliminated products and their corresponding first clamping force and first clamping position, and it is divided into a clamping training set and a clamping test set. Construct a second convolutional neural network, using the first dataset of different removed products in the clamping training set as the input data of the second convolutional neural network, and using the first clamping force and the first clamping position corresponding to the clamping training set as the output data of the second convolutional neural network; The second convolutional neural network is trained using a clamping training set to obtain an initial second convolutional neural network. The initial second convolutional neural network is then validated using a clamping test set. The initial second convolutional neural network whose output is less than or equal to a preset clamping test error threshold is used as the clamping control model.

Images