Feeding equipment control method and system capable of preventing PET sheet from being carried
Through image acquisition and convolutional neural network, the uneven parameters of PET sheets are analyzed, and the flattening control parameters are optimized, which solves the problem that uneven areas cannot be identified in the PET sheet loading equipment, and improves the loading accuracy and production efficiency. High-quality flattening and electrostatic control of PET sheets is realized.
Patent Information
- Application Number
- CN202411590955.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-11-08
- Publication Date
- 2024-12-17
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
The prior art cannot accurately identify the uneven areas and degrees of the surface of PET sheets, resulting in poor loading accuracy, low product quality and low production efficiency.
The uneven parameters of PET sheets are analyzed by image acquisition and convolutional neural network recognizer, and the flattening control parameters are optimized to improve flatness and reduce static electricity. The optimal flattening control parameters are used for the loading equipment control.
It improves feeding accuracy and product quality, improves production efficiency, and ensures the flatness and electrostatic performance of PET sheets.
Smart Images
Figure CN119142889A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the technical field related to intelligent control, and in particular to a control method and system for feeding equipment capable of preventing PET sheet from carrying material. Background Art
[0002] In modern manufacturing, PET (polyethylene terephthalate) sheets are an important packaging material. They are widely used in food, medicine, electronics and other fields due to their excellent transparency, air tightness, tensile strength and impact resistance. However, during the production, transportation, processing and storage of PET sheets, they often carry materials due to various factors (such as temperature, humidity, pressure, etc.), resulting in uneven surfaces. That is, other impurities, particles, dust or foreign matter are attached or adsorbed on the PET sheets, which not only affects the aesthetics of the product, but also may have adverse effects on subsequent processing procedures such as printing, cutting, and molding. In addition, PET sheets are prone to static electricity during the flattening process, which not only affects the stable operation of the equipment, but also poses a threat to the safety of operators. Traditional feeding equipment control methods often rely on manual observation and manual control adjustment, which is not only inefficient, but also unable to accurately detect and identify various surface unevenness of PET sheets, difficult to identify the precise uneven areas and unevenness of the surface of PET sheets, and difficult to ensure the accuracy and consistency of each feeding, thus affecting the subsequent processing quality.
[0003] Therefore, in the current technology related to the control of PET sheet feeding equipment, there is a problem that the uneven area and degree of the PET sheet surface cannot be accurately identified, which makes it impossible to accurately adjust and optimize based on the uneven parameter distribution, which in turn leads to technical problems such as poor feeding accuracy, low product quality and low overall production efficiency. Summary of the invention
[0004] The present application provides a feeding equipment control method and system with the function of preventing PET sheets from carrying materials, thereby solving the technical problems in the prior art that the uneven areas and degrees of the surface of the PET sheet cannot be accurately identified, making it impossible to accurately adjust and optimize the uneven parameter distribution, which in turn leads to poor feeding accuracy, low product quality and low overall production efficiency, thereby achieving the technical effect of improving feeding accuracy and product quality and improving production efficiency.
[0005] The present application provides a method for controlling a feeding device capable of preventing PET sheets from being brought along, the method comprising: acquiring images of PET sheets to be fed, obtaining PET images, performing unevenness analysis on the PET sheets, and obtaining unevenness parameter distribution; obtaining a flattening control parameter space for flattening and feeding the PET sheets; within the flattening control parameter space, optimizing the flattening control parameters according to the unevenness parameter distribution for the purpose of improving the flatness of the PET sheets and reducing the electrostatic parameters of the PET sheets after being flattened, and obtaining optimal flattening control parameters, wherein the electrostatic parameters of the PET sheets after being flattened are predicted and optimized by carrying out a material bringing simulation; and using the optimal flattening control parameters to perform flattening and feeding control on the feeding device for the PET sheets.
[0006] In a possible implementation, image acquisition is performed on the PET sheet to be loaded to obtain a PET image, and an unevenness analysis is performed on the PET sheet to obtain an unevenness parameter distribution. The following processing is also performed: a PET image recognizer for PET image recognition is trained; a plurality of acquisition positions are randomly selected for image acquisition on the PET sheet to be loaded to obtain a plurality of PET images; the plurality of PET images are input into the PET image recognizer to obtain a plurality of unevenness parameters, and an unevenness parameter distribution is constructed based on the maximum and minimum values of the plurality of unevenness parameters.
[0007] In a possible implementation, a PET image recognizer that is trained to perform PET image recognition also performs the following processing: based on historical flatness detection data of PET sheets, a set of sample PET images is collected, and the unevenness size in each sample PET image is marked to obtain a set of sample unevenness parameters; the sample PET image set and the sample unevenness parameter set are used as supervised training data, and a convolutional neural network is used to train the PET image recognizer; the trained PET image recognizer is verified, and when the requirements are met, the PET image recognizer is obtained.
[0008] In a possible implementation, a flattening control parameter space for flattening and loading PET sheets is obtained, and the following processing is also performed: a pressure control parameter range of a flattening device for flattening and loading PET sheets is obtained; and the pressure control parameter range is constructed as a flattening control parameter space.
[0009] In a possible implementation, in the flattening control parameter space, according to the unevenness parameter distribution, the flattening control parameters are optimized for the purpose of improving the flatness of the PET sheet and reducing the electrostatic parameters of the PET sheet after flattening, and the optimal flattening control parameters are obtained. The following processing is also performed: randomly selecting a first flattening control parameter in the flattening control parameter space; performing simulated flattening prediction according to the first flattening control parameter and the unevenness parameter distribution to obtain a first flattening ratio and a first electrostatic parameter; collecting environmental parameters, and performing material carrying prediction for the PET sheet in combination with the first electrostatic parameter to obtain a first material carrying parameter, and calculating and obtaining a first feeding fitness in combination with the first flattening ratio; continuing to optimize the flattening control parameters until the optimization converges, outputting the flattening control parameters with the largest feeding fitness, and obtaining the optimal flattening control parameters.
[0010] In a possible implementation, a simulated flattening prediction is performed according to the first flattening control parameter and the uneven parameter distribution to obtain a first flattening ratio and a first electrostatic parameter, and the following processing is also performed: according to the historical data of PET sheet feeding and flattening, a sample flattening control parameter set and a sample uneven parameter distribution set are collected, and the flattening area ratio and surface resistivity of the PET sheet after flattening under different sample flattening control parameters and sample uneven parameter distributions are obtained, and marked as a sample flattening ratio set and a sample electrostatic parameter set; the sample flattening control parameter set and the sample uneven parameter distribution set are used as input data, and the sample flattening ratio set and the sample electrostatic parameter set are used as output supervision data to train a flattening simulation predictor; the first flattening control parameter and the uneven parameter distribution are input into the flattening simulation predictor to predict and obtain a first flattening ratio and a first electrostatic parameter; In a possible implementation, environmental parameters are collected, and the material carrying prediction of the PET sheet is performed in combination with the first electrostatic parameters to obtain the first material carrying parameters, and the first loading fitness is calculated in combination with the first flattening ratio, and the following processing is also performed: the dust concentration in the environment is collected as an environmental parameter; the environmental parameter and the first electrostatic parameter are input into a material carrying predictor to predict and obtain the first material carrying parameter, wherein the material carrying predictor is obtained by pre-training with sample environmental parameters, sample electrostatic parameters and sample material carrying parameters, and each sample material carrying parameter includes the attached particle density data of the PET sheet; according to the first flattening ratio and the first material carrying parameter, the first loading fitness is calculated as follows: ; Among them, LA is the feeding adaptability, is the weight, N is a positive integer, K is the flattening ratio, Belt parameters, is the density threshold of attached particles.
[0011] The present application also provides a feeding equipment control system with the function of preventing PET sheets from belting, including: an unevenness parameter distribution acquisition module, used to collect images of PET sheets to be fed, obtain PET images, perform unevenness analysis on the PET sheets, and obtain unevenness parameter distribution; a flattening control parameter space acquisition module, used to obtain the flattening control parameter space for flattening and feeding the PET sheets; an optimal flattening control parameter acquisition module, used to optimize the flattening control parameters within the flattening control parameter space according to the unevenness parameter distribution, with the purpose of improving the flatness of the PET sheets and reducing the electrostatic parameters of the PET sheets after flattening, and obtain the optimal flattening control parameters, wherein the electrostatic parameters of the PET sheets after flattening are predicted and optimized by belting simulation; a flattening feeding control module, used to use the optimal flattening control parameters to perform flattening and feeding control on the feeding equipment of the PET sheets.
[0012] Through the control method and system of feeding equipment with the function of preventing PET sheet from carrying material proposed in this application, it is intended to obtain PET images, analyze the unevenness of PET sheets, and obtain unevenness parameter distribution; obtain the flattening control parameter space of PET sheet flattening and feeding; optimize the flattening control parameters for the purpose of improving the flatness of PET sheets and reducing the electrostatic parameters of PET sheets after flattening, and obtain the optimal flattening control parameters; and control the flattening and feeding of PET sheet feeding equipment. This solves the technical problem that the uneven area and degree of the surface of PET sheets cannot be accurately identified in the prior art, making it impossible to accurately adjust and optimize in combination with the unevenness parameter distribution, which leads to poor feeding accuracy, low product quality and low overall production efficiency, and achieves the technical effect of improving feeding accuracy and product quality and improving production efficiency. BRIEF DESCRIPTION OF THE DRAWINGS
[0013] In order to more clearly illustrate the technical solution of the embodiment of the present disclosure, the accompanying drawings of the embodiment of the present disclosure will be briefly introduced below. A flow chart is used in the present application to illustrate the operations performed by the system according to the embodiment of the present application. It should be understood that the preceding or following operations are not necessarily performed accurately in order. On the contrary, various steps can be processed in reverse order or simultaneously as needed. At the same time, other operations can also be added to these processes, or one or more operations can be removed from these processes.
[0014] Figure 1 A schematic flow chart of a method for controlling a feeding device with anti-PET sheet material belting provided in an embodiment of the present application; Figure 2 A schematic diagram of the control system structure of a feeding equipment with anti-PET sheet material belt provided in an embodiment of the present application.
[0015] Explanation of the reference numerals: unevenness parameter distribution acquisition module 10 , flattening control parameter space acquisition module 20 , optimal flattening control parameter acquisition module 30 , flattening feeding control module 40 . DETAILED DESCRIPTION
[0016] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below.
[0017] In order to make the objectives, technical solutions and advantages of the present application clearer, the present application will be further described in detail below in conjunction with the accompanying drawings. The described embodiments should not be regarded as limiting the present application. All other embodiments obtained by ordinary technicians in the field without making creative work are within the scope of protection of this application.
[0018] In the following description, reference is made to "some embodiments", which describe a subset of all possible embodiments, but it is understood that "some embodiments" may be the same subset or different subsets of all possible embodiments, and may be combined with each other without conflict, and the terms "first\second" involved are merely to distinguish similar objects and do not represent a specific ordering of objects. The terms "including" and "having" and any variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product, or server that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or modules that are not clearly listed or inherent to these processes, methods, products, or devices. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those generally understood by technicians in the technical field of this application. The terms used herein are for the purpose of describing the embodiments of the present application only.
[0019] The embodiment of the present application provides a method for controlling a feeding device with a PET sheet material protection function, such as Figure 1 As shown, the method includes: In the processing of PET sheets, due to the possible uneven deviation, it is necessary to use a pressure roller to perform precise flattening. However, the choice of flattening pressure is crucial and directly related to the processing quality and production efficiency. When the flattening pressure is too large, it helps to flatten the PET sheet, but it may also cause a large amount of static electricity on the surface of the sheet. The accumulation of static electricity will aggravate the phenomenon of carrying materials, that is, adsorbing more particles, dust and other impurities, which will not only affect the cleanliness and processing quality of the PET sheet, but also may have an adverse effect on subsequent processes; on the contrary, if the flattening pressure is too small, the PET sheet may not be fully flattened, and the uneven deviation on its surface will still exist, which will not only reduce the processing quality, but also may affect the appearance and performance of the product. Therefore, it is particularly important to optimize the flattening pressure, a key flattening control parameter. Through precise measurement and debugging, the optimal pressure value that can ensure that the PET sheet is fully flattened without generating too much static electricity is found, and continuous testing and adjustment are required during the production process to achieve the best flattening effect and processing quality.
[0020] Step S100 , collecting images of the PET sheet to be loaded, obtaining a PET image, and performing unevenness analysis on the PET sheet to obtain unevenness parameter distribution.
[0021] Preferably, a high-definition image acquisition device is used to acquire images of the PET sheet to be loaded and analyze the unevenness of the PET sheet to obtain unevenness parameter distribution, which refers to obtaining various unevenness characteristics of the PET sheet at a microscopic scale and their distribution patterns. Specifically, the unevenness characteristics may include but are not limited to surface roughness, thickness unevenness, bubbles and holes, warping and deformation, as well as waviness, orange peel phenomenon and other parameters related to the unevenness of the PET sheet. Among them, the surface roughness reflects the unevenness of the sheet surface, which can usually be obtained by the surface texture analysis method in the image analysis technology. Excessive surface roughness may be caused by excessively high melt temperature or excessively fast extrusion speed; uneven thickness It may cause deformation or rupture of the sheet during subsequent processing. Through image acquisition and analysis, the thickness of each part of the PET sheet can be accurately measured and its uniformity can be evaluated; bubbles and holes are usually caused by excessive moisture content or insufficient vacuum in the raw materials. After image acquisition, image analysis technology can be used to clearly identify bubbles and holes in the PET sheet, and their number, size and distribution can be counted; PET sheets may warp and deform due to various reasons during processing and storage, which will seriously affect the subsequent processing and use of the sheet. Through image acquisition and analysis technology, the degree of warping and deformation of the PET sheet can be accurately measured, and its impact on product quality can be evaluated.
[0022] In a possible implementation, step S100 further includes step S110, training a PET image recognizer for PET image recognition; step S120, randomly selecting multiple acquisition positions for the PET sheet to be loaded for image acquisition, to obtain multiple PET images; step S130, inputting the multiple PET images into the PET image recognizer to obtain multiple unevenness parameters, and constructing an unevenness parameter distribution based on the maximum and minimum values of the multiple unevenness parameters.
[0023] Preferably, a large number of PET image data sets are collected, covering PET sheets of different types and different flatnesses, to ensure that the data sets are diverse and representative, and then image processing techniques such as image enhancement and edge detection are used to extract key features from the PET images, which can reflect the flatness of the PET sheets. A recognition model is then constructed based on a convolutional neural network (CNN), and the extracted uneven features are used as training inputs to train the recognition model to obtain a PET image recognizer to accurately identify the flatness of the PET sheets while ensuring recognition accuracy. For the PET sheets to be loaded, multiple acquisition positions are randomly selected for image acquisition to ensure that the acquired images can fully reflect the flatness of the sheets, and then the acquired images are used as inputs to train the recognition model. A plurality of PET images are input into a trained PET image recognizer, which analyzes the input images and outputs a plurality of unevenness parameters for quantifying the flatness of the PET sheet. Based on the plurality of unevenness parameters output by the recognizer, the maximum and minimum values are calculated. The unevenness parameters include the size of the unevenness, such as warping of 1 mm or depression of -1 mm. An unevenness parameter distribution is constructed to intuitively display the flatness of the PET sheet at different acquisition positions. The unevenness parameter distribution includes the interval range of the unevenness size. Finally, according to the distribution of the unevenness parameters, the overall flatness of the PET sheet is analyzed, thereby realizing precise control and optimized processing of the PET sheet and improving product quality and production efficiency.
[0024] In a possible implementation, step S110 further includes step S111, collecting a set of sample PET images based on historical flatness detection data of the PET sheet, marking the unevenness size in each sample PET image, and obtaining a set of sample unevenness parameters; step S112, using the sample PET image set and the sample unevenness parameter set as supervised training data, and using a convolutional neural network to train the PET image recognizer; step S113, verifying the trained PET image recognizer, and obtaining the PET image recognizer when the requirements are met.
[0025] Preferably, a plurality of sample PET images are collected from the historical flatness detection data of the PET sheet to form a sample PET image set, which includes unevenness of different degrees and types. For each collected PET image, the size of the uneven area (for example, length, width or area) is automatically annotated, and these annotation information constitutes a sample unevenness parameter set; then these annotated sample PET image sets and sample unevenness parameter sets are used as supervised training data, that is, each PET image has a corresponding unevenness parameter label, and a convolutional neural network (CNN) is used as a model, and the supervised training data is input into the network for supervised training, which can be learned through learning. The features in the learning image are used to identify the uneven areas of the PET sheet. Finally, after the training is completed, an independent validation dataset (usually does not contain any images in the training data to ensure the objectivity of the test results) is used to test the performance of the PET image recognizer. By comparing the predicted results of the recognizer with the true annotations in the validation dataset, the accuracy of the recognizer can be evaluated. If the validation results show that the performance of the recognizer meets the requirements (for example, the prediction accuracy reaches a predetermined threshold), the trained CNN model is regarded as a qualified PET image recognizer and used in future PET sheet flatness detection tasks to automatically identify and measure the size of uneven areas.
[0026] Step S200, obtaining the flattening control parameter space for flattening and loading the PET sheet.
[0027] Furthermore, step S200 also includes step S210, obtaining a pressure control parameter range of a flattening device for flattening and feeding the PET sheet; and step S220, constructing the pressure control parameter range into a flattening control parameter space.
[0028] Preferably, in the production and processing of PET sheets, flattening and feeding is a key step, including converting the PET sheets from a raw material state to a flat state suitable for subsequent processing, obtaining a flattening control parameter space for flattening and feeding the PET sheets, specifically determining the flattening control parameters to ensure that the PET sheets can achieve the expected flattening effect during the flattening process. The parameter space usually includes a pressure control parameter range. Specifically, the pressure control parameter range required for the flattening device to flatten the PET sheets is determined according to the equipment type and performance, the characteristics of the PET sheets, and the flattening effect requirements, wherein different types of flattening equipment may have different pressure control methods and parameter ranges; the thickness, hardness, toughness and other characteristics of the PET sheets will also affect the pressure required during the flattening process. Thicker sheets may require greater pressure to ensure their flatness, while thinner sheets may require more sophisticated pressure control to avoid excessive indentations; according to the needs of subsequent processing, there are different requirements for the flattening effect of the PET sheets. For example, some application scenarios may require higher flatness, while others may pay more attention to the integrity and surface finish of the sheets. After determining the pressure control parameter range, it is constructed into a flattening control parameter space. Specifically, the pressure control parameter range is represented to more intuitively understand the relationship between the parameters and their impact on the flattening effect. The parameter space is then optimized through simulation, including adjusting the parameter range, determining the optimal parameter combination, and evaluating the impact of different parameter combinations on the flattening effect. At the same time, the optimized parameter space is verified to ensure that it can meet the expected requirements in actual applications. For example, the pressure control parameters are automatically adjusted according to the input PET sheet characteristics and flattening effect requirements to achieve the best flattening effect.
[0029] Step S300, within the flattening control parameter space, according to the unevenness parameter distribution, the flattening control parameters are optimized to obtain the optimal flattening control parameters with the purpose of improving the flatness of the PET sheet and reducing the electrostatic parameters of the PET sheet after flattening, wherein the electrostatic parameters of the PET sheet after flattening are predicted and optimized by stripping simulation.
[0030] Preferably, within the flattening control parameter space, the flattening control parameters are optimized according to the unevenness parameter distribution, aiming to improve the flatness of the PET sheet and reduce the electrostatic parameters after flattening, wherein the optimization goal is to improve the flatness of the PET sheet, that is, by adjusting the flattening control parameters, the PET sheet can achieve a higher flatness during the flattening process to meet the requirements of subsequent processing or application, and reduce the electrostatic parameters of the PET sheet after flattening. Static electricity is a problem that is easily generated during the processing of PET sheets, which will lead to the phenomenon of material banding, that is, adsorbing more impurities such as particles and dust. By optimizing the flattening control parameters, the generation of static electricity is reduced, thereby reducing the occurrence of material banding. Specifically, an electrostatic parameter prediction model for PET sheets is constructed based on the machine learning model, and the electrostatic parameters that may be generated in the flattening process of the PET sheets are predicted. The prediction results can reflect the changing trend of the electrostatic parameters under different flattening control parameters. Then, based on the predicted electrostatic parameters, a stripping simulation experiment is carried out, that is, by simulating the flattening process of the PET sheet under laboratory conditions, the occurrence of the stripping phenomenon is observed and recorded. The results of the stripping simulation experiment can intuitively reflect the influence of the electrostatic parameters on the stripping phenomenon, as well as the improvement of the stripping phenomenon under different flattening control parameters. Finally, according to the predicted electrostatic parameters and the results of the stripping simulation experiment, the flattening control parameters are adjusted and optimized, including adjusting the flattening pressure control parameters, to ensure that the optimal flattening control parameters (i.e., the optimal pressure control parameters) are found to improve the flatness of the PET sheet and reduce the electrostatic parameters after flattening.
[0031] In a possible implementation, step S300 further includes step S310, randomly selecting a first flattening control parameter in the flattening control parameter space; step S320, performing simulated flattening prediction based on the first flattening control parameter and the unevenness parameter distribution to obtain a first flattening ratio and a first electrostatic parameter; step S330, collecting environmental parameters, and performing material belting prediction of the PET sheet in combination with the first electrostatic parameter to obtain a first material belting parameter, and calculating a first feeding fitness in combination with the first flattening ratio; step S340, continuing to optimize the flattening control parameters until the optimization converges, outputting the flattening control parameters with the largest feeding fitness, and obtaining the optimal flattening control parameters.
[0032] Preferably, the optimal flattening control parameters are obtained through simulation, prediction and optimization in the flattening control parameter space. Specifically, in the determined flattening control parameter space, an initial flattening control parameter (flattening pressure) is randomly selected as the first flattening control parameter, and then a simulated flattening prediction is performed based on the first flattening control parameter and the unevenness parameter distribution (i.e., the unevenness of the PET sheet at different positions), that is, the first flattening control parameter is simulated to predict its flattening effect, and a first flattening ratio and a first electrostatic parameter are obtained, wherein the first flattening ratio refers to the ratio of the PET sheet to a preset flatness under given parameters, and the first electrostatic parameter refers to the static electricity level generated during the flattening process; then, parameters related to the current production environment are collected, such as humidity, temperature, air flow rate, etc., which may affect the material carrying condition of the PET sheet, and then the material carrying prediction model is used to perform material carrying prediction on the PET sheet in combination with the first electrostatic parameter and environmental parameters to obtain. The first feeding parameter, that is, predicting the degree or probability of feeding phenomenon of PET sheet under given conditions, and then combining with the first flattening ratio using the fitness function to calculate the corresponding first feeding fitness, reflects the comprehensive performance of the flatness and feeding situation of the PET sheet under given flattening control parameters. Generally speaking, the higher the flattening ratio and the lower the feeding parameter (that is, the less feeding phenomenon), the higher the feeding fitness; finally, according to the result of the first feeding fitness, a certain optimization algorithm (such as genetic algorithm, particle swarm algorithm, etc.) is used to adjust the flattening control parameters, that is, repeating the simulation of flattening prediction, feeding prediction and fitness calculation, and using new flattening control parameters for prediction and calculation in each iteration. If the feeding fitness is no longer significantly improved or reaches the preset convergence condition, the flattening control parameter with the largest feeding fitness is output as the optimal flattening control parameter, that is, under given conditions, the best parameter that can simultaneously meet the requirements of high flatness and low feeding.
[0033] In a possible implementation, step S320 further includes step S321, collecting a sample flattening control parameter set and a sample unevenness parameter distribution set based on historical data of PET sheet loading and flattening, and obtaining the flattening area ratio and surface resistivity of the PET sheet after flattening under different sample flattening control parameters and sample unevenness parameter distributions, which are marked as sample flattening ratio set and sample electrostatic parameter set; step S322, using the sample flattening control parameter set and the sample unevenness parameter distribution set as input data, and using the sample flattening ratio set and the sample electrostatic parameter set as output supervision data to train a flattening simulation predictor; step S323, inputting the first flattening control parameter and the unevenness parameter distribution into the flattening simulation predictor to predict a first flattening ratio and a first electrostatic parameter.
[0034] Preferably, a sample flattening control parameter set (including a plurality of different flattening control parameters, i.e., different flattening pressures) and a sample unevenness parameter distribution set (a parameter set describing the initial uneven state of the PET sheet, and the unevenness parameters may include the thickness unevenness, warping degree, etc. of the sheet) are collected from the historical data of the flattening of the PET sheet, wherein the historical data of the flattening of the PET sheet refers to the data recorded in the past during the flattening process of the PET sheet, including the flattening control parameters, the unevenness parameter distribution, the effect after flattening (such as the flattening area ratio and the surface resistivity), etc.; the flattening area ratio and the surface resistivity of the PET sheet after flattening under different sample flattening control parameters and sample unevenness parameter distributions are obtained, wherein the flattening area ratio is used to evaluate the flattening effect, and are respectively marked as a sample flattening ratio set and a sample electrostatic parameter set, and the sample flattening ratio set is obtained under different flattening control parameters and unevenness parameter distributions. The sample electrostatic parameter set is a set of surface resistivities of the PET sheet after flattening under different flattening conditions, wherein the surface resistivity is used to measure the electrostatic performance of the material; then a predictor is constructed based on a machine learning model, and a sample flattening control parameter set and a sample unevenness parameter distribution set are used as input data, and a sample flattening ratio set and a sample electrostatic parameter set are used as output supervision data, and the predictor is supervised and trained to learn the relationship between the input parameters and the output effects of the PET sheet, and a flattening simulation predictor is obtained, and the first flattening control parameter and the unevenness parameter distribution are input into the flattening simulation predictor for prediction, and the flattening simulation predictor outputs a prediction result, namely a first flattening ratio and a first electrostatic parameter, which respectively represent the predicted flattening area ratio and surface resistivity of the PET sheet under such flattening control parameter and unevenness parameter distribution; In a possible implementation, step S330 further includes step S331, collecting the dust concentration in the environment as an environmental parameter; step S332, inputting the environmental parameter and the first electrostatic parameter into a material belt predictor to predict and obtain the first material belt parameter, wherein the material belt predictor is obtained by pre-training with sample environmental parameters, sample electrostatic parameters and sample material belt parameters, and each sample material belt parameter includes the attached particle density data of the PET sheet; step S333, according to the first flattening ratio and the first material belt parameter, calculating and obtaining the first feeding fitness, as shown in the following formula: ; Among them, LA is the feeding adaptability, is the weight, N is a positive integer, K is the flattening ratio, Belt parameters, is the density threshold of attached particles; Preferably, the dust concentration data in the current environment is obtained, which may affect the material carrying condition of the PET sheet during the flattening process, and is used as an environmental parameter for subsequent material carrying prediction. Then the environmental parameter and the first electrostatic parameter are input into a material carrying predictor for prediction, wherein the first electrostatic parameter refers to the static electricity level generated by the PET sheet during the flattening process, which is usually obtained by measurement. The material carrying predictor is a machine learning model, which is trained by sample environmental parameters, sample electrostatic parameters and sample material carrying parameters, and can predict the material carrying condition of the PET sheet according to the input environmental parameters and electrostatic parameters. Specifically, the sample environmental parameters include dust concentration data collected at different time points or in different production batches in the past; the sample electrostatic ... The number is the electrostatic level of the PET sheet at the corresponding time point or batch; the sample feeding parameter is the attached particle density data of the PET sheet obtained by some means (such as experimental measurement, image processing, etc.), which reflects the quantity and distribution of dust or particles attached to the PET sheet; finally, the first feeding parameter is predicted by the feeding predictor, that is, the possible feeding situation of the PET sheet under the current environmental parameters and the first electrostatic parameter, wherein the first feeding parameter includes the attached particle density data of the PET sheet, that is, the predicted quantity and distribution of dust or particles attached to the PET sheet; finally, according to the first flattening ratio and the first feeding parameter, the first feeding fitness is calculated by the fitness formula, as shown in the following formula: ; Among them, LA is the feeding adaptability, is the weight, N is a positive integer, such as 3, K is the flattening ratio, Belt parameters, is the density threshold of attached particles; the larger the flattening ratio K is, the larger the material parameters are. The smaller the density of attached particles, the greater the feeding adaptability LA, and the better the feeding control effect.
[0035] Step S400, using the optimal flattening control parameters to perform flattening and feeding control on the feeding equipment of the PET sheet.
[0036] Preferably, the feeding equipment of the PET sheet is flattened and controlled by the optimal flattening control parameters (optimal flattening pressure) to achieve the best flattening effect of the PET sheet, so as to ensure that the PET sheet can achieve the expected flatness and electrostatic performance during the flattening process. The feeding equipment refers to the equipment used to feed the PET sheet into the flattening process, which may include a feeding mechanism, a guiding mechanism, a clamping mechanism and other parts. These parts work together to ensure that the PET sheet can smoothly and accurately enter the flattening area. Specifically, the optimal flattening control parameters are applied to the feeding equipment, that is, the parameter settings of the feeding equipment, such as the clamping force, are adjusted according to the optimal flattening pressure to ensure that the PET sheet can be fed and flattened in the expected manner during the flattening process. The flattening of the PET sheet is monitored in real time, and the control parameters are fine-tuned according to the actual situation to ensure that the final flattening effect is optimal, thereby ensuring that the final product quality meets the requirements.
[0037] In the above, refer to Figure 1 A method for controlling a feeding device with a PET sheet material according to an embodiment of the present invention is described in detail. Figure 2 A feeding equipment control system with anti-PET sheet material belting function according to an embodiment of the present invention is described.
[0038] According to an embodiment of the present invention, a feeding equipment control system with anti-PET sheet material belting is used to solve the technical problem that the uneven area and degree of the PET sheet surface cannot be accurately identified in the prior art, so that the uneven parameter distribution cannot be accurately adjusted and optimized, which leads to poor feeding accuracy, low product quality and low overall production efficiency, thereby achieving the technical effect of improving feeding accuracy and product quality and improving production efficiency. A feeding equipment control system with anti-PET sheet material belting includes: an uneven parameter distribution acquisition module 10, a flattening control parameter space acquisition module 20, an optimal flattening control parameter acquisition module 30, and a flattening feeding control module 40.
[0039] The unevenness parameter distribution acquisition module 10 is used to capture images of the PET sheet to be loaded, obtain PET images, perform unevenness analysis on the PET sheet, and obtain unevenness parameter distribution; the flattening control parameter space acquisition module 20 is used to obtain the flattening control parameter space for flattening and loading the PET sheet; the optimal flattening control parameter acquisition module 30 is used to optimize the flattening control parameters within the flattening control parameter space according to the unevenness parameter distribution, with the purpose of improving the flatness of the PET sheet and reducing the electrostatic parameters of the PET sheet after flattening, and obtain the optimal flattening control parameters, wherein the electrostatic parameters of the PET sheet after flattening are predicted and optimized by belt simulation; the flattening and loading control module 40 is used to use the optimal flattening control parameters to control the flattening and loading of the PET sheet loading equipment.
[0040] The specific configuration of the unevenness parameter distribution acquisition module 10 will be described in detail below. The unevenness parameter distribution acquisition module 10 further includes: training a PET image recognizer for PET image recognition; randomly selecting multiple acquisition positions for the PET sheet to be loaded for image acquisition to obtain multiple PET images; inputting the multiple PET images into the PET image recognizer to obtain multiple unevenness parameters, and constructing an unevenness parameter distribution based on the maximum and minimum values of the multiple unevenness parameters.
[0041] The specific configuration of the unevenness parameter distribution acquisition module 10 will be described in detail below. The unevenness parameter distribution acquisition module 10 further includes: collecting a sample PET image set according to the historical flatness detection data of the PET sheet, marking the unevenness size in each sample PET image, and obtaining a sample unevenness parameter set; using the sample PET image set and the sample unevenness parameter set as supervised training data, using a convolutional neural network to train the PET image recognizer; verifying the trained PET image recognizer, and obtaining the PET image recognizer when the requirements are met.
[0042] The following will describe in detail the specific configuration of the flattening control parameter space acquisition module 20. The flattening control parameter space acquisition module 20 may further include: acquiring a pressure control parameter range of a flattening device for flattening and feeding PET sheets; and constructing the pressure control parameter range into a flattening control parameter space.
[0043] The specific configuration of the optimal flattening control parameter acquisition module 30 will be described in detail below. The optimal flattening control parameter acquisition module 30 may further include: randomly selecting a first flattening control parameter in the flattening control parameter space; performing simulated flattening prediction according to the first flattening control parameter and the unevenness parameter distribution to obtain a first flattening ratio and a first electrostatic parameter; collecting environmental parameters, combining the first electrostatic parameter to perform material belting prediction of the PET sheet to obtain a first material belting parameter, and combining the first flattening ratio to calculate and obtain a first feeding fitness; continuing to optimize the flattening control parameters until the optimization converges, outputting the flattening control parameters with the maximum feeding fitness, and obtaining the optimal flattening control parameters.
[0044] The specific configuration of the optimal flattening control parameter acquisition module 30 will be described in detail below. The optimal flattening control parameter acquisition module 30 may further include: collecting a sample flattening control parameter set and a sample unevenness parameter distribution set according to the historical data of the PET sheet loading and flattening, and obtaining the flattening area ratio and surface resistivity of the PET sheet after flattening under different sample flattening control parameters and sample unevenness parameter distributions, which are marked as sample flattening ratio set and sample electrostatic parameter set; using the sample flattening control parameter set and the sample unevenness parameter distribution set as input data, and using the sample flattening ratio set and the sample electrostatic parameter set as output supervision data to train a flattening simulation predictor; inputting the first flattening control parameter and the unevenness parameter distribution into the flattening simulation predictor to predict and obtain a first flattening ratio and a first electrostatic parameter; The specific configuration of the optimal flattening control parameter acquisition module 30 will be described in detail below. The optimal flattening control parameter acquisition module 30 may further include: collecting dust concentration in the environment as an environmental parameter; inputting the environmental parameter and the first electrostatic parameter into a belt material predictor to predict and obtain a first belt material parameter, wherein the belt material predictor is obtained by pre-training with sample environmental parameters, sample electrostatic parameters and sample belt material parameters, and each sample belt material parameter includes attached particle density data of the PET sheet; and calculating and obtaining a first feeding fitness according to the first flattening ratio and the first belt material parameter, as shown in the following formula: ; Among them, LA is the feeding adaptability, is the weight, N is a positive integer, K is the flattening ratio, Belt parameters, is the density threshold of attached particles.
[0045] A feeding equipment control system with anti-PET sheet material belt provided in an embodiment of the present invention can execute a feeding equipment control method with anti-PET sheet material belt provided in any embodiment of the present invention, and has functional modules and beneficial effects corresponding to the execution method.
[0046] Although the present application makes various references to certain modules in the system according to the embodiments of the present application, any number of different modules may be used and run on the user terminal and / or server, and the various units and modules included are only divided according to functional logic, but are not limited to the above division, as long as the corresponding functions can be achieved; in addition, the specific names of the functional units are only for the convenience of distinguishing each other, and are not used to limit the scope of protection of the present invention.
[0047] The above specific implementations do not constitute a limitation on the protection scope of this application. It should be understood by those skilled in the art that various modifications, combinations and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions and improvements made within the spirit and principles of this application should be included in the protection scope of this application.
Claims
1. A method for controlling a feeding device with anti-PET sheet material belting, characterized in that: The method comprises: Capture images of the PET sheet to be loaded to obtain a PET image, perform unevenness analysis on the PET sheet, and obtain unevenness parameter distribution; Obtain the flattening control parameter space for flattening and loading PET sheets; In the flattening control parameter space, according to the unevenness parameter distribution, the flattening control parameters are optimized to obtain the optimal flattening control parameters for the purpose of improving the flatness of the PET sheet and reducing the electrostatic parameters of the PET sheet after flattening, wherein the electrostatic parameters of the PET sheet after flattening are predicted and optimized by stripping simulation; The optimal flattening control parameters are used to perform flattening and feeding control on the feeding equipment of the PET sheet.
2. A method for controlling a feeding device with anti-PET sheet material belt according to claim 1, characterized in that: The image of the PET sheet to be loaded is collected to obtain the PET image, and the unevenness of the PET sheet is analyzed to obtain the unevenness parameter distribution, including: training a PET image recognizer for PET image recognition; Randomly select multiple acquisition positions for the PET sheet to be loaded to acquire images, and obtain multiple PET images; The plurality of PET images are input into the PET image identifier to obtain a plurality of roughness parameters, and a roughness parameter distribution is constructed based on maximum values and minimum values of the plurality of roughness parameters.
3. A method for controlling a feeding device with anti-PET sheet material belt according to claim 2, characterized in that: A PET image recognizer is trained to perform PET image recognition, including: According to the historical flatness detection data of the PET sheet, a set of sample PET images is collected, and the unevenness size in each sample PET image is marked to obtain a set of sample unevenness parameters; Using the sample PET image set and the sample roughness parameter set as supervised training data, and using a convolutional neural network to train the PET image recognizer; The trained PET image recognizer is verified, and when the requirements are met, the PET image recognizer is obtained.
4. A method for controlling a feeding device with anti-PET sheet material belt according to claim 1, characterized in that: Get the flattening control parameter space for flattening and loading PET sheets, including: Obtain the pressure control parameter range of the flattening equipment for flattening and feeding the PET sheet; The pressure control parameter range is constructed as a flattened control parameter space.
5. The method for controlling a feeding device with anti-PET sheet material belt according to claim 1, characterized in that: In the flattening control parameter space, according to the unevenness parameter distribution, the flattening control parameters are optimized for the purpose of improving the flatness of the PET sheet and reducing the electrostatic parameters of the PET sheet after flattening, and the optimal flattening control parameters are obtained, including: Randomly selecting a first flattening control parameter in the flattening control parameter space; Perform simulated flattening prediction according to the first flattening control parameter and the uneven parameter distribution to obtain a first flattening ratio and a first electrostatic parameter; Collect environmental parameters, combine the first electrostatic parameters to predict the material carrying of the PET sheet, obtain the first material carrying parameters, and calculate and obtain the first material feeding adaptability in combination with the first flattening ratio; Continue to optimize the flattening control parameters until the optimization converges, output the flattening control parameters with the maximum feeding fitness, and obtain the optimal flattening control parameters.
6. A method for controlling a feeding device with anti-PET sheet material belt according to claim 5, characterized in that: Performing simulated flattening prediction according to the first flattening control parameter and the uneven parameter distribution to obtain a first flattening ratio and a first electrostatic parameter includes: According to the historical data of PET sheet feeding and flattening, the sample flattening control parameter set and the sample unevenness parameter distribution set are collected, and the flattening area ratio and surface resistivity of the PET sheet after flattening under different sample flattening control parameters and sample unevenness parameter distributions are obtained, which are marked as the sample flattening ratio set and the sample electrostatic parameter set; The sample flattening control parameter set and the sample unevenness parameter distribution set are used as input data, and the sample flattening ratio set and the sample electrostatic parameter set are used as output supervision data to train a flattening simulation predictor; The first flattening control parameter and the uneven parameter distribution are input into the flattening simulation predictor to predict and obtain a first flattening ratio and a first electrostatic parameter.
7. A method for controlling a feeding device with anti-PET sheet material belt according to claim 5, characterized in that: Collecting environmental parameters, combining the first electrostatic parameters to predict the material carrying of the PET sheet, obtaining first material carrying parameters, combining the first flattening ratio, and calculating the first material feeding adaptability, including: Collect dust concentration in the environment as environmental parameter; Inputting the environmental parameter and the first electrostatic parameter into a material carrying predictor to predict the first material carrying parameter, wherein the material carrying predictor is pre-trained by using sample environmental parameters, sample electrostatic parameters and sample material carrying parameters, each sample material carrying parameter includes attached particle density data of the PET sheet; According to the first flattening ratio and the first material-carrying parameter, the first material-carrying adaptability is calculated as follows: ； Among them, LA is the feeding adaptability, and is the weight, N is a positive integer, K is the flattening ratio, Belt material parameters, is the density threshold of attached particles.
8. A feeding equipment control system with the function of preventing PET sheet from carrying material, characterized in that: The system is used to implement a feeding device control method with anti-PET sheet material belt according to any one of claims 1 to 7, and the system comprises: An unevenness parameter distribution acquisition module is used to collect images of the PET sheet to be loaded, obtain PET images, perform unevenness analysis on the PET sheet, and obtain unevenness parameter distribution; A flattening control parameter space acquisition module is used to acquire the flattening control parameter space of the PET sheet flattening and feeding; An optimal flattening control parameter acquisition module is used to optimize the flattening control parameters within the flattening control parameter space according to the unevenness parameter distribution, with the purpose of improving the flatness of the PET sheet and reducing the electrostatic parameters of the PET sheet after flattening, and obtain the optimal flattening control parameters, wherein the electrostatic parameters of the PET sheet after flattening are predicted and optimized by stripping simulation; The flattening and feeding control module is used to adopt the optimal flattening control parameters to perform flattening and feeding control on the feeding equipment of the PET sheet.