A collaborative control method and system for an aluminum-plastic composite tape production line

Abstract

The present application relates to the technical field of production line control, and particularly relates to a kind of aluminum-plastic composite tape production line's collaborative control method and system.In the cleaning process in aluminum-plastic composite tape production line, aluminum foil image is collected, aluminum foil pollution identification is carried out, and aluminum foil pollution characteristics are obtained;According to aluminum foil pollution characteristics, aluminum foil cleaning parameters and aluminum foil moving speed optimization are carried out, and optimized aluminum foil cleaning parameter set and optimized aluminum foil moving speed set are obtained;Real-time coating parameters in coating process are obtained, and optimized aluminum foil cleaning parameter set and optimized aluminum foil moving speed set are combined to optimize cleaning parameters and moving speed, and optimal cleaning parameters, optimal moving speed are obtained;According to optimal cleaning parameters, optimal moving speed, cleaning process and coating process are collaboratively digitally controlled.It significantly improves the quality of aluminum-plastic composite tape product, and reduces the unqualified rate of product.

Description

Technical Field

[0001] This invention relates to the field of production line control technology, and in particular to a collaborative control method and system for an aluminum-plastic composite belt production line. Background Technology

[0002] In traditional aluminum-plastic composite strip production lines, the cleaning and coating processes are controlled independently. The cleaning process relies on manual experience to preset fixed parameters and visually identifies aluminum foil contamination, making it difficult to accurately obtain contamination characteristics. The coating process operates according to preset parameters without considering the actual condition of the aluminum foil after cleaning and lacks effective predictive optimization methods, resulting in low levels of production intelligence and automation. Existing technologies suffer from inaccurate and inefficient contamination identification, as well as unstable coating quality, leading to low production efficiency and poor product quality consistency. Summary of the Invention

[0003] This invention addresses the technical problems of inaccurate and inefficient pollution identification and unstable coating quality in existing technologies by providing a collaborative control method and system for aluminum-plastic composite strip production lines.

[0004] The technical solution of the present invention to solve the above-mentioned technical problems is as follows:

[0005] In a first aspect, the present invention provides a collaborative control method for an aluminum-plastic composite strip production line, comprising: acquiring aluminum foil images and identifying aluminum foil contamination during the cleaning process within the aluminum-plastic composite strip production line to obtain aluminum foil contamination characteristics; optimizing aluminum foil cleaning parameters and aluminum foil moving speed based on the aluminum foil contamination characteristics to obtain an optimized set of aluminum foil cleaning parameters and an optimized set of aluminum foil moving speeds; acquiring real-time coating parameters during the coating process, and optimizing the cleaning parameters and moving speed in conjunction with the optimized set of aluminum foil cleaning parameters and the optimized set of aluminum foil moving speeds to obtain optimal cleaning parameters and optimal moving speeds, wherein optimization is performed at the cost of reducing coating quality; and performing collaborative digital control of the cleaning process and the coating process according to the optimal cleaning parameters and optimal moving speeds.

[0006] Optionally, during the cleaning process in the aluminum-plastic composite strip production line, aluminum foil images are collected, aluminum foil contamination is identified, and aluminum foil contamination characteristics are obtained. This includes: collecting aluminum foil images during the cleaning process in the aluminum-plastic composite strip production line; inputting the aluminum foil images into the aluminum foil contamination identification path; and identifying and outputting aluminum foil contamination characteristics, wherein the aluminum foil contamination characteristics include the area of ​​aluminum foil contamination.

[0007] The aluminum foil contamination identification path is constructed using the following steps: Based on historical processing data from the cleaning process, a set of sample aluminum foil images is collected, and the aluminum foil contamination features within each sample aluminum foil image are labeled to obtain a set of sample aluminum foil contamination features; based on a convolutional neural network, the aluminum foil images and aluminum foil contamination features are used as input and output data to construct the structure of the aluminum foil contamination identification path; using the set of sample aluminum foil images and the set of sample aluminum foil contamination features, the aluminum foil contamination identification path is iteratively trained in a supervised manner until the test converges, completing the construction and training.

[0008] Optionally, based on the aluminum foil contamination characteristics, the aluminum foil cleaning parameters and aluminum foil movement rate are optimized to obtain an optimized aluminum foil cleaning parameter set and an optimized aluminum foil movement rate set. This includes: obtaining a cleaning parameter space and a movement rate space; randomly generating multiple first aluminum foil cleaning parameters and multiple first aluminum foil movement rates; performing cleaning prediction based on the aluminum foil contamination characteristics and combining the multiple first aluminum foil cleaning parameters and multiple first aluminum foil movement rates to obtain multiple first cleaning rates; calculating the ratio of the multiple first aluminum foil movement rates to a preset movement rate to obtain multiple first speed coefficients; calculating multiple first cleaning fitness based on the multiple first cleaning rates and multiple first speed coefficients; and continuing iterative optimization of the aluminum foil cleaning parameters and aluminum foil movement rate to obtain an optimized aluminum foil cleaning parameter set and an optimized aluminum foil movement rate set.

[0009] Specifically, based on the aluminum foil contamination characteristics, multiple first aluminum foil cleaning parameters and multiple first aluminum foil movement rates are combined to perform cleaning prediction and obtain multiple first cleaning rates. This includes: obtaining a set of aluminum foil contamination characteristics, a set of sample aluminum foil cleaning parameters, and a set of sample aluminum foil movement rates based on historical processing data of the cleaning process, and collecting cleaning rates under different cleaning conditions, and labeling them to obtain a set of sample cleaning rates.

[0010] A machine learning-based cleaning prediction path is constructed. The cleaning prediction path is trained in a supervised iterative manner using the aluminum foil contamination feature set, sample aluminum foil cleaning parameter set, sample aluminum foil movement rate set, and sample cleaning rate set. After convergence, the path is configured in the factory control center of the aluminum-plastic composite belt. The aluminum foil contamination features are combined with each first aluminum foil cleaning parameter and first aluminum foil movement rate and input into the cleaning prediction path to obtain multiple first cleaning rates.

[0011] Optionally, the aluminum foil cleaning parameters and aluminum foil moving rate are iteratively optimized to obtain an optimized set of aluminum foil cleaning parameters and an optimized set of aluminum foil moving rates. This includes: continuing to generate second aluminum foil cleaning parameters and multiple second aluminum foil moving rates, and predictively processing to obtain multiple second cleaning fitness values; retaining the multiple aluminum foil cleaning parameters and aluminum foil moving rates with the highest cleaning fitness values; continuing to generate and iteratively optimize the aluminum foil cleaning parameters and aluminum foil moving rates until the optimization converges, and obtaining the final retained multiple aluminum foil cleaning parameters and aluminum foil moving rates to obtain an optimized set of aluminum foil cleaning parameters and an optimized set of aluminum foil moving rates.

[0012] Optionally, real-time coating parameters in the coating process are obtained, and the cleaning parameters and movement speed are optimized in combination with the optimized aluminum foil cleaning parameter set and the optimized aluminum foil moving speed set to obtain the optimal cleaning parameters and optimal movement speed. This includes: obtaining the optimized cleaning rate set and optimized cleaning fitness set corresponding to the optimized aluminum foil cleaning parameter set and the optimized aluminum foil moving speed set; obtaining real-time coating parameters in the coating process; inputting the real-time coating parameters into the coating prediction path in combination with each optimized aluminum foil moving speed and optimized cleaning rate, and outputting a predicted coating accuracy parameter set. The coating prediction path is obtained by training with a sample coating parameter set, a sample aluminum foil moving speed set, a sample cleaning rate set, and a sample coating accuracy parameter set. Each predicted coating accuracy parameter includes the average similarity between the coating thickness and the standard thickness.

[0013] Obtain preset coating accuracy parameters; calculate the reduction rate of each predicted coating accuracy parameter in the predicted coating accuracy parameter set compared with the preset coating accuracy parameter to obtain multiple coating quality cost parameters;

[0014] Based on the optimized cleaning fitness set and multiple coating quality cost parameters, a cooperative fitness set is calculated. The optimized aluminum foil cleaning parameters and optimized aluminum foil moving speed corresponding to the maximum cooperative fitness value are selected as the optimal cleaning parameters and optimal moving speed. Specifically, calculating the cooperative fitness set based on the optimized cleaning fitness set and multiple coating quality cost parameters includes subtracting the corresponding coating quality cost parameter from each optimized cleaning fitness value within the optimized cleaning fitness set.

[0015] Secondly, the present invention provides a collaborative control system for an aluminum-plastic composite strip production line, comprising:

[0016] The contamination feature recognition module is used to collect aluminum foil images during the cleaning process in the aluminum-plastic composite belt production line, identify aluminum foil contamination, and obtain aluminum foil contamination features.

[0017] The optimization parameter acquisition module is used to optimize the aluminum foil cleaning parameters and aluminum foil movement rate based on the aluminum foil contamination characteristics, and obtain an optimized aluminum foil cleaning parameter set and an optimized aluminum foil movement rate set.

[0018] The optimal parameter acquisition module is used to acquire real-time coating parameters in the coating process, and to optimize the cleaning parameters and moving speed by combining the optimized aluminum foil cleaning parameter set and the optimized aluminum foil moving speed set, so as to obtain the optimal cleaning parameters and the optimal moving speed, wherein the optimization is carried out at the cost of reducing coating quality.

[0019] The optimal parameter application module is used to perform coordinated digital control of the cleaning process and the coating process according to the optimal cleaning parameters and the optimal moving speed.

[0020] By implementing this invention, aluminum foil images can be collected during the cleaning process in the aluminum-plastic composite belt production line to identify aluminum foil contamination and obtain aluminum foil contamination characteristics. This provides accurate data for optimizing subsequent cleaning parameters and moving speed, avoiding insufficient or excessive cleaning caused by inaccurate judgment of aluminum foil contamination. At the same time, the aluminum foil contamination identification path based on convolutional neural networks can significantly improve the accuracy and efficiency of aluminum foil contamination identification. Compared with traditional manual identification methods, it reduces human error and enables real-time and continuous contamination monitoring.

[0021] By implementing this invention, it is possible to optimize aluminum foil cleaning parameters and aluminum foil moving speed based on the characteristics of aluminum foil contamination, thereby obtaining an optimized set of aluminum foil cleaning parameters and an optimized set of aluminum foil moving speeds. Through multiple rounds of random generation of parameters and speeds and iterative optimization, a better combination of cleaning parameters and moving speeds can be selected from a large number of parameter combinations to form an optimized parameter set, laying the foundation for obtaining the optimal parameters in the future.

[0022] By implementing this invention, it is possible to obtain real-time coating parameters in the coating process, and optimize the cleaning parameters and moving speed by combining the optimized aluminum foil cleaning parameter set and the optimized aluminum foil moving speed set, thereby obtaining the optimal cleaning parameters and the optimal moving speed. This fully considers the correlation between the cleaning process and the coating process, breaks the limitation of independent control of each process in traditional production, and optimizes the parameters of the cleaning process by combining them with the real-time parameters of the coating process, thus ensuring the quality of subsequent coating processes.

[0023] By implementing this invention, it is possible to achieve coordinated digital control of the cleaning process and the coating process according to the optimal cleaning parameters and the optimal moving speed. This enables precise coordinated control of the cleaning process and the coating process, avoids errors caused by human operation, and ensures the stability and consistency of the production process.

[0024] In summary, by implementing this invention, it is possible to accurately identify aluminum foil contamination characteristics, scientifically optimize cleaning parameters and moving speed, and fully integrate real-time parameters of the coating process to achieve cross-process collaborative optimization. Ultimately, collaborative digital control ensures that the cleaning and coating processes operate according to optimal parameters, thereby improving the intelligence and automation level of the production line, significantly improving the quality of aluminum-plastic composite belt products, and reducing the product defect rate. Attached Figure Description

[0025] Figure 1 A flowchart illustrating a collaborative control method for an aluminum-plastic composite strip production line provided by the present invention;

[0026] Figure 2 This invention provides a schematic diagram of the structure of a collaborative control system for an aluminum-plastic composite belt production line.

[0027] In the attached diagram, the components represented by each number are as follows:

[0028] Pollution feature identification module 11, optimization parameter acquisition module 12, optimal parameter acquisition module 13, and optimal parameter application module 14. Detailed Implementation

[0029] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0030] In the description of this invention, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more features. In the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0031] In the description of this invention, the term "for example" is used to mean "used as an example, illustration, or description." Any embodiment described as "for example" in this invention is not necessarily to be construed as being more preferred or advantageous than other embodiments. The following description is provided to enable any person skilled in the art to make and use the invention. Details are set forth in the following description for purposes of explanation. It should be understood that those skilled in the art will recognize that the invention can be made without using these specific details. In other instances, well-known structures and processes will not be described in detail to avoid obscuring the description of the invention with unnecessary detail. Therefore, the invention is not intended to be limited to the embodiments shown, but is consistent with the broadest scope of the principles and features disclosed herein.

[0032] Example 1, as Figure 1 As shown, this embodiment of the invention provides a collaborative control method for an aluminum-plastic composite strip production line, including:

[0033] S100: During the cleaning process in the aluminum-plastic composite belt production line, aluminum foil images are collected to identify aluminum foil contamination and obtain aluminum foil contamination characteristics.

[0034] S200: Based on the aluminum foil contamination characteristics, optimize the aluminum foil cleaning parameters and aluminum foil moving speed to obtain an optimized set of aluminum foil cleaning parameters and an optimized set of aluminum foil moving speeds;

[0035] S300: Obtain real-time coating parameters in the coating process, and optimize the cleaning parameters and moving speed by combining the optimized aluminum foil cleaning parameter set and the optimized aluminum foil moving speed set to obtain the optimal cleaning parameters and the optimal moving speed, wherein the optimization is performed at the cost of reducing coating quality.

[0036] S400: Perform coordinated digital control on the cleaning process and the coating process according to the optimal cleaning parameters and the optimal moving speed.

[0037] In step S100 of this application embodiment, during the cleaning process in the aluminum-plastic composite strip production line, aluminum foil images are acquired to identify aluminum foil contamination and obtain aluminum foil contamination characteristics, including:

[0038] Images of aluminum foil are captured during the cleaning process in the aluminum-plastic composite belt production line.

[0039] The aluminum foil image is input into the aluminum foil contamination identification path, and the identification output obtains the aluminum foil contamination features, wherein the aluminum foil contamination features include the area of ​​aluminum foil contamination.

[0040] In this embodiment, step S100 is a fundamental step in the collaborative control of the aluminum-plastic composite strip production line. Its purpose is to provide data support for subsequent cleaning parameter optimization and process collaborative control by accurately collecting and identifying aluminum foil contamination information.

[0041] First, images of the aluminum foil need to be collected during the cleaning process in the aluminum-plastic composite strip production line. Specifically, this can be achieved by using image acquisition equipment such as high-definition cameras installed at appropriate locations on the production line to capture real-time images of the aluminum foil surface during the cleaning process and obtain visual information about the contaminated areas of the aluminum foil.

[0042] Furthermore, the aluminum foil image needs to be input into the aluminum foil contamination identification path, and the identification output will obtain the aluminum foil contamination characteristics.

[0043] In step S100 of this application embodiment, the aluminum foil contamination identification path is constructed using the following steps:

[0044] Based on historical processing data of the cleaning process, a set of sample aluminum foil images was collected, and the aluminum foil contamination features in each sample aluminum foil image were labeled to obtain a set of sample aluminum foil contamination features;

[0045] Based on a convolutional neural network, an aluminum foil contamination identification path is constructed by using aluminum foil images and aluminum foil contamination features as input and output data.

[0046] Using the sample aluminum foil image set and sample aluminum foil contamination feature set, the aluminum foil contamination recognition path is trained iteratively in a supervised manner until the test converges, thus completing the construction and training.

[0047] In this embodiment, the aluminum foil contamination identification path can replace the traditional manual visual identification method, avoid human error, and realize quantitative and real-time identification of aluminum foil contamination characteristics. This provides reliable contamination data for the subsequent optimization of cleaning parameters and moving speed in step S200 and process collaboration optimization in step S300, ensuring the scientific and accurate collaborative control of the aluminum-plastic composite belt production line from a technical perspective.

[0048] First, based on historical processing data from the cleaning process, a set of sample aluminum foil images needs to be collected. The aluminum foil contamination features within each sample image are then labeled to obtain a sample aluminum foil contamination feature set, where the main feature is the area of ​​contamination. Specifically, from historical aluminum foil images collected during the cleaning process's historical timeframe, images covering different degrees and types of contamination are selected to form a sample aluminum foil image set. Subsequently, the contamination features in each sample aluminum foil image are precisely labeled manually, forming a sample aluminum foil contamination feature set that corresponds one-to-one with each sample aluminum foil image. This provides matching input and output supervisory data for subsequent aluminum foil contamination identification path training.

[0049] Then, based on a convolutional neural network, it is necessary to construct the structure of the aluminum foil contamination identification path, using aluminum foil images and aluminum foil contamination features as input and output data.

[0050] Optionally, the main structure of the convolutional neural network used to construct the aluminum foil contamination identification path consists of an input layer, a convolutional layer, a pooling layer, and a fully connected layer.

[0051] Specifically, the input data is an aluminum foil image, the image size can be set to 320 pixels × 320 pixels, the number of image channels is 3, and the format is an RGB color image.

[0052] The convolutional layer for identifying aluminum foil contamination contains a total of 3 convolutional layers.

[0053] The first convolutional layer has 32 convolutional kernels, each with a size of 3×3 and a stride of 1. The padding method is "same," using edge padding to ensure that the output feature map size is consistent with the input. The activation function is ReLU, which is used to initially extract the basic texture and edge features of the aluminum foil image, laying the groundwork for subsequent identification of contaminated areas.

[0054] The second convolutional layer has 64 convolutional kernels, a kernel size of 3×3, a stride of 1, and a padding method of "same". The activation function is ReLU, which further refines feature extraction and focuses on the features of possible contamination areas on the aluminum foil surface.

[0055] The third convolutional layer has 128 convolutional kernels, a kernel size of 3×3, a stride of 1, and a padding method of "same". The activation function is ReLU, which deeply extracts the unique features of the contaminated area, such as the contour of the contaminated area and gray-level differences.

[0056] Furthermore, a max pooling layer is set after each convolutional layer, with a pooling kernel size of 2×2 and a stride of 2, to compress the feature map dimension, reduce computation, and retain key contamination features to avoid overfitting.

[0057] Two fully connected layers are set in the aluminum foil contamination identification path.

[0058] The first fully connected layer has 256 neurons, and the activation function is ReLU.

[0059] The second fully connected layer is the output layer: it has one neuron, uses the sigmoid activation function, and the output is the aluminum foil contamination area in the aluminum foil contamination feature.

[0060] For the hyperparameter settings of the aluminum foil contamination identification path, the learning rate was set to 0.001. This ensures stable parameter updates during model training while avoiding parameter oscillations due to an excessively high learning rate or inefficient training due to an excessively low learning rate. The batch size was set to 32, with 32 sample aluminum foil images input for parameter updates each training iteration. This balances training efficiency and memory usage, ensuring the model can stably learn the contamination characteristics in the samples. The regularization coefficient was set to 0.0001, employing L2 regularization to suppress overfitting.

[0061] Furthermore, the aluminum foil contamination recognition path needs to be trained in a supervised manner using the sample aluminum foil image set and the sample aluminum foil contamination feature set until the test converges, thus completing the construction and training.

[0062] The sample aluminum foil image set and sample aluminum foil contamination feature set include aluminum foil images recorded by image acquisition equipment during the cleaning process in historical production, covering different production batches, different aluminum foil raw materials, and different degrees of contamination. The training sample size is set to 20,000, of which 16,000 (80%) are used as the training set for iterative updates of the aluminum foil contamination recognition path parameters, and 4,000 (20%) are used as the validation set for real-time verification of the aluminum foil contamination recognition path performance during training.

[0063] In training the aluminum foil contamination identification path, the initial training rounds were set to 100 rounds, and a dynamic adjustment strategy was adopted. If the performance of the aluminum foil contamination identification path on the validation set, such as the contamination area identification error, did not decrease for three consecutive rounds during the training process, the training was stopped early. If the aluminum foil contamination identification path still did not reach the convergence criterion after 100 rounds of training, the number of training rounds was appropriately increased.

[0064] The convergence criterion for the aluminum foil contamination identification path can be set as follows: after five consecutive training rounds, the average absolute difference between the model's output contamination area and the labeled sample contamination area on the validation set should be stable within 5%. When the convergence condition is met, the model is considered to have converged, and the construction and training of the aluminum foil contamination identification path are complete.

[0065] In step S200 of this application embodiment, based on the aluminum foil contamination characteristics, the aluminum foil cleaning parameters and aluminum foil movement rate are optimized to obtain an optimized set of aluminum foil cleaning parameters and an optimized set of aluminum foil movement rates, including:

[0066] Obtain the cleaning parameter space and the moving speed space, and randomly generate multiple first aluminum foil cleaning parameters and multiple first aluminum foil moving speeds;

[0067] Based on the aluminum foil contamination characteristics, and combined with multiple first aluminum foil cleaning parameters and multiple first aluminum foil movement rates, cleaning prediction is performed to obtain multiple first cleaning rates;

[0068] Calculate the ratio of multiple first aluminum foil moving speeds to a preset moving speed to obtain multiple first speed coefficients;

[0069] Multiple first cleaning fitness values ​​are calculated based on multiple first cleaning rates and multiple first speed coefficients;

[0070] Continue iterative optimization of aluminum foil cleaning parameters and aluminum foil moving speed to obtain an optimized set of aluminum foil cleaning parameters and an optimized set of aluminum foil moving speed.

[0071] In this embodiment of the application, the purpose of step S200 is to select a preliminary high-quality parameter combination that balances cleaning effect and production efficiency from the possible combinations of cleaning parameters and moving speed based on the aluminum foil contamination characteristics obtained in S100, forming a first-order optimized cleaning parameter set and a first-order moving speed set, laying the foundation for cross-process collaborative optimization in the subsequent step S300 combined with the coating process.

[0072] First, it is necessary to obtain the cleaning parameter space and the moving speed space, and randomly generate multiple first aluminum foil cleaning parameters and multiple first aluminum foil moving speeds.

[0073] First, the cleaning parameter space is defined, including the value range of key parameters such as cleaning solution concentration, cleaning temperature, and cleaning time. Then, the movement rate space is obtained, which is the reasonable range of values ​​for the aluminum foil's movement speed during the cleaning process. Based on the aforementioned space range, multiple first aluminum foil cleaning parameters and multiple first aluminum foil movement rates are obtained through random generation, such as cleaning solutions of different concentrations, different temperature settings, and different aluminum foil transport speeds. This ensures that the initial aluminum foil cleaning parameters and aluminum foil movement rates cover a wide range of values, providing a sufficient candidate basis for subsequent optimization.

[0074] Furthermore, based on the aluminum foil contamination characteristics, multiple first aluminum foil cleaning parameters and multiple first aluminum foil movement rates are combined to perform cleaning prediction and obtain multiple first cleaning rates.

[0075] In step S200 of this application embodiment, cleaning prediction is performed based on the aluminum foil contamination characteristics combined with multiple first aluminum foil cleaning parameters and multiple first aluminum foil movement rates to obtain multiple first cleaning rates, including:

[0076] Based on historical processing data of the cleaning process, a set of aluminum foil contamination characteristics, a set of sample aluminum foil cleaning parameters, and a set of sample aluminum foil movement rates were obtained. The cleanliness rate under different cleaning conditions was collected and labeled to obtain a set of sample cleanliness rates.

[0077] Construct a clean prediction path based on machine learning;

[0078] Using the aforementioned aluminum foil contamination feature set, sample aluminum foil cleaning parameter set, sample aluminum foil movement rate set, and sample cleanliness rate set, the cleaning prediction path is trained iteratively under supervision and configured in the factory control center of the aluminum-plastic composite belt after convergence.

[0079] The aluminum foil contamination characteristics are combined with each first aluminum foil cleaning parameter and the first aluminum foil movement rate, and then input into the cleaning prediction path to obtain multiple first cleaning rates.

[0080] In this embodiment, to perform cleaning prediction and obtain multiple first cleanliness rates, it is first necessary to prepare a sample dataset. Specifically, three types of core data need to be extracted from the historical production records of the cleaning process of the aluminum-plastic composite strip production line: first, a set of aluminum foil contamination characteristics covering different levels of contamination; second, a set of sample aluminum foil cleaning parameters actually used in historical production; and third, a set of sample aluminum foil movement rates corresponding to different aluminum foil cleaning parameters. Simultaneously, the actual cleanliness rate data after aluminum foil cleaning is collected under the above combinations of different aluminum foil contamination characteristics, aluminum foil cleaning parameters, and aluminum foil movement rates, and each cleanliness rate data is accurately labeled to form a sample cleanliness rate set corresponding one-to-one with the aforementioned three sets. This provides supervised data with matching input and output for subsequent cleaning prediction path training.

[0081] Wherein, when the aluminum foil contamination characteristic is the contaminated area, the cleaning rate = (initial contaminated area of ​​aluminum foil before cleaning - residual contaminated area of ​​aluminum foil after cleaning) ÷ initial contaminated area of ​​aluminum foil before cleaning × 100%. If the cleaning rate is 100%, it means that the surface contamination of the aluminum foil has been completely removed; the higher the cleaning rate value, the better the cleaning effect, and vice versa.

[0082] Furthermore, it is necessary to construct a clean prediction path based on machine learning.

[0083] Based on the task type of the cleaning prediction path, the Gradient Boosting Decision Tree (GBDT) can be selected to construct the cleaning prediction path.

[0084] The structure of the clean prediction path includes an input layer, a decision tree-based model, and an output layer.

[0085] The input data for the input layer consists of a combination of aluminum foil contamination features, sample aluminum foil cleaning parameters, and sample aluminum foil movement rate, comprising 2+n dimensions: 1 dimension for aluminum foil contamination features; n dimensions for sample aluminum foil cleaning parameters, where n depends on the number of cleaning parameters (e.g., n=4 when cleaning parameters include cleaning solution concentration, cleaning temperature, cleaning time, and cleaning pressure); and 1 dimension for sample aluminum foil movement rate. Furthermore, all input features need to be standardized to ensure a balanced weighting of each feature's influence on the training of the cleaning prediction path.

[0086] In the decision tree-based model, the maximum depth of each decision tree is set to 5 to avoid overfitting due to excessive tree depth, while ensuring that key correlations between features are captured. The minimum number of splits per decision tree is set to 20, meaning that splitting stops when the number of node samples is less than 20, ensuring that the split nodes are statistically significant. The minimum number of leaf nodes per decision tree is set to 10 to avoid leaf nodes with too few samples, ensuring the stability of prediction results. The feature sampling ratio is set to 0.8, meaning that 80% of the input features are randomly selected during the training of each decision tree, enhancing the model's generalization ability and reducing the risk of overfitting.

[0087] The output data of the output layer is a continuous sample cleanliness rate, with a value range of [0,100] and a unit of %. Through regression calculation of the GBDT model, the predicted cleanliness rate value under the corresponding input parameter combination is directly output.

[0088] In the parameter settings for the cleanliness prediction path, the learning rate is set to 0.05 to control the correction magnitude of each decision tree to the total model error. This ensures that the model gradually approaches the optimal solution while avoiding training oscillations due to an excessively high learning rate or low training efficiency due to an excessively low learning rate. The number of decision trees is set to 200. Through iterative ensemble of multiple decision trees, the prediction error is gradually reduced, while avoiding excessive model complexity and training time due to an excessive number of decision trees. The mean squared error (MSE) is used as the loss function, calculating the squared difference between the cleanliness rate predicted by the cleanliness prediction path and the sample cleanliness rate. Minimizing this loss function optimizes the parameters of the cleanliness prediction path, aligning with the regression prediction requirement of cleanliness rate as a continuous value. The regularization coefficient is set to 0.001, employing L2 regularization to constrain the decision tree parameters of the cleanliness prediction path and suppress overfitting.

[0089] Furthermore, the cleaning prediction path needs to be trained in a supervised iterative manner using the aluminum foil contamination feature set, sample aluminum foil cleaning parameter set, sample aluminum foil movement rate set, and sample cleaning rate set, and then configured in the factory control center of the aluminum-plastic composite belt after convergence.

[0090] For training the cleaning prediction path, the training data consisted of the aforementioned set of aluminum foil contamination features, sample aluminum foil cleaning parameters, sample aluminum foil movement rates, and sample cleanliness rates. The training sample size was set to 15,000 samples, divided into training and validation sets in an 8:2 ratio. The initial training epochs were set to 100 epochs, dynamically adjusted using an "early stop mechanism." The cleaning prediction path was considered to have converged when the loss value on the validation set remained stable below 0.5 for five consecutive epochs, and the fluctuation range of the loss value in each epoch was less than 0.001.

[0091] After training is completed, the cleaning prediction path is configured in the factory control center of the aluminum-plastic composite belt to ensure that it can respond to subsequent prediction requests in real time.

[0092] Finally, the aluminum foil contamination characteristics are combined with each first aluminum foil cleaning parameter and the first aluminum foil movement rate, and input into the cleaning prediction path to output multiple first cleaning rates.

[0093] Furthermore, it is necessary to calculate the ratios of multiple first aluminum foil moving speeds to a preset moving speed to obtain multiple first speed coefficients. The preset moving speed can be set as a benchmark aluminum foil moving speed based on the design standards, historical production experience, or capacity requirements of the aluminum-plastic composite belt production line. Next, for each first aluminum foil moving speed randomly generated in S200, a division operation is performed with the preset moving speed to obtain the first speed coefficient, i.e., first speed coefficient = first aluminum foil moving speed ÷ preset moving speed, resulting in multiple first speed coefficients corresponding to the first aluminum foil moving speeds. When the first speed coefficient is greater than 1, it indicates that the first aluminum foil moving speed is higher than the preset value, and the production efficiency is higher; when it equals 1, it perfectly matches the preset efficiency; when it is less than 1, the production efficiency is lower than the preset standard, thus quantifying the impact of different moving speeds on production efficiency.

[0094] Next, multiple first cleaning fitness scores need to be calculated based on multiple first cleaning rates and multiple first speed coefficients. Specifically, a weighted summation method can be used, combining the first cleaning rate reflecting cleaning effectiveness with the first speed coefficient reflecting production efficiency. For example, first cleaning fitness score = (first cleaning rate × 0.6) + (first speed coefficient × 0.4). The cleaning rate is weighted at 0.6, and the speed coefficient is weighted at 0.4. These weights can be adjusted according to the emphasis placed on cleaning effectiveness and production efficiency in actual production.

[0095] Then, by substituting each first cleaning rate and its corresponding first speed coefficient into the above formula, multiple first cleaning fitness values ​​are calculated respectively.

[0096] In step S200 of this embodiment, the aluminum foil cleaning parameters and aluminum foil moving speed are iteratively optimized to obtain an optimized set of aluminum foil cleaning parameters and an optimized set of aluminum foil moving speeds, including:

[0097] Continue to generate second aluminum foil cleaning parameters and multiple second aluminum foil movement rates, and predict the processing to obtain multiple second cleaning fitnesss;

[0098] It retains the maximum range of aluminum foil cleaning parameters and aluminum foil movement rates for optimal cleaning performance;

[0099] Continue generating and iteratively optimizing aluminum foil cleaning parameters and aluminum foil moving speed until optimization converges, and obtain the final retained multiple aluminum foil cleaning parameters and aluminum foil moving speed sets to obtain the optimized aluminum foil cleaning parameter set and the optimized aluminum foil moving speed set.

[0100] Specifically, all combinations of first aluminum foil cleaning parameters and first aluminum foil moving speed are sorted according to the magnitude of the first cleaning adaptability, and the top 30% of the combinations with the first cleaning adaptability are retained as candidate aluminum foil cleaning parameters and candidate aluminum foil moving speeds for the first round of optimization.

[0101] Then, based on the retained candidate aluminum foil cleaning parameters and candidate aluminum foil moving rates, a local random search is performed in the cleaning parameter space and moving rate space to generate new second aluminum foil cleaning parameters and second aluminum foil moving rates. Specifically, the generation method can be to take values ​​that fluctuate within a small range around the candidate parameter values. The number of parameter pairs of the finally generated second aluminum foil cleaning parameters and second aluminum foil moving rates is the same as the number of the initially generated first aluminum foil cleaning parameters and first aluminum foil moving rates.

[0102] Next, the newly generated second aluminum foil cleaning parameters and second aluminum foil moving rate are combined with the aluminum foil contamination characteristics obtained through step S100. The second cleaning rate is obtained through the cleaning prediction path, the second speed coefficient and the second cleaning fitness are calculated, and the top 30% of the combinations with the second cleaning fitness are selected again.

[0103] Then, the process of generating candidate aluminum foil cleaning parameters and candidate aluminum foil movement rates, calculating the cleaning rate, calculating the speed coefficient, calculating the cleaning fitness, and screening based on the cleaning fitness is repeated continuously until the improvement of the highest cleaning fitness is less than 0.01 in three consecutive iterations, at which point the iteration stops. The final set of retained aluminum foil cleaning parameters and aluminum foil movement rates are then used as the optimized aluminum foil cleaning parameter set and the optimized aluminum foil movement rate set.

[0104] In step S300 of this application embodiment, real-time coating parameters in the coating process are obtained, and the cleaning parameters and moving speed are optimized by combining the optimized aluminum foil cleaning parameter set and the optimized aluminum foil moving speed set to obtain the optimal cleaning parameters and optimal moving speed, including:

[0105] Obtain the optimized cleaning rate set and optimized cleaning fitness set corresponding to the optimized aluminum foil cleaning parameter set and optimized aluminum foil moving speed set;

[0106] Obtain real-time coating parameters during the coating process;

[0107] The real-time coating parameters are combined with each optimized aluminum foil moving speed and optimized cleaning rate input into the coating prediction path to output a set of predicted coating accuracy parameters. The coating prediction path is obtained by training with a set of sample coating parameters, a set of sample aluminum foil moving speeds, a set of sample cleaning rates and a set of sample coating accuracy parameters. Each predicted coating accuracy parameter includes the average similarity between the coating thickness and the standard thickness.

[0108] Obtain the preset coating accuracy parameters;

[0109] The reduction rate of each predicted coating accuracy parameter in the predicted coating accuracy parameter set compared with the preset coating accuracy parameter is calculated to obtain multiple coating quality cost parameters.

[0110] Based on the optimized cleaning fitness set and multiple coating quality cost parameters, a cooperative fitness set is calculated. The optimized aluminum foil cleaning parameters and optimized aluminum foil moving rate corresponding to the maximum cooperative fitness are selected as the optimal cleaning parameters and optimal moving rate.

[0111] In step S300 of this application embodiment, a cooperative fitness set is calculated based on the optimized cleaning fitness set and multiple coating quality cost parameters, including:

[0112] The co-fitness set is obtained by subtracting the corresponding coating quality cost parameter from each optimized cleaning fitness within the optimized cleaning fitness set.

[0113] In the embodiments of this application, the purpose of step S300 is to break the independent control limitations of the aluminum foil cleaning process and the coating process. By introducing real-time parameters of the coating process, the optimized parameter set obtained in step S200 is optimized across processes, and finally the optimal cleaning parameters and the optimal moving speed that can simultaneously take into account the cleaning effect, production efficiency and coating quality are determined.

[0114] First, it is necessary to extract the optimized cleaning rate set and optimized cleaning fitness set corresponding to the optimized aluminum foil cleaning parameter set and optimized aluminum foil moving speed set in S200. That is, to obtain the cleaning rate and cleaning fitness corresponding to each optimized parameter combination as the basic data for collaborative optimization.

[0115] Next, it is necessary to obtain the real-time coating parameters during the coating process. Specifically, this can be achieved by collecting real-time coating parameters through sensors or a control system in the coating process. These coating parameters include key parameters that directly affect the coating quality, such as coating speed, coating liquid concentration, coating temperature, and coating pressure.

[0116] Furthermore, the real-time coating parameters need to be combined with each optimized aluminum foil moving speed and optimized cleaning rate input into the coating prediction path to output a set of predicted coating accuracy parameters. The coating prediction path is obtained by training with a set of sample coating parameters, a set of sample aluminum foil moving speeds, a set of sample cleaning rates, and a set of sample coating accuracy parameters. Each predicted coating accuracy parameter includes the average similarity between the coating thickness and the standard thickness.

[0117] The task type of the coating prediction path is the same as that of the cleaning prediction path mentioned above. The cleaning prediction path outputs a continuous value of cleaning rate, while the coating prediction path outputs a continuous value of the average similarity between coating thickness and standard thickness. Both require establishing a non-linear relationship between input features and continuous output. Furthermore, the inputs of the cleaning prediction path are aluminum foil contamination features + aluminum foil cleaning parameters + aluminum foil movement speed, while the inputs of the coating prediction path are real-time coating parameters + optimized aluminum foil movement speed + optimized cleaning rate. Both are numerical features. Therefore, the coating prediction path can adopt the same method as the cleaning prediction path, building and training based on gradient boosting decision trees. It is only necessary to change the dimension of the input features and replace the training data with sample coating parameter sets, sample aluminum foil movement speed sets, sample cleaning rate sets, and sample coating accuracy parameter sets. The specific operation method will not be elaborated here.

[0118] Furthermore, the real-time coating parameters are combined with each rate in the optimized aluminum foil moving rate set and the corresponding cleaning rate in the optimized cleaning rate set, respectively, and input into the coating prediction path. The output is the predicted coating accuracy parameter corresponding to each optimized parameter combination, forming a predicted coating accuracy parameter set.

[0119] Next, a preset coating accuracy parameter needs to be set as the benchmark for coating quality. This preset coating accuracy parameter can be set based on product quality standards, such as requiring the average similarity between the coating thickness and the standard thickness to reach 95%. Then, the reduction rate of each predicted coating accuracy parameter within the predicted coating accuracy parameter set compared to the preset coating accuracy parameter is calculated to obtain multiple coating quality cost parameters. The specific calculation method can be: Coating quality cost parameter = (Preset coating accuracy - Predicted coating accuracy) ÷ Preset coating accuracy × 100%. Finally, multiple coating quality cost parameters are obtained; the higher the value, the greater the negative impact of the optimized parameter combination on coating quality.

[0120] Furthermore, a cooperative fitness set needs to be calculated based on the optimized cleaning fitness set and multiple coating quality cost parameters. Specifically, this can be achieved by subtracting the corresponding coating quality cost parameter from each optimized cleaning fitness value within the optimized cleaning fitness set. That is, Cooperative Fitness = Optimized Cleaning Fitness - Coating Quality Cost Parameter.

[0121] Finally, the cooperative fitness with the largest value is selected from the cooperative fitness set. The corresponding optimized aluminum foil cleaning parameters and optimized aluminum foil moving speed are the final optimal cleaning parameters and optimal moving speed.

[0122] Example 2, as Figure 2 As shown, based on the same inventive concept as the collaborative control method for an aluminum-plastic composite strip production line provided in Embodiment 1, this embodiment of the invention also provides a collaborative control system for an aluminum-plastic composite strip production line, comprising:

[0123] The contamination feature recognition module 11 is used to collect aluminum foil images during the cleaning process in the aluminum-plastic composite belt production line, identify aluminum foil contamination, and obtain aluminum foil contamination features.

[0124] The optimization parameter acquisition module 12 is used to optimize the aluminum foil cleaning parameters and aluminum foil moving speed according to the aluminum foil contamination characteristics, and obtain an optimized aluminum foil cleaning parameter set and an optimized aluminum foil moving speed set;

[0125] The optimal parameter acquisition module 13 is used to acquire real-time coating parameters in the coating process, and to optimize the cleaning parameters and moving speed by combining the optimized aluminum foil cleaning parameter set and the optimized aluminum foil moving speed set, so as to obtain the optimal cleaning parameters and the optimal moving speed, wherein the optimization is carried out at the cost of reducing coating quality.

[0126] The optimal parameter application module 14 is used to perform coordinated digital control of the cleaning process and the coating process according to the optimal cleaning parameters and the optimal moving speed.

[0127] Furthermore, the pollution feature identification module 11 includes the following execution steps:

[0128] Images of aluminum foil are captured during the cleaning process in the aluminum-plastic composite belt production line.

[0129] The aluminum foil image is input into the aluminum foil contamination identification path, and the identification output obtains the aluminum foil contamination features, wherein the aluminum foil contamination features include the area of ​​aluminum foil contamination.

[0130] The aluminum foil contamination identification path is constructed using the following steps:

[0131] Based on historical processing data of the cleaning process, a set of sample aluminum foil images was collected, and the aluminum foil contamination features in each sample aluminum foil image were labeled to obtain a set of sample aluminum foil contamination features;

[0132] Based on a convolutional neural network, an aluminum foil contamination identification path is constructed by using aluminum foil images and aluminum foil contamination features as input and output data.

[0133] Using the sample aluminum foil image set and sample aluminum foil contamination feature set, the aluminum foil contamination recognition path is trained iteratively in a supervised manner until the test converges, thus completing the construction and training.

[0134] Furthermore, the parameter acquisition module 12 includes the following execution steps:

[0135] Obtain the cleaning parameter space and the moving speed space, and randomly generate multiple first aluminum foil cleaning parameters and multiple first aluminum foil moving speeds;

[0136] Based on the aluminum foil contamination characteristics, and combined with multiple first aluminum foil cleaning parameters and multiple first aluminum foil movement rates, cleaning prediction is performed to obtain multiple first cleaning rates;

[0137] Calculate the ratio of multiple first aluminum foil moving speeds to a preset moving speed to obtain multiple first speed coefficients;

[0138] Multiple first cleaning fitness values ​​are calculated based on multiple first cleaning rates and multiple first speed coefficients;

[0139] Continue iterative optimization of aluminum foil cleaning parameters and aluminum foil moving speed to obtain an optimized set of aluminum foil cleaning parameters and an optimized set of aluminum foil moving speed.

[0140] Specifically, based on the aluminum foil contamination characteristics, and combining multiple first aluminum foil cleaning parameters and multiple first aluminum foil movement rates, cleaning prediction is performed to obtain multiple first cleaning rates, including:

[0141] Based on historical processing data of the cleaning process, a set of aluminum foil contamination characteristics, a set of sample aluminum foil cleaning parameters, and a set of sample aluminum foil movement rates were obtained. The cleanliness rate under different cleaning conditions was collected and labeled to obtain a set of sample cleanliness rates.

[0142] Construct a clean prediction path based on machine learning;

[0143] Using the aforementioned aluminum foil contamination feature set, sample aluminum foil cleaning parameter set, sample aluminum foil movement rate set, and sample cleanliness rate set, the cleaning prediction path is trained iteratively under supervision and configured in the factory control center of the aluminum-plastic composite belt after convergence.

[0144] The aluminum foil contamination characteristics are combined with each first aluminum foil cleaning parameter and the first aluminum foil movement rate, and then input into the cleaning prediction path to obtain multiple first cleaning rates.

[0145] This involves continuing iterative optimization of aluminum foil cleaning parameters and aluminum foil movement speed to obtain an optimized set of aluminum foil cleaning parameters and an optimized set of aluminum foil movement speeds, including:

[0146] Continue to generate second aluminum foil cleaning parameters and multiple second aluminum foil movement rates, and predict the processing to obtain multiple second cleaning fitnesss;

[0147] Retain the aluminum foil cleaning parameters and aluminum foil movement rate that offer the greatest cleaning adaptability;

[0148] Continue generating and iteratively optimizing aluminum foil cleaning parameters and aluminum foil moving speed until optimization converges, and obtain the final retained multiple aluminum foil cleaning parameters and aluminum foil moving speed sets to obtain the optimized aluminum foil cleaning parameter set and the optimized aluminum foil moving speed set.

[0149] Furthermore, the optimal parameter acquisition module 13 includes the following execution steps:

[0150] Obtain the optimized cleaning rate set and optimized cleaning fitness set corresponding to the optimized aluminum foil cleaning parameter set and optimized aluminum foil moving speed set;

[0151] Obtain real-time coating parameters during the coating process;

[0152] The real-time coating parameters are combined with each optimized aluminum foil moving speed and optimized cleaning rate input into the coating prediction path to output a set of predicted coating accuracy parameters. The coating prediction path is obtained by training with a set of sample coating parameters, a set of sample aluminum foil moving speeds, a set of sample cleaning rates and a set of sample coating accuracy parameters. Each predicted coating accuracy parameter includes the average similarity between the coating thickness and the standard thickness.

[0153] Obtain the preset coating accuracy parameters;

[0154] The reduction rate of each predicted coating accuracy parameter in the predicted coating accuracy parameter set compared with the preset coating accuracy parameter is calculated to obtain multiple coating quality cost parameters.

[0155] Based on the optimized cleaning fitness set and multiple coating quality cost parameters, a cooperative fitness set is calculated. The optimized aluminum foil cleaning parameters and optimized aluminum foil moving rate corresponding to the maximum cooperative fitness are selected as the optimal cleaning parameters and optimal moving rate.

[0156] Specifically, based on the optimized cleaning fitness set and multiple coating quality cost parameters, a collaborative fitness set is calculated, including:

[0157] The co-fitness set is obtained by subtracting the corresponding coating quality cost parameter from each optimized cleaning fitness within the optimized cleaning fitness set.

[0158] It should be noted that the descriptions of each embodiment in the above embodiments have different focuses. For parts that are not described in detail in a certain embodiment, please refer to the relevant descriptions in other embodiments.

[0159] Those skilled in the art will understand that embodiments of the present invention can provide methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0160] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, as well as combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded computer, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0161] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0162] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0163] Although preferred embodiments of the invention have been described, those skilled in the art, once they have learned the basic inventive concept, can make other changes and modifications to these embodiments.

[0164] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of this invention and its equivalents, this invention also intends to include these modifications and variations.

Claims

1. A collaborative control method for an aluminum-plastic composite belt production line, characterized in that, The method includes: During the cleaning process in the aluminum-plastic composite belt production line, aluminum foil images are collected to identify aluminum foil contamination and obtain aluminum foil contamination characteristics. Based on the aluminum foil contamination characteristics, the aluminum foil cleaning parameters and aluminum foil movement rate are optimized to obtain an optimized set of aluminum foil cleaning parameters and an optimized set of aluminum foil movement rates. The real-time coating parameters in the coating process are obtained, and the cleaning parameters and moving speed are optimized by combining the optimized aluminum foil cleaning parameter set and the optimized aluminum foil moving speed set to obtain the optimal cleaning parameters and the optimal moving speed, wherein the optimization is carried out at the cost of reducing coating quality. The cleaning process and the coating process are controlled digitally in a coordinated manner according to the optimal cleaning parameters and the optimal moving speed. Specifically, based on the characteristics of aluminum foil contamination, aluminum foil cleaning parameters and aluminum foil movement speed are optimized to obtain an optimized set of aluminum foil cleaning parameters and an optimized set of aluminum foil movement speeds, including: Obtain the cleaning parameter space and the moving speed space, and randomly generate multiple first aluminum foil cleaning parameters and multiple first aluminum foil moving speeds; Based on the aluminum foil contamination characteristics, and combined with multiple first aluminum foil cleaning parameters and multiple first aluminum foil movement rates, cleaning prediction is performed to obtain multiple first cleaning rates; Calculate the ratio of multiple first aluminum foil moving speeds to a preset moving speed to obtain multiple first speed coefficients; Multiple first cleaning fitness values ​​are calculated based on multiple first cleaning rates and multiple first speed coefficients; Continue to iteratively optimize the aluminum foil cleaning parameters and aluminum foil moving speed to obtain an optimized set of aluminum foil cleaning parameters and an optimized set of aluminum foil moving speed. Specifically, the process involves acquiring real-time coating parameters during the coating process, and then optimizing the cleaning parameters and movement speed using the optimized aluminum foil cleaning parameter set and optimized aluminum foil moving speed set to obtain the optimal cleaning parameters and optimal movement speed. This includes: Obtain the optimized cleaning rate set and optimized cleaning fitness set corresponding to the optimized aluminum foil cleaning parameter set and optimized aluminum foil moving speed set; Obtain real-time coating parameters during the coating process; By combining real-time coating parameters with each optimized aluminum foil moving speed and optimized cleaning rate input into the coating prediction path, the output obtains a set of predicted coating accuracy parameters. Obtain the preset coating accuracy parameters; The reduction rate of each predicted coating accuracy parameter in the predicted coating accuracy parameter set compared with the preset coating accuracy parameter is calculated to obtain multiple coating quality cost parameters. Based on the optimized cleaning fitness set and multiple coating quality cost parameters, a cooperative fitness set is calculated. The optimized aluminum foil cleaning parameters and optimized aluminum foil moving rate corresponding to the maximum cooperative fitness are selected as the optimal cleaning parameters and optimal moving rate.

2. The collaborative control method for the aluminum-plastic composite strip production line according to claim 1, characterized in that, During the cleaning process in the aluminum-plastic composite belt production line, aluminum foil images are captured to identify aluminum foil contamination and obtain its characteristics, including: Images of aluminum foil are captured during the cleaning process in the aluminum-plastic composite belt production line. The aluminum foil image is input into the aluminum foil contamination identification path, and the identification output obtains the aluminum foil contamination features, wherein the aluminum foil contamination features include the area of ​​aluminum foil contamination.

3. The collaborative control method for the aluminum-plastic composite strip production line according to claim 2, characterized in that, The aluminum foil contamination identification path is constructed using the following steps: Based on historical processing data of the cleaning process, a set of sample aluminum foil images was collected, and the aluminum foil contamination features in each sample aluminum foil image were labeled to obtain a set of sample aluminum foil contamination features. Based on a convolutional neural network, an aluminum foil contamination identification path is constructed by using aluminum foil images and aluminum foil contamination features as input and output data. Using the sample aluminum foil image set and sample aluminum foil contamination feature set, the aluminum foil contamination recognition path is trained iteratively in a supervised manner until the test converges, thus completing the construction and training.

4. The collaborative control method for the aluminum-plastic composite strip production line according to claim 1, characterized in that, Based on the aluminum foil contamination characteristics, and combining multiple first aluminum foil cleaning parameters and multiple first aluminum foil movement rates, cleaning prediction is performed to obtain multiple first cleaning rates, including: Based on historical processing data of the cleaning process, a set of aluminum foil contamination characteristics, a set of sample aluminum foil cleaning parameters, and a set of sample aluminum foil movement rates were obtained. The cleanliness rate under different cleaning conditions was collected and labeled to obtain a set of sample cleanliness rates. Construct a clean prediction path based on machine learning; Using the aforementioned aluminum foil contamination feature set, sample aluminum foil cleaning parameter set, sample aluminum foil movement rate set, and sample cleanliness rate set, the cleaning prediction path is trained iteratively under supervision and configured in the factory control center of the aluminum-plastic composite belt after convergence. The aluminum foil contamination characteristics are combined with each first aluminum foil cleaning parameter and the first aluminum foil movement rate, and then input into the cleaning prediction path to obtain multiple first cleaning rates.

5. The collaborative control method for the aluminum-plastic composite strip production line according to claim 1, characterized in that, Continue iterative optimization of aluminum foil cleaning parameters and aluminum foil movement speed to obtain an optimized set of aluminum foil cleaning parameters and an optimized set of aluminum foil movement speeds, including: Continue to generate second aluminum foil cleaning parameters and multiple second aluminum foil movement rates, and predict the processing to obtain multiple second cleaning fitnesss; Retain the aluminum foil cleaning parameters and aluminum foil movement rate that offer the greatest cleaning adaptability; Continue generating and iteratively optimizing aluminum foil cleaning parameters and aluminum foil moving speed until the optimization converges. Obtain the final set of multiple aluminum foil cleaning parameters and aluminum foil moving speeds, thus obtaining the optimized aluminum foil cleaning parameter set and the optimized aluminum foil moving speed set.

6. The collaborative control method for the aluminum-plastic composite strip production line according to claim 1, characterized in that, The coating prediction path is obtained by training a set of sample coating parameters, a set of sample aluminum foil moving speeds, a set of sample cleaning rates, and a set of sample coating accuracy parameters. Each predicted coating accuracy parameter includes the average similarity between the coating thickness and the standard thickness.

7. The collaborative control method for the aluminum-plastic composite strip production line according to claim 1, characterized in that, Based on the optimized cleaning fitness set and multiple coating quality cost parameters, a cooperative fitness set is calculated, including: The co-fitness set is obtained by subtracting the corresponding coating quality cost parameter from each optimized cleaning fitness within the optimized cleaning fitness set.

8. A collaborative control system for an aluminum-plastic composite strip production line, characterized in that, The system is used to implement the collaborative control method for the aluminum-plastic composite strip production line as described in any one of claims 1-7, the system comprising: The contamination feature recognition module is used to collect aluminum foil images during the cleaning process in the aluminum-plastic composite belt production line, identify aluminum foil contamination, and obtain aluminum foil contamination features. The optimization parameter acquisition module is used to optimize the aluminum foil cleaning parameters and aluminum foil moving speed based on the aluminum foil contamination characteristics, and obtain an optimized aluminum foil cleaning parameter set and an optimized aluminum foil moving speed set; The optimal parameter acquisition module is used to acquire real-time coating parameters in the coating process, and to optimize the cleaning parameters and moving speed by combining the optimized aluminum foil cleaning parameter set and the optimized aluminum foil moving speed set, so as to obtain the optimal cleaning parameters and the optimal moving speed, wherein the optimization is carried out at the cost of reducing coating quality. The optimal parameter application module is used to perform coordinated digital control of the cleaning process and the coating process according to the optimal cleaning parameters and the optimal moving speed.

Images