Intelligent Control System and Method for Manufacturing Cooling Fabrics Based on Data Processing

Abstract

This invention relates to the technical field of manufacturing control, and discloses an intelligent control system and method for manufacturing cooling fabrics based on data processing. The invention first acquires an image of the cooling fabric surface, performs image preprocessing to obtain an initial set of cooling fabric manufacturing parameters, and then performs data cleaning to obtain a further set of cooling fabric manufacturing parameters. Next, through deviation analysis, the cooling fabric manufacturing parameters are adjusted to achieve real-time control of cooling fabric manufacturing. Then, an RBF neural network is trained to obtain an RBF neural network model, which predicts the cooling sensation of the cooling fabric and outputs a predicted cooling value. Finally, a gazelle optimization algorithm is used to optimize and adjust the cooling fabric yarn spacing, cooling fabric density, and yarn twist in the cooling fabric manufacturing parameter set, achieving intelligent control of cooling fabric manufacturing. This invention achieves intelligent control of cooling fabric manufacturing through data processing of cooling fabric manufacturing parameters, and the method is objective and accurate.

Description

Technical Field

[0001] This invention relates to the field of manufacturing control technology, specifically to an intelligent control system and method for manufacturing cool-feeling fabrics based on data processing. Background Technology

[0002] Traditional methods for manufacturing cooling fabrics often rely on production lines, which cannot flexibly allocate resources and are limited in terms of energy consumption and space utilization. They also cannot intelligently control production parameters such as yarn spacing and fabric density during fabric manufacturing. Furthermore, they do not utilize advanced technologies such as neural networks and artificial intelligence, which brings inconvenience to fabric manufacturing. Summary of the Invention

[0003] To address the problems in related technologies, this invention provides an intelligent control system and method for manufacturing cooling fabrics based on data processing, thereby overcoming the aforementioned technical problems in existing related technologies.

[0004] To solve the above-mentioned technical problems, the present invention is achieved through the following technical solution:

[0005] This invention relates to an intelligent control method for manufacturing cooling fabrics based on data processing, comprising the following steps:

[0006] S1. Obtain a surface image of the cooling fabric, preprocess the surface image of the cooling fabric, calculate the yarn spacing and density of the cooling fabric, obtain the manufacturing process parameters of the cooling fabric, form an initial set of manufacturing parameters for the cooling fabric, perform abnormal data cleaning on the initial set of manufacturing parameters for the cooling fabric, and obtain the final set of manufacturing parameters for the cooling fabric.

[0007] S2. Compare the deviation between the set of manufacturing parameters for the cooling fabric and the set of target parameters for the manufacturing of the cooling fabric, adjust the manufacturing parameters for the cooling fabric, and achieve real-time control of the manufacturing of the cooling fabric.

[0008] S3. After dimensionality reduction of the manufacturing parameters of the cooling fabric in the set of manufacturing parameters of the cooling fabric, train the RBF neural network to obtain the RBF neural network model, output the cooling prediction value, and complete the cooling prediction of the cooling fabric.

[0009] S4. Based on the predicted cooling value, the Gazelle Optimization Algorithm is used to continuously optimize and adjust the yarn spacing, density, and twist of the cooling fabric manufacturing parameter set to achieve the optimal cooling value and realize intelligent control of cooling fabric manufacturing.

[0010] This invention acquires an image of the cooling fabric surface, performs grayscale processing, then Wiener filtering on the grayscale image, followed by decomposition and reconstruction, and finally smoothing to complete image preprocessing. This method reduces the computational load and improves processing speed, effectively reducing image noise and simplifying complex images into computable binary images. Density clustering is then used to clean outlier data in the fabric manufacturing parameters, eliminating interference from abnormal data and facilitating subsequent neural network processing. Furthermore, the manufacturing parameters and target parameters of the cooling fabric are compared for deviation, and real-time control is implemented for any deviations. Finally, an RBF neural network is trained to obtain the RBF... The neural network model performs dimensionality reduction on the manufacturing parameters of the cooling fabric before inputting them into the RBF neural network model, outputting a predicted cooling value. Dimensionality reduction effectively preserves the original data features, removes redundant parts, and accelerates neural network processing. RBF neural networks have fast learning speeds and are widely used in large-sample nonlinear problems. Finally, the yarn spacing, density, and twist of the cooling fabric manufacturing parameters are extracted and optimized using a gazelle optimization algorithm. By finding the optimal fitness function value, the optimal cooling value is achieved, realizing intelligent control of cooling fabric manufacturing. This algorithm has strong global search capabilities and can effectively solve situations where the cooling effect does not reach its optimal level during the manufacturing process.

[0011] Preferably, step S1 includes the following steps:

[0012] S11. During the manufacturing process of the cooling fabric, an initial image of the cooling fabric surface is acquired, forming an initial cooling fabric surface image set. Sample points are selected from the initial cooling fabric surface images, quantized, and treated as pixels. The initial cooling fabric surface images are then transformed into a cooling fabric surface image matrix A as follows:

[0013]

[0014] Among them, b mn This represents the m-th pixel horizontally and the n-th pixel vertically within the image matrix of the cool-feeling fabric surface.

[0015] Based on the cooling fabric surface image matrix, the initial cooling fabric surface images in the initial cooling fabric surface image set are preprocessed to obtain a processed cooling fabric surface image set. The specific steps are as follows:

[0016] S111. Set the signal intensity values ​​of the red channel, green channel, and blue channel of the initial cooling fabric surface image, and assign weights to them respectively. Perform a weighted average of the three channel signal intensity values ​​of the initial cooling fabric surface image according to the weights to obtain a grayscale cooling fabric surface image.

[0017] S112. Select a local neighborhood of size α×α in the image matrix of the cooling fabric surface, and calculate the mean and variance of the gray values ​​of the pixels in the local neighborhood, denoted as a1 and a2 respectively; select any pixel with a gray value of a3 in the local neighborhood, and set the noise variance of the grayscale cooling fabric surface image to a4. Then, use a Wiener filter for filtering, and the calculation formula is as follows:

[0018]

[0019] Where a′ represents the grayscale value of the pixel after filtering;

[0020] Replace the gray values ​​of local pixels with the gray values ​​of the filtered pixels until all pixels in the cool fabric surface image matrix are replaced, and the filtered cool fabric surface image matrix is ​​obtained, thus generating the filtered cool fabric surface image.

[0021] S113. Perform wavelet transform on the filtered cool fabric surface image, set the wavelet basis function to F′, the decomposition level to b′, the decomposition coefficient of each level to b″, and the length of each decomposition coefficient to b″′, perform b′ level decomposition and reconstruction on the filtered cool fabric surface image to obtain the decomposed and reconstructed cool fabric surface image.

[0022] S114. The decomposed and reconstructed surface image of the cooling fabric is binarized to obtain a binarized surface image of the cooling fabric. The number of pixels with grayscale values ​​of 1 and 0 in each row of the binarized surface image of the cooling fabric is counted, and recorded as the number of pixels in the first row and the number of pixels in the second row, respectively. If the number of pixels in the first row is greater than the total number of pixels in each row of the binarized surface image of the cooling fabric, the pixels in each row of the binarized surface image of the cooling fabric are set to 1; if the number of pixels in the second row is greater than the total number of pixels in each row of the binarized surface image of the cooling fabric, the pixels in each row of the binarized surface image of the cooling fabric are set to 0. Simultaneously, the number of pixels with grayscale values ​​of 1 and 0 in each column of the binarized cooling fabric surface image is counted, and recorded as the number of pixels in the first column and the number of pixels in the second column, respectively. If the number of pixels in the first column is greater than the total number of pixels in each column of the binarized cooling fabric surface image, the number of pixels in each column of the binarized cooling fabric surface image is set to 1; if the number of pixels in the second column is greater than the total number of pixels in each column of the binarized cooling fabric surface image, the number of pixels in each column of the binarized cooling fabric surface image is set to 0. The image smoothing process is then completed to obtain the processed cooling fabric surface image, forming a set of processed cooling fabric surface images.

[0023] S12. Set the width and length of the processed cool-feeling fabric surface images in the processed cool-feeling fabric surface image set to c′ and c″ respectively, the number of yarns in the horizontal direction to c1, the number of yarns in the vertical direction to c2, and the number of pixels per unit area to c3. Then the formulas for calculating the yarn spacing and density of the cool-feeling fabric are as follows:

[0024]

[0025] Wherein, B1 represents the yarn spacing of the cooling fabric, and B2 represents the density of the cooling fabric;

[0026] Parameters from the manufacturing process of the cooling fabric are obtained, including yarn twist, yarn strength, etc., and combined with the yarn spacing and density of the cooling fabric to form an initial set of manufacturing parameters for the cooling fabric, C = {C1, C2, C3, ..., C...}. d}, where C d This represents the d-th cooling fabric manufacturing parameter, and each cooling fabric manufacturing parameter includes e cooling fabric manufacturing parameter data.

[0027] S13. Density-based clustering is used to clean the outlier data of the cooling fabric manufacturing parameters to obtain a set of cooling fabric manufacturing parameters. The specific steps are as follows:

[0028] S131, regarding the first Given e cooling fabric manufacturing parameter data under a given cooling fabric manufacturing parameter, and setting the neighborhood radius to β, select any data in the cooling fabric manufacturing parameter data and record it as a cluster data point. Using the cluster data point as the core, mark all density-connected data points within the neighborhood radius. Select any density-connected data point as the core, traverse all data points within the neighborhood radius, and mark all density-connected data points within the neighborhood radius.

[0029] S132. Repeat S131 until there are no new density-connected data points as the core. Treat the unmarked cooling fabric manufacturing parameter data as outliers and delete the outliers to obtain the cooling fabric manufacturing parameter set.

[0030] This invention reduces the computational load and improves the computational speed by acquiring images of the cool-feeling fabric surface and performing grayscale processing; Wiener filtering is used to effectively reduce image noise; the filtered cool-feeling fabric surface image is then decomposed and reconstructed, and smoothing is applied to simplify the complex image into a computable binary image; density clustering is used to clean abnormal data in the fabric manufacturing parameter data, thus eliminating abnormal data interference.

[0031] Preferably, step S2 includes the following steps:

[0032] S21. Define the target parameter set for manufacturing the cooling fabric, and the deviation threshold set is D = {ω1, ω2, ω3, ..., ω...} d}, where ω d This represents the d-th deviation threshold. The set of manufacturing parameters for the cooling fabric and the set of target parameters for manufacturing the cooling fabric are compared, and the d-th value in the set of manufacturing parameters for the cooling fabric is selected. The manufacturing parameters for cooling fabrics are selected from the set of target parameters for cooling fabric manufacturing. There are target parameters for manufacturing cooling fabrics, and there exists a first... The absolute value of the difference between the manufacturing parameter data corresponding to the first cooling fabric manufacturing parameter and the manufacturing target parameter data corresponding to the dth to dth cooling fabric manufacturing target parameters is greater than the first Then the first deviation threshold, then the second If there is a deviation in the manufacturing parameters of a cooling fabric, the manufacturing process of the cooling fabric should be adjusted.

[0033] S22. Calculate the relationship between climate factors such as temperature and humidity and the yarn spacing and density of the cooling fabric. As temperature and humidity rise and fall, automatically control the yarn spacing and density of the cooling fabric to achieve real-time control of the manufacturing of the cooling fabric.

[0034] Preferably, step S3 includes the following steps:

[0035] S31. For the cooling fabric manufacturing parameters in the set of cooling fabric manufacturing parameters, set the average value and standard deviation of the cooling fabric manufacturing parameter data corresponding to the d1th cooling fabric manufacturing parameter as e1 and f1 respectively, and set the average value and standard deviation of the cooling fabric manufacturing parameter data corresponding to the d2th cooling fabric manufacturing parameter as e2 and f2 respectively, and set the number of cooling fabric manufacturing parameter data as follows: The formula for calculating the correlation coefficient is as follows:

[0036]

[0037] Where δ represents the correlation coefficient, and δ∈[-1,1], D g This represents the manufacturing parameter data for the g-th cooling fabric.

[0038] Obtain the correlation coefficients of the manufacturing parameters for cooling fabrics from the set of manufacturing parameters for cooling fabrics. Sort the correlation coefficients in descending order and select the top ones. The dimension is used as the parameter feature after dimensionality reduction, resulting in the set of parameter features after dimensionality reduction;

[0039] S32. Obtain a new image of the cooling fabric surface. After preprocessing and normalization, extract a set of parametric feature samples. Use the set of parametric feature samples to train an RBF (Radial Basis Function) neural network to obtain the RBF neural network model. The specific steps are as follows:

[0040] S321. Set the mapping function of the RBF neural network to a Gaussian kernel function, the radial basis function to F″, the number of input layers to ε1, the number of neurons to ε2, and the number of output layers to 1; divide the parameter feature sample set into a parameter feature sample training set and a parameter feature sample test set; input the parameter feature sample training set into the RBF neural network and iterate continuously until the RBF neural network converges to obtain a trained RBF neural network;

[0041] S322. Input the parameter feature sample test set into the trained RBF neural network, set the accuracy threshold to ξ, and obtain the RBF neural network model when the accuracy of the output result is less than the accuracy threshold; otherwise, adjust the weights until the accuracy of the output result is less than the accuracy threshold.

[0042] S33. Normalize the reduced parameter feature set to obtain a normalized parameter feature set. Input the normalized parameter feature set into the RBF neural network model and output the cooling prediction value. The cooling prediction value is between (0, φ1). When the cooling prediction value is between (0, φ1), it indicates no cooling sensation. When the cooling prediction value is between (φ1, φ2), it indicates moderate cooling sensation. When the cooling prediction value is between (φ2, φ3), it indicates obvious cooling sensation. When the cooling prediction value is between (φ3, φ4), it indicates strong cooling sensation. When the cooling prediction value is between (φ4, φ1), it indicates strong cooling sensation. The cooling sensation prediction of the cooling fabric is completed.

[0043] This invention obtains an RBF neural network model by training an RBF neural network. RBF neural networks have a fast learning speed and are suitable for large-sample complex problems. After dimensionality reduction processing of the manufacturing parameters of the cooling fabric, the original data features are effectively preserved, redundant parts in the data are deleted, the processing speed of the neural network is accelerated, and the cooling sensation of the cooling fabric is predicted.

[0044] Preferably, step S4 includes the following steps:

[0045] S41. Select yarn spacing, density, and twist as adjustment parameters from the set of manufacturing parameters for the cooling fabric. Assign different weights to each of the yarn spacing, density, and twist to calculate the cooling effect of the cooling fabric, thereby obtaining a cooling effect function. Use this cooling effect function as a fitness function. Set the optimal cooling effect value as... When the predicted cooling value cannot reach the optimal cooling value, the Gazelle optimization algorithm is used to adjust and optimize the yarn spacing, density, and twist of the cooling fabric to achieve the optimal cooling value. The specific steps are as follows:

[0046] S411. In the gazelle population initialization phase, the number of gazelles in the population is set to i, and the dimension of each gazelle is j. The initial population matrix A1 is generated as follows:

[0047]

[0048] Among them, h ij This represents the position of the i-th gazelle individual in the j-th dimension;

[0049] Let p1 represent a random number and p1∈[0,1]. Let l1 be the upper bound of the dimension of the gazelle population and l2 be the lower bound. Then the position of the i′ gazelle in the gazelle population is E(i′)=p1·(l1-l2)+l2;

[0050] The positions of individual gazelles in the gazelle population are iterated, and the gazelles whose positions correspond to the best fitness function values ​​obtained in each iteration are denoted as elite gazelles, thus obtaining an elite gazelle matrix. The process of updating the positions of individual gazelles is the process of adjusting and optimizing the yarn spacing, density, and twist of the cooling fabric.

[0051] S412. In the gazelle population development phase, let the current iteration number be k, and the elite gazelle individual matrix in the k-th iteration be A. k Let E(i′,k) be the position of the i′-th gazelle in the k-th iteration. Assume that the gazelle moves by Brownian motion while feeding, and its speed is q. Let p2 and p3 represent random number vectors of Brownian motion, and p3∈[0,1]. Then, the formula for calculating the position E(i′,k+1) of the i′-th gazelle in the (k+1)-th iteration is as follows:

[0052] E(i′,k+1)=E(i′,k)+p2·p3·q(A k -p2·E(i′,k));

[0053] By comparing the fitness function values ​​after each iteration, the optimal fitness function value is found, and the elite gazelle individual matrix is ​​updated.

[0054] S413. During the gazelle population exploration phase, Levy flight is introduced. When the number of iterations is odd, individual gazelles in the population move in one direction; when the number of iterations is even, they move in the other direction. The movement follows Brownian motion before Levy flight. Let p4 be the random number vector for Levy flight, Q be the fastest speed of an individual gazelle, and p5 represent a constant. When a gazelle spots a predator, the position E(i′, k+1) of the i′ gazelle in the (k+1)th iteration is updated using the following formula:

[0055] E(i′,k+1)=E(i′,k)+p3·p4·p5·Q(A k -p4·E(i′,k));

[0056] The maximum number of iterations is set to K, and the predator begins chasing the gazelle. The chasing coefficient is... At this point, individual gazelles in the population begin to move. The position E(i′,k+1) of the i′ gazelle in the (k+1)th iteration is updated again, and the calculation formula is as follows:

[0057] E(i′,k+1)=E(i′,k)+p2·p4·p5·Q·γ(A k -p4·E(i′,k));

[0058] S414. Set the predation success rate. Individual gazelles in the gazelle population move according to the predation success rate. When the current iteration count reaches the maximum iteration count, stop the iteration and obtain the final gazelle position. The three-dimensional coordinate values ​​of the final gazelle position are the adjusted and optimized yarn spacing, density, and yarn twist of the cooling fabric, respectively. At this time, the optimal cooling value is achieved.

[0059] S42. Using the adjusted and optimized yarn spacing, density, and twist of the cooling fabric as the optimal manufacturing parameters for the cooling fabric, a set of manufacturing parameters for the cooling fabric is obtained, thereby realizing intelligent control of the manufacturing of the cooling fabric.

[0060] This invention extracts the yarn spacing, density, and twist of cooling fabric from the set of manufacturing parameters for cooling fabrics, and uses the Gazelle Optimization Algorithm to optimize and adjust them to achieve the optimal cooling value, thus realizing intelligent control of cooling fabric manufacturing. The algorithm has strong global search capabilities and can effectively solve the problem of the cooling sensation not reaching the optimal level during the manufacturing process of cooling fabrics.

[0061] This embodiment also discloses a system for intelligent control of cooling fabric manufacturing based on data processing, specifically including: a cooling fabric manufacturing parameter processing module, a cooling fabric manufacturing real-time control module, a cooling fabric cooling prediction module, and a cooling fabric manufacturing parameter adjustment and optimization module.

[0062] The cooling fabric manufacturing parameter processing module is used to preprocess the surface image of the cooling fabric to obtain the manufacturing parameters of the cooling fabric.

[0063] The real-time control module for manufacturing cooling fabric is used to compare the deviation between the manufacturing parameters and the target parameters for manufacturing cooling fabric, and to control the manufacturing of cooling fabric in real time.

[0064] The cooling fabric cooling prediction module is used to predict the cooling sensation of cooling fabrics using an RBF neural network model.

[0065] The cooling fabric manufacturing parameter adjustment and optimization module is used to optimize and adjust the cooling fabric manufacturing parameters using the Gazelle Optimization Algorithm.

[0066] The present invention has the following beneficial effects:

[0067] 1. This invention reduces the computational load and improves the computational speed by acquiring an image of the surface of the cooling fabric and then converting it to grayscale; Wiener filtering is used to filter the grayscale image of the cooling fabric surface, which effectively reduces image noise.

[0068] 2. This invention decomposes and reconstructs the filtered image of the cool-feeling fabric surface, uses wavelet decomposition to obtain detail components, which facilitates subsequent processing, and then simplifies the complex image into a computable binary image through smoothing.

[0069] 3. This invention uses density clustering to clean abnormal data in the fabric manufacturing parameter data, identify and remove outliers, eliminate abnormal data interference, and facilitate subsequent processing.

[0070] 4. This invention obtains an RBF neural network model by training an RBF neural network. The RBF neural network has a fast learning speed and is suitable for large-sample complex problems. After the manufacturing parameters of the cooling fabric are reduced in dimension, the original data features are effectively preserved, redundant parts in the data are deleted, and the processing speed of the neural network is accelerated.

[0071] 5. This invention optimizes and adjusts the yarn spacing, density, and twist of the cooling fabric manufacturing parameters by using the Gazelle Optimization Algorithm. The algorithm has strong global search capabilities and can effectively solve the problem of the cooling sensation not reaching the optimal level during the manufacturing process of cooling fabric.

[0072] Of course, any product implementing this invention does not necessarily need to achieve all of the advantages described above at the same time. Attached Figure Description

[0073] To more clearly illustrate the technical solutions of the embodiments of the invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the invention. For those skilled in the art, the drawings can be obtained from these drawings without creative effort.

[0074] Figure 1 This is a schematic diagram illustrating the process of intelligent control of cool-feel fabric manufacturing based on data processing provided by the present invention. Detailed Implementation

[0075] 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.

[0076] In the description of this invention, it should be understood that the terms "opening", "upper", "lower", "top", "middle", "inner", etc., which indicate orientation or positional relationship, are only for the convenience of describing the invention and simplifying the description, and do not indicate or imply that the components or elements referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation on the invention.

[0077] Example 1

[0078] Please see Figure 1 This implementation discloses an intelligent control method for manufacturing cooling fabrics based on data processing, specifically including the following:

[0079] S1. Obtain a surface image of the cooling fabric, preprocess the surface image of the cooling fabric, calculate the yarn spacing and density of the cooling fabric, obtain the manufacturing process parameters of the cooling fabric, form an initial set of manufacturing parameters for the cooling fabric, perform abnormal data cleaning on the initial set of manufacturing parameters for the cooling fabric, and obtain the final set of manufacturing parameters for the cooling fabric.

[0080] S1 includes the following steps:

[0081] S11. During the manufacturing process of the cooling fabric, an initial image of the cooling fabric surface is acquired, forming an initial cooling fabric surface image set. Sample points are selected from the initial cooling fabric surface images, quantized, and treated as pixels. The initial cooling fabric surface images are then transformed into a cooling fabric surface image matrix A as follows:

[0082]

[0083] Among them, b mn This represents the m-th pixel horizontally and the n-th pixel vertically within the image matrix of the cool-feeling fabric surface.

[0084] Based on the cooling fabric surface image matrix, the initial cooling fabric surface images in the initial cooling fabric surface image set are preprocessed to obtain a processed cooling fabric surface image set. The specific steps are as follows:

[0085] S111. Set the signal intensity values ​​of the red channel, green channel, and blue channel of the initial cooling fabric surface image, and assign weights to them respectively. Perform a weighted average of the three channel signal intensity values ​​of the initial cooling fabric surface image according to the weights to obtain a grayscale cooling fabric surface image.

[0086] S112. Select a local neighborhood of size α×α in the image matrix of the cooling fabric surface, and calculate the mean and variance of the gray values ​​of the pixels in the local neighborhood, denoted as a1 and a2 respectively; select any pixel with a gray value of a3 in the local neighborhood, and set the noise variance of the grayscale cooling fabric surface image to a4. Then, use a Wiener filter for filtering, and the calculation formula is as follows:

[0087]

[0088] Where a′ represents the grayscale value of the pixel after filtering;

[0089] Replace the gray values ​​of local pixels with the gray values ​​of the filtered pixels until all pixels in the cool fabric surface image matrix are replaced, and the filtered cool fabric surface image matrix is ​​obtained, thus generating the filtered cool fabric surface image.

[0090] S113. Perform wavelet transform on the filtered cool fabric surface image, set the wavelet basis function to F′, the decomposition level to b′, the decomposition coefficient of each level to b″, and the length of each decomposition coefficient to b″′, perform b′ level decomposition and reconstruction on the filtered cool fabric surface image to obtain the decomposed and reconstructed cool fabric surface image.

[0091] S114. The decomposed and reconstructed surface image of the cooling fabric is binarized to obtain a binarized surface image of the cooling fabric. The number of pixels with grayscale values ​​of 1 and 0 in each row of the binarized surface image of the cooling fabric is counted, and recorded as the number of pixels in the first row and the number of pixels in the second row, respectively. If the number of pixels in the first row is greater than the total number of pixels in each row of the binarized surface image of the cooling fabric, the pixels in each row of the binarized surface image of the cooling fabric are set to 1; if the number of pixels in the second row is greater than the total number of pixels in each row of the binarized surface image of the cooling fabric, the pixels in each row of the binarized surface image of the cooling fabric are set to 0. Simultaneously, the number of pixels with grayscale values ​​of 1 and 0 in each column of the binarized cooling fabric surface image is counted, and recorded as the number of pixels in the first column and the number of pixels in the second column, respectively. If the number of pixels in the first column is greater than the total number of pixels in each column of the binarized cooling fabric surface image, the number of pixels in each column of the binarized cooling fabric surface image is set to 1; if the number of pixels in the second column is greater than the total number of pixels in each column of the binarized cooling fabric surface image, the number of pixels in each column of the binarized cooling fabric surface image is set to 0. The image smoothing process is then completed to obtain the processed cooling fabric surface image, forming a set of processed cooling fabric surface images.

[0092] S12. Set the width and length of the processed cool-feeling fabric surface images in the processed cool-feeling fabric surface image set to c′ and c″ respectively, the number of yarns in the horizontal direction to c1, the number of yarns in the vertical direction to c2, and the number of pixels per unit area to c3. Then the formulas for calculating the yarn spacing and density of the cool-feeling fabric are as follows:

[0093]

[0094] Wherein, B1 represents the yarn spacing of the cooling fabric, and B2 represents the density of the cooling fabric;

[0095] Parameters from the manufacturing process of the cooling fabric are obtained, including yarn twist, yarn strength, etc., and combined with the yarn spacing and density of the cooling fabric to form an initial set of manufacturing parameters for the cooling fabric, C = {C1, C2, C3, ..., C...}. d}, where C d This represents the d-th cooling fabric manufacturing parameter, and each cooling fabric manufacturing parameter includes e cooling fabric manufacturing parameter data.

[0096] S13. Density-based clustering is used to clean the outlier data of the cooling fabric manufacturing parameters to obtain a set of cooling fabric manufacturing parameters. The specific steps are as follows:

[0097] S131, regarding the first Given e cooling fabric manufacturing parameter data under a given cooling fabric manufacturing parameter, and setting the neighborhood radius to β, select any data in the cooling fabric manufacturing parameter data and record it as a cluster data point. Using the cluster data point as the core, mark all density-connected data points within the neighborhood radius. Select any density-connected data point as the core, traverse all data points within the neighborhood radius, and mark all density-connected data points within the neighborhood radius.

[0098] S132. Repeat S131 until there are no new density-connected data points as the core. Treat the unmarked cooling fabric manufacturing parameter data as outliers and delete the outliers to obtain the cooling fabric manufacturing parameter set.

[0099] S2. Compare the deviation between the set of manufacturing parameters for the cooling fabric and the set of target parameters for the manufacturing of the cooling fabric, adjust the manufacturing parameters for the cooling fabric, and achieve real-time control of the manufacturing of the cooling fabric.

[0100] S2 includes the following steps:

[0101] S21. Define the target parameter set for manufacturing the cooling fabric, and the deviation threshold set is D = {ω1, ω2, ω3, ..., ω...} d}, where ω d This represents the d-th deviation threshold. The set of manufacturing parameters for the cooling fabric and the set of target parameters for manufacturing the cooling fabric are compared, and the d-th value in the set of manufacturing parameters for the cooling fabric is selected. The manufacturing parameters for cooling fabrics are selected from the set of target parameters for cooling fabric manufacturing. There are target parameters for manufacturing cooling fabrics, and there exists a first... The manufacturing parameters of the cooling fabric correspond to the manufacturing parameter data of the cooling fabric and the first... The absolute value of the difference between the target parameters for manufacturing cooling fabrics for each target parameter is greater than that for the first target parameter. Then the first deviation threshold, then the second If there is a deviation in the manufacturing parameters of a cooling fabric, the manufacturing process of the cooling fabric should be adjusted.

[0102] S22. Calculate the relationship between climate factors such as temperature and humidity and the yarn spacing and density of the cooling fabric. As temperature and humidity rise and fall, automatically control the yarn spacing and density of the cooling fabric to achieve real-time control of the manufacturing of the cooling fabric.

[0103] S3. After dimensionality reduction of the manufacturing parameters of the cooling fabric in the set of manufacturing parameters of the cooling fabric, train the RBF neural network to obtain the RBF neural network model, output the cooling prediction value, and complete the cooling prediction of the cooling fabric.

[0104] S3 includes the following steps:

[0105] S31. For the cooling fabric manufacturing parameters in the set of cooling fabric manufacturing parameters, set the average value and standard deviation of the cooling fabric manufacturing parameter data corresponding to the d1th cooling fabric manufacturing parameter as e1 and f1 respectively, and set the average value and standard deviation of the cooling fabric manufacturing parameter data corresponding to the d2th cooling fabric manufacturing parameter as e2 and f2 respectively, and set the number of cooling fabric manufacturing parameter data as follows: The formula for calculating the correlation coefficient is as follows:

[0106]

[0107] Where δ represents the correlation coefficient, and δ∈[-1,1], D g This represents the manufacturing parameter data for the g-th cooling fabric.

[0108] Obtain the correlation coefficients of the manufacturing parameters for cooling fabrics from the set of manufacturing parameters for cooling fabrics. Sort the correlation coefficients in descending order and select the top ones. The dimension is used as the parameter feature after dimensionality reduction, resulting in the set of parameter features after dimensionality reduction;

[0109] S32. Obtain a new image of the cool-feeling fabric surface. After preprocessing and normalization, extract a set of parametric feature samples. Use the set of parametric feature samples to train an RBF neural network to obtain the RBF neural network model. The specific steps are as follows:

[0110] S321. Set the mapping function of the RBF neural network to a Gaussian kernel function, the radial basis function to F″, the number of input layers to ε1, the number of neurons to ε2, and the number of output layers to 1; divide the parameter feature sample set into a parameter feature sample training set and a parameter feature sample test set; input the parameter feature sample training set into the RBF neural network and iterate continuously until the RBF neural network converges to obtain a trained RBF neural network;

[0111] S322. Input the parameter feature sample test set into the trained RBF neural network, set the accuracy threshold to ξ, and obtain the RBF neural network model when the accuracy of the output result is less than the accuracy threshold; otherwise, adjust the weights until the accuracy of the output result is less than the accuracy threshold.

[0112] S33. Normalize the reduced parameter feature set to obtain a normalized parameter feature set. Input the normalized parameter feature set into the RBF neural network model and output the cooling prediction value. The cooling prediction value is between (0, φ1). When the cooling prediction value is between (0, φ1), it indicates no cooling sensation. When the cooling prediction value is between (φ1, φ2), it indicates moderate cooling sensation. When the cooling prediction value is between (φ2, φ3), it indicates obvious cooling sensation. When the cooling prediction value is between (φ3, φ4), it indicates strong cooling sensation. When the cooling prediction value is between (φ4, φ1), it indicates strong cooling sensation. Complete the cooling sensation prediction of the cooling fabric.

[0113] S4. Based on the predicted cooling value, the Gazelle optimization algorithm is used to continuously optimize and adjust the yarn spacing, density and twist of the cooling fabric manufacturing parameter set to achieve the optimal cooling value and realize intelligent control of cooling fabric manufacturing.

[0114] S4 includes the following steps:

[0115] S41. Select yarn spacing, density, and twist as adjustment parameters from the set of manufacturing parameters for the cooling fabric. Assign different weights to each of the yarn spacing, density, and twist to calculate the cooling effect of the cooling fabric, thereby obtaining a cooling effect function. Use this cooling effect function as a fitness function. Set the optimal cooling effect value as... When the predicted cooling value cannot reach the optimal cooling value, the Gazelle optimization algorithm is used to adjust and optimize the yarn spacing, density, and twist of the cooling fabric to achieve the optimal cooling value. The specific steps are as follows:

[0116] S411. In the gazelle population initialization phase, the number of gazelles in the population is set to i, and the dimension of each gazelle is j. The initial population matrix A1 is generated as follows:

[0117]

[0118] Among them, h ij This represents the position of the i-th gazelle individual in the j-th dimension;

[0119] Let p1 represent a random number and p1∈[0,1]. Let l1 be the upper bound of the dimension of the gazelle population and l2 be the lower bound. Then the position of the i′ gazelle in the gazelle population is E(i′)=p1·(l1-l2)+l2;

[0120] The positions of individual gazelles in the gazelle population are iterated, and the gazelles whose positions correspond to the best fitness function values ​​obtained in each iteration are denoted as elite gazelles, thus obtaining an elite gazelle matrix. The process of updating the positions of individual gazelles is the process of adjusting and optimizing the yarn spacing, density, and twist of the cooling fabric.

[0121] S412. In the gazelle population development phase, let the current iteration number be k, and the elite gazelle individual matrix in the k-th iteration be A. k Let E(i′,k) be the position of the i′-th gazelle in the k-th iteration. Assume that the gazelle moves by Brownian motion while feeding, and its speed is q. Let p2 and p3 represent random number vectors of Brownian motion, and p3∈[0,1]. Then, the formula for calculating the position E(i′,k+1) of the i′-th gazelle in the (k+1)-th iteration is as follows:

[0122] E(i′,k+1)=E(i′,k)+p2·p3·q(A k -p2·E(i′,k));

[0123] By comparing the fitness function values ​​after each iteration, the optimal fitness function value is found, and the elite gazelle individual matrix is ​​updated.

[0124] S413. During the gazelle population exploration phase, Levy flight is introduced. When the number of iterations is odd, individual gazelles in the population move in one direction; when the number of iterations is even, they move in the other direction. The movement follows Brownian motion before Levy flight. Let p4 be the random number vector for Levy flight, Q be the fastest speed of an individual gazelle, and p5 represent a constant. When a gazelle spots a predator, the position E(i′, k+1) of the i′ gazelle in the (k+1)th iteration is updated using the following formula:

[0125] E(i′,k+1)=E(i′,k)+p3·p4·p5·Q(A k -p4·E(i′,k));

[0126] The maximum number of iterations is set to K, and the predator begins chasing the gazelle. The chasing coefficient is... At this point, individual gazelles in the population begin to move. The position E(i′,k+1) of the i′ gazelle in the (k+1)th iteration is updated again, and the calculation formula is as follows:

[0127] E(i′,k+1)=E(i′,k)+p2·p4·p5·Q·γ(A k -p4·E(i′,k));

[0128] S414. Set the predation success rate. Individual gazelles in the gazelle population move according to the predation success rate. When the current iteration count reaches the maximum iteration count, stop the iteration and obtain the final gazelle position. The three-dimensional coordinate values ​​of the final gazelle position are the adjusted and optimized yarn spacing, density, and yarn twist of the cooling fabric, respectively. At this time, the optimal cooling value is achieved.

[0129] S42. Using the adjusted and optimized yarn spacing, density, and twist of the cooling fabric as the optimal manufacturing parameters for the cooling fabric, a set of manufacturing parameters for the cooling fabric is obtained, thereby realizing intelligent control of the manufacturing of the cooling fabric.

[0130] Example 2

[0131] This embodiment also discloses a system for intelligent control of cooling fabric manufacturing based on data processing, specifically including: a cooling fabric manufacturing parameter processing module, a cooling fabric manufacturing real-time control module, a cooling fabric cooling prediction module, and a cooling fabric manufacturing parameter adjustment and optimization module.

[0132] The cooling fabric manufacturing parameter processing module is used to preprocess the surface image of the cooling fabric to obtain the manufacturing parameters of the cooling fabric.

[0133] The real-time control module for manufacturing cooling fabric is used to compare the deviation between the manufacturing parameters and the target parameters for manufacturing cooling fabric, and to control the manufacturing of cooling fabric in real time.

[0134] The cooling fabric cooling prediction module is used to predict the cooling sensation of cooling fabrics using an RBF neural network model.

[0135] The cooling fabric manufacturing parameter adjustment and optimization module is used to optimize and adjust the cooling fabric manufacturing parameters using the Gazelle Optimization Algorithm.

[0136] In the description of this specification, references to terms such as "an embodiment," "example," "specific example," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0137] The preferred embodiments of the invention disclosed above are merely illustrative of the invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the invention to the specific implementations described. Clearly, many modifications and variations can be made based on the content of this specification. This specification selects and specifically describes these embodiments to better explain the principles and practical applications of the invention, thereby enabling those skilled in the art to better understand and utilize the invention.

Claims

1. A smart control method for manufacturing cooling fabrics based on data processing, characterized in that, Includes the following steps: S1. Obtain a surface image of the cooling fabric, preprocess the surface image of the cooling fabric, calculate the yarn spacing and density of the cooling fabric, obtain the manufacturing process parameters of the cooling fabric, form an initial set of manufacturing parameters for the cooling fabric, perform abnormal data cleaning on the initial set of manufacturing parameters for the cooling fabric, and obtain the final set of manufacturing parameters for the cooling fabric. S2. Compare the deviation between the set of manufacturing parameters for the cooling fabric and the set of target parameters for the manufacturing of the cooling fabric, adjust the manufacturing parameters for the cooling fabric, and achieve real-time control of the manufacturing of the cooling fabric. S3. After dimensionality reduction of the manufacturing parameters of the cooling fabric in the set of manufacturing parameters of the cooling fabric, train the RBF neural network to obtain the RBF neural network model, output the cooling prediction value, and complete the cooling prediction of the cooling fabric. S4. Based on the predicted cooling value, optimize and adjust the yarn spacing, density, and twist of the cooling fabric manufacturing parameter set to achieve the optimal cooling value and realize intelligent control of cooling fabric manufacturing. S41. Select yarn spacing, density, and twist as adjustment parameters from the set of manufacturing parameters for the cooling fabric. Assign different weights to each of the yarn spacing, density, and twist to calculate the cooling effect of the cooling fabric, thereby obtaining a cooling effect function. Use this cooling effect function as a fitness function. Set the optimal cooling effect value as... When the predicted cooling value cannot reach the optimal cooling value, the Gazelle optimization algorithm is used to adjust and optimize the yarn spacing, density and twist of the cooling fabric to achieve the optimal cooling value. S42. Using the adjusted and optimized yarn spacing, density and twist of the cooling fabric as the best manufacturing parameters for the cooling fabric, a set of manufacturing parameters for the cooling fabric is obtained, and intelligent control of the manufacturing of the cooling fabric is realized. The process of using the Gazelle Optimization Algorithm to adjust and optimize the yarn spacing, density, and twist of the cooling fabric includes the following steps: In the initialization phase of the gazelle population, the number of gazelles in the population is set to i, the dimension of each gazelle is j, and an initial population matrix is ​​generated. p1 represents a random number and p1∈[0,1], the upper bound of the gazelle population dimension is l1, and the lower bound of the dimension is l2. Then the position of the i′th gazelle in the population is E(i′)=p1·(l1-l2)+l2. The positions of individual gazelles in the gazelle population are iterated, and the gazelles whose positions correspond to the best fitness function values ​​obtained in each iteration are denoted as elite gazelles, thus obtaining an elite gazelle matrix. The process of updating the positions of individual gazelles is the process of adjusting and optimizing the yarn spacing, density, and twist of the cooling fabric. In the gazelle population development phase, let the current iteration number be k, and let A be the matrix of elite gazelle individuals in the k-th iteration. k Let E(i′,k) be the position of the i′-th gazelle in the k-th iteration. Assume that the gazelle moves by Brownian motion while feeding, and its speed is q. Let p2 and p3 represent random number vectors of Brownian motion, and p3∈[0,1]. Then, the formula for calculating the position E(i′,k+1) of the i′-th gazelle in the (k+1)-th iteration is as follows: E(i′,k+1)=E(i′,k)+p2·p3·q(A k -p2·E(i′,k)); By comparing the fitness function values ​​after each iteration, the optimal fitness function value is found, and the elite gazelle individual matrix is ​​updated. During the gazelle population exploration phase, Levy flight was introduced. When the number of iterations was odd, individual gazelles in the population moved in one direction, and when the number of iterations was even, individual gazelles in the population moved in another direction. The movement followed Brownian motion and then Levy flight. Let p4 be the random number vector for Levi's flight, Q be the fastest speed of the gazelle individual, and p5 represent a constant. When the gazelle spots a predator, the position E(i′,k+1) of the i′ gazelle individual in the (k+1)th iteration is updated, and the calculation formula is as follows: E(i′,k+1)=E(i′,k)+p3·p4·p5·Q(A k -p4·E(i′,k)); The maximum number of iterations is set to K, and the predator begins chasing the gazelle. The chasing coefficient is... At this point, individual gazelles in the population begin to move. The position E(i′,k+1) of the i′ gazelle in the (k+1)th iteration is updated again, and the calculation formula is as follows: E(i′,k+1)=E(i′,k)+p2 p4 p5 Q γ(A k -p4·E(i′,k)); A predation success rate is set, and individual gazelles in the gazelle population move according to the predation success rate. When the current iteration count reaches the maximum iteration count, the iteration stops, and the final gazelle position is obtained. The three-dimensional coordinate values ​​of the final gazelle position are the adjusted and optimized yarn spacing, density, and twist of the cooling fabric, respectively, at which point the optimal cooling value is achieved.

2. The intelligent control method for manufacturing cooling fabrics based on data processing according to claim 1, characterized in that, S1 includes the following steps: S11. Obtain the initial cool-feeling fabric surface image, form an initial cool-feeling fabric surface image set, perform image preprocessing on the initial cool-feeling fabric surface image in the initial cool-feeling fabric surface image set, and obtain the processed cool-feeling fabric surface image set. S12. Calculate the yarn spacing and density of the processed cool-feeling fabric in the processed cool-feeling fabric surface image set to obtain the initial cool-feeling fabric manufacturing parameter set. S13. Density-based clustering is used to clean the abnormal data of the manufacturing parameters of the cooling fabric to obtain a set of manufacturing parameters for the cooling fabric.

3. The intelligent control method for manufacturing cooling fabrics based on data processing according to claim 2, characterized in that, The image preprocessing of the initial cool-feeling fabric surface images in the initial cool-feeling fabric surface image set includes the following steps: The initial image of the cooling fabric surface is converted to grayscale to obtain a grayscale cooling fabric surface image. A Wiener filter is then used to filter the grayscale cooling fabric surface image to generate a filtered cooling fabric surface image. A wavelet transform is performed on the filtered cooling fabric surface image to obtain a decomposed and reconstructed cooling fabric surface image. The decomposed and reconstructed cooling fabric surface image is then binarized to obtain a binary cooling fabric surface image. Finally, a smoothing process is applied to obtain a processed cooling fabric surface image, forming a set of processed cooling fabric surface images.

4. The intelligent control method for manufacturing cooling fabrics based on data processing according to claim 3, characterized in that, S2 includes the following steps: S21. Set a target parameter set for manufacturing cool-feeling fabric, compare the deviation between the set of manufacturing parameters for cool-feeling fabric and the target parameter set for manufacturing cool-feeling fabric, and adjust the parameters for the manufacturing process of cool-feeling fabric. S22. As the temperature and humidity rise and fall, the yarn spacing and density of the cooling fabric are automatically controlled to achieve real-time control of the manufacturing of the cooling fabric.

5. The intelligent control method for manufacturing cooling fabrics based on data processing according to claim 4, characterized in that, S3 includes the following steps: S31. Perform dimensionality reduction processing on the cooling fabric manufacturing parameters in the set of cooling fabric manufacturing parameters to obtain a set of parameter features after dimensionality reduction. S32. Obtain a new image of the cool-feeling fabric surface. After preprocessing and normalization, extract a set of parameter feature samples. Use the set of parameter feature samples to train the RBF neural network to obtain the RBF neural network model. S33. Normalize the reduced parameter feature set to obtain a normalized parameter feature set. Input the normalized parameter feature set into the RBF neural network model to output the cooling prediction value and complete the cooling prediction of the cooling fabric.

6. The intelligent control method for manufacturing cooling fabrics based on data processing according to claim 5, characterized in that, S32 includes the following steps: S321. Set the mapping function of the RBF neural network to a Gaussian kernel function, the radial basis function to F″, the number of input layers to ε1, the number of neurons to ε2, and the number of output layers to 1; divide the parameter feature sample set into a parameter feature sample training set and a parameter feature sample test set; input the parameter feature sample training set into the RBF neural network and iterate continuously until the RBF neural network converges to obtain a trained RBF neural network; S322. Input the parameter feature sample test set into the trained RBF neural network, set the accuracy threshold to ξ, and obtain the RBF neural network model when the accuracy of the output result is less than the accuracy threshold; otherwise, adjust the weights until the accuracy of the output result is less than the accuracy threshold.

7. A system for implementing the intelligent control method for manufacturing cooling fabrics based on data processing as described in any one of claims 1-6, characterized in that, Specifically, it includes: Cooling fabric manufacturing parameter processing module, cooling fabric manufacturing real-time control module, cooling fabric cooling prediction module, and cooling fabric manufacturing parameter adjustment and optimization module; The cooling fabric manufacturing parameter processing module is used to preprocess the surface image of the cooling fabric to obtain the manufacturing parameters of the cooling fabric. The real-time control module for manufacturing cooling fabric is used to compare the deviation between the manufacturing parameters and the target parameters for manufacturing cooling fabric, and to control the manufacturing of cooling fabric in real time. The cooling fabric cooling prediction module is used to predict the cooling sensation of cooling fabrics using an RBF neural network model. The cooling fabric manufacturing parameter adjustment and optimization module is used to optimize and adjust the cooling fabric manufacturing parameters using the Gazelle Optimization Algorithm.

Images