Intelligent control system for processing differentiated fiber high-density composite yarn

Abstract

The invention discloses an intelligent control system for processing differentiated fiber high-density composite yarns, and relates to the field of processing flow control systems. The invention provides adjustment objects and data basis for optimizing and adjusting the performance parameters of the pretreatment in the subsequent process by setting the initial parameters and performance parameter thresholds of each pretreatment step corresponding to each differentiated fiber raw material; by predicting the performance parameter data matrix, the parameters in the pretreatment step can be optimized and adjusted in time, so that the efficiency in the pretreatment step is higher, thereby improving the efficiency of the entire processing process of the differentiated fiber high-density composite yarn; and by constructing a final CNN neural network classification model, a classification model is provided for the subsequent classification and acquisition of color grade data of a differentiated fiber raw material set after the first comprehensive treatment, and the color grade data that does not meet the requirements is continued to be dyed until the requirements are met.

Description

Technical Field

[0001] The present invention belongs to the field of process control systems, and in particular, relates to an intelligent control system for processing differentiated fiber high-density composite yarns. Background Art

[0002] Chinese patent CN115993806B discloses a monitoring control system and method based on yarn spinning, which collects weight, humidity and temperature information through various data acquisition devices; analyzes and calculates the collected weight information to determine whether adjustment and control can be performed. If the result is yes, all data are substituted into the temperature and humidity adjustment model for simulation calculation to obtain temperature or humidity adjustment information, and automatically adjusts its own working state according to the received instruction information; automatically records the collected information and adjustment information, and stores them in the database.

[0003] When processing differentiated fiber high-density composite yarns, special attention should be paid to the performance of the fibers in each step. If the performance of the fibers does not meet the corresponding performance requirements in a certain step of the processing, it will affect the production process, final quality and sales of the finished yarns, thereby affecting the profits of the manufacturing company. Summary of the invention

[0004] In view of the problems in the related art, the present invention proposes an intelligent control system for processing differentiated fiber high-density composite yarns to overcome the above-mentioned technical problems existing in the existing related art.

[0005] In order to solve the above technical problems, the present invention is achieved through the following technical solutions:

[0006] The present invention is an intelligent control method for processing differentiated fiber high-density composite yarns, comprising the following steps:

[0007] S1. Collecting several different types of differentiated fiber raw materials and setting the initial parameters and performance parameter thresholds corresponding to each pretreatment step to obtain a differentiated fiber raw material set, a pretreatment step initial parameter set matrix, and a pretreatment step performance parameter threshold set matrix;

[0008] S2, collecting performance parameter data at several time points in each pretreatment step of the differentiated fiber raw material set and constructing a final BP neural network prediction model; using the final BP neural network prediction model to predict the performance parameter data matrix to obtain a performance parameter prediction data matrix;

[0009] S3, determining whether it is necessary to optimize the initial parameter set matrix of the preprocessing step according to the performance parameter prediction data matrix and the preprocessing step performance parameter threshold set matrix; when optimization is required, optimizing the initial parameter set matrix of the preprocessing step; and preprocessing the differentiated fiber raw material set after optimization to obtain a differentiated fiber raw material set after preprocessing;

[0010] S4, using multiple differentiated fiber high-density composite yarn samples and corresponding color grade data to build the final CNN neural network classification model;

[0011] S5. Using the final CNN neural network classification model to classify the pre-treated differentiated fiber raw material set after the first comprehensive treatment, to obtain color grade data of the yarn to be classified; and determining whether dyeing is successful according to the color grade data of the yarn to be classified;

[0012] High-density composite yarn is composed of many different types of differentiated fibers, so several different types of differentiated fiber raw materials are first collected; before using these differentiated fiber raw materials to make composite yarns, it is necessary to pre-treat according to the characteristics of the differentiated fiber raw materials to ensure the stability of the performance of the fiber in subsequent processing; by setting the initial parameters and performance parameter thresholds of each pre-treatment step corresponding to each differentiated fiber raw material, the adjustment object and data basis are provided for the subsequent optimization and adjustment of the performance parameters of the pre-treatment; by constructing the final BP neural network prediction model, a prediction model is provided for the subsequent prediction of the performance parameters of each pre-treatment step, so as to facilitate the subsequent analysis process; by predicting the performance parameter data matrix, the performance parameter data that can be achieved in the pre-treatment step can be obtained in advance, so as to optimize and adjust the parameters in the pre-treatment step in time, so that the efficiency in the pre-treatment step is higher, thereby improving the efficiency of the entire differential fiber high-density composite yarn processing process; and then by constructing the final CNN neural network classification model, a classification model is provided for the subsequent classification and acquisition of the color grade data of the differentiated fiber raw material set after the first comprehensive treatment, and the color grade data that does not meet the requirements continues to be dyed until the requirements are met.

[0013] Preferably, the S1 comprises the following steps:

[0014] S11. collecting corresponding differentiated fiber raw materials of different types according to the production requirements of high-density composite yarn to obtain a differentiated fiber raw material set;

[0015] S12, setting a pretreatment step for each differentiated fiber raw material according to the production requirements of the high-density composite yarn and the differentiated fiber raw material set, to obtain a pretreatment step matrix;

[0016] Set the preprocessing step matrix The initial parameters of each preprocessing step in , and the initial parameter set matrix of the preprocessing step is obtained;

[0017] S13, setting a plurality of parameter types that can reflect differentiated fiber performance to obtain a differentiated fiber performance parameter type set;

[0018] According to the differentiated fiber performance parameter type set and the preprocessing requirements, the data value that the performance parameter of the differentiated fiber needs to reach after each preprocessing step in the preprocessing step matrix is ​​completed is set to obtain a preprocessing step performance parameter threshold set matrix;

[0019] Differentiated fiber raw materials include special-shaped fibers, hollow fibers, fine fibers, antistatic fibers, high shrinkage fibers, flame-retardant fibers, ultrafine fibers, easily dyed fibers, hydrophilic synthetic fibers, colored fibers, and anti-pilling fibers; the pretreatment step matrix includes cleaning, drying, softening, etc.; the differentiated fiber performance parameter types include strength, elongation, hygroscopicity, etc.; different types of differentiated fibers have different corresponding pretreatment steps, therefore, each pretreatment step of each type of differentiated fiber is analyzed and processed separately, making the analysis process more specific and accurate.

[0020] Preferably, S2 comprises the following steps:

[0021] S21, setting a number of equal number and equally spaced parameter data collection time points for each preprocessing step in the preprocessing step matrix to obtain a collection time point set matrix;

[0022] According to the differentiated fiber performance parameter type set and the collection time point set matrix, the performance parameter data of each differentiated fiber raw material in the differentiated fiber raw material set at the first several collection time points in each corresponding pretreatment step are collected to obtain a performance parameter data matrix;

[0023] S22, constructing an initial BP neural network prediction model and setting a first training data ratio and a test data ratio; dividing the performance parameter data matrix according to the first training data ratio and the test data ratio to obtain a performance parameter training data matrix, a performance parameter test data matrix, and a performance parameter to-be-predicted data matrix;

[0024] The performance parameter training data matrix and the performance parameter test data matrix are used to train and test the initial BP neural network prediction model; after the training and testing are completed, the final BP neural network prediction model is obtained;

[0025] S23, inputting the performance parameter to-be-predicted data matrix into the final BP neural network prediction model for prediction operation to obtain a performance parameter prediction data matrix;

[0026] By setting multiple collection time points and collecting the corresponding differentiated fiber performance parameters according to the collection time points, data support is provided for the subsequent construction of the final BP neural network prediction model.

[0027] Preferably, in S22, the performance parameter training data matrix and the performance parameter test data matrix are used to train and test the initial BP neural network prediction model; after the training and testing are completed, obtaining the final BP neural network prediction model includes the following steps:

[0028] S221, setting a first training error threshold; inputting the performance parameter training data matrix into the initial BP neural network prediction model for training; during the training process, when the training error is less than the first training error threshold, stopping the training to obtain a trained BP neural network prediction model; otherwise, continuing the training until the training error is less than the first training error threshold;

[0029] S222, set a first test accuracy threshold; input the performance parameter test data matrix into the trained BP neural network prediction model for testing; after the test is completed, obtain a first test accuracy; when the first test accuracy is greater than or equal to the first test accuracy threshold, use the trained BP neural network prediction model as the final BP neural network prediction model; otherwise, set a first training iteration number, return to S221 according to the first training iteration number to continue training the trained BP neural network prediction model until the first test accuracy is greater than or equal to the first test accuracy threshold, and obtain the final BP neural network prediction model;

[0030] By training and testing the initial BP neural network prediction model using the performance parameter training data matrix and the performance parameter test data matrix respectively, the final BP neural network prediction model obtained has better prediction ability for the differentiated fiber performance parameter data based on time series.

[0031] Preferably, S3 comprises the following steps:

[0032] S31, the performance parameter prediction data matrix and the pre-processing step performance parameter threshold set matrix c Compare; obtain the acquisition time point corresponding to the first performance parameter prediction data greater than or equal to the corresponding performance parameter threshold in each preprocessing step in the performance parameter prediction data matrix, and obtain an acquisition time point matrix;

[0033] S32, performing a difference operation between each acquisition time point in the acquisition time point matrix and the first acquisition time point in each corresponding preprocessing step in the acquisition time point set matrix, to obtain an initial step time-consuming data matrix;

[0034] Setting a time-consuming data threshold for each preprocessing step in the preprocessing step matrix to obtain a time-consuming data threshold matrix;

[0035] S33, when the time-consuming data of the preprocessing step in the initial step time-consuming data matrix is ​​greater than or equal to the time-consuming data threshold in the corresponding time-consuming data threshold matrix, the initial parameter set corresponding to the preprocessing step in the initial parameter set matrix of the preprocessing step is optimized until the time-consuming data matrix of the initial step does not contain the time-consuming data of the preprocessing step greater than or equal to the corresponding time-consuming data threshold matrix e Until the time-consuming data threshold in, the final parameter set matrix of the preprocessing step is obtained;

[0036] Otherwise, there is no need to optimize the initial parameter set corresponding to the preprocessing step in the initial parameter set matrix of the preprocessing step, and the initial parameter set matrix of the preprocessing step is used as the final parameter set matrix of the preprocessing step;

[0037] S34, applying the final parameter set matrix of the preprocessing step to each preprocessing step in the preprocessing step matrix, and preprocessing each differentiated fiber raw material in the differentiated fiber raw material set to obtain a differentiated fiber raw material set after preprocessing;

[0038] By performing a difference operation between the acquisition time point corresponding to the first performance parameter prediction data greater than or equal to the corresponding performance parameter threshold in each pretreatment step in the performance parameter prediction data matrix and the first acquisition time point in each corresponding pretreatment step in the acquisition time point set matrix, the time consumed from the start of each pretreatment step to the performance parameter of the corresponding differentiated fiber raw material reaching the performance parameter threshold is calculated; the calculated consumption time is then compared with a pre-set consumption time threshold. When the consumption time does not meet the requirements, the parameters of the corresponding pretreatment step are adjusted to reduce the consumption time, thereby improving the efficiency of each pretreatment step and ultimately improving the efficiency of the entire pretreatment process.

[0039] Preferably, in S33, the initial parameter set matrix of the preprocessing step b Optimizing the initial set of parameters for the preprocessing step includes the following steps:

[0040] S331, setting the initial parameter set to be optimized and the corresponding time-consuming data;

[0041] Construct a krill population; set the maximum number of iterations corresponding to the krill population to , the current number of iterations is And the search space dimension is ;

[0042] S332, setting a fitness function of the krill population according to the time-consuming data;

[0043] S333, constructing a forward population initial position matrix according to the initial parameter set to be optimized; constructing a reverse population initial position matrix according to the forward population initial position matrix;

[0044] The forward population initial position matrix and the reverse population initial position matrix are combined to obtain a combined population initial position matrix; the fitness function of the krill population is used to calculate the fitness value of each initial position in the combined population initial position matrix to obtain a first fitness value set; the fitness values ​​in the first fitness value set are sorted from small to large and the initial position corresponding to the first fitness value is selected as the initial position of the krill population to obtain an initial position matrix;

[0045] S334, start iteration, and set the current iteration number to 1 before iteration; in the first round of iteration, use the fitness function of the krill population to calculate the fitness value of each initial position in the initial position matrix, and in each other round of iteration, use the fitness function of the krill population to calculate the fitness value of each position in the position matrix updated in the previous round of iteration to obtain a second fitness set; use the maximum fitness value and the corresponding krill position in the second fitness set as the global optimal fitness and the global optimal position respectively;

[0046] updating the position of each krill in the krill population according to the global optimal fitness and the global optimal position;

[0047] S335, when When , stop the iteration and output the optimized parameter set; apply each parameter in the optimized parameter set to the corresponding preprocessing step, and measure the time-consuming data to obtain the optimized time-consuming data; otherwise, continue to iterate until until the time;

[0048] When the optimized time-consuming data is less than the corresponding time-consuming data threshold in the time-consuming data threshold matrix, the optimized parameter set is used to replace the corresponding initial parameter set in the initial parameter set matrix of the preprocessing step to obtain the final parameter set matrix of the preprocessing step; otherwise, return to S334 and continue to iterate until the optimized time-consuming data is less than the corresponding time-consuming data threshold in the time-consuming data threshold matrix, and obtain the final parameter set matrix of the preprocessing step;

[0049] The krill swarm optimization algorithm has strong search diversity, simple operation and few adjustable parameters; therefore, in this scheme, the krill swarm optimization algorithm is used to iterate and optimize the parameters in the pretreatment step multiple times, and the time-consuming data of the corresponding differentiated fiber pretreatment to reach the corresponding performance parameter threshold is used as the fitness function, so that with continuous iterations, the time-consuming data of the differentiated fiber pretreatment to reach the corresponding performance parameter threshold will continue to decrease, thereby meeting the requirements.

[0050] Preferably, S4 comprises the following steps:

[0051] S41, collecting a plurality of differentiated fiber high-density composite yarn samples and corresponding color grade data to obtain a composite yarn sample set and a color grade data set;

[0052] Acquiring image data of each composite yarn sample in the composite yarn sample set to obtain a yarn image data set;

[0053] S42, constructing an initial CNN neural network classification model; using the yarn image data set and the color grade data set to train and test the initial CNN neural network classification model; after the training and testing are completed, a final CNN neural network classification model is obtained;

[0054] By constructing the final CNN neural network classification model, a classification model is provided for the subsequent classification of the color level data of the dyed composite yarn.

[0055] Preferably, S5 comprises the following steps:

[0056] S51, setting a first comprehensive processing step set; performing a first comprehensive processing on the pretreated differentiated fiber raw material set according to the first comprehensive processing step set to obtain a composite yarn to be dyed; setting a color grade requirement for the composite yarn to be dyed after dyeing to obtain color grade requirement data;

[0057] S52, performing dyeing operation on the composite yarn to be dyed; obtaining the dyed composite yarn after dyeing; collecting image data of the dyed composite yarn to obtain image data of the yarn to be classified; inputting the image data of the yarn to be classified into the final CNN neural network classification model for classification operation, and obtaining color grade data of the yarn to be classified;

[0058] If the color grade data of the yarn to be classified is the same as the color grade requirement data, the dyeing is successful; if the color grade data of the yarn to be classified is not the same as the color grade requirement data, the dyeing fails; continue dyeing the dyed composite yarn until the color grade data of the yarn to be classified is the same as the color grade requirement data;

[0059] S53, setting a second comprehensive processing step set; performing a second comprehensive processing on the dyed composite yarn according to the second comprehensive processing step set; after the processing is completed, a finished composite yarn is obtained;

[0060] By comparing the color grade data of the yarn to be classified with the color grade requirement data, it is determined whether the color of the current composite yarn meets the production requirements. If it does not meet the requirements, it is dyed again to ensure that the final composite yarn product meets the requirements.

[0061] The intelligent control system used for the processing of differentiated fiber high-density composite yarns includes a differentiated fiber raw material collection module, a pretreatment step initial parameter setting module, a performance parameter threshold setting module, a performance parameter data collection module, a prediction model construction module, a prediction module, a performance parameter optimization judgment module, a performance parameter optimization module, a pretreatment module, a classification model construction module, a classification module and a color judgment module.

[0062] The present invention has the following beneficial effects:

[0063] 1. In the present invention, by setting the initial parameters and performance parameter thresholds of each pretreatment step corresponding to each differentiated fiber raw material, an adjustment object and a data basis are provided for the subsequent optimization and adjustment of the pretreatment performance parameters; by predicting the performance parameter data matrix, the performance parameter data that can be achieved in the pretreatment step is obtained in advance, and the parameters in the pretreatment step are optimized and adjusted in time, so that the efficiency in the pretreatment step is higher, thereby improving the efficiency of the entire differential fiber high-density composite yarn processing process; and by constructing the final CNN neural network classification model, a classification model is provided for the subsequent classification and acquisition of the color grade data of the differentiated fiber raw material set after the first comprehensive treatment, and the color grade data that does not meet the requirements continues to be dyed until the requirements are met.

[0064] 2. In the present invention, the time consumed from the beginning of each pretreatment step to the performance parameter of the corresponding differentiated fiber raw material reaching the performance parameter threshold is calculated and compared with the pre-set consumption time threshold. When the consumption time does not meet the requirements, the parameters of the corresponding pretreatment step are adjusted to reduce the consumption time, thereby improving the efficiency of each pretreatment step and ultimately improving the efficiency of the entire pretreatment process.

[0065] 3. In the present invention, the parameters in the pretreatment step are iterated and optimized multiple times by using the krill swarm optimization algorithm, and the time-consuming data of the corresponding differentiated fiber pretreatment to reach the corresponding performance parameter threshold is used as the fitness function, so that the time-consuming data is continuously reduced, thereby meeting the requirements.

[0066] Of course, any product implementing the present invention does not necessarily need to achieve all of the advantages described above at the same time. BRIEF DESCRIPTION OF THE DRAWINGS

[0067] In order to more clearly illustrate the technical solutions of the embodiments of the invention, the drawings required for describing the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the invention. For ordinary technicians in this field, they can also obtain drawings based on these drawings without paying creative work.

[0068] Figure 1 The present invention is a flow chart of the intelligent control system for controlling the differential fiber high-density composite yarn processing process. DETAILED DESCRIPTION

[0069] The following will be combined with the drawings in the embodiments of the invention to clearly and completely describe the technical solutions in the embodiments of the invention. Obviously, the described embodiments are only part of the embodiments of the invention, not all of the embodiments. Based on the embodiments in the invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the invention.

[0070] In the description of the present invention, it is necessary to understand that the terms "opening", "upper", "lower", "top", "middle", "inside" and the like indicating 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, be constructed and operate in a specific orientation, and therefore cannot be understood as limiting the invention.

[0071] Embodiment 1

[0072] This embodiment is an intelligent control method for processing differentiated fiber high-density composite yarns, comprising the following steps:

[0073] S1. Collecting several different types of differentiated fiber raw materials and setting the initial parameters and performance parameter thresholds corresponding to each pretreatment step to obtain a differentiated fiber raw material set, a pretreatment step initial parameter set matrix, and a pretreatment step performance parameter threshold set matrix;

[0074] The S1 comprises the following steps:

[0075] S11. According to the production requirements of high-density composite yarn, several different types of corresponding differentiated fiber raw materials are collected to obtain a differentiated fiber raw material set. , Indicates the collected i types of differentiated fiber raw materials, Indicates the total number of differentiated fiber raw material types collected;

[0076] S12, according to the production requirements of high-density composite yarn and the differentiated fiber raw material set Set the pretreatment steps for each differentiated fiber raw material to obtain the pretreatment step matrix ;as follows,

[0077] ;

[0078] in, Indicates the differentiated fiber raw material set Middle i Differentiated fiber raw materials set j A preprocessing step, Indicates the differentiated fiber raw material set Middle i The total number of pretreatment steps set for the differentiated fiber raw materials;

[0079] Set the preprocessing step matrix The initial parameters of each preprocessing step in the matrix are obtained. b ;as follows,

[0080] ;

[0081] in, Indicates the differentiated fiber raw material set Middle i Differentiated fiber raw materials set j The initial parameter set for the preprocessing steps; , express The k Initial parameters, Indicates the differentiated fiber raw material set Middle i Differentiated fiber raw materials set j The total number of parameters in the preprocessing steps;

[0082] S13, setting a plurality of parameter types that can reflect the performance of differentiated fibers, and obtaining a set of differentiated fiber performance parameter types , Indicates the setting i Differentiated fiber performance parameter types, Indicates the total number of set differential fiber performance parameter types;

[0083] According to the differentiated fiber performance parameter type set And preprocessing requirements set the preprocessing steps matrix After each preprocessing step is completed, the corresponding performance parameter of the differentiated fiber needs to reach the data value, and the preprocessing step performance parameter threshold set matrix is ​​obtained. c ;as follows,

[0084] ;

[0085] in, Indicates the differentiated fiber raw material set Middle i Differentiated fiber raw materials set j The performance parameter threshold set corresponding to the preprocessing steps; , express The k performance parameter thresholds;

[0086] S2, collecting performance parameter data at several time points in each pretreatment step of the differentiated fiber raw material set and constructing a final BP neural network prediction model; using the final BP neural network prediction model to predict the performance parameter data matrix to obtain a performance parameter prediction data matrix;

[0087] The S2 comprises the following steps:

[0088] S21, for the pre-processing step matrix In each preprocessing step, a number of equal and equally spaced parameter data collection time points are set to obtain the collection time point set matrix ;as follows,

[0089] ;

[0090] in, Indicates the differentiated fiber raw material set Middle i Differentiated fiber raw materials set j The set of acquisition time points corresponding to the preprocessing steps; , express The i The collection time points, Indicates the differentiated fiber raw material set The total number of collection time points in each pretreatment step set for each differentiated fiber raw material;

[0091] Set the number of times to be predicted , according to the differentiated fiber performance parameter type set And the collection time point set matrix Collection of differentiated fiber raw materials Each differentiated fiber raw material in each corresponding pretreatment step The performance parameter data of each acquisition time point is obtained to obtain the performance parameter data matrix d ;as follows,

[0092] ; ;

[0093] in, Indicates the differentiated fiber raw material set Middle i Differentiated fiber raw materials set j The performance parameter data matrix corresponding to the preprocessing steps; express Middle k The first collection time point Performance parameter data;

[0094] S22, constructing an initial BP neural network prediction model and setting a first training data ratio and a test data ratio; and calculating a performance parameter data matrix according to the first training data ratio and the test data ratio. d Perform data division to obtain a performance parameter training data matrix, a performance parameter test data matrix, and a performance parameter to-be-predicted data matrix;

[0095] The performance parameter training data matrix and the performance parameter test data matrix are used to train and test the initial BP neural network prediction model; after the training and testing are completed, the final BP neural network prediction model is obtained;

[0096] In S22, the initial BP neural network prediction model is trained and tested using the performance parameter training data matrix and the performance parameter test data matrix; after the training and testing are completed, the final BP neural network prediction model is obtained, including the following steps:

[0097] S221, setting a first training error threshold; inputting the performance parameter training data matrix into the initial BP neural network prediction model for training; during the training process, when the training error is less than the first training error threshold, stopping the training to obtain a trained BP neural network prediction model; otherwise, continuing the training until the training error is less than the first training error threshold;

[0098] S222, set a first test accuracy threshold; input the performance parameter test data matrix into the trained BP neural network prediction model for testing; after the test is completed, obtain a first test accuracy; when the first test accuracy is greater than or equal to the first test accuracy threshold, use the trained BP neural network prediction model as the final BP neural network prediction model; otherwise, set a first training iteration number, return to S221 according to the first training iteration number to continue training the trained BP neural network prediction model until the first test accuracy is greater than or equal to the first test accuracy threshold, and obtain the final BP neural network prediction model;

[0099] S23, inputting the performance parameter to-be-predicted data matrix into the final BP neural network prediction model for prediction operation, and obtaining the performance parameter prediction data matrix ;as follows,

[0100] ; ;

[0101] in, Indicates the differentiated fiber raw material set Middle i Differentiated fiber raw materials set j The performance parameter prediction data matrix corresponding to the preprocessing steps is express The predicted k The first collection time point Performance parameter data;

[0102] S3, determining whether it is necessary to optimize the initial parameter set matrix of the preprocessing step according to the performance parameter prediction data matrix and the preprocessing step performance parameter threshold set matrix; when optimization is required, optimizing the initial parameter set matrix of the preprocessing step; and preprocessing the differentiated fiber raw material set after optimization to obtain a differentiated fiber raw material set after preprocessing;

[0103] The S3 comprises the following steps:

[0104] S31, the performance parameter prediction data matrix The matrix of threshold sets of performance parameters in the preprocessing step c Compare; obtain the performance parameter prediction data matrix The acquisition time point corresponding to the first performance parameter prediction data greater than or equal to the corresponding performance parameter threshold in each preprocessing step is obtained to obtain the acquisition time point matrix ;as follows,

[0105] ;

[0106] in, Indicates the differentiated fiber raw material set Middle i Differentiated fiber raw materials set j The acquisition time point corresponding to the first performance parameter prediction data greater than or equal to the corresponding performance parameter threshold value in the preprocessing step;

[0107] S32, the acquisition time point matrix Each acquisition time point and the acquisition time point set matrix The first acquisition time point in each corresponding preprocessing step is subtracted to obtain the initial step time-consuming data matrix ;as follows,

[0108] ;

[0109] in, Indicates the differentiated fiber raw material set Middle i Differentiated fiber raw materials set j Initial time consumption data corresponding to the preprocessing steps;

[0110] Setting the preprocessing step matrix The time-consuming data threshold of each preprocessing step in the process is obtained to obtain the time-consuming data threshold matrix e ;as follows,

[0111] ;

[0112] in, Indicates the differentiated fiber raw material set Middle i Differentiated fiber raw materials set j The time-consuming data threshold for each preprocessing step;

[0113] S33, when the initial step consumes time data matrix There are time-consuming data of the preprocessing step that is greater than or equal to the corresponding time-consuming data threshold matrix e When the time-consuming data threshold is , the initial parameter set matrix of the preprocessing step is b The initial set of parameters for the preprocessing step is optimized until the initial step takes time for the data matrix There is no time-consuming data of the preprocessing step greater than or equal to the corresponding time-consuming data threshold matrix e The final parameter set matrix of the preprocessing step is obtained until the time-consuming data threshold in ;as follows,

[0114] ;

[0115] in, Indicates the differentiated fiber raw material set Middle i Differentiated fiber raw materials set j The final set of parameters for the preprocessing steps; , express The k final parameters;

[0116] Otherwise, there is no need to initialize the parameter set matrix for the preprocessing step b The initial parameter set of the preprocessing step is optimized, and the matrix of the initial parameter set of the preprocessing step is b As a preprocessing step, the final parameter set matrix ;

[0117] In S33, the initial parameter set matrix of the preprocessing step is b Optimizing the initial set of parameters for the preprocessing step includes the following steps:

[0118] S331. Setting the initial parameter set to be optimized And the corresponding time consumption data , Indicates the first i Initial parameters, Indicates the total number of initial parameters to be optimized;

[0119] Building krill populations , Indicates the first i Krill, represents the size of the krill population; the maximum number of iterations corresponding to the krill population is set to , the current number of iterations is And the search space dimension is ;

[0120] S332: According to the time-consuming data Setting the fitness function of the krill population ;as follows,

[0121] ;

[0122] S333, according to the initial parameter set to be optimized Construct the initial position matrix of the forward population ;as follows,

[0123] ;

[0124] in, Indicates the krill population i The initial positive position of the krill is j The component in dimension; the calculation formula is as follows,

[0125] ;

[0126] In the formula, , Respectively represent the initial parameter set to be optimized Middle i The lower and upper limits of the initial parameters; Indicates for Generate a random number between 0 and 1;

[0127] According to the forward population initial position matrix Construct the reverse population initial position matrix ;as follows,

[0128] ;

[0129] in, Indicates the krill population i The reverse initial position of the krill is j The component in dimension; the calculation formula is as follows,

[0130] ;

[0131] The forward population initial position matrix And the reverse population initial position matrix Merge to obtain the initial position matrix of the combined population; adopt the fitness function of the krill population Calculate the fitness value of each initial position in the initial position matrix of the combined population to obtain a first fitness value set; sort the fitness values ​​in the first fitness value set from small to large and select the top fitness values. The initial position corresponding to the fitness value is taken as the initial position of the krill population, and the initial position matrix is ​​obtained. ;as follows,

[0132] ;

[0133] in, Indicates the krill population i The initial position of the krill is j The weight in the dimensions;

[0134] S334, start iteration, before iteration, the current number of iterations is Set to 1; the fitness function of the krill population is used in the first iteration Calculate the initial position matrix The fitness value of each initial position in the process, and the fitness function of the krill population is used in each other iteration process Calculate the fitness value of each position in the position matrix updated in the previous iteration process to obtain a second fitness set; take the maximum fitness value and the corresponding krill position in the second fitness set as the global optimal fitness and the global optimal position respectively;

[0135] The position of each krill in the krill population is updated according to the global optimal fitness and the global optimal position; after the update is completed, the current number of iterations is Add 1 and enter the next iteration;

[0136] S335, when When , stop the iteration and output the optimized parameter set; apply each parameter in the optimized parameter set to the corresponding preprocessing step, and measure the time-consuming data to obtain the optimized time-consuming data; otherwise, continue to iterate until until the time;

[0137] When the optimized time-consuming data is less than the time-consuming data threshold matrix e When the corresponding time-consuming data threshold is reached, the optimized parameter set is used to replace the initial parameter set matrix of the preprocessing step. b The corresponding initial parameter set in the preprocessing step is obtained Otherwise, return to S334 and continue iterating until the optimized time-consuming data is less than the time-consuming data threshold matrix e The corresponding time-consuming data threshold in the preprocessing step is obtained. ;

[0138] S34, the final parameter set matrix of the preprocessing step Apply to the preprocessing step matrix In each pretreatment step, and for differentiated fiber raw material sets Each differentiated fiber raw material is pretreated to obtain a differentiated fiber raw material set after pretreatment. , Indicates i Types of differentiated fiber raw materials after pretreatment;

[0139] S4, using multiple differentiated fiber high-density composite yarn samples and corresponding color grade data to build the final CNN neural network classification model;

[0140] The S4 comprises the following steps:

[0141] S41, collecting multiple differentiated fiber high-density composite yarn samples and corresponding color grade data to obtain a composite yarn sample set And the color grade dataset , , Respectively represent the collected i Differentiated fiber high-density composite yarn samples and corresponding color grade data, It indicates the total number of differentiated fiber high-density composite yarn samples collected;

[0142] Obtaining the composite yarn sample set The image data of each composite yarn sample in is obtained to obtain the yarn image dataset , Indicates i Image data of differentiated fiber high-density composite yarn samples;

[0143] S42, constructing an initial CNN neural network classification model; using the yarn image data set And the color grade dataset Train and test the initial CNN neural network classification model; after the training and testing are completed, the final CNN neural network classification model is obtained;

[0144] The S42 comprises the following steps:

[0145] S421, setting a second training data ratio; and performing a multi-step training on the yarn image data set according to the second training data ratio. And the color grade dataset Performing data division to obtain a yarn image training data set, a yarn image test data set, a color grade training data set, and a color grade test data set;

[0146] S422, setting a second training error threshold; inputting the yarn image training data set and the color grade training data set into the initial CNN neural network classification model for training; during the training process, when the training error is less than the second training error threshold, stopping the training to obtain a trained CNN neural network classification model; otherwise, continuing the training until the training error is less than the second training error threshold;

[0147] S423, set a second test accuracy threshold; input the yarn image test data set and the color grade test data set into the trained CNN neural network classification model for testing; after the test is completed, obtain a second test accuracy; when the second test accuracy is greater than or equal to the second test accuracy threshold, use the trained CNN neural network classification model as the final CNN neural network classification model; otherwise, set a second training iteration number, and return to S432 according to the second training iteration number to continue training the trained CNN neural network classification model until the second test accuracy is greater than or equal to the second test accuracy threshold, and obtain the final CNN neural network classification model;

[0148] By using the yarn image training data set and the color grade training data set as well as the yarn image test data set and the color grade test data set to train and test the initial CNN neural network classification model, the final CNN neural network classification model has a good classification ability for the color grade data of the composite yarn;

[0149] S5. Using the final CNN neural network classification model to classify the pre-treated differentiated fiber raw material set after the first comprehensive treatment, to obtain color grade data of the yarn to be classified; and determining whether dyeing is successful according to the color grade data of the yarn to be classified;

[0150] The S5 comprises the following steps:

[0151] S51, setting a first comprehensive processing step set; treating the pretreated differentiated fiber raw material set according to the first comprehensive processing step set Performing a first comprehensive treatment to obtain a composite yarn to be dyed; setting a color grade requirement for the composite yarn to be dyed after dyeing to obtain color grade requirement data;

[0152] The first comprehensive treatment step includes combining woolen spinning treatment and worsted spinning treatment, etc.;

[0153] S52, performing dyeing operation on the composite yarn to be dyed; obtaining the dyed composite yarn after dyeing; collecting image data of the dyed composite yarn to obtain image data of the yarn to be classified; inputting the image data of the yarn to be classified into the final CNN neural network classification model for classification operation, and obtaining color grade data of the yarn to be classified;

[0154] If the color grade data of the yarn to be classified is the same as the color grade requirement data, the dyeing is successful; if the color grade data of the yarn to be classified is not the same as the color grade requirement data, the dyeing fails; continue dyeing the dyed composite yarn until the color grade data of the yarn to be classified is the same as the color grade requirement data;

[0155] S53, setting a second comprehensive processing step set; performing a second comprehensive processing on the dyed composite yarn according to the second comprehensive processing step set; after the processing is completed, a finished composite yarn is obtained;

[0156] The second comprehensive treatment step collectively includes shaping treatment and softening treatment, etc.

[0157] Embodiment 2

[0158] This embodiment discloses an intelligent control system for differential fiber high-density composite yarn processing, the system can implement the method of the above embodiment, including a differential fiber raw material collection module, a pretreatment step initial parameter setting module, a performance parameter threshold setting module, a performance parameter data collection module, a prediction model construction module, a prediction module, a performance parameter optimization determination module, a performance parameter optimization module, a pretreatment module, a classification model construction module, a classification module and a color determination module;

[0159] The differentiated fiber raw material collection module is used to collect several different types of differentiated fiber raw materials to obtain a differentiated fiber raw material set;

[0160] The pre-processing step initial parameter setting module is used to set the initial parameters of each pre-processing step corresponding to the differentiated fiber raw material set to obtain the pre-processing step initial parameter set matrix;

[0161] The performance parameter threshold setting module is used to set the performance parameter threshold of each pretreatment step corresponding to the individualized fiber raw material set, and obtain the pretreatment step performance parameter threshold set matrix;

[0162] The performance parameter data collection module is used to collect performance parameter data at several time points in each pretreatment step of the differentiated fiber raw material set to obtain a performance parameter data matrix;

[0163] The prediction model building module is used to construct a final BP neural network prediction model using the performance parameter data matrix;

[0164] The prediction module is used to predict the performance parameter data matrix using the final BP neural network prediction model to obtain the performance parameter prediction data matrix;

[0165] The performance parameter optimization determination module is used to determine whether it is necessary to optimize the initial parameter set matrix of the preprocessing step according to the performance parameter prediction data matrix and the preprocessing step performance parameter threshold set matrix to obtain a determination result;

[0166] The performance parameter optimization module is used to optimize the initial parameter set matrix of the preprocessing step to obtain the final parameter set matrix of the preprocessing step when the judgment result shows that optimization is needed;

[0167] The preprocessing module is used to preprocess the differentiated fiber raw material set according to the final parameter set matrix of the preprocessing step to obtain the differentiated fiber raw material set after preprocessing;

[0168] The classification model building module is used to construct a final CNN neural network classification model using a plurality of differentiated fiber high-density composite yarn samples and corresponding color grade data;

[0169] The classification module is used to classify the pre-treated differentiated fiber raw material set after the first comprehensive treatment using the final CNN neural network classification model to obtain the color grade data of the yarn to be classified;

[0170] The color determination module is used to determine whether dyeing is successful according to the color grade data of the yarn to be classified.

[0171] In the description of this specification, the description with reference to the terms "one embodiment", "example", "specific example", etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the invention. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any one or more embodiments or examples in a suitable manner.

[0172] The preferred embodiments of the invention disclosed above are only used to help explain the invention. The preferred embodiments do not describe all the details in detail, nor do they limit the invention to the specific implementation methods described. Obviously, many modifications and changes can be made according to the content of this specification. This specification selects and specifically describes these embodiments in order to better explain the principles and practical applications of the invention, so that those skilled in the art can understand and use the invention well.

Claims

1. An intelligent control method for processing differentiated fiber high-density composite yarns, characterized in that: The following steps are involved: S1. Collecting several different types of differentiated fiber raw materials and setting the initial parameters and performance parameter thresholds corresponding to each pretreatment step to obtain a differentiated fiber raw material set, a pretreatment step initial parameter set matrix, and a pretreatment step performance parameter threshold set matrix; S2, collecting performance parameter data at several time points in each pretreatment step of the differentiated fiber raw material set and constructing a final BP neural network prediction model; using the final BP neural network prediction model to predict the performance parameter data matrix to obtain a performance parameter prediction data matrix; S3, determining whether it is necessary to optimize the initial parameter set matrix of the preprocessing step according to the performance parameter prediction data matrix and the preprocessing step performance parameter threshold set matrix; when optimization is required, optimizing the initial parameter set matrix of the preprocessing step; After optimization, the differentiated fiber raw material set is pretreated to obtain a differentiated fiber raw material set after pretreatment; S4, using multiple differentiated fiber high-density composite yarn samples and corresponding color grade data to build the final CNN neural network classification model; S5. Using the final CNN neural network classification model to classify the pre-treated differentiated fiber raw material set after the first comprehensive treatment, to obtain color grade data of the yarn to be classified; and determining whether dyeing is successful according to the color grade data of the yarn to be classified; The S1 comprises the following steps: S11. collecting corresponding differentiated fiber raw materials of different types according to the production requirements of high-density composite yarn to obtain a differentiated fiber raw material set; S12, setting a pretreatment step for each differentiated fiber raw material according to the production requirements of the high-density composite yarn and the differentiated fiber raw material set, to obtain a pretreatment step matrix; Setting the initial parameters of each preprocessing step in the preprocessing step matrix to obtain a preprocessing step initial parameter set matrix; S13, setting a plurality of parameter types that can reflect the performance of differentiated fibers to obtain a set of differentiated fiber performance parameter types; setting the data values ​​that the performance parameters of the differentiated fibers corresponding to each preprocessing step in the preprocessing step matrix need to reach after the preprocessing step is completed according to the set of differentiated fiber performance parameter types and the preprocessing requirements, to obtain a preprocessing step performance parameter threshold set matrix; The S2 comprises the following steps: S21, for each preprocessing step in the preprocessing step matrix, set a number of parameter data collection time points to obtain a collection time point set matrix; set the number of times to be predicted According to the differentiated fiber performance parameter type set and the collection time point set matrix, the differential fiber raw material set is collected according to the differential fiber performance parameter type set and the collection time point set matrix. The performance parameter data of each acquisition time point is obtained to obtain a performance parameter data matrix; S22, constructing an initial BP neural network prediction model; performing data partitioning on the performance parameter data matrix to obtain a performance parameter training data matrix, a performance parameter test data matrix, and a performance parameter to-be-predicted data matrix; using the performance parameter training data matrix and the performance parameter test data matrix to train and test the initial BP neural network prediction model; after the training and testing are completed, obtaining the final BP neural network prediction model; S23, inputting the performance parameter to-be-predicted data matrix into the final BP neural network prediction model for prediction operation to obtain a performance parameter prediction data matrix.

2. The intelligent control method for differential fiber high-density composite yarn processing according to claim 1 is characterized in that: The training and testing of the initial BP neural network prediction model using the performance parameter training data matrix and the performance parameter test data matrix in S22 includes the following steps: S221, setting a first training error threshold; inputting the performance parameter training data matrix into the initial BP neural network prediction model for training; during the training process, when the training error is less than the first training error threshold, stopping the training to obtain a trained BP neural network prediction model; otherwise, continuing the training until the training error is less than the first training error threshold; S222, set a first test accuracy threshold; input the performance parameter test data matrix into the trained BP neural network prediction model for testing; after the test is completed, obtain a first test accuracy; when the first test accuracy is greater than or equal to the first test accuracy threshold, use the trained BP neural network prediction model as the final BP neural network prediction model; otherwise, set a first training iteration number, and return to S221 according to the first training iteration number to continue training the trained BP neural network prediction model until the first test accuracy is greater than or equal to the first test accuracy threshold, to obtain the final BP neural network prediction model.

3. The intelligent control method for differential fiber high-density composite yarn processing according to claim 2 is characterized in that: The S3 comprises the following steps: S31, obtaining the collection time point corresponding to the first performance parameter prediction data greater than or equal to the corresponding performance parameter threshold in each preprocessing step in the performance parameter prediction data matrix, and obtaining a collection time point matrix; S32, performing a difference operation between the first acquisition time point in each preprocessing step of the acquisition time point matrix and the acquisition time point set matrix to obtain an initial step time-consuming data matrix; setting a time-consuming data threshold for each preprocessing step in the preprocessing step matrix; S33, when there is time-consuming data of the pre-processing step in the time-consuming data matrix of the initial step that is greater than or equal to the corresponding time-consuming data threshold, the initial parameter set is optimized until there is no time-consuming data of the pre-processing step in the time-consuming data matrix of the initial step that is greater than or equal to the corresponding time-consuming data threshold, and the final parameter set matrix of the pre-processing step is obtained; otherwise, there is no need to optimize the initial parameter set; S34, preprocessing the differentiated fiber raw material set according to the final parameter set matrix of the preprocessing step to obtain a differentiated fiber raw material set after preprocessing.

4. The intelligent control method for differential fiber high-density composite yarn processing according to claim 3 is characterized in that: Optimizing the initial parameter set in S33 includes the following steps: S331, setting the initial parameter set to be optimized and the corresponding time-consuming data ; Construct a krill population; Set the maximum number of iterations corresponding to the krill population to , the current number of iterations is ; S332, according to Set the fitness function of the krill population; S333, constructing an initial position matrix according to the initial parameter set to be optimized; S334, start iteration; in each round of iteration, use the fitness function of the krill population to calculate the fitness value of each position in the position matrix updated in the previous round of iteration and update the position of each krill in the krill population; S335, when When , stop the iteration, output the optimized parameter set and measure the time consumption data to obtain the optimized time consumption data; otherwise, continue to iterate until until the time-consuming data after optimization is less than the corresponding time-consuming data threshold, the optimized parameter set is used to replace the corresponding initial parameter set in the initial parameter set matrix of the preprocessing step to obtain the final parameter set matrix of the preprocessing step; otherwise, return to S334 and continue to iterate until the time-consuming data after optimization is less than the corresponding time-consuming data threshold to obtain the final parameter set matrix of the preprocessing step.

5. The intelligent control method for differential fiber high-density composite yarn processing according to claim 4 is characterized in that: The S4 comprises the following steps: S41, collecting a plurality of differentiated fiber high-density composite yarn samples and corresponding color grade data to obtain a composite yarn sample set and a color grade data set; obtaining image data of each composite yarn sample in the composite yarn sample set to obtain a yarn image data set; S42, constructing an initial CNN neural network classification model; using the yarn image dataset and the color grade dataset to train and test the initial CNN neural network classification model; after the training and testing are completed, obtaining the final CNN neural network classification model.

6. The intelligent control method for differential fiber high-density composite yarn processing according to claim 5 is characterized in that: The S5 comprises the following steps: S51, setting a first comprehensive processing step set; performing a first comprehensive processing on the pretreated differentiated fiber raw material set according to the first comprehensive processing step set to obtain a composite yarn to be dyed; setting a color grade requirement for the composite yarn to be dyed after dyeing to obtain color grade requirement data; S52, performing dyeing operation on the composite yarn to be dyed; obtaining the dyed composite yarn after dyeing; collecting image data of the dyed composite yarn to obtain image data of the yarn to be classified; inputting the image data of the yarn to be classified into the final CNN neural network classification model for classification operation, and obtaining color grade data of the yarn to be classified; If the color grade data of the yarn to be classified is the same as the color grade requirement data, the dyeing is successful; if the color grade data of the yarn to be classified is different from the color grade requirement data, the dyeing fails; continue dyeing the dyed composite yarn until the color grade data of the yarn to be classified is the same as the color grade requirement data.

7. A system for realizing the intelligent control method for differential fiber high-density composite yarn processing as described in any one of claims 1 to 6.

Images