Regenerated polyester processing performance analysis method and system based on artificial intelligence

Abstract

The invention discloses a regenerated polyester processing performance analysis method and system based on artificial intelligence, and relates to the field of regenerated polyester processing. According to the scheme, the spinning sample is dyed by adopting the control variable method, the dyeing performance evaluation values of the spinning sample under different influence indexes are obtained, and the neural network model taking the influence index parameters as input data and the dyeing performance evaluation values as output data is constructed; outputting a mapping relationship between the influence index parameter and the dyeing performance evaluation value, and solving an initial value of each coefficient of the linear regression equation by taking an output data group of the influence index parameter and the dyeing performance evaluation value obtained through the mapping relationship as test data, and carrying out iterative adjustment on the equation coefficient by taking a data set collected in the dyeing treatment process as verification data until the dyeing performance evaluation value prediction accuracy is not lower than a set threshold value, completing construction of a linear regression equation, and taking the obtained regression equation as an analysis basis of the dyeing performance.

Description

Technical Field

[0001] The present invention relates to the field of recycled polyester processing, and particularly to a method and system for analyzing the processing performance of recycled polyester based on artificial intelligence. Background Technique

[0002] Recycled polyester, also known as recycled PET fiber, is a synthetic fiber made by recycling waste polyester raw materials. In its production process, there is no need to extract new raw materials, which greatly reduces the environmental load and waste of raw materials. The production process of recycled polyester mainly includes waste recycling, sorting, washing, regeneration, processing, testing and other links. At present, recycled polyester has become one of the most widely used environmentally friendly fibers in the textile industry. It has the characteristics of wear resistance, easy care, breathability, moisture absorption, corrosion resistance, etc., and is widely used in the fields of textiles, clothing, home textiles, automotive interiors, etc.

[0003] The dyeability of recycled polyester is an important performance during its processing and use. At present, relevant research shows that there is a strong linear relationship between the heat color temperature, dyeing time, and fuel consumption during the dyeing process and the dye uptake percentage after the recycled polyester is spun and dyed. In the existing technology, the control variable method is usually used to analyze the relevant relationship, and it can only show the relationship between a single influencing factor and the dyeing performance when certain influencing factors are fixed values, and cannot reflect the dyeability of the recycled polyester spun under the combined action of multiple influencing factors. For this reason, we propose a method and system for analyzing the processing performance of recycled polyester based on artificial intelligence. Summary of the Invention

[0004] The main purpose of the present invention is to provide a method and system for analyzing the processing performance of recycled polyester based on artificial intelligence, which can effectively solve the problems in the background technique.

[0005] To achieve the above purpose, the technical solution adopted by the present invention is as follows:

[0006] A method for analyzing the processing performance of recycled polyester based on artificial intelligence includes:

[0007] Obtain a sample of recycled polyester spun to be detected, and use the control variable method to dye the spun sample, and obtain the dyeing performance evaluation value of the spun sample after dyeing under different influencing indicators. The dyeing performance evaluation value is one of the apparent color depth K / S value or the dye uptake percentage of the spun sample, and the influencing indicators include at least one of the heating temperature, heat preservation treatment time, dye dosage per unit weight of the spun, and pH value of the dye bath during the dyeing process;

[0008] The influencing index parameters in the dyeing process of the spinning sample are obtained as input data, and the dyeing performance evaluation value of the sample spinning is used as output data to construct a neural network model. The mapping relationship between the influencing index parameters and the dyeing performance evaluation value is output through the constructed network model: f:(x1,x2,...,x i ) T →ds, where x i It is represented as the i-th type influencing index parameter of the dyeing performance evaluation value, and ds is represented as the dyeing performance evaluation value of the sample spinning;

[0009] A linear regression equation was constructed between the dyeing performance evaluation value Ds of regenerated polyester spinning and the influencing index parameters, where the equation is expressed as follows:

[0010]

[0011] Where β0 is the constant term in the equation; β i It is expressed as the partial regression coefficient of the i-th type of impact indicator; ε is the error term of the equation;

[0012] Obtaining an output data set of influencing index parameters and corresponding sample spinning dyeing performance evaluation values through a mapping relationship, and using the output data set as test data to solve the initial values of various coefficients of the linear regression equation;

[0013] The data set collected during the dyeing process was used as validation data to iteratively adjust the equation coefficients until the prediction accuracy of the dyeing performance evaluation value was no less than the set threshold. The linear regression equation between the dyeing performance evaluation value Ds of the regenerated polyester spinning and the influencing index parameters was obtained. The calculation formula for the prediction accuracy was:

[0014]

[0015] Where p is the prediction accuracy of the linear regression equation; n is the number of times the prediction deviation is within a reasonable threshold range (0, a max ], where a max is the maximum value of the prediction deviation; N represents the total amount of the validation data set, and the maximum value of the prediction deviation a max Determine according to the following calculation formula:

[0016]

[0017] Where, dr j Denotes the validation data of the jth group’s dyeing performance evaluation value; Ds j It is expressed as the calculated data of the dyeing performance evaluation value of the jth group.

[0018] Artificial Intelligence-based Analysis System for the Processing Performance of Recycled Polyester, including an Influence Index Parameter Acquisition Module, a Dyeing Performance Testing Module, a Mapping Relationship Acquisition Module, and a Regression Model Construction Module;

[0019] The dyeing performance testing module is used to obtain the dyeing performance evaluation value after the spinning sample is dyed under different influence indexes. Among them, the dyeing performance evaluation value is any one of the apparent color depth K / S value or the dye uptake percentage of the spinning sample;

[0020] The influence index parameter acquisition module is used to collect the influence index parameters during the dyeing process of the spinning sample;

[0021] The mapping relationship acquisition module is used to construct a neural network model with the influence index parameters during the dyeing process of the spinning sample as input data and the dyeing performance evaluation value of the sample spinning as output data, and output the mapping relationship between the influence index parameters and the dyeing performance evaluation value according to the constructed network model;

[0022] The regression model construction module is used to perform linear regression analysis on the relationship between the dyeing performance evaluation value Ds of recycled polyester spinning and the influence index parameters, and obtain the linear regression equation between the dyeing performance evaluation value Ds and the influence index parameters. The expression of the equation is:

[0023]

[0024] In the formula, β0 is the constant term in the equation; β i represents the partial regression coefficient of the i-th type of influence index; ε is the error term of the equation. Among them, the specific steps of the implementation process of the regression model construction module are:

[0025] S1: Obtain the output data group of the influence index parameters and the corresponding dyeing performance evaluation value of the sample spinning through the mapping relationship;

[0026] S2: Use the output data group as test data to solve the initial values of the coefficients β0, β i , ε of the linear regression equation;

[0027] S3: Use the data group collected during the dyeing process as verification data, and calculate the prediction accuracy of the linear regression equation for the dyeing performance evaluation value Ds when the coefficients β0, β i , ε are the initial values. Among them, the calculation formula of the prediction accuracy is:

[0028]

[0029] In the formula, p represents the prediction accuracy of the linear regression equation; n represents the number of data groups in the verification data where the prediction deviation is within the reasonable threshold range (0, a max , where amax is the maximum value of the prediction deviation; N represents the total amount of the verification data set, and the maximum value a of the prediction deviation max is determined according to the following calculation formula:

[0030]

[0031] In the formula, dr j represents the verification data of the dyeing performance evaluation value of the j-th group; Ds j represents the calculated data of the dyeing performance evaluation value of the j-th group;

[0032] S4: Determine whether the prediction accuracy rate of the linear regression equation for the dyeing performance evaluation value Ds reaches the expected value. When the prediction accuracy rate is less than the set threshold, return to step S1 to obtain the first iteration values of the initial values of the coefficients β0, β i , and ε;

[0033] S5: Calculate the prediction accuracy rate of the linear regression equation for the dyeing performance evaluation value Ds when the coefficients β0, β i , and ε are iteration values. When the prediction accuracy rate is not lower than the set threshold, obtain the linear regression equation between the dyeing performance evaluation value Ds of the regenerated polyester spinning and the influencing index parameters;

[0034] The system further includes a memory, a processor, and an electronic program stored in the memory and capable of running on the processor.

[0035] The present invention has the following beneficial effects

[0036] Compared with the prior art, by obtaining the regenerated polyester spinning samples to be detected, using the method of controlling variables to dye the spinning samples, obtaining the dyeing performance evaluation values of the spinning samples after dyeing treatment under different influencing indicators, taking the influencing indicator parameters in the process of dyeing the spinning samples as input data, and taking the dyeing performance evaluation values of the sample spinning as output data, a neural network model is constructed. By outputting the mapping relationship between the influencing indicator parameters and the dyeing performance evaluation values through the constructed network model, a linear regression equation between the dyeing performance evaluation value Ds of the regenerated polyester spinning and the influencing indicator parameters is constructed. Through the mapping relationship, an output data set of the influencing indicator parameters and the corresponding dyeing performance evaluation values of the sample spinning is obtained. Taking the output data set as test data to solve the initial values of the coefficients of the linear regression equation, and taking the data set collected during the dyeing process as verification data to iteratively adjust the equation coefficients. When the prediction accuracy of the dyeing performance evaluation value is not lower than the set threshold, a linear regression equation between the dyeing performance evaluation value Ds of the regenerated polyester spinning and the influencing indicator parameters is obtained. Through the obtained regression equation, the dyeing performance of the regenerated polyester spinning under the combined action of various influencing factors can be reflected. By precisely controlling each influencing factor during the dyeing process, the dyeing treatment process flow of the regenerated polyester spinning is optimized. Description of the Drawings

[0037] Figure 1 It is a schematic flow chart of the method for analyzing the processing performance of regenerated polyester based on artificial intelligence according to the present invention;

[0038] Figure 2 It is a structural block diagram of the system for analyzing the processing performance of regenerated polyester based on artificial intelligence according to the present invention;

[0039] Figure 3 It is a schematic diagram of the coefficient iteration process of the linear regression equation. Detailed Embodiments

[0040] The following further describes the present invention in conjunction with the detailed embodiments. Among them, the drawings are only for illustrative purposes, showing only schematic diagrams, not physical diagrams, and should not be construed as limiting the present invention. In order to better illustrate the detailed embodiments of the present invention, some components in the drawings will be omitted, enlarged or reduced, and do not represent the dimensions of the actual product.

[0041] Example 1

[0042] As Figures 1-3 shown, the specific implementation process of the technical solution of the present invention includes the following steps:

[0043] Step 1: Obtain the regenerated polyester spinning samples to be detected, use the method of controlling variables to dye the spinning samples, and obtain the dyeing performance evaluation values of the spinning samples after dyeing treatment under different influencing indicators;

[0044] Among them, the influencing indicators include at least one of the heating temperature during the dyeing process, the heat preservation treatment time, the dye dosage per unit weight of the spinning, and the pH value of the dye bath; the evaluation value of the dyeing performance is one of the apparent color depth K / S value or the percentage of dye uptake of the spinning sample. Among them,

[0045] The apparent color depth K / S value can be obtained by testing with a colorimetric and color matching instrument;

[0046] The percentage of dye uptake value can be obtained by dissolving the disperse dyes in the dyeing stock solution and the dyeing residual solution during the dyeing process with DMF respectively, testing the absorbance of the solution at the maximum wavelength of the dye by an ultraviolet-visible-near-infrared spectrophotometer, and calculating the percentage of dye uptake M according to the following formula. The formula is:

[0047]

[0048] In the formula, E0 is the absorbance of the dyeing stock solution; E k is the absorbance of the dyeing residual solution; n0 is the dilution multiple of the dyeing stock solution; n k is the dilution multiple of the dyeing residual solution;

[0049] Step 2: Obtain the influencing index parameters during the dyeing treatment process of the spinning sample as input data, and use the evaluation value of the dyeing performance of the sample spinning as output data to construct a neural network model, and output the mapping relationship between the influencing index parameters and the evaluation value of the dyeing performance through the constructed network model: f:(x1,x2,...,x i ) T →ds, where, x i represents the i-th type of influencing index parameter of the dyeing performance evaluation value, and ds represents the evaluation value of the dyeing performance of the sample spinning;

[0050] In this embodiment, taking the constructed network model as an example of a BP neural network model, it is a three-layer structure of an input layer-hidden layer-output layer. Among them, the number of neurons in the input layer is equal to the type of influencing indicators, the number of neurons in the output layer is 1, and the number of neurons s in the hidden layer can be calculated according to the following formula. Specifically:

[0051]

[0052] In the formula, u represents the number of neurons in the input layer, v represents the number of neurons in the output layer. In this model, v = 1, and a is a constant from 1 to 9;

[0053] Step 3: Construct a linear regression equation between the evaluation value Ds of the dyeing performance of the regenerated polyester spinning and the influencing index parameters. Among them, the expression of the equation is:

[0054]

[0055] In the formula, β0 is the constant term in the equation; β i represents the partial regression coefficient of the i-th type of influencing index; ε is the error term of the equation;

[0056] Step 4: Obtain the output data group of the influencing index parameters and the corresponding dyeing performance evaluation values of the sample spinning through the mapping relationship, and use the output data group as the test data to solve the initial values of the coefficients of the linear regression equation;

[0057] Step 5: Use the data group collected during the dyeing process as the verification data, and calculate the prediction accuracy of the linear regression equation for the dyeing performance evaluation value Ds when the coefficients β0, β i , ε are the initial values. Among them, the calculation formula for the prediction accuracy is:

[0058]

[0059] In the formula, p represents the prediction accuracy of the linear regression equation; n represents the number of data groups in the verification data where the prediction deviation is within the reasonable threshold range (0, a max ; where a max is the maximum value of the prediction deviation; N represents the total amount of verification data groups, and the maximum value a max of the prediction deviation is determined according to the following calculation formula:

[0060]

[0061] In the formula, dr j represents the j-th group of verification data of the dyeing performance evaluation value; Ds j represents the calculated data of the j-th group of dyeing performance evaluation value;

[0062] Step 6: Determine whether the prediction accuracy of the linear regression equation for the dyeing performance evaluation value Ds reaches the expected value. When the prediction accuracy is less than the set threshold, return to Step 4 to obtain the first iteration values of the coefficients β0, β i , ε;

[0063] Step 7: Calculate the prediction accuracy of the linear regression equation for the dyeing performance evaluation value Ds when the coefficients β0, β i , ε are the iteration values. When the prediction accuracy is not lower than the set threshold, obtain the linear regression equation between the dyeing performance evaluation value Ds of the regenerated polyester spinning and the influencing index parameters.

[0064] The foregoing has shown and described the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited by the above embodiments, and what is described in the above embodiments and the specification is only to illustrate the principle of the present invention. Without departing from the spirit and scope of the present invention, the present invention will have various changes and improvements, and these changes and improvements fall within the scope of the present invention claimed. The scope of the present invention claimed is defined by the appended claims and their equivalents.

Claims

1. A method for analyzing the processing performance of regenerated polyester based on artificial intelligence, characterized in that, Including: Obtain the regenerated polyester spinning sample to be detected, and use the method of controlling variables to dye the spinning sample, and obtain the dyeing performance evaluation values of the spinning sample after dyeing treatment under different influencing indicators; Obtain the influencing index parameters during the dyeing process of the spun sample as input data, and use the dyeing performance evaluation value of the sample spinning as output data to construct a neural network model. Through the constructed network model, output the mapping relationship between the influencing index parameters and the dyeing performance evaluation value: f:(x1,x2,...,x i ) T →ds, where x i represents the i-th type of influencing index parameter of the dyeing performance evaluation value, and ds represents the dyeing performance evaluation value of the sample spinning; Construct a linear regression equation between the dyeing performance evaluation value Ds of the regenerated polyester spinning and the influencing indicator parameters. Among them, the expression of the equation is: where β0 is the constant term in the equation; β i represents the partial regression coefficient of the i-th type of impact index; ε is the error term of the equation; Obtain the output data group of the influencing indicator parameters and the corresponding dyeing performance evaluation value of the sample spinning through the mapping relationship, and use the output data group as the test data to solve the initial values of the coefficients of the linear regression equation; Use the data group collected during the dyeing process as the verification data to iteratively adjust the equation coefficients. When the prediction accuracy rate of its dyeing performance evaluation value is not lower than the set threshold, obtain the linear regression equation between the dyeing performance evaluation value Ds of the regenerated polyester spinning and the influencing indicator parameters.

2. The method for analyzing the processing performance of recycled polyester based on artificial intelligence according to claim 1, characterized in that The dyeing performance evaluation value is one of the apparent color depth K / S value or the dye uptake percentage of the spinning sample.

3. The method for analyzing the processing performance of recycled polyester based on artificial intelligence according to claim 1, wherein, The calculation formula of the prediction accuracy rate is: Wherein, p represents the prediction accuracy of the linear regression equation; n represents the number of data groups in the verification data where the prediction deviation is within a reasonable threshold range (0, a max , where a max is the maximum value of the prediction deviation; N represents the total amount of verification data groups.

4. The method for analyzing the processing performance of recycled polyester based on artificial intelligence according to claim 3, wherein The maximum value a of the prediction deviation max It is determined according to the following calculation formula: where dr j represents the verification data of the dyeing performance evaluation value for the j-th group; Ds j represents the calculated data of the dyeing performance evaluation value for the j-th group.

5. The method for analyzing the processing performance of recycled polyester based on artificial intelligence according to claim 1, wherein The influencing indicators include at least one of the heating temperature, heat preservation treatment time, dye dosage per unit weight of spinning, and dye bath pH value during the dyeing process.

6. An artificial intelligence-based regenerated polyester processing performance analysis system, characterized in that, Including an influencing indicator parameter acquisition module, a dyeing performance test module, a mapping relationship acquisition module, and a regression model construction module; The dyeing performance test module is used to obtain the dyeing performance evaluation value of the spinning sample after dyeing treatment under different influencing indicators. Among them, the dyeing performance evaluation value is any one of the apparent color depth K / S value or the dye uptake percentage of the spinning sample; The influencing indicator parameter acquisition module is used to collect the influencing indicator parameters during the dyeing process of the spinning sample; The mapping relationship acquisition module is used to construct a neural network model with the influencing indicator parameters during the dyeing process of the spinning sample as the input data and the dyeing performance evaluation value of the sample spinning as the output data, and output the mapping relationship between the influencing indicator parameters and the dyeing performance evaluation value according to the constructed network model; The regression model construction module is used to perform a linear regression analysis on the relationship between the dyeing performance evaluation value Ds of the regenerated polyester spinning and the influencing indicator parameters, and obtain a linear regression equation between the dyeing performance evaluation value Ds and the influencing indicator parameters. The expression of the equation is: In the formula, β0 is the constant term in the equation; β i is expressed as the partial regression coefficient of the i-th type of influence index; ε is the error term of the equation. Among them, the specific steps of the implementation process of the regression model construction module are as follows: S1: Obtain the output data group of the influencing indicator parameters and the corresponding dyeing performance evaluation value of the sample spinning through the mapping relationship; S2: Using the output data set as test data to solve the coefficients β0, β i of the linear regression equation and the initial values of ε; S3: Using the data set collected during the dyeing process as the verification data, calculate the prediction accuracy of the linear regression equation for the dyeing performance evaluation value Ds when the coefficients β0, β i , and ε are the initial values. The calculation formula for the prediction accuracy is as follows: Wherein, p represents the prediction accuracy rate of the linear regression equation; n represents the number of data groups in the verification data where the prediction deviation is within a reasonable threshold range (0, a max , where a max is the maximum value of the prediction deviation; N represents the total amount of verification data groups, and the maximum value a of the prediction deviation max is determined according to the following calculation formula: where dr j represents the verification data for the j-th group of dyeing performance evaluation values; Ds j represents the calculated data for the j-th group of dyeing performance evaluation values; S4: Determine whether the prediction accuracy of the linear regression equation for the dyeing performance evaluation value Ds reaches the expected value. When the prediction accuracy is less than the set threshold, return to step S1 to obtain the first iteration values of the initial values of the coefficients β0, β i , and ε; S5: Calculate the prediction accuracy of the linear regression equation for the dyeing performance evaluation value Ds when the coefficients β0, β i , and ε are iteration values. When the prediction accuracy is not lower than the set threshold, obtain the linear regression equation between the dyeing performance evaluation value Ds of regenerated polyester spinning and the influencing index parameters.

7. The system for analyzing the processing performance of regenerated polyester based on artificial intelligence according to claim 6, wherein, The system further includes a memory, a processor, and an electronic program stored in the memory and capable of running on the processor. Among them, when the processor runs the electronic program, it can implement the steps of the method described in any one of claims 1-5.

Images