Comprehensive performance evaluation and intelligent prediction method for multi-component fiber composite yarn

Abstract

The invention relates to the technical field of performance evaluation of fiber composite yarns, in particular to a comprehensive performance evaluation and intelligent prediction method for multi-component fiber composite yarns. By combining a fuzzy analytic hierarchy process, a fuzzy ideal solution sorting method Fuzzy-AHP-TOPSIS and an adaptive neural fuzzy inference system (ANFIS) algorithm, quantitative evaluation of yarn performance and accurate prediction of key indexes are realized. A Fuzzy-AHP algorithm is utilized to reasonably distribute the weight of each performance index, and a Fuzzy-TOPSIS method is utilized to compare the distance between a sample and an ideal optimal solution and the distance between the sample and a worst solution so as to carry out dynamic sorting. An ANFIS algorithm is adopted to learn a complex nonlinear relation between fiber characteristics and yarn performance, so that a high-precision prediction model is established. The research and development efficiency and scientificity of textile enterprises are improved, an innovative solution is provided for evaluating and predicting the performance of the multi-component fiber composite yarn in an actual scene, and the method has wide application prospects in the fields of textile material research and development and yarn production.

Description

Technical Field

[0001] The present invention relates to the technical field of fiber composite yarn performance evaluation, and particularly to a comprehensive performance evaluation and intelligent prediction method for multi-component fiber composite yarns. Background Art

[0002] In the textile industry, during the research and development and production process of multi-component fiber composite yarns, the evaluation and prediction of yarn performance are of crucial importance. Traditional evaluation methods mainly rely on manual experience and experimental tests. For example, single-factor analysis is carried out on indicators such as roving spacing, traveler weight, yarn twist coefficient, and spinning spindle speed in the spinning process, or empirical formulas are fitted through a large amount of experimental data, or finite element models of fibers and yarns are established using the finite element method for simulation prediction, etc. However, these traditional methods have obvious limitations. For example, it is difficult to comprehensively reflect the comprehensive performance of composite yarns, the ability to handle uncertainty and ambiguity is limited, the data processing efficiency is low and it is difficult to achieve automation and intelligence, and the prediction accuracy is limited.

[0003] With the development of computer technology and artificial intelligence, it has become possible to evaluate and predict yarn performance using intelligent algorithms. Most existing intelligent evaluation systems use a single intelligent algorithm, such as artificial neural network (ANN) or fuzzy logic, to predict performance indicators such as the breaking strength and elongation rate of yarns. However, these studies mostly focus on single performance indicators and require a large amount of training data; the application of fuzzy logic in the foreign textile field mainly focuses on control and optimization, such as yarn tension control and fabric quality evaluation, and there is no research on comprehensively evaluating the performance of multi-component fiber composite yarns.

[0004] The following problems exist in traditional multi-component fiber composite yarn performance evaluation and prediction methods:

[0005] (1) It is impossible to comprehensively reflect the comprehensive performance of composite yarns and difficult to achieve accurate performance prediction. Traditional methods only focus on single indicators and cannot comprehensively reflect the comprehensive performance of composite yarns, resulting in inaccurate evaluation results. For example, the single-factor analysis method ignores the interaction between various factors and it is difficult to obtain the optimal combination of process parameters.

[0006] (2) The ability to handle uncertainty and ambiguity is limited and it is difficult to adapt to the complex situation of multi-component fiber composite yarns. The performance of textile materials is affected by multiple factors, and the relationships between these factors are complex and non-linear. Traditional methods are difficult to describe with precise mathematical models, resulting in insufficient reliability of evaluation results. For example, in yarn quality evaluation, traditional methods are difficult to handle fuzzy and uncertain information.

[0007] (3) Low data processing efficiency, making it difficult to achieve automation and intelligence. In the research and development of multi-component fiber composite yarns, a large amount of experimental data needs to be analyzed and processed. Traditional manual methods are inefficient and prone to errors, making it difficult to achieve automation and intelligence.

[0008] (4) The prediction accuracy is limited and cannot accurately capture the complex nonlinear relationship between the performance of fiber composite yarns. The performance of multi-component fiber composite yarns has a complex nonlinear relationship with the fiber characteristics and spinning process. Traditional linear regression or statistical methods are difficult to accurately capture this relationship, resulting in low prediction accuracy.

[0009] In view of this, there is an urgent need for a comprehensive performance evaluation and intelligent prediction method to achieve accurate evaluation and intelligent prediction of multi-component fiber composite yarns. Summary of the Invention

[0010] In view of the shortcomings of the above-mentioned prior art and in order to solve the above-mentioned technical problems, the present invention discloses a method for comprehensive performance evaluation and intelligent prediction of multi-component fiber composite yarns. The intelligent system is based on the Fuzzy-AHP-TOPSIS (Fuzzy Hierarchy Analysis Method-Ideal Solution Sorting Method) algorithm and the ANFIS (Adaptive Neuro-Fuzzy Inference System) algorithm, which fully utilizes the advantages of the intelligent algorithm to realize the quantitative performance evaluation of multi-component fiber composite yarns and the accurate prediction of key indicators.

[0011] The present invention adopts the following technical solutions:

[0012] In a first aspect, the present application provides a method for comprehensive performance evaluation and intelligent prediction of a multi-component fiber composite yarn, the method comprising:

[0013] Step 1: Construct a structured database containing different composite yarn sample schemes and sample performance levels;

[0014] The yarn sample scheme includes fiber materials and their mixing ratios in the yarn; the yarn sample properties include yarn mechanical property parameters and yarn production process parameters;

[0015] Step 2: Set weights for yarn key performance indicators based on expert prior fuzzy ratings;

[0016] The expert prior fuzzy rating is defined as a method for assigning weights to key performance indicators of yarns by combining expert experience and fuzzy mathematics theory. The core is to quantify the fuzzy judgment of the experts on the importance of the indicators in a structured way;

[0017] Step 3: Evaluate the sample performance by the Technique for Order Preference by Similarity to an Ideal Solution (TOPSIS) and output the sample ranking results to obtain the optimal formulation plan; the Technique for Order Preference by Similarity to an Ideal Solution (TOPSIS) is defined as a multi-attribute decision-making analysis method based on fuzzy mathematics, which comprehensively evaluates the sample performance and outputs the ranking results by comparing each sample with the ideal solution; the ideal solution is defined as the optimal performance standard.

[0018] Step 4: Through the Adaptive Neuro-Fuzzy Inference System (ANFIS) algorithm, automatically learn the non-linear relationship between fiber characteristics and yarn performance indicators, establish a high-precision prediction model, and use the trained model to predict the performance of the yarn samples to be tested.

[0019] In one embodiment, the comprehensive performance evaluation and intelligent prediction method for a multi-component fiber composite yarn further includes Step 5: Visualize and display the comparison between the dynamic regression graph and the prediction results.

[0020] In one embodiment, after Step 1, it includes preprocessing the data and storing the associated data of fiber ratio and performance in a relational database.

[0021] In one embodiment, the preprocessing of the data in Step 1 includes normalization, missing value processing, and consistency verification.

[0022] In one embodiment, the key performance indicators of the yarn in Step 2 include: breaking strength, elongation at break, evenness, hairiness index, and flame retardant performance.

[0023] In one embodiment, the process of setting weights in Step 2 is as follows: construct the hierarchical structure of yarn selection, convert the expert language evaluation into triangular fuzzy numbers, compare the criteria pairwise, output the fuzzy judgment matrix, and calculate the fuzzy weight vector of each criterion using the geometric mean method.

[0024] In one embodiment, the construction of the hierarchical structure of yarn selection in Step 2 includes an input of the target layer, criterion layer, and scheme layer for selecting the best blended yarn, and an output of a tree-shaped hierarchical diagram.

[0025] In one embodiment, in Step 3, the sample ranking results are reflected by the proximity coefficient, and the proximity coefficient is defined as the degree of proximity of the sample alternative solutions to the best and worst solution.

[0026] In one embodiment, the algorithm process in step 3 is as follows: construct a fuzzy decision matrix by inputting expert scores in the form of fuzzy numbers, normalize the fuzzy decision matrix, determine the weights of each evaluation criterion by the fuzzy analytic hierarchy process, multiply the normalized fuzzy decision matrix by the weight matrix to obtain a weighted normalized fuzzy decision matrix, calculate the distance of each alternative using the Euclidean distance formula of fuzzy numbers, sort the alternatives according to the proximity coefficient to obtain an evaluation ranking result, and select the alternative with the largest proximity coefficient as the optimal solution; the calculation formula for the proximity coefficient Ci is as follows:

[0027] ;

[0028] wherein, is defined as the distance between the sample and the worst solution, is defined as the distance between the sample and the ideal solution.

[0029] In one embodiment, the algorithm process of the adaptive neuro-fuzzy inference system in step 4 is as follows:

[0030] Define a fuzzy logic system: determine the fiber characteristics and yarn performance characteristics as input variables and output variables respectively, and define fuzzy sets and corresponding fuzzy rules for each input variable;

[0031] Construct a neural network structure: the neural network structure consists of an input layer, a fuzzification layer, a rule layer, a normalization layer, a defuzzification layer and an output layer;

[0032] Train the network: enhance the robustness of the data through Bootstrap sampling, calculate the node output using forward propagation, and adjust the network parameters through the error backpropagation algorithm to minimize the error, and iterate and optimize until the error converges.

[0033] In a second aspect, an embodiment of the present application provides a multi-component fiber composite yarn comprehensive performance evaluation and intelligent prediction system, and the system includes:

[0034] A data acquisition module, used to construct a structured database including different composite yarn sample schemes and sample performance grades; the yarn sample scheme includes fiber materials and their mixing ratios in the yarn; the yarn sample performance includes yarn mechanical property parameters and yarn production process parameters;

[0035] A weight setting module, used to set weights for the key performance indicators of the yarn according to the expert's prior fuzzy ratings;

[0036] A performance evaluation module, used to evaluate the sample performance by the fuzzy TOPSIS method and output the sample ranking result to obtain the optimal formulation scheme;

[0037] A performance prediction module, which is used to automatically learn the non-linear relationship between fiber characteristics and yarn performance indicators through the Adaptive Neuro-Fuzzy Inference System (ANFIS) algorithm, establish a high-precision prediction model, and use the trained model to predict the performance of new yarn samples;

[0038] A visualization module, which is used to visually display the comparison between the dynamic regression graph and the prediction results.

[0039] In a third aspect, an embodiment of the present application provides a control device, which includes a processor and a memory. The memory is used to store programs, instructions or codes, and the processor is used to execute the programs, instructions or codes in the memory to complete the above-mentioned comprehensive performance evaluation and intelligent prediction method for multi-component fiber composite yarns.

[0040] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium, which stores a computer program. The computer program is loaded by a processor to execute the above-mentioned comprehensive performance evaluation and intelligent prediction method for multi-component fiber composite yarns.

[0041] Compared with the prior art, the present invention has significant advantages in terms of the comprehensiveness of evaluation indicators, the ability to handle uncertainties, prediction accuracy, and automation level, etc.; compared with single intelligent algorithm methods, the present invention has advantages in terms of the comprehensiveness of functions and the integrity of the system; compared with existing combined methods, the present invention has advantages in terms of innovation and practicality; compared with existing artificial intelligence applications in the textile field, the present invention has advantages in terms of the depth and breadth of application. A comprehensive performance evaluation and intelligent prediction method for multi-component fiber composite yarns provided by the present invention can comprehensively consider multiple performance indicators of multi-component fiber composite yarns through the Fuzzy-AHP-TOPSIS (Fuzzy Analytic Hierarchy Process - Technique for Order Preference by Similarity to Ideal Solution) algorithm, achieve a comprehensive performance evaluation, effectively handle the uncertainties and subjectivities in the evaluation process, and improve the reliability and consistency of evaluation results. At the same time, through the ANFIS (Adaptive Neuro-Fuzzy Inference System) algorithm, it can automatically learn the complex non-linear relationship between fiber characteristics and yarn performance indicators, establish a high-precision prediction model, real-time predict the key performance indicators of the yarn, and visually display the comparison between the dynamic regression graph and the prediction results. It has the following technical effects:

[0042] 1. Combining the dual-engine drive mode of the Fuzzy-AHP-TOPSIS algorithm and the ANFIS algorithm to achieve dual functions of performance evaluation and prediction, which can not only comprehensively evaluate the performance of existing yarns but also accurately predict the performance of new formulations.

[0043] 2. High degree of automation and intelligence. Different composite yarn sample solutions can be quickly constructed through the built-in fiber database, and performance indicators can be automatically calculated and predicted through the Fuzzy-AHP-TOPSIS algorithm and the ANFIS algorithm, greatly improving the R & D efficiency and scientific nature.

[0044] 3. Introduce expert prior fuzzy rating, assign importance levels to key performance indicators, and provide a basis for subsequent intelligent evaluation, making the evaluation process more scientific and reasonable.

[0045] 4. The ANFIS algorithm has the ability of adaptive learning, can continuously adjust and optimize the model according to new data, and improve the accuracy and adaptability of prediction. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Figure 1 Schematic diagram of a comprehensive performance evaluation and intelligent prediction method for multi-component fiber composite yarns provided by an embodiment of the present invention;

[0047] Figure 2 Schematic diagram of the process of a comprehensive performance evaluation and intelligent prediction system for multi-component fiber composite yarns provided by an embodiment of the present invention;

[0048] Figure 3 Schematic diagram of a hierarchical structure of yarn selection provided by an embodiment of the present invention;

[0049] Figure 4 Schematic diagram of a Fuzzy-AHP-TOPSIS process provided by an embodiment of the present invention;

[0050] Figure 5 Schematic diagram of an ANFIS adaptive neuro-fuzzy inference algorithm provided by an embodiment of the present invention;

[0051] Figure 6 Schematic diagram of the prediction fitting of an ANFIS adaptive neuro-fuzzy inference system provided by an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0052] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments.

[0053] The relevant concept definitions are as follows:

[0054] 1. Expert prior fuzzy rating is a method of combining expert experience and fuzzy mathematics theory to assign weights to the key performance indicators of yarns. Its core is to quantify the fuzzy judgment of experts on the importance of indicators in a structured manner.

[0055] 2. The Technique for Order Preference by Similarity to Ideal Solution (TOPSIS) is a multi-attribute decision-making analysis method based on fuzzy mathematics. It comprehensively evaluates the performance of samples and outputs a ranking result by comparing each sample with the ideal solution (i.e., the optimal performance standard). The steps and related definitions of the TOPSIS are as follows:

[0056] Determine evaluation indicators: First, determine each indicator used to evaluate the performance of samples, such as breaking strength, elongation at break, etc.

[0057] Establish the ideal solution: According to the attribute of the indicator, set the ideal value of each performance indicator, that is, the optimal performance standard.

[0058] Fuzzification: Make a fuzzy comparison between the actual performance indicators of each sample and the ideal solution, calculate the similarity or closeness between them, and use the membership function to quantify this similarity.

[0059] Comprehensive evaluation: Conduct a comprehensive evaluation of all indicators of each sample to obtain the comprehensive performance score of each sample.

[0060] Ranking: Rank the samples according to the comprehensive score. The sample with a higher score has a performance closer to the ideal solution, which is the better formulation scheme.

[0061] Select the optimal formulation: Select the sample with the optimal performance from the ranking result as the optimal formulation scheme.

[0062] A method for comprehensively evaluating and intelligently predicting the comprehensive performance of a multi-component fiber composite yarn disclosed in the present invention, as Figure 1-2 shown, the method includes:

[0063] Step 1: Construct a structured database containing different composite yarn sample schemes and sample performance grades;

[0064] The yarn sample scheme includes fiber materials and their mixing ratios in the yarn; the yarn sample performance includes yarn mechanical property parameters and yarn production process parameters;

[0065] Integrate experimental tests (such as mechanical property tests), production process parameters (such as fiber ratio, twist), and relevant literature data to establish a structured database containing more than 30 key indicators. Perform preprocessing on the data, including normalization, missing value processing, and consistency verification, to provide a standardized input for subsequent dynamic evaluation and intelligent prediction.

[0066] Step 2: Set weights for the key performance indicators of the yarn according to the expert's prior fuzzy rating;

[0067] The Fuzzy-AHP (Fuzzy Analytic Hierarchy Process) is used to determine the weights of various performance indicators. The key performance indicators of yarn include breaking strength, elongation at break, evenness, hairiness index, and flame retardancy performance, etc., to ensure that it meets the design requirements and actual application needs.

[0068] Figure 4 The working process of Fuzzy AHP-TOPSIS (Fuzzy Analytic Hierarchy Process - Technique for Order Preference by Similarity to Ideal Solution) is detailed, including the following key steps:

[0069] 2.1 Hierarchical structure construction

[0070] The hierarchical structure construction for yarn selection is carried out. As Figure 3 shown, the hierarchical structure construction for yarn selection includes the goal layer, criterion layer, and scheme layer with the goal of selecting the best blended yarn as the input, and the output is a tree-like hierarchical diagram.

[0071] 2.2 Definition of linguistic variables

[0072] Next, the definition of linguistic variables is carried out, and the expert language evaluation is transformed into triangular fuzzy numbers, which can be described as follows:

[0073]

[0074] Among them l and u respectively represent the lower and upper limits of the fuzzy number, m is the median.

[0075] 2.3 Construction of fuzzy judgment matrix

[0076] After obtaining the fuzzy numbers, the criteria are compared pairwise, and the fuzzy judgment matrix is output, which can be described as follows:

[0077]

[0078] Among them , n represents the number of criteria.

[0079] 2.4 Calculation of fuzzy weight vector

[0080] Subsequently, the geometric mean method is used to calculate the fuzzy weight vector of each criterion, which can be expressed by the following formula:

[0081]

[0082]

[0083] Among them, w i represents the iThe fuzzy weights of the criteria are expressed as triangular fuzzy numbers. The fuzzy analytic hierarchy process can handle the uncertainty and fuzziness of the evaluation indicators, making the weight distribution more reasonable.

[0084] Step 3: Evaluate sample performance through the algorithm and output the sample ranking results to obtain the optimal formulation solution;

[0085] The sample ranking results are reflected by the closeness coefficient, which is defined as the closeness of the sample alternatives to the best and worst solutions.

[0086] 3.1 Construction and normalization of fuzzy decision matrix

[0087] The fuzzy decision matrix is constructed by inputting expert scores in the form of fuzzy numbers. The process can be described as follows:

[0088] ;

[0089] Among them, is the fuzzy evaluation value of the ith option under the jth criterion, m represents the number of alternative options, and n represents the number of evaluation indicators; each element represents the fuzzy score of the ith option under the jth indicator.

[0090] In order to make the evaluation criteria of different dimensions comparable, it is necessary to normalize the fuzzy decision matrix. For the benefit criterion (the bigger the better), the following formula can be used:

[0091]

[0092] in It is j The maximum fuzzy value under the criterion. For the cost criterion (the smaller the better), the following formula can be used:

[0093]

[0094] in, It is j The minimum fuzzy value under the criteria.

[0095] 3.2 Weighted Normalized Fuzzy Decision Matrix Calculation

[0096] Multiply the normalized fuzzy decision matrix by the weight matrix to obtain the weighted normalized fuzzy decision matrix. The whole process is described as follows:

[0097]

[0098] 3.3 Determination of ideal solution and negative ideal solution

[0099] To determine the ideal solution of each alternative, the FPIS is used to represent the best value under each criterion, and the FNIS is used to represent the worst value under each criterion, which can be expressed by the following formula:

[0100]

[0101]

[0102] Wherein, and are the maximum and minimum values under the n -th criterion, respectively.

[0103] The Euclidean distance formula of fuzzy numbers is used to calculate the distance of each alternative between the FPIS and the FNIS, and the process is described as follows:

[0104]

[0105]

[0106] These two distances respectively reflect the closeness of the alternative to the best and worst solutions.

[0107] 3.4 Calculation and Sorting of Closeness Coefficient

[0108] Closeness Coefficient C i is a key index to measure the advantages and disadvantages of alternatives, and the calculation formula is as follows:

[0109] ;

[0110] Wherein, is defined as the distance between the sample and the worst solution, is defined as the distance between the sample and the ideal solution; the distance between the sample and the ideal solution The smaller it is, the closer the sample performance is to the ideal state, and the distance between the sample and the worst solution The larger it is, the farther the sample performance is from the worst state.

[0111] The larger the closeness coefficient is, the closer the alternative is to the ideal solution. Finally, the alternatives are sorted according to the closeness coefficient, and the alternative with the largest proximity coefficient is selected as the optimal solution.

[0112] The expert scores are input in the form of fuzzy numbers to construct a fuzzy decision matrix, which is normalized. After the weights of each evaluation criterion are determined by the fuzzy hierarchical analysis method, the normalized fuzzy decision matrix is multiplied by the weight matrix to obtain a weighted normalized fuzzy decision matrix. The Euclidean distance formula of fuzzy numbers is used to calculate the distance of each alternative solution. The samples are ranked by comparing their distances to the ideal best solution and the worst solution. The alternative solutions are ranked according to the proximity coefficient to obtain the evaluation ranking results, and the solution with the largest proximity coefficient is selected as the optimal solution.

[0113] The Fuzzy-AHP-TOPSIS model uses Fuzzy-AHP (Fuzzy Analytic Hierarchy Process) to determine the weights of various performance indicators. After obtaining the weights of each criterion, the fuzzy ideal solution ranking method (Fuzzy-TOPSIS) is applied to dynamically sort the yarn samples.

[0114] Step 4: Through the ANFIS adaptive neural fuzzy inference system algorithm, the nonlinear relationship between fiber characteristics and yarn performance indicators is automatically learned, a high-precision prediction model is established, and the trained model is used to predict the performance of the yarn samples to be tested.

[0115] Figure 5 The structure and workflow of ANFIS (Adaptive Neuro-Fuzzy Inference System) are presented, including the following key parts:

[0116] 4.1 Definition of Fuzzy Logic System

[0117] When studying the impact of yarn composition on yarn performance, we can use fuzzy sets and fuzzy rules to establish relationships between input and output variables. Input variables include yarn composition (such as fiber type and fiber ratio), while output variables include yarn performance indicators such as breaking strength and flame retardancy. For each input variable, we define a corresponding fuzzy set and formulate fuzzy rules based on these sets. These fuzzy rules enable a more intuitive understanding and prediction of the impact of fiber characteristics on yarn performance, providing a theoretical basis for yarn design and production.

[0118] 4.2 Neural Network Structure Construction

[0119] When constructing a neural network architecture for predicting yarn performance indicators, we designed a system consisting of multiple functional layers. The network's input layer receives the yarn composition as input variables. Next, the fuzzification layer maps the input variables into predefined fuzzy sets and calculates the membership degree for each input value, transforming the precise input data into a fuzzy representation. Subsequently, the rule layer calculates the activation strength of each rule based on predefined fuzzy rules. These rules derive fuzzy outputs of yarn performance indicators based on the fuzzy sets of the input variables. To ensure that the rule activation strengths are within a reasonable range, the normalization layer normalizes these activation strengths for comparability and consistency. The defuzzification layer then converts the fuzzy outputs into specific numerical values for practical application and interpretation. Finally, the output layer outputs the processed results as predicted yarn performance indicators, providing a quantitative basis for yarn design and performance evaluation. This architecture combines the learning capabilities of neural networks with the processing power of fuzzy logic to effectively handle the ambiguity and uncertainty of input variables, thereby achieving accurate prediction of yarn performance indicators.

[0120] 4.3 Network training process

[0121] During neural network training, data is first sampled using bootstrap sampling. This method generates multiple subsets of the original dataset with replacement, enhancing data diversity and robustness, effectively alleviating issues caused by insufficient or unevenly distributed data. The forward propagation phase then begins, with input data passing through layers, including the fuzzification layer, regularization layer, normalization layer, and defuzzification layer. The output value of each node is calculated, ultimately resulting in the predicted yarn performance indicators at the output layer. Backpropagation then occurs, propagating the error between the predicted and true values from the output layer back to the input layer. By calculating the error gradient, network parameters (such as weights and biases) are adjusted to gradually reduce the difference between the predicted and true values. This process continues iteratively until the error converges to an acceptable range or the preset number of iterations is reached, completing network training and enabling the model to accurately predict yarn performance indicators.

[0122] 4.4 Model Evaluation and Prediction

[0123] In the model development process, model evaluation and prediction are crucial steps. First, conduct a comprehensive evaluation of the trained model. By using methods such as independent validation datasets or cross-validation, examine the performance of the model on unseen data, thus ensuring its good generalization ability to accurately handle new and unknown input data. Subsequently, use the trained model to predict the performance of new yarn samples, analyze the deviation between the prediction results and the actual values, and evaluate the accuracy and reliability of the prediction to verify the effectiveness of the model in practical applications.

[0124] In one embodiment, a comprehensive performance evaluation and intelligent prediction method for multi-component fiber composite yarns further includes step 5: visually display the comparison between the dynamic regression graph and the prediction results.

[0125] To more intuitively display the model performance, adopt visualization means to draw a dynamic regression graph, compare the prediction results with the true values, and show the fitting effect and prediction accuracy of the model through intuitive images to help users better understand the performance and application value of the model.

[0126] Figure 6 The figure shows the results of predicting the flame retardancy grade of multi-component fiber composite yarns by the ANFIS (Adaptive Neuro-Fuzzy Inference System) model. The horizontal axis represents the actual flame retardancy grade, and the vertical axis represents the flame retardancy grade predicted by the model. The data points represent the prediction results of the model for each test case, and the dashed line represents the ideal prediction line, that is, the case where the predicted value is exactly the same as the actual value. Figure 6 The R2 value in the figure is 0.9035, and the accuracy reaches more than 90%, indicating that the model has high prediction accuracy and further verifying the stability and reliability of the model.

[0127] Through the Fuzzy-AHP-TOPSIS (Fuzzy Analytic Hierarchy Process - Technique for Order Preference by Similarity to Ideal Solution) algorithm, the present invention can comprehensively consider multiple performance indicators of multi-component fiber composite yarns, achieve a comprehensive performance evaluation, effectively handle the uncertainty and subjectivity in the evaluation process, and improve the reliability and consistency of the evaluation results. At the same time, through the ANFIS (Adaptive Neuro-Fuzzy Inference System) algorithm, it can automatically learn the complex non-linear relationship between yarn compatibility and yarn performance indicators, establish a high-precision prediction model, real-time predict the key performance indicators of the yarn, and visually display the comparison of the dynamic regression graph and the prediction results. Compared with traditional methods, the present invention has significant advantages in terms of the comprehensiveness of evaluation indicators, the ability to handle uncertainty, prediction accuracy, and automation level; compared with single intelligent algorithm methods, these single algorithms may not be as effective as the dual-engine algorithm of the present invention in dealing with complex non-linear relationships and uncertainties. The present invention has advantages in terms of the comprehensiveness of functions and the integrity of the system; compared with existing combined methods, the present invention has advantages in terms of innovation and practicality; compared with existing artificial intelligence applications in the textile field, the present invention has advantages in terms of the depth and breadth of application.

[0128] In one embodiment, a comprehensive performance evaluation and intelligent prediction system for multi-component fiber composite yarns is provided, and the system includes:

[0129] A data acquisition module, used to construct a structured database containing different composite yarn sample schemes and sample performance grades; the yarn sample scheme includes fiber materials and their mixing ratios in the yarn; the yarn sample performance includes yarn mechanical property parameters and yarn production process parameters;

[0130] A weight setting module, used to set weights for the key performance indicators of the yarn according to the prior fuzzy ratings of experts;

[0131] A performance evaluation module, used to evaluate the sample performance through an algorithm and output the sample ranking results to obtain the optimal formulation scheme;

[0132] A performance prediction module, used to automatically learn the non-linear relationship between fiber characteristics and yarn performance indicators through the Adaptive Neuro-Fuzzy Inference System algorithm, establish a high-precision prediction model, and use the trained model to predict the performance of new yarn samples;

[0133] A visualization module, used to visually display the comparison of the dynamic regression graph and the prediction results.

[0134] Those skilled in the art can clearly understand that for the convenience and conciseness of description, the specific working processes of the above-described system, system, and units can refer to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0135] When the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on such an understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of this technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in various embodiments of this application. The foregoing storage medium includes: various media that can store program codes, such as USB flash drives, mobile hard disks, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical discs.

[0136] As described above, the above is only a preferred specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. The substitution may be the substitution of part of the structure, device, or method steps, or a complete technical solution. Any equivalent substitution or change made according to the technical solution of the present invention and its inventive concept should be covered within the protection scope of the present invention.

Claims

1. A comprehensive performance evaluation and intelligent prediction method for multi-component fiber composite yarns, characterized in that, The method includes: Step 1: Construct a structured database containing different composite yarn sample schemes and sample performance grades; The yarn sample scheme includes fiber materials and their mixing ratios in the yarn; the yarn sample performance includes yarn mechanical property parameters and yarn production process parameters; Step 2: Set weights for key yarn performance indicators according to the expert prior fuzzy rating; the expert prior fuzzy rating is defined as a method of allocating weights to key yarn performance indicators by combining expert experience and fuzzy mathematics theory. The core is to quantify the fuzzy judgment of experts on the importance of indicators in a structured way; Step 3: Evaluate the sample performance by the Technique for Order Preference by Similarity to an Ideal Solution (TOPSIS) method and output the sample ranking result to obtain the optimal formulation scheme; the TOPSIS method is defined as a multi-attribute decision-making analysis method based on fuzzy mathematics. By comparing each sample with the ideal solution, the sample performance is comprehensively evaluated and the ranking result is output; the ideal solution is defined as the optimal performance standard; Step 4: Through the Adaptive Neuro-Fuzzy Inference System (ANFIS) algorithm, automatically learn the non-linear relationship between fiber characteristics and yarn performance indicators, establish a high-precision prediction model, and use the trained model to predict the performance of the yarn samples to be measured.

2. The comprehensive performance evaluation and intelligent prediction method for a multi-component fiber composite yarn according to claim 1, characterized in that The method further includes Step 5: Visualize and display the comparison between the dynamic regression graph and the prediction result.

3. The comprehensive performance evaluation and intelligent prediction method for a multi-component fiber composite yarn according to claim 1, characterized in that After Step 1, it includes preprocessing the data and storing the associated data of fiber ratio and performance in a relational database; the preprocessing of the data includes normalization, missing value processing, and consistency verification.

4. A comprehensive performance evaluation and intelligent prediction method for a multi-component fiber composite yarn according to claim 1, characterized in that, The key yarn performance indicators in Step 2 include: breaking strength, elongation at break, evenness, hairiness index, and flame retardancy performance; the process of setting weights in Step 2 includes: constructing a hierarchical structure for yarn selection, converting expert language evaluations into triangular fuzzy numbers, comparing the criteria pairwise, outputting a fuzzy judgment matrix, and calculating the fuzzy weight vector of each criterion using the geometric mean method.

5. The comprehensive performance evaluation and intelligent prediction method for a multi-component fiber composite yarn according to claim 1, characterized in that The construction of the hierarchical structure for yarn selection in Step 2 includes an input of the target layer, criterion layer, and scheme layer for selecting the best blended yarn, and an output of a tree-like hierarchical graph.

6. The comprehensive performance evaluation and intelligent prediction method for a multi-component fiber composite yarn according to claim 1, characterized in that The process in step 3 includes: constructing a fuzzy decision matrix by inputting expert scores in the form of fuzzy numbers, normalizing the fuzzy decision matrix, determining the weights of each evaluation criterion through the fuzzy analytic hierarchy process, multiplying the normalized fuzzy decision matrix by the weight matrix to obtain a weighted normalized fuzzy decision matrix, calculating the distance of each alternative using the Euclidean distance formula of fuzzy numbers, sorting the alternatives according to the proximity coefficient to obtain an evaluation ranking result, and selecting the alternative with the largest proximity coefficient as the optimal solution; the calculation of the proximity coefficient C i The calculation formula is as follows: ； Among them, is defined as the distance between the sample and the worst solution, is defined as the distance between the sample and the ideal solution.

7. A comprehensive performance evaluation and intelligent prediction method for a multi-component fiber composite yarn according to claim 1, characterized in that, The process of the ANFIS algorithm in Step 4 includes: Defining a fuzzy logic system: determining fiber characteristics and yarn performance characteristics as input variables and output variables respectively, and defining fuzzy sets and corresponding fuzzy rules for each input variable; Constructing a neural network structure: The neural network structure consists of an input layer, a fuzzification layer, a rule layer, a normalization layer, a defuzzification layer, and an output layer; Training the network: Enhance the robustness of the data through Bootstrap sampling, calculate the node output using forward propagation, and adjust the network parameters through the error backpropagation algorithm to minimize the error, and iterate and optimize until the error converges.

8. A comprehensive performance evaluation and intelligent prediction system for multi-component fiber composite yarns, which is used to implement the comprehensive performance evaluation and intelligent prediction method for multi-component fiber composite yarns according to any one of claims 1-7, characterized in that, The system includes: A data acquisition module for constructing a structured database containing different composite yarn sample schemes and sample performance grades; the yarn sample scheme includes fiber materials and their mixing ratios in the yarn; the yarn sample performance includes yarn mechanical property parameters and yarn production process parameters; A weight setting module for setting weights for key yarn performance indicators according to the expert prior fuzzy rating; A performance evaluation module, which is used to evaluate the performance of samples by the Technique for Order of Preference by Similarity to Ideal Solution (TOPSIS) and output the sample ranking results to obtain the optimal formulation scheme; A performance prediction module, which is used to automatically learn the non-linear relationship between fiber characteristics and yarn performance indicators through the Adaptive Neuro-Fuzzy Inference System (ANFIS) algorithm, establish a high-precision prediction model, and use the trained model to predict the performance of new yarn samples; A visualization module, which is used to visually display the comparison between the dynamic regression graph and the prediction results.

9. A control device, characterized in that, It includes a processor and a memory. The memory is used to store programs, instructions or codes, and the processor is used to execute the programs, instructions or codes in the memory to complete the comprehensive performance evaluation and intelligent prediction method of multi-component fiber composite yarns as described in any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, A computer program is stored, and the computer program is loaded through a processor to execute the comprehensive performance evaluation and intelligent prediction method of multi-component fiber composite yarns as described in any one of claims 1-7.

Images