Preparation process and comprehensive quality detection method of high-performance textile fiber material

Abstract

The application belongs to the technical field of fiber material preparation, and discloses a preparation process and a comprehensive quality detection method of high-performance textile fiber material, which comprises the following steps: a. raw material treatment, to prepare functional premix; the material aggregation tension is monitored in real time through a tension sensor during the mixing process; b. melt blending: after extrusion, granulation and grading screening, a functional master batch is prepared, and the melt material flow tension is monitored through a built-in tension sensor of a machine cylinder; c. spinning forming: to prepare a nascent fiber; meanwhile, three tension monitoring points are arranged in a spinning path, and the fiber diameter, crystallinity and tension parameters are detected in real time through an online monitoring system; d. post-treatment: the nascent fiber is stretched and shaped, washed and dried after stretching, and the high-performance textile fiber material finished product is obtained after airing. Through precise regulation and control of key links such as raw material pretreatment, gradient melt blending, intelligent spinning forming and post-treatment, the tension control of the whole process is strengthened, and the multi-dimensional mechanical property is synergistically optimized.

Description

Technical Field

[0001] This specification relates to the field of fiber material preparation technology, and in particular to a preparation process and comprehensive quality testing method for high-performance textile fiber materials. Background Technology

[0002] With the diversification of textile product applications, the market demands increasingly higher comprehensive performance from textile fiber materials. These materials not only require excellent mechanical stability but also integrate multifunctional properties such as antibacterial, antistatic, and UV protection. Simultaneously, precise quality control during production is crucial for ensuring large-scale application, and tension control and mechanical performance testing are core aspects determining product stability. Existing technologies suffer from the following key challenges: 1. Limited testing dimensions: mostly limited to a few indicators such as diameter and fracture strength, lacking systematic testing of key characteristics such as functional component dispersion, environmental adaptability, and long-term durability; 2. Inefficient data processing: Faced with high-dimensional parameters in the production process, ineffective dimensionality reduction and noise removal are not performed, resulting in low detection accuracy and many interfering factors; 3. Weak predictive ability: Relying solely on a single model or empirical formula, it cannot achieve real-time performance prediction, early warning of quality anomalies, or long-term stability assessment. Therefore, developing a comprehensive quality testing method that covers all aspects of the testing process and possesses high accuracy and creativity has become crucial for promoting the upgrading of the textile fiber materials industry. Summary of the Invention

[0003] This invention addresses the problems existing in the prior art by proposing a preparation process and comprehensive quality testing method for high-performance textile fiber materials. It solves the problems of lack of tension control, single mechanical property testing, and blind parameter adjustment in traditional processes, significantly improving the mechanical stability, functional durability, and production qualification rate of textile fiber materials. It is suitable for diverse scenarios such as medical, outdoor, and home use, and has important industrial application value.

[0004] To achieve the above objectives, this application provides the following technical solution: In a first aspect, a preparation process for a high-performance textile fiber material is characterized by the following steps: a. Raw material treatment: placing the basic fiber substrate into a plasma treatment device, under an argon-oxygen mixed atmosphere, with power controlled at 90-130W, for a treatment time of 6-12 minutes, to enhance the surface activity of the fiber; mixing the composite functional agent and compatibilizer at a mass ratio of 9:1-7:1, adding them to a high-speed mixer, rotating at 250-350 rpm, and mixing for 18-25 minutes, during which time the mixture is treated by an ultrasonic device. a. Process for 10-15 minutes to prepare the functional premix; during the mixing process, the agglomeration tension of the material is monitored in real time by a tension sensor; b. Melt blending: The pretreated base fiber matrix, functional premix, antioxidant, modifier and functional additives are added to a twin-screw extruder. The extrusion temperature is controlled in stages: Zone 1 155-165℃, Zone 2 170-180℃, Zone 3 180-190℃, Zone 4 185-195℃, and the screw speed is 200-240 rpm. The molten material is monitored by a tension sensor built into the barrel. The process involves: 1. Extrusion and pelletizing to produce functional masterbatch; 2. Spinning: The functional masterbatch is added to a spinning machine at a spinning temperature of 190-200℃, a spinneret orifice diameter of 0.25-0.35mm, and a winding speed of 2800-3200m / min to produce nascent fibers. Simultaneously, three tension monitoring points are set along the spinning path, controlling the tension at 3.5-5.0cN, 5.0-7.0cN, and 7.0-9.0cN respectively. An online monitoring system is used to detect the fiber diameter and crystallinity in real time. and tension parameters; d. Post-treatment: The nascent fibers are stretched and shaped using a three-stage stretching process of "pre-stretching-low temperature relaxation-secondary stretching". The pre-stretching ratio is 1.5-2.0 times, the low temperature relaxation temperature is 35-45℃, the heat preservation time is 6-10 minutes, the secondary stretching ratio is 1.2-1.8 times, and the total stretching ratio is 2.8-3.6 times. After stretching, the fibers are washed and dried. An anti-UV agent is injected using in-situ composite technology. Finally, an antibacterial finishing liquid is sprayed with ultrasonic assistance. After drying, a high-performance textile fiber material is obtained.

[0005] Optionally, in step a, after plasma treatment, the surface hydrophilicity of the base fiber substrate is tested to ensure that the contact angle is ≤55°.

[0006] Optionally, the base fiber substrate is a blended system of polylactic acid fiber and cotton fiber; the composite functional agent is a compound system of nano zinc oxide and honeysuckle flavonoid-artemisia essential oil compound extract; the compatibilizer is maleic anhydride grafted polyethylene; the antioxidant is a compound system of tea polyphenols and vitamin E; the modifier is a compound of organosilane coupling agents KH-550 and KH-560; and the functional additives include antistatic agents and softeners.

[0007] Secondly, this invention provides a comprehensive quality testing method for high-performance textile fiber materials, based on the preparation process of high-performance textile fiber materials in the first aspect, including the following steps: S1. Multi-dimensional data acquisition and preprocessing: Simultaneously acquire raw material characteristics, process parameters, real-time performance and environmental parameters throughout the entire preparation process, perform dimensionality reduction and standardization on the acquired high-dimensional data, and remove noisy data and redundant information; S2. Real-time monitoring and performance prediction: Based on the preprocessed core data, a multi-parameter collaborative prediction model is established to predict the dispersion uniformity of composite functional agents and the core mechanical properties of fibers in real time; S3. Quality Anomaly Diagnosis and Early Warning: Establish an anomaly identification model to determine the quality status of the production process, explore the causes of anomalies and form source tracing results, and at the same time use a time-series prediction model to warn of the changing trends of key process parameters. S4. Long-term stability assessment: Establish environmental adaptability prediction model and durability prediction model to evaluate the performance degradation law of fiber under different environmental conditions and the performance retention rate after long-term use; S5. Comprehensive Quality Judgment: Based on the dispersion prediction results, mechanical property prediction results, anomaly diagnosis results, and long-term stability assessment results from multiple models, a weighted judgment mechanism is established to comprehensively determine whether the product quality is qualified; if it is not qualified, the parameter adjustment mechanism or the non-qualified product recycling process is initiated.

[0008] Optionally, in step S1, principal component analysis is used for dimensionality reduction of high-dimensional data; Z-score standardization is used for standardization, and the specific formula is shown below: (1) in, Let be the standardized value of the j-th principal component of the i-th sample; This is the original data; Let be the mean of the j-th principal component; Let be the standard deviation of the j-th principal component.

[0009] Optionally, in step S2, the dispersion uniformity of the composite functional agent is predicted using a partial small squares regression model, as shown in the following formula: (2) in, This is the predicted dispersion value; The intercept; These are the regression coefficients; is the k-th principal component after dimensionality reduction; m is the number of principal components.

[0010] Optionally, in step S2, the core mechanical property prediction uses the spinning crystallinity X, the average tension T throughout the process, the fiber diameter D, and the principal components after PCA dimensionality reduction as inputs to establish a random forest model, simultaneously predicting the three core mechanical indicators: breaking strength, elongation at break, and elastic recovery rate. The specific formulas are shown below: (3) in, This is the predicted value for fracture strength; This is the predicted value of elongation at break; This is the predicted value for the elastic recovery rate.

[0011] Optionally, in step S3, the anomaly detection model adopts a support vector machine binary classification model, and the kernel function adopts a radial basis function, as shown in the following formula: (4) in, For sample vectors; The parameters are kernel function parameters; anomaly causes are obtained through association rule mining algorithms; key process parameters include tension parameters; the time series prediction model uses a long short-term memory network, and the specific formula is shown below: (5) in, Here is the predicted tension value at time t+k; k is the prediction time step. This represents the time step for historical data.

[0012] Optionally, in step S4, the environmental adaptability prediction model uses a recurrent neural network, with the input being a sequence of parameter changes under different environmental conditions and the output being the mechanical performance degradation rate; the durability prediction model uses a grey prediction model, and the specific formula is shown below: (6) in, This is the cumulative sequence prediction value; denoted as the first term of the original sequence; a is the model evolution coefficient; u is the gray action of the model; and k is the prediction step size.

[0013] Optionally, in step S1, the parameters collected include, but are not limited to, substrate mixing ratio, functional agent particle size, compatibilizer grafting rate, temperature at each stage, rotation speed, tension, time, fiber diameter, crystallinity, surface resistance, contact angle, ambient temperature and humidity, and equipment vibration value.

[0014] The beneficial effects of this invention are as follows: 1. The preparation process of this application achieves synergistic improvement of mechanical properties, antibacterial properties, antistatic properties and other functions through precise tension control and multi-dimensional process optimization throughout the entire process; 2. The detection method of this application integrates multiple algorithm models such as principal component analysis, partial least squares regression, random forest, support vector machine and recurrent neural network to establish a full-chain detection system of "data preprocessing-real-time monitoring-performance prediction-anomaly diagnosis-long-term stability assessment". It solves the problems of traditional detection methods being single, low in accuracy and lacking creativity, and significantly improves the controllability of product quality and the production qualification rate. It is applicable to diverse scenarios such as medical, outdoor and home use. Attached Figure Description

[0015] This specification will be further described by way of exemplary embodiments, which will be described in detail with reference to the accompanying drawings. The same numbers in the drawings denote the same structures or steps.

[0016] Figure 1 This is a schematic diagram of the preparation process of the high-performance textile fiber material of Embodiment 1 of this application.

[0017] Figure 2 This is a schematic diagram of the comprehensive quality testing method for high-performance textile fiber materials according to Embodiment 2 of this application. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description of this application is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely one preferred embodiment of this application and are only used to explain this application. They do not limit the scope of protection of this application. All other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0019] This application pertains to textile fiber materials, which include 60-80 parts of a base fiber substrate, 5-15 parts of a composite antibacterial agent, 1-3 parts of a compatibilizer, 0.2-0.8 parts of an antioxidant, 2-5 parts of a modifier, and 3-8 parts of functional additives. The composite antibacterial agent is a mixture of nano-zinc oxide and plant-derived antibacterial extracts, wherein the nano-zinc oxide particle size is 20-50 nm, and the plant-derived antibacterial extract is a mixture of honeysuckle flavonoids and artemisia argyi essential oil, with a mass ratio of 3:1-2:1. The base fiber substrate is a blend of polylactic acid fiber and cotton fiber, with polylactic acid fiber comprising 40%-60%. The compatibilizer is maleic anhydride-grafted polyethylene with a grafting rate of 1.5%-2.5%, enhancing the interfacial bonding between the composite antibacterial agent and the base fiber substrate. The antioxidant is a compound system of tea polyphenols and vitamin E with a mass ratio of 1:1 to 1:2; the modifier is organosilane coupling agent KH-550, which can improve the hydrophilicity of the fiber surface; the functional additives include antistatic agent and softener with a mass ratio of 2:1; the antistatic agent is lauryl dimethyl betaine and the softener is lanolin.

[0020] Example 1: like Figure 1 As shown, a preparation process for a high-performance textile fiber material includes the following steps: a. Raw material processing: The basic fiber substrate is placed in a plasma treatment device under an argon-oxygen mixed atmosphere with a volume ratio of 8:2, a power control of 90-130W, and a treatment time of 6-12 minutes. The composite functional agent and compatibilizer are mixed at a mass ratio of 9:1-7:1 and added to a high-speed mixer at a speed of 250-350 rpm for 18-25 minutes. During this period, the mixture is treated with an ultrasonic device for 10-15 minutes to produce a functional premix. During the mixing process, the agglomeration tension of the material is monitored in real time by a tension sensor and controlled to be ≤3.0cN. b. Melt blending: The pretreated base fiber substrate, functional premix, antioxidant, modifier, and functional additives are added to a twin-screw extruder. Four-stage temperature control is employed: Zone 1 155-165℃, Zone 2 170-180℃, Zone 3 180-190℃, and Zone 4 185-195℃. The screw speed is 200-240 rpm. During extrusion, the vacuum degree is controlled to ≤-0.09MPa. The flow tension of the molten material is monitored by a tension sensor built into the barrel and controlled within the range of 4.5-6.5 cN. After extrusion, the material is pelletized and graded to produce functional masterbatch. The particle size range of the graded and screened material is 0.8-1.2 mm. c. Spinning and forming: The functional masterbatch is added to the spinning machine, the spinning temperature is 190-200℃, the spinneret orifice diameter is 0.25-0.35mm, and the winding speed is 2800-3200m / min to produce nascent fibers; three tension monitoring points are set in the spinning path, and the tension is controlled at 3.5-5.0cN, 5.0-7.0cN, and 7.0-9.0cN respectively. The fiber diameter, crystallinity and tension parameters are detected in real time through an online monitoring system. When the tension fluctuation at any monitoring point exceeds ±8%, the winding speed and blowing intensity are automatically adjusted. d. Post-treatment: The nascent fibers are stretched and shaped using a three-stage stretching process of "pre-stretching - low-temperature relaxation - secondary stretching". The pre-stretching ratio is 1.5-2.0 times, the tension is controlled at 8.0-10.0 cN, the low-temperature relaxation temperature is 35-45℃, the holding time is 6-10 minutes, the relaxation tension is ≤2.0 cN, the secondary stretching ratio is 1.2-1.8 times, the tension is controlled at 10.0-12.0 cN, and the total stretching ratio is 2.8-3.6 times. After stretching, the fibers are washed and dried at 85-95℃ for 35-45 minutes. An anti-UV agent is injected using in-situ composite technology. Finally, an antibacterial finishing liquid is sprayed with ultrasonic assistance at a frequency of 40-50 kHz. After drying, the finished product is obtained.

[0021] The preparation process of this application continues the core advantage of precise tension control throughout the entire process. Through the synergistic optimization of four stages—raw material pretreatment, gradient melt blending, intelligent spinning, and post-treatment—it provides a stable process foundation for multi-model testing. Specifically, the compounding system of composite functional agents and the interface compatibility modification technology ensure the uniform dispersion and stable binding of functional components; the three-stage stretching and in-situ composite technology achieve a synergistic improvement in mechanical properties and multifunctional characteristics, providing rich performance data support for multi-model testing.

[0022] In step a. Raw material processing, the base fiber substrate is a blended system of polylactic acid fiber and cotton fiber, with polylactic acid fiber accounting for 45%-65%, possessing both high strength and flexibility; the composite functional agent is a compound system with honeysuckle flavonoid-artemisia essential oil compound extract, with nano zinc oxide particles of 25-45nm, forming a physical-chemical synergistic antibacterial system with the plant-derived extract, expanding the antibacterial spectrum and improving durability, with a mass ratio of 2.5:1-2:1; the compatibilizer is maleic anhydride grafted polyethylene, with a grafting rate of 1.8%-2.6%, enhancing the interfacial bonding force between the functional agent and the substrate.

[0023] Specifically, the plasma treatment uses an argon-oxygen mixed atmosphere, which, compared to a single argon atmosphere, can further enhance the surface activity and hydrophilicity of the substrate, ensuring a contact angle ≤55°. During the preparation of the functional premix, the synergistic effect of high-speed mixing, ultrasonic dispersion, and polyethylene glycol 400 dispersant is combined to break up functional agent agglomerates. Simultaneously, polycaprolactone oligomers are added as interfacial compatibility modifiers to form a transition layer, improving dispersion uniformity and binding stability.

[0024] Specifically, 0.5%-1.0% of polycaprolactone oligomer is added to the functional premix as an interfacial compatibility modifier, and 0.3%-0.6% of polyethylene glycol 400 is added as a dispersant. The mixing system is monitored in real time using a high-speed camera until no obvious particulate matter is found. During the mixing process of the functional premix, agglomeration tension monitoring is introduced. The agglomeration state of the functional agent is detected in real time by a tension sensor. When the agglomeration tension exceeds 3.0 cN, the ultrasonic power is automatically increased or the mixing time is extended to break up the agglomerates and ensure uniform dispersion of the functional agent, laying the foundation for the uniformity of subsequent mechanical properties.

[0025] Specifically, the ultrasonic equipment uses a 35-45kHz model.

[0026] In step b. melt blending, the antioxidant is a compound system of tea polyphenols and vitamin E with a mass ratio of 1:1.2-1:1.8; the modifier is a compound of organosilane coupling agents KH-550 and KH-560 with a mass ratio of 3:1; the functional additives include antistatic agent (lauryl dimethyl betaine) and softener (lanolin) with a mass ratio of 2.5:1.

[0027] In step b. melt blending, a four-stage temperature gradient control is adopted, which is more in line with the melting characteristics of the raw materials than the traditional three-stage temperature control, and avoids the decomposition of functional components caused by local overheating; the vacuum degree is strictly controlled to ≤-0.09MPa, which effectively eliminates air bubbles in the melt system and improves fiber density; after pelleting, the functional masterbatch is graded and screened to ensure uniform particle size and avoid clogging of the spinneret during spinning.

[0028] In step b. melt blending, a tension sensor is built into the barrel of the twin-screw extruder to monitor the flow tension of the molten material and control it within the range of 4.5-6.5 cN. This avoids material degradation due to excessive flow tension and uneven mixing due to insufficient flow tension. At the same time, it is linked with the segmented temperature control and screw speed to ensure the stability of the melt system.

[0029] In step c. spinning, the spinneret is laser-cleaned before spinning to remove residual impurities and aged materials, ensuring unobstructed spinneret holes. The spinning temperature is precisely controlled at 190-200℃ to match the melting characteristics of the functional masterbatch. During the cooling stage, a variable-direction blowing technology is used, and the blowing direction is periodically adjusted to avoid stress concentration during fiber cooling and improve mechanical property stability. The online monitoring system integrates a laser diameter gauge, X-ray diffractometer, tension sensor, and near-infrared spectrometer to achieve real-time monitoring of multiple indicators, providing data support for dynamic parameter adjustment.

[0030] In step c. spinning, three tension monitoring points are set at the spinneret, cooling section, and winding section. The tension range of each section is controlled in a targeted manner: 3.5-5.0 cN, 5.0-7.0 cN, and 7.0-9.0 cN. This avoids excessive tension at the spinneret leading to fiber necking, uneven tension in the cooling section leading to crystallization imbalance, and tension fluctuations in the winding section leading to linear density deviation. A high-precision fiber optic tension sensor with a detection frequency of 200-300 Hz is used to ensure real-time feedback of tension data and parameter adjustment.

[0031] Specifically, the spinneret undergoes laser cleaning pretreatment with a pulse frequency of 10-15Hz and a cleaning time of 3-5 minutes. Spinning cooling employs a variable-direction blowing technology with a blowing speed of 0.8-1.2m / s and periodic adjustment of the blowing direction for 3-5 seconds. The online monitoring system includes a laser diameter gauge, an X-ray diffractometer, and a tension sensor, with a data sampling frequency of 100-150Hz.

[0032] In step d. post-processing, the three-stage stretching process effectively releases the internal stress of the fiber through a combination of pre-stretching, low-temperature relaxation, and secondary stretching, avoiding the embrittlement problem caused by single stretching. The total stretching ratio is controlled at 2.8-3.6 times to ensure a breaking strength ≥3.8cN / dtex. The UV protectant is a compound system of nano-titanium dioxide and zinc oxide in a mass ratio of 1:1. It is mixed with the binder (polyvinyl acetate) in a mass ratio of 5:1 and then injected into the micropores inside the fiber through high-pressure injection (pressure 5-8MPa). The UV protectant is injected into the micropores of the fiber using in-situ composite technology under high pressure. Compared with surface spraying, the bonding is more stable, with a UPF value ≥50+. The antibacterial finishing solution is a chitosan-sodium alginate composite solution with a mass fraction of 0.8%-1.2% and a mass ratio of 3:1. It is sprayed with ultrasonic assistance to promote penetration. After 50 water washes, the antibacterial rate is still ≥90%. During the stretching process, a tension feedback adjustment mechanism is adopted. When the secondary stretching tension exceeds 12.5cN, the stretching rate is automatically reduced by 0.5-1.0mm / s to avoid fiber breakage.

[0033] In step d. post-processing, the tension at each stage of the three-stage stretching process is precisely controlled: pre-stretch tension 8.0-10.0 cN, relaxation tension ≤2.0 cN, and secondary stretch tension 10.0-12.0 cN. Simultaneously, a tension overload protection mechanism is implemented to prevent fiber breakage during stretching and improve mechanical property stability. Pre-stretch tension ensures initial fiber orientation, relaxation tension fully releases internal stress, and secondary stretch tension achieves high fiber orientation and perfect crystallization.

[0034] Example 2: like Figure 2 As shown, a comprehensive quality testing method for high-performance textile fiber materials, based on the preparation process of high-performance textile fiber materials described in Example 1, includes the following steps: S1. Multi-dimensional data acquisition and preprocessing: Simultaneously acquire raw material characteristics, process parameters, real-time performance and environmental parameters throughout the entire preparation process, perform dimensionality reduction and standardization on the acquired high-dimensional data, and remove noisy data and redundant information; S2. Real-time monitoring and performance prediction: Based on the preprocessed core data, a multi-parameter collaborative prediction model is established to predict the dispersion uniformity of composite functional agents and the core mechanical properties of fibers in real time; S3. Quality Anomaly Diagnosis and Early Warning: Establish an anomaly identification model to determine the quality status of the production process, explore the causes of anomalies and form source tracing results, and at the same time use a time-series prediction model to warn of the changing trends of key process parameters. S4. Long-term stability assessment: Establish environmental adaptability prediction model and durability prediction model to evaluate the performance degradation law of fiber under different environmental conditions and the performance retention rate after long-term use; S5. Comprehensive Quality Judgment: Based on the dispersion prediction results, mechanical property prediction results, anomaly diagnosis results, and long-term stability assessment results from multiple models, a weighted judgment mechanism is established to comprehensively determine whether the product quality is qualified; if it is not qualified, the parameter adjustment mechanism or the non-qualified product recycling process is initiated.

[0035] This embodiment integrates multiple algorithm models, including Principal Component Analysis (PCA), Partial Least Squares Regression (PLSR), Random Forest (RF), Support Vector Machine (SVM), and Recurrent Neural Network (RNN), to establish a full-chain detection system encompassing "data preprocessing - real-time monitoring - performance prediction - anomaly diagnosis - long-term stability assessment." This addresses the limitations of traditional detection methods, such as their reliance on single methods, low accuracy, and insufficient creativity, significantly improving product quality control and production pass rates. It is applicable to diverse scenarios such as medical, outdoor, and home applications. Through multiple algorithm models, a collaborative detection system covering the entire process, multiple dimensions, and the entire lifecycle is constructed, overcoming the limitations of traditional single detection methods and achieving full-chain quality control from real-time monitoring to long-term prediction. By using LSTM for early warning of tension anomalies and Apriori to trace the causes of anomalies, the response time for adjusting production process parameters is shortened to within 5 seconds, and the product pass rate is increased to over 99%. This system is suitable for scenarios with high requirements for the comprehensive performance and stability of textile fiber materials, such as medical, outdoor, and home applications, providing technical support for the large-scale production of high-performance textile fiber materials.

[0036] In this embodiment, the parameters for multi-dimensional data acquisition in step S1 include, but are not limited to, substrate mixing ratio, functional agent particle size, compatibilizer grafting rate, temperature, rotation speed, tension, time, fiber diameter, crystallinity, surface resistance, contact angle, ambient temperature and humidity, and equipment vibration value. Specifically, 18 core parameters are collected simultaneously throughout the entire process of step a (raw material mixing), step b (melt blending), step c (spinning and forming), and step d (post-processing). Raw material characteristic parameters include substrate mixing ratio, functional agent particle size, and compatibilizer grafting rate; process parameters include temperature / rotation speed / tension / time at each stage; real-time performance parameters include fiber diameter, crystallinity, surface resistance, and contact angle; and environmental parameters include temperature and humidity and equipment vibration value.

[0037] In this embodiment, in step S1, data preprocessing uses Principal Component Analysis (PCA) to reduce the dimensionality of the 18 high-dimensional parameters, retaining the top 5-7 principal components with a cumulative contribution rate ≥95%, and removing noisy data and redundant information. Standardization uses Z-score standardization to normalize the principal component data, eliminating dimensional differences. The specific formula is shown below: (1) in, Let be the standardized value of the j-th principal component of the i-th sample; This is the original data; Let be the mean of the j-th principal component; Let be the standard deviation of the j-th principal component.

[0038] In this embodiment, in step S2, the dispersion of the functional agent is predicted based on near-infrared spectral data from the melt blending stage and the principal components after PCA dimensionality reduction. A partial least squares regression (PLSR) model is established to predict the dispersion uniformity of the composite functional agent in real time. The specific formula is as follows: (2) in, This is the predicted dispersion value; The intercept; These are the regression coefficients; is the k-th principal component after dimensionality reduction; m is the number of principal components, and the prediction error is ≤ ±3%.

[0039] Near-infrared spectral data were collected in wavelengths ranging from 800 to 2500 nm with a sampling interval of 2 nm. Each sample was collected three times and the average value was taken. The PLSR model optimized the regression coefficients through cross-validation (leave-one-out method) to ensure prediction accuracy.

[0040] In this embodiment, in step S2, the core mechanical properties are predicted by using spinning crystallinity X, average tension T throughout the process, fiber diameter D, and principal components after PCA dimensionality reduction as inputs to establish a random forest model, which simultaneously predicts three core mechanical indicators: breaking strength, elongation at break, and elastic recovery rate. The specific formulas are shown below: (3) in, This is the predicted value for fracture strength; This is the predicted value of elongation at break; This is the predicted value for the elastic recovery rate; These are random forest sub-models trained for fracture strength, elongation at break, and elastic recovery rate, respectively. The 5-7 core principal components are obtained after dimensionality reduction from PCA. The feature importance of the random forest model is calculated using the Gini coefficient, and input parameters with a feature importance ≥ 0.6 are retained first. The model training data is updated periodically every 100 batches to dynamically optimize the model parameters.

[0041] In this embodiment, step S3 includes: S31. Anomaly Pattern Recognition: Based on the full-process parameter data of historical qualified / unqualified products, a Support Vector Machine (SVM) binary classification model is trained. Using the principal components after PCA dimensionality reduction as input, it determines whether there are quality anomalies in the current production process. The kernel function adopts the radial basis function (RBF), and the specific formula is shown below: (4) in, The sample vectors are the feature vectors of samples from different production batches, which are composed of principal components after dimensionality reduction. The kernel function parameters are defined as follows. The ratio of the training set, validation set, and test set of the SVM model is 7:1:2. The grid search method is used to optimize the penalty coefficient C (1-10) and the kernel function parameter γ (0.1-0.5) to ensure classification stability under different production conditions.

[0042] S32. Anomaly Source Analysis: When the SVM model determines that an anomaly exists, the association rule mining algorithm (Apriori) is used to analyze the correlation between 18 original parameters and the anomaly type, and output the top 3 key anomaly causes (confidence ≥ 0.85).

[0043] S33. Dynamic Early Warning of Tension Anomalies: This system integrates a Long Short-Term Memory (LSTM) network with a sliding window algorithm to predict tension changes over the next 5-10 seconds based on tension data from the previous 100-200 time steps. Anomaly causes are identified using an association rule mining algorithm. Key process parameters, including tension parameters, are used in the time-series prediction model, which employs a LSTM network. The specific formula is shown below: (5) in, Here is the predicted tension value at time t+k; k is the prediction time step. For historical data time steps; It is a long short-term memory network model, specifically designed to process time-series data and capture long-term dependencies of tension changes.

[0044] In this embodiment, in step S4, the environmental adaptability prediction model employs a recurrent neural network (RNN). Using parameter data from normal temperature and humidity, high temperature and high humidity, and low temperature and low humidity environments as input, it predicts the mechanical property degradation rate of fibers after 24 hours of placement in different environments. The model has 64-128 hidden layer nodes, 500-800 iterations, and a prediction error ≤ ±2%. The input sequence length of the RNN model is 24, corresponding to 24-hour environmental parameter changes, and the output is the mechanical property degradation rate under the three environments. It is trained using the Adam optimizer with a learning rate of 0.001-0.005. The RNN addresses the problem of long environmental adaptability detection cycles by predicting the mechanical property degradation rate under different temperature and humidity environments based on a 24-hour environmental parameter change sequence, with an error ≤ ±2%, significantly shortening the detection cycle.

[0045] In this embodiment, in step S4, the durability prediction model is based on the mechanical property change data during 50 water washing cycles, establishing a grey prediction model. The specific formula is shown below: (6) in, This is the cumulative sequence prediction value; denoted as the first term of the original sequence; 'a' is the model evolution coefficient; 'u' is the model gray action; and 'k' is the prediction step size. A model is built using a small amount of washing data to predict the performance retention rate after 100 washes, eliminating the need for a full-cycle washing test and improving detection efficiency.

[0046] In this embodiment, in step S5, if the output results of the multiple models meet the following conditions, the product is judged to be qualified: ① PLSR model prediction dispersion deviation ≤ ±5%; ② RF model prediction mechanical properties are within the qualified range; ③ SVM model has no abnormality judgment; ④ LSTM tension warning has no continuous abnormality; ⑤ RNN and grey prediction GM model prediction long-term stability meets the standard; otherwise, the parameter adjustment mechanism or the non-conforming product recycling process is initiated. Specifically, a model output weight allocation mechanism is established: RF mechanical property prediction weight 0.3, SVM abnormality diagnosis weight 0.25, PLSR dispersion prediction weight 0.2, and long-term stability prediction weight 0.25. A weighted score ≥ 85 points is judged as qualified. The model training data is updated regularly, and the model parameters are dynamically optimized to ensure the detection stability and accuracy under different production conditions.

[0047] Example 3: A preparation process for a high-performance textile fiber material, comprising the following steps: A. Raw material preparation: The composition includes: 70 parts of basic fiber substrate (50% polylactic acid fiber and 50% cotton fiber), 11 parts of composite functional agent (nano zinc oxide and honeysuckle flavonoid-artemisia essential oil extract in a mass ratio of 2.5:1), 2.2 parts of compatibilizer (maleic anhydride grafted polyethylene, grafting rate 2.2%), 0.5 parts of antioxidant (tea polyphenols and vitamin E in a mass ratio of 1:1.5), 3.5 parts of modifier (KH-550 and KH-560 in a mass ratio of 3:1), and 5.5 parts of functional additive (lauryl dimethyl betaine and lanolin in a mass ratio of 2.5:1).

[0048] B. Preparation process: a. Raw material pretreatment: plasma treatment (argon-oxygen 8:2, power 110W, time 8 minutes), functional premix mixing (speed 300 rpm, time 20 minutes), ultrasonic treatment (40kHz, 12 minutes), agglomeration tension controlled at 2.5cN; b. Melt blending: four-stage temperature 160℃ / 175℃ / 185℃ / 190℃, screw speed 220 rpm, vacuum degree -0.095MPa, melt flow tension controlled at 5.5cN, and masterbatch with a screening particle size of 0.9-1.1mm; c. Spinning and forming: spinning temperature 195℃, spinneret orifice diameter 0.3mm, winding speed 3000m / min, spinneret tension 4.2cN, cooling section tension 6.0cN, winding section tension 8.0cN, and change-direction blowing cycle 4 seconds. d. Post-treatment: Pre-stretch 1.8 times (tension 9.0 cN) → Relax at 40℃ for 8 minutes (tension 1.5 cN) → Second stretch 1.5 times (tension 11.0 cN), drying temperature 90℃ (40 minutes), high pressure injection of nano titanium dioxide-zinc oxide composite UV protectant, ultrasonic spraying of 1.0% chitosan-sodium alginate solution (3:1).

[0049] C. Quality Inspection: S1. Multi-dimensional data acquisition and preprocessing: 18 core parameters were collected, and 6 principal components were retained after PCA dimensionality reduction, with a cumulative contribution rate of 96.2%. After Z-score standardization, the data were used as model input. S2. Real-time monitoring and performance prediction: Partial Least Squares Regression (PLSR) model, near-infrared spectral data predicts a dispersion deviation of 3.1% for the composite functional agent, which meets the requirements; Random Forest (RF) model predicts a breaking strength of 4.1 cN / dtex, an elongation at break of 20.5%, and an elastic recovery rate of 78%, with deviations from offline detection values ​​of ≤1.5%. S3. Quality Anomaly Diagnosis and Early Warning: Support Vector Machine (SVM) model determines that there are no anomalies in the production process; Long Short-Term Memory (LSTM) network tension early warning predicts that tension fluctuations ≤5% in the next 8 seconds, and no warning is given. S4. Long-term stability assessment: The recurrent neural network (RNN) model predicts a mechanical property degradation rate of 5.2% in high temperature and high humidity environment and 4.7% in low temperature and low humidity environment, which meets the requirements. The grey prediction GM model predicts a fracture strength retention rate of 88% after 100 water washes, which meets the usage requirements. S5. Overall quality assessment: Weighted score of 92 points, deemed qualified.

[0050] The above-described specific embodiments are preferred embodiments of the preparation process and comprehensive quality testing method for high-performance textile fiber materials of this application, and are not intended to limit the specific scope of this application. The scope of this application includes but is not limited to the specific embodiments described herein. All equivalent changes made in accordance with the shape and structure of this application are within the protection scope of this application.

Claims

1. A preparation process for a high-performance textile fiber material, characterized in that, Includes the following steps: a. Raw material processing: The basic fiber substrate is placed in a plasma treatment device under an argon-oxygen mixed atmosphere, with power controlled at 90-130W and a treatment time of 6-12 minutes, to enhance the surface activity of the fiber; the composite functional agent and compatibilizer are mixed at a mass ratio of 9:1-7:1 and added to a high-speed mixer at a speed of 250-350 rpm for 18-25 minutes, during which time the mixture is treated with ultrasonic equipment for 10-15 minutes to produce a functional premix; during the mixing process, the agglomeration tension of the material is monitored in real time by a tension sensor. b. Melt blending: The pretreated base fiber substrate, functional premix, antioxidant, modifier and functional additives are added to a twin-screw extruder. The extrusion temperature is controlled in stages: Zone 1 155-165℃, Zone 2 170-180℃, Zone 3 180-190℃, Zone 4 185-195℃, and the screw speed is 200-240 rpm. The flow tension of the molten material is monitored by a tension sensor built into the barrel. After extrusion, the material is pelletized and graded to produce functional masterbatch. c. Spinning and forming: The functional masterbatch is added to the spinning machine, the spinning temperature is 190-200℃, the spinneret orifice diameter is 0.25-0.35mm, and the winding speed is 2800-3200m / min to produce nascent fibers; at the same time, three tension monitoring points are set in the spinning path, and the tension is controlled at 3.5-5.0cN, 5.0-7.0cN, and 7.0-9.0cN respectively. The fiber diameter, crystallinity and tension parameters are detected in real time through an online monitoring system. d. Post-treatment: The nascent fibers are stretched and shaped using a three-stage stretching process of "pre-stretching - low-temperature relaxation - secondary stretching". The pre-stretching ratio is 1.5-2.0 times, the low-temperature relaxation temperature is 35-45℃, the holding time is 6-10 minutes, the secondary stretching ratio is 1.2-1.8 times, and the total stretching ratio is 2.8-3.6 times. After stretching, the fibers are washed and dried. An anti-UV agent is injected using in-situ composite technology. Finally, an antibacterial finishing liquid is sprayed with ultrasonic assistance, and the finished high-performance textile fiber material is obtained after drying.

2. The preparation process of the high-performance textile fiber material according to claim 1, characterized in that, In step a, after plasma treatment, the surface hydrophilicity of the basic fiber substrate is tested to ensure that the contact angle is ≤55°.

3. The preparation process of the high-performance textile fiber material according to claim 1, characterized in that, The basic fiber substrate is a blended system of polylactic acid fiber and cotton fiber; the composite functional agent is a compound system of nano zinc oxide and honeysuckle flavonoid-artemisia essential oil compound extract; the compatibilizer is maleic anhydride grafted polyethylene; the antioxidant is a compound system of tea polyphenols and vitamin E; the modifier is a compound of organosilane coupling agents KH-550 and KH-560; the functional additives include antistatic agents and softeners.

4. A comprehensive quality testing method for high-performance textile fiber materials based on the preparation process described in any one of claims 1-3, characterized in that, Includes the following steps: S1. Multi-dimensional data acquisition and preprocessing: Simultaneously acquire raw material characteristics, process parameters, real-time performance and environmental parameters throughout the entire preparation process, perform dimensionality reduction and standardization on the acquired high-dimensional data, and remove noisy data and redundant information; S2. Real-time monitoring and performance prediction: Based on the preprocessed core data, a multi-parameter collaborative prediction model is established to predict the dispersion uniformity of composite functional agents and the core mechanical properties of fibers in real time; S3. Quality Anomaly Diagnosis and Early Warning: Establish an anomaly identification model to determine the quality status of the production process, explore the causes of anomalies and form source tracing results, and at the same time use a time-series prediction model to warn of the changing trends of key process parameters. S4. Long-term stability assessment: Establish environmental adaptability prediction model and durability prediction model to evaluate the performance degradation law of fiber under different environmental conditions and the performance retention rate after long-term use; S5. Comprehensive Quality Judgment: Based on the dispersion prediction results, mechanical property prediction results, anomaly diagnosis results, and long-term stability assessment results from multiple models, a weighted judgment mechanism is established to comprehensively determine whether the product quality is qualified; if it is not qualified, the parameter adjustment mechanism or the non-qualified product recycling process is initiated.

5. The preparation process of the high-performance textile fiber material according to claim 4, characterized in that, In step S1, principal component analysis is used for dimensionality reduction of high-dimensional data; Z-score standardization is used for standardization, and the specific formula is shown below: （1） in, Let be the standardized value of the j-th principal component of the i-th sample; This is the original data; Let be the mean of the j-th principal component; Let be the standard deviation of the j-th principal component.

6. The preparation process of the high-performance textile fiber material according to claim 4, characterized in that, In step S2, the dispersion uniformity of the composite functional agent is predicted using a partial small squares regression model, as shown in the following formula: （2） in, This is the predicted dispersion value; The intercept; These are the regression coefficients; is the k-th principal component after dimensionality reduction; m is the number of principal components.

7. The comprehensive quality testing method for high-performance textile fiber materials according to claim 4, characterized in that, In step S2, the core mechanical properties are predicted by using spinning crystallinity X, average tension T throughout the process, fiber diameter D, and principal components after PCA dimensionality reduction as inputs to establish a random forest model, which simultaneously predicts three core mechanical indicators: breaking strength, elongation at break, and elastic recovery rate. The specific formulas are shown below: （3） in, This is the predicted value for fracture strength; This is the predicted value of elongation at break; This is the predicted value for the elastic recovery rate.

8. The preparation process of the high-performance textile fiber material according to claim 4, characterized in that, In step S3, the anomaly detection model adopts a support vector machine binary classification model, and the kernel function adopts a radial basis function, as shown in the following formula: （4） in, For sample vectors; The parameters are kernel function parameters; anomaly causes are obtained through association rule mining algorithms; key process parameters include tension parameters; the time series prediction model uses a long short-term memory network, and the specific formula is shown below: （5） in, Here is the predicted tension value at time t+k; k is the prediction time step. This represents the time step for historical data.

9. The preparation process of the high-performance textile fiber material according to claim 4, characterized in that, In step S4, the environmental adaptability prediction model uses a recurrent neural network, with the input being a sequence of parameter changes under different environmental conditions and the output being the mechanical performance degradation rate; the durability prediction model uses a grey prediction model, and the specific formula is shown below: （6） in, This is the cumulative sequence prediction value; denoted as the first term of the original sequence; a is the model evolution coefficient; u is the gray action of the model; and k is the prediction step size.

10. The comprehensive quality testing method for high-performance textile fiber materials according to claim 4, characterized in that, In step S1, the parameters collected include, but are not limited to, substrate mixing ratio, functional agent particle size, compatibilizer grafting rate, temperature at each stage, rotation speed, tension, time, fiber diameter, crystallinity, surface resistance, contact angle, ambient temperature and humidity, and equipment vibration value.

Images