Method and system for predicting shrinkage performance of knitted fabric
By constructing a multi-layer model from fiber to fabric scale, the problem of unstable shrinkage rate of knitted fabrics was solved, enabling rapid and accurate shrinkage rate prediction, thereby improving production efficiency and product quality.
Patent Information
- Application Number
- CN202511006900.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-07-22
- Publication Date
- 2025-10-31
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
The shrinkage rate of knitted fabrics is unstable during the production process, resulting in uneven quality of finished products. Traditional testing methods are time-consuming, rely on experience, and lack accuracy.
By constructing a multi-layered model from fiber to fabric scale, including molecular dynamics model, mechanical response model and shrinkage prediction model, and combining finite element analysis and iterative training, the deformation process of knitted fabrics in humid and hot environments is accurately simulated.
It enables rapid and accurate prediction of shrinkage rate of knitted fabrics, improves production efficiency and product quality, and provides guidance for process optimization.
Smart Images
Figure CN120877959A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of data processing technology, and specifically to a method and system for predicting the shrinkage performance of knitted fabrics. Background Technology
[0002] Knitted fabrics are susceptible to various factors during production, leading to unstable shrinkage rates and affecting finished product quality. Knitted fabrics are also prone to dimensional changes during washing, causing garment deformation and impacting performance. Traditional shrinkage testing requires multiple washes and measurements, with a cycle exceeding 48 hours. Furthermore, traditional methods rely on experience and lack precision. Therefore, developing an efficient and accurate shrinkage prediction method and system is crucial for improving the production efficiency and product quality of knitted fabrics. Summary of the Invention
[0003] To address the aforementioned issues, this invention provides a method and system for predicting the shrinkage performance of knitted fabrics, aiming to achieve rapid and accurate prediction of the shrinkage rate of knitted fabrics, reduce product quality problems caused by unstable shrinkage rates, and improve production efficiency.
[0004] To achieve the above objectives, the present invention provides the following technical solution: On one hand, embodiments of the present invention provide a method for predicting the shrinkage properties of knitted fabrics, the method comprising the following steps: S100: Obtain the physical structure parameters, fiber material parameters, historical shrinkage data, and finishing process parameters of the target knitted fabric. S200, Based on the physical structure parameters and fiber material parameters, construct a fiber-scale molecular dynamics model; S300, based on the molecular dynamics model, a humid and hot environment simulation was performed to obtain hygroscopic expansion data at the fiber scale; S400, based on the physical structure parameters and moisture absorption expansion data, construct a mechanical response model at the yarn scale; S500, based on the mechanical response model, yarn shrinkage simulation is performed to obtain shrinkage deformation data at the yarn scale; S600, Based on the physical structure parameters and shrinkage deformation data, construct a fabric-scale shrinkage prediction model; S700, input the historical shrinkage rate data into the shrinkage rate prediction model for iterative training to obtain the trained shrinkage rate prediction model; S800, based on the post-finishing process parameters and the trained shrinkage prediction model, outputs the predicted shrinkage rate of the target knitted fabric.
[0005] Optionally, in step S100, the physical structure parameters include coil density, yarn count, yarn twist, and fabric structure; the fiber material parameters are obtained by Fourier transform infrared spectroscopy testing of fiber samples, including fiber crystallinity, orientation, and moisture regain; the finishing process parameters include setting temperature, setting time, mechanical pre-shrinking force, and overfeed rate.
[0006] Optionally, in step S200, constructing the fiber-scale molecular dynamics model includes, S201, establish the initial conformation of cellulose macromolecular chain based on fiber chemical structure; S202, set the water molecule distribution model and temperature gradient field; S203, setting the intermolecular interaction potential function based on CHARMM force field parameters; S204 uses periodic boundary conditions to construct a fiber unit cell model.
[0007] Optionally, in step S300, the step of simulating the hygrothermal environment based on the molecular dynamics model to obtain fiber-scale hygroscopic expansion data includes, S301, under specific temperature and relative humidity conditions, performs moisture absorption simulation and records the curves of fiber axial and radial expansion rates over time; S302, extract the maximum expansion rate under equilibrium state as the hygroscopic expansion characteristic value.
[0008] Optionally, in step S400, constructing a mechanical response model at the yarn scale based on the physical structure parameters and moisture absorption and expansion data includes: S401, Based on the yarn geometry and fiber moisture absorption and expansion data, establish a stress distribution model inside the yarn; S402 combines the elastic modulus and Poisson's ratio of the yarn to transform the stress distribution model into a mechanical response model.
[0009] Optionally, in step S500, the step of simulating yarn shrinkage based on the mechanical response model to obtain shrinkage deformation data at the yarn scale includes: S501 uses the finite element method to numerically solve the mechanical response model, simulates the shrinkage behavior of yarn under different stress conditions, and obtains shrinkage deformation data at the yarn scale. S502, based on the Peirce loop geometric model, establishes the knitted loop topology, sets the boundary constraint condition to fix the four corners of the fabric, and constructs the yarn-scale shrinkage deformation equation based on the yarn-scale shrinkage deformation data and the principle of energy minimization: ;in, Total system energy , is the yarn elasticity coefficient. This represents the change in yarn length. The included angle of the coil, Let i be the initial coil angle, i be the yarn number, n be the total number of yarns, j be the coil number, m be the total number of coils, and λ be the balance factor, with a value between 0 and 1. S503 uses the shrinkage deformation equation to numerically simulate the yarn shrinkage process, solves for the amount of yarn shrinkage and the change of the coil angle at different times, and obtains the shrinkage deformation data at the yarn scale.
[0010] Optionally, in S600, the step of constructing a fabric-scale shrinkage prediction model based on the physical structure parameters and shrinkage deformation data includes: S601, based on different fabric structures, determines the interaction between yarn shrinkage deformation data and fabric loops; S602, based on the principle of energy conservation, combined with interaction relationships and yarn shrinkage deformation data, establishes a fabric-scale shrinkage prediction model; S603, compare the shrinkage rate predicted by the shrinkage rate prediction model with the actual measured shrinkage rate, adjust the parameters of the shrinkage rate prediction model according to the comparison results, until the comparison results are lower than the deviation threshold, and obtain the adjusted shrinkage rate prediction model.
[0011] Optionally, in S700, the step of inputting the historical shrinkage rate data into the shrinkage rate prediction model for iterative training to obtain the trained shrinkage rate prediction model includes: S701 divides the historical shrinkage rate dataset into a training set and a validation set according to a set ratio; S702, Obtain the hyperparameters of the shrinkage rate prediction model, and optimize the hyperparameters of the shrinkage rate prediction model through grid search. The range of grid search includes the range of tree depth and the range of the number of trees. S703: Stop iterative training when the error of the shrinkage rate prediction model on the validation set is lower than the set error threshold, and obtain the trained shrinkage rate prediction model.
[0012] On the other hand, embodiments of the present invention provide a shrinkage performance prediction system for knitted fabrics, comprising: At least one processor; At least one memory for storing at least one program; When the at least one program is executed by the at least one processor, the at least one processor performs the method described above.
[0013] The beneficial effects of this invention are as follows: This invention discloses a method and system for predicting the shrinkage performance of knitted fabrics. By acquiring multi-dimensional parameters, a model is constructed from the fiber to the yarn to the fabric scale, which can comprehensively consider various factors affecting the shrinkage rate of knitted fabrics. This progressive modeling approach has higher accuracy and efficiency compared to traditional methods that rely on experience-based judgment and long-term washing measurements. Attached Figure Description
[0014] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0015] Figure 1 This is a flowchart illustrating a method for predicting the shrinkage properties of knitted fabrics according to an embodiment of the present invention.
[0016] Figure 2 This is a schematic diagram of the structure of a shrinkage performance prediction system for knitted fabrics provided in an embodiment of the present invention. Detailed Implementation
[0017] The following will provide a clear and complete description of the concept, specific structure, and technical effects of the present invention in conjunction with embodiments and accompanying drawings, so as to fully understand the purpose, solution, and effects of the present invention. It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other.
[0018] In related technologies, methods for predicting the shrinkage properties of knitted fabrics mainly suffer from the following drawbacks: Single-scale models are insufficient to reflect the cross-scale deformation transfer mechanism of fiber-yarn-fabric; The dynamic changes in fiber molecular structure under humid and hot conditions were not considered; The modeling of the nonlinear relationship between process parameters and shrinkage rate is inaccurate; The prediction results lack interpretability and are difficult to guide process optimization.
[0019] To address the aforementioned technical problems, this invention provides a method and system for predicting the shrinkage properties of knitted fabrics, by referring to... Figure 1 ,like Figure 1 The figure shows a method for predicting the shrinkage properties of knitted fabrics according to an embodiment of the present invention. The method includes the following steps: S100: Obtain the physical structure parameters, fiber material parameters, historical shrinkage data, and finishing process parameters of the target knitted fabric. S200, Based on the physical structure parameters and fiber material parameters, construct a fiber-scale molecular dynamics model; S300, based on the molecular dynamics model, a humid and hot environment simulation was performed to obtain hygroscopic expansion data at the fiber scale; S400, based on the physical structure parameters and moisture absorption expansion data, construct a mechanical response model at the yarn scale; S500, based on the mechanical response model, yarn shrinkage simulation is performed to obtain shrinkage deformation data at the yarn scale; S600, Based on the physical structure parameters and shrinkage deformation data, construct a fabric-scale shrinkage prediction model; S700, input the historical shrinkage rate data into the shrinkage rate prediction model for iterative training to obtain the trained shrinkage rate prediction model; S800, based on the post-finishing process parameters and the trained shrinkage prediction model, outputs the predicted shrinkage rate of the target knitted fabric.
[0020] In the embodiments provided by this invention, the cross-scale deformation transfer mechanism of knitted fabrics from the micro to the macro scale is fully considered through the gradual establishment of a molecular dynamics model at the fiber scale, a mechanical response model at the yarn scale, and a shrinkage prediction model at the fabric scale. Simulating a humid and hot environment using a molecular dynamics model can accurately capture the dynamic changes in the molecular structure of fibers under humid and hot conditions, providing accurate basic data for subsequent models.
[0021] In the process of constructing yarn shrinkage simulation and fabric shrinkage prediction models, combining physical structural parameters with data such as moisture absorption expansion and shrinkage deformation allows for a relatively accurate modeling of the nonlinear relationship between process parameters and shrinkage rate. By inputting historical shrinkage rate data into the prediction model for iterative training, and dividing the model into training and validation sets, techniques such as random forest algorithms and grid search are used to optimize hyperparameters, making the prediction model more accurate and reliable. Furthermore, iteration stops when the prediction error on the validation set reaches a certain standard, ensuring the model's stability.
[0022] Finally, based on the trained shrinkage prediction model and the finishing process parameters, the predicted shrinkage rate is output, making the prediction results interpretable. This provides strong guidance for optimizing the production process of knitted fabrics, effectively improving the production efficiency and product quality of knitted fabrics, thereby solving the main defects in related technologies.
[0023] As an improvement to the above embodiment, in step S100, the physical structure parameters include coil density, yarn count, yarn twist, and fabric structure; the fiber material parameters are obtained by Fourier transform infrared spectroscopy testing of fiber samples, including fiber crystallinity, orientation, and moisture regain; the finishing process parameters include setting temperature, setting time, mechanical pre-shrinking force, and overfeed rate.
[0024] In this step, by clearly defining the specific details of various parameters, the information obtained regarding the target knitted fabric becomes more accurate and comprehensive. For example, different loop densities result in variations in the tightness of the knitted fabric, thus affecting its shrinkage performance; yarn count reflects the fineness of the yarn, and yarns of different fineness exhibit different behaviors during moisture absorption, expansion, and shrinkage. Obtaining fiber material parameters such as crystallinity, orientation, and moisture regain through Fourier transform infrared spectroscopy allows for a deeper understanding of fiber characteristics at the microscopic level, providing crucial data support for constructing fiber-scale molecular dynamics models. Furthermore, in the finishing process parameters, different settings for setting temperature and setting time alter the internal structure of the fabric, and adjustments to mechanical pre-shrinkage force and overfeed rate also significantly impact the fabric's shrinkage performance. Accurately obtaining these parameters is a crucial foundation for accurately predicting the shrinkage performance of knitted fabrics, helping subsequent model construction and simulation processes to better reflect reality and improve the accuracy and reliability of predictions.
[0025] As an improvement to the above embodiment, the construction of the fiber-scale molecular dynamics model in step S200 includes, S201, establish the initial conformation of cellulose macromolecular chain based on fiber chemical structure; S202, set the water molecule distribution model and temperature gradient field; S203, setting the intermolecular interaction potential function based on CHARMM force field parameters; S204 uses periodic boundary conditions to construct a fiber unit cell model.
[0026] In some embodiments, the fiber unit cell model has a size of 10 nm × 10 nm × 20 nm. By establishing the initial conformation of the cellulose macromolecular chain based on the fiber's chemical structure, a foundation is laid for the molecular dynamics model, enabling the model to accurately reflect the microstructure within the fiber. The water molecule distribution model and temperature gradient field settings simulate the interaction between the fiber and water molecules in a real environment, as well as the impact of temperature changes on the fiber. In actual humid and hot environments, the distribution of water molecules and temperature changes are crucial factors affecting the fiber's hygroscopic expansion. Setting the intermolecular interaction potential function based on the CHARMM force field parameters accurately describes the intermolecular forces, which is essential for accurately simulating the dynamic changes of fiber molecules. Furthermore, using periodic boundary conditions to construct the fiber unit cell model allows for the simulation of an infinitely large fiber system with limited computational resources, improving computational efficiency and simulation accuracy. The fiber-scale molecular dynamics model constructed through this series of steps can more realistically simulate the behavior of fibers in humid and hot environments, ensuring accurate hygroscopic expansion data for subsequent acquisition. This provides reliable data support for the construction of yarn-scale mechanical response models and fabric-scale shrinkage prediction models, further enhancing the scientific rigor and accuracy of the entire method for predicting the shrinkage performance of knitted fabrics.
[0027] As an improvement to the above embodiment, step S300 includes, S301, under specific temperature and relative humidity conditions, performs moisture absorption simulation and records the curves of fiber axial and radial expansion rates over time; S302, extract the maximum expansion rate under equilibrium state as the hygroscopic expansion characteristic value.
[0028] In this step, moisture absorption simulation is conducted under specific temperature and relative humidity conditions. The simulated temperature and humidity environment closely resembles the humid and hot environment that actual knitted fabrics may encounter during storage or use, thus more realistically reflecting the moisture absorption characteristics of the fibers. In one embodiment, moisture absorption simulation is performed at a temperature of 40±5℃ and a relative humidity of 65±5%. The changes in the axial and radial expansion rates of the fibers over time are recorded. These curves visually demonstrate the dynamic process of fiber expansion in different directions during moisture absorption, providing detailed data for in-depth research on the moisture absorption and expansion behavior of fibers. The maximum expansion rate under equilibrium conditions is extracted as the moisture absorption and expansion characteristic value. This characteristic value is a key indicator for measuring the degree of fiber moisture absorption and expansion, comprehensively reflecting the fiber's moisture absorption and expansion capacity under a given humid and hot environment. Accurately obtaining this characteristic value provides important input parameters for subsequently constructing a yarn-scale mechanical response model, as the mechanical response of the yarn is closely related to the moisture absorption and expansion of the fibers. The yarn-scale mechanical response model constructed based on this moisture absorption and expansion characteristic value can more accurately simulate the mechanical changes of yarn caused by fiber expansion after moisture absorption, thereby further improving the accuracy of the fabric-scale shrinkage prediction model. This makes the entire method for predicting the shrinkage performance of knitted fabrics more practical and reliable, helping manufacturers to more accurately predict fabric shrinkage performance, take corresponding measures in advance to optimize production processes, and reduce product quality problems caused by fabric shrinkage.
[0029] As an improvement to the above embodiments, the step of constructing a yarn-scale mechanical response model based on the physical structure parameters and moisture absorption expansion data includes: S401, Based on the yarn geometry and fiber moisture absorption and expansion data, establish a stress distribution model inside the yarn; Because the geometry of yarn determines the arrangement of its internal fibers, and the expansion of fibers after absorbing moisture will generate stress inside the yarn, accurately establishing a stress distribution model can better understand the mechanical changes of yarn after absorbing moisture.
[0030] The expression for the stress distribution model inside the yarn is: σ = E × ε, where σ is the stress, E is the elastic modulus of the yarn, and ε is the strain, which can be calculated based on the fiber's moisture absorption and expansion data. This model allows for the quantification of the stress generated inside the yarn due to the fiber's moisture absorption and expansion.
[0031] S402 combines the elastic modulus and Poisson's ratio of the yarn to transform the stress distribution model into a mechanical response model.
[0032] Elastic modulus and Poisson's ratio are key parameters describing the mechanical properties of yarn. By incorporating them into the mechanical response model, the equations can more accurately reflect the actual response of the yarn under stress. The mechanical response model is expressed as: σx = E × (εx + ν × εy), σy = E × (εy + ν × εx); where σx is the stress in the x-direction, σy is the stress in the y-direction, E is the elastic modulus, εx and εy are the strains in the x and y directions, respectively, and ν is Poisson's ratio. This equation fully considers the stress-strain relationship of the yarn in different directions, as well as the influence of yarn material properties on the mechanical response. By constructing such equations, the mechanical changes of the yarn in each direction under stress caused by fiber moisture absorption and expansion can be more comprehensively and accurately reflected. This is of great significance for subsequent solution using the finite element analysis method to accurately simulate the yarn shrinkage process. Only by accurately describing the mechanical response of the yarn can boundary conditions and material properties be reasonably set in the finite element analysis, thereby obtaining yarn shrinkage and deformation data that are closer to reality. These data will provide a more reliable foundation for the construction of fabric-scale shrinkage prediction models, further improving the accuracy and practicality of the entire knitted fabric shrinkage performance prediction system.
[0033] As an improvement to the above embodiment, step S500, which involves simulating yarn shrinkage based on the mechanical response model to obtain yarn-scale shrinkage deformation data, includes: S501 uses the finite element method to numerically solve the mechanical response model, simulates the shrinkage behavior of yarn under different stress conditions, and obtains shrinkage deformation data at the yarn scale. The finite element method (FEM) discretizes complex mechanical problems, facilitating numerical solutions by computers. This allows for the acquisition of yarn shrinkage and deformation data under various conditions, providing detailed and accurate data support for subsequent construction of fabric-scale shrinkage prediction models. When solving the mechanical response model using the finite element analysis method, the yarn is divided into numerous tiny finite element elements. For each element, its stress and strain states at different times are determined based on the force response equation. By integrating and analyzing these element states, the overall shrinkage process of the yarn after moisture absorption can be simulated in detail. For example, during the simulation, the different degrees of shrinkage caused by stress distribution differences between the yarn surface and internal elements, as well as details such as the sequence of shrinkage in different parts of the yarn, can be observed. This accurately simulated yarn-scale shrinkage and deformation data provides more detailed and reliable data support for constructing fabric-scale shrinkage prediction models.
[0034] S502, based on the Peirce loop geometric model, establishes the knitted loop topology, sets the boundary constraint condition to fix the four corners of the fabric, and constructs the yarn-scale shrinkage deformation equation based on the yarn-scale shrinkage deformation data and the principle of energy minimization: ;in, The total energy of the system. The elastic modulus of the yarn. This represents the change in yarn length. The included angle of the coil, Let i be the initial coil angle, i be the yarn number, n be the total number of yarns, j be the coil number, m be the total number of coils, and λ be the balance factor, with a value between 0 and 1. S503 uses the shrinkage deformation equation to numerically simulate the yarn shrinkage process, solves for the amount of yarn shrinkage and the change of the coil angle at different times, and obtains the shrinkage deformation data at the yarn scale.
[0035] In this step, the knitting loop topology is established based on the Peirce loop geometric model, which accurately describes the shape of the knitting loop from a geometric perspective, providing a clear structural foundation for subsequent calculations. The loop transfer matrix is calculated based on yarn shrinkage deformation data. This step quantifies the relationship between yarn shrinkage and loops, enabling the model to dynamically reflect the structural changes of the fabric under the influence of yarn shrinkage. The fabric deformation equation is constructed using the energy minimization principle, comprehensively considering factors such as yarn elasticity coefficient, length change, and loop angle change, describing the fabric deformation trend in the form of total system energy, revealing the inherent laws of fabric shrinkage from an energy perspective. Boundary constraints are set to fix the four corners of the fabric, simulating possible boundary states of the fabric during testing or use, making the model closer to reality and ensuring the accuracy and reliability of the calculation results. These steps are closely linked, jointly constructing a relatively complete and scientifically reasonable fabric-scale shrinkage prediction model, which can effectively predict fabric shrinkage under various conditions.
[0036] During the simulation, the knitted loop topology established based on the Peirce loop geometry model accurately reflects the spatial layout of yarns in the knitted fabric. The boundary constraints of fixing the four corners of the fabric simulate the actual boundary limitations of the fabric when subjected to external forces or deformation. The shrinkage deformation equation, constructed using the energy minimization principle, comprehensively considers the influence of yarn elasticity, length changes, and loop angle changes on the total system energy. The balance factor λ, ranging from 0 to 1, balances the contributions of yarn elasticity and loop angle changes to the system energy, making the simulation results more consistent with reality. By accurately solving this shrinkage deformation equation, detailed data on the dynamic changes of yarn during shrinkage are obtained. This data provides crucial information for the subsequent construction of fabric-scale shrinkage prediction models, helping to more accurately predict the shrinkage of knitted fabrics during actual use.
[0037] As an improvement to the above embodiment, the step S600 of constructing a fabric-scale shrinkage prediction model includes: S601, based on different fabric structures, determines the interaction between yarn shrinkage deformation data and fabric loops; The interaction relationship is expressed as F = k × Δx, where F is the interaction force between the loops, k is the interaction coefficient, and Δx is the relative displacement of the loops caused by yarn shrinkage. This expression quantifies the relationship between yarn shrinkage and loop interaction, laying the foundation for subsequent fabric-scale shrinkage prediction models based on the principle of energy conservation. For plain knit weaves, due to their relatively simple structure and regular loop arrangement, the interaction force between loops during yarn shrinkage is mainly concentrated at the longitudinal and transverse connection points, resulting in a relatively small and stable interaction coefficient k. However, for rib knit weaves, due to their unique warp and transverse interlacing structure, the relative displacement of loops caused by yarn shrinkage is more significant in the warp direction, and the interaction coefficient k differs from that of plain knit weaves in the warp direction. Specific analysis and determination are needed based on the structural characteristics of rib knit weaves. For different fabric structures, the interaction coefficient k can be determined through a combination of experimental measurement and data analysis. First, for specific fabric structures, such as plain knit, rib, and double-sided knit, a series of fabric samples with different degrees of yarn shrinkage are prepared. During sample preparation, other influencing factors, such as fiber material, yarn count, and twist, were strictly controlled to remain consistent, with only the yarn shrinkage parameters being varied. Then, using high-precision force sensors and displacement measurement equipment, the relative displacement Δx of the coils caused by yarn shrinkage and the corresponding interaction force F between the coils were precisely measured in each sample. After collecting a large amount of such data, the expression F=k×Δx was solved using mathematical methods such as least squares fitting, thus obtaining the value of the interaction coefficient k under the specific fabric structure. Simultaneously, considering that the fabric may be affected by various environmental factors during actual use, such as temperature and humidity, the above experimental process needs to be repeated under different temperature and humidity conditions to analyze the influence of environmental factors on the interaction coefficient k, thereby establishing a functional relationship between the interaction coefficient k and environmental factors. In this way, when predicting fabric shrinkage, the value of the interaction coefficient k can be determined more accurately based on specific environmental conditions and fabric structure, improving the accuracy of the fabric-scale shrinkage prediction model and providing a more reliable basis for quality control and process optimization in the production of knitted fabrics. By accurately determining the interaction relationships under different fabric structures, the influence of yarn shrinkage on fabric loops can be precisely described. This allows for a more accurate prediction of shrinkage rate at the fabric scale, better reflecting the actual deformation of the fabric and improving the accuracy and reliability of the prediction model.
[0038] S602, based on the principle of energy conservation, combined with interaction relationships and yarn shrinkage deformation data, establishes a fabric-scale shrinkage prediction model.
[0039] The principle of energy conservation is crucial in establishing fabric-scale shrinkage prediction models. Energy remains conserved during the interaction between yarn shrinkage and fabric loops, providing a solid theoretical foundation for model construction. Taking plain knit fabric as an example, when the yarn shrinks after absorbing moisture and expanding, its shrinkage deformation data reflects the change in energy within the yarn. This energy is transferred to the overall fabric structure through the interaction between the yarn and fabric loops. Based on the interaction relationship expression F=k×Δx and the yarn shrinkage deformation data, we can analyze how energy is distributed and transformed among the loops in plain knit fabric.
[0040] For rib knit fabrics, due to their structural characteristics, the energy transfer and conversion paths differ from those of plain knit fabrics. The relative displacement of loops caused by yarn shrinkage is more pronounced in the warp direction, making the energy changes in this direction more significant. When determining the interaction coefficient k, considering the unique energy transfer mechanism of rib knit fabrics, experimental measurements and data analysis can yield a k value that better reflects their structural characteristics. Using this data, a fabric-scale shrinkage prediction model based on the principle of energy conservation can more accurately reflect the shrinkage of rib knit fabrics under the influence of yarn shrinkage.
[0041] For double-sided weaves, the complex coil structure makes energy transfer and interaction more complex and diverse. When building a model based on the principle of energy conservation, it is necessary not only to consider yarn shrinkage deformation data and the interaction between coils, but also to deeply analyze the unique energy dissipation and conversion mechanisms of double-sided weaves. For example, the tilt angle and nesting method of the coils in double-sided weaves will lead to special situations in the distribution and conversion of energy in different directions during yarn shrinkage. By comprehensively considering these complex factors, and combining experimental measurements and theoretical analysis, an accurate shrinkage prediction model for double-sided weaves can be established.
[0042] After establishing a fabric-scale shrinkage prediction model, extensive practical testing is required to verify its accuracy and reliability. Knitted fabric samples with different fiber materials, yarn counts, twists, and finishing process parameters are selected, covering various fabric structures such as plain knit, rib knit, and double-sided knit. The samples are placed in a specific temperature and humidity environment to simulate the damp and heat conditions in actual use, and the actual shrinkage rate of the samples is measured. Simultaneously, using the established prediction model, corresponding physical structure parameters, fiber material parameters, finishing process parameters, and moisture absorption and expansion data are input to calculate the predicted shrinkage rate. The actual shrinkage rate and the predicted shrinkage rate are compared, and the difference between the two is analyzed. If the difference is within an acceptable range, the model has high accuracy and reliability; if the difference is large, further optimization and adjustment of the model are needed. Through continuous verification and optimization, the fabric-scale shrinkage prediction model can be better applied to actual production, providing more accurate shrinkage performance predictions for knitted fabric manufacturers, helping them optimize production processes, improve product quality, and reduce production costs.
[0043] The final shrinkage prediction model expression is S=f(D,T,M,H,α,β), where D represents yarn diameter, T represents twist, M represents fiber material properties, H represents finishing process parameters, and α and β are energy transfer and dissipation coefficients. This model can accurately predict the shrinkage rate of fabrics under different conditions, providing a scientific basis for production. Further research revealed that the α and β coefficients in the model are significantly affected by environmental temperature and humidity, especially under high temperature and high humidity conditions, where the changes in α and β values are more drastic, necessitating corresponding corrections to the prediction results. By introducing temperature and humidity correction factors, the model's adaptability is significantly improved, ensuring prediction accuracy under different environments. The calculation formula for each parameter in the expression is D=d0. (1+ε),T=t0 (1-δ),M=m0 (1+η),H=h0 (1-θ), α=α0 (1+κ ΔT), β=β0 (1+λ ΔH), where ε, δ, η, and θ are the coefficients for yarn deformation, twist change, fiber property change, and finishing parameter change, respectively; κ and λ are the temperature and humidity influence coefficients; and ΔT and ΔH are the temperature and humidity changes. Through precise measurement and calculation of each parameter, combined with extensive experimental data, the model's prediction error is significantly reduced. In practical applications, companies can adjust production parameters according to the model to ensure that fabric shrinkage is controlled within the ideal range.
[0044] S603, compare the shrinkage rate predicted by the shrinkage rate prediction model with the actual measured shrinkage rate, adjust the parameters of the shrinkage rate prediction model according to the comparison results, until the comparison results are lower than the deviation threshold, and obtain the adjusted shrinkage rate prediction model.
[0045] The model was validated using sample data of knitted fabrics from actual production, and the predicted shrinkage rate was compared with the actual measured shrinkage rate. If there were discrepancies between the predicted results and the actual data, the causes of the discrepancies were analyzed, which may be due to inaccurate determination of the interaction relationships or the neglect of certain factors in the energy calculation process. To address these issues, the model was optimized and adjusted, such as recalibrating the interaction coefficient k and considering more factors affecting the energy state of the fabric. Through continuous validation and optimization, the accuracy and stability of the fabric-scale shrinkage prediction model were improved.
[0046] For example, in plain knit fabrics, due to their relatively simple loop structure, yarn shrinkage mainly causes stretching and compression of the loops in the horizontal and vertical directions. By analyzing the degree and direction of yarn shrinkage, the magnitude and direction of the interaction force between the loops can be determined relatively intuitively. However, for rib knit fabrics, due to their unique longitudinal elasticity, yarn shrinkage not only generates direct tension and pressure between the loops but also triggers longitudinal deformation and adjustment of the loop structure. Therefore, the interaction relationship needs to be determined by comprehensively considering the amount of yarn shrinkage, loop density, and the elastic characteristics of the rib knit fabric. Accurately determining these relationships is the foundation for establishing a fabric-scale shrinkage prediction model based on the principle of energy conservation. Only by clearly understanding the interaction between the yarn and the loops can the energy changes of the fabric during the shrinkage process be reasonably calculated, thereby accurately predicting the shrinkage rate.
[0047] As an improvement to the above embodiment, step S700 includes, S701 divides the historical shrinkage rate dataset into a training set and a validation set according to a set ratio; S702, Obtain the hyperparameters of the shrinkage rate prediction model, and optimize the hyperparameters of the shrinkage rate prediction model through grid search. The range of grid search includes the range of tree depth and the range of the number of trees. S703: Stop iterative training when the error of the shrinkage rate prediction model on the validation set is lower than the set error threshold, and obtain the trained shrinkage rate prediction model.
[0048] In this step, the historical shrinkage rate dataset is divided into training and validation sets in a 7:3 ratio. This division ensures sufficient data for model training while reserving a certain amount for validating model performance. Hyperparameters are optimized using a grid search, with a tree depth range of 5-15 and a tree count range of 100-500. Searching within this range finds the most suitable hyperparameter combination for the current dataset, further improving the model's prediction accuracy. Iteration stops when the validation set prediction error is ≤3%. This provides a clear termination condition for the model training process, avoiding overfitting caused by overtraining and ensuring high prediction accuracy and stability on the validation set. This training process makes the trained shrinkage rate prediction model more accurate and reliable, providing stronger support for predicting the shrinkage performance of knitted fabrics, helping companies better control product quality and improve production efficiency. Furthermore, in practical applications, new historical shrinkage rate data can be collected periodically, and the model can be retrained and optimized according to the above process to adapt to the impact of improvements in production processes and changes in raw materials on the shrinkage performance of knitted fabrics, ensuring the model maintains good predictive performance. Furthermore, the trained model can be visualized to intuitively show the relationship between each parameter and the shrinkage rate, providing a more intuitive reference for optimizing the production process.
[0049] In some examples, when the predicted shrinkage rate exceeds a preset threshold, an optimized process parameter scheme is generated; after adjusting and reorganizing the process parameters according to the optimized scheme, S700-S800 are re-executed; the optimized process parameter scheme includes: When the predicted warp shrinkage rate is >5%, it is recommended to increase the overfeed rate by 2-5%. When the predicted latitudinal shrinkage rate is >8%, it is recommended to reduce the setting temperature by 10-15℃. When the predicted difference in bidirectional shrinkage rate is >3%, it is recommended to adjust the distribution of mechanical pre-shrinkage force.
[0050] After predicting the shrinkage rate, output a shrinkage rate distribution cloud map and mark the coordinates of the region with the largest shrinkage; generate a cross-scale correlation report containing fiber expansion rate, yarn shrinkage rate and fabric shrinkage rate; and store the prediction model parameters and optimization path to the process knowledge base.
[0051] refer to Figure 2 This invention also provides a shrinkage prediction system for knitted fabrics, comprising: At least one processor; At least one memory for storing at least one program; When the at least one program is executed by the at least one processor, the at least one processor performs the method described above.
[0052] The content of the above method embodiments is applicable to this embodiment. The specific functions implemented in this embodiment are the same as those in the above method embodiments, and the beneficial effects achieved are also the same as those achieved in the above method embodiments. Therefore, they will not be repeated here.
[0053] Although the description of this disclosure has been quite detailed and particularly focused on several of the described embodiments, it is not intended to limit itself to any of these details or embodiments or any particular embodiment, but should be considered as effectively covering the intended scope of this disclosure by referring to the appended claims and taking into account the broad possible interpretations of these claims provided by the prior art. Furthermore, the foregoing description of this disclosure with respect to embodiments foreseeable by the inventors is intended to provide a useful description, and non-substantial modifications to this disclosure that have not yet been foreseen may still represent equivalent modifications.
Claims
1. A method for predicting the shrinkage properties of knitted fabrics, characterized in that, Includes the following steps, S100: Obtain the physical structure parameters, fiber material parameters, historical shrinkage data, and finishing process parameters of the target knitted fabric. S200, Based on the physical structure parameters and fiber material parameters, construct a fiber-scale molecular dynamics model; S300, based on the molecular dynamics model, a humid and hot environment simulation was performed to obtain hygroscopic expansion data at the fiber scale; S400, based on the physical structure parameters and moisture absorption expansion data, construct a mechanical response model at the yarn scale; S500, based on the mechanical response model, yarn shrinkage simulation is performed to obtain shrinkage deformation data at the yarn scale; S600, Based on the physical structure parameters and shrinkage deformation data, construct a fabric-scale shrinkage prediction model; S700, input the historical shrinkage rate data into the shrinkage rate prediction model for iterative training to obtain the trained shrinkage rate prediction model; S800, based on the post-finishing process parameters and the trained shrinkage prediction model, outputs the predicted shrinkage rate of the target knitted fabric.
2. The method according to claim 1, characterized in that, In step S100, the physical structure parameters include coil density, yarn count, yarn twist, and fabric structure; the fiber material parameters are obtained by Fourier transform infrared spectroscopy testing of fiber samples, including fiber crystallinity, orientation, and moisture regain; the finishing process parameters include setting temperature, setting time, mechanical pre-shrinking force, and overfeed rate.
3. The method according to claim 1, characterized in that, In step S200, constructing the fiber-scale molecular dynamics model includes, S201, establish the initial conformation of cellulose macromolecular chain based on fiber chemical structure; S202, set the water molecule distribution model and temperature gradient field; S203, setting the intermolecular interaction potential function based on CHARMM force field parameters; S204 uses periodic boundary conditions to construct a fiber unit cell model.
4. The method according to claim 1, characterized in that, In step S300, the simulation of the hygrothermal environment based on the molecular dynamics model to obtain fiber-scale hygroscopic expansion data includes, S301, under specific temperature and relative humidity conditions, performs moisture absorption simulation and records the curves of fiber axial and radial expansion rates over time; S302, extract the maximum expansion rate under equilibrium state as the hygroscopic expansion characteristic value.
5. The method according to claim 1, characterized in that, In step S400, constructing a mechanical response model at the yarn scale based on the physical structure parameters and moisture absorption and expansion data includes: S401, Based on the yarn geometry and fiber moisture absorption and expansion data, establish a stress distribution model inside the yarn; S402 combines the elastic modulus and Poisson's ratio of the yarn to transform the stress distribution model into a mechanical response model.
6. The method according to claim 1, characterized in that, In step S500, the yarn shrinkage simulation based on the mechanical response model to obtain yarn-scale shrinkage deformation data includes: S501 uses the finite element method to numerically solve the mechanical response model, simulates the shrinkage behavior of yarn under different stress conditions, and obtains shrinkage deformation data at the yarn scale. S502, based on the Peirce loop geometric model, establishes the knitted loop topology, sets the boundary constraint condition to fix the four corners of the fabric, and constructs the yarn-scale shrinkage deformation equation based on the yarn-scale shrinkage deformation data and the principle of energy minimization: ;in, The total energy of the system. The elastic modulus of the yarn. This represents the change in yarn length. The included angle of the coil, Let i be the initial coil angle, i be the yarn number, n be the total number of yarns, j be the coil number, m be the total number of coils, and λ be the balance factor, with a value between 0 and 1. S503 uses the shrinkage deformation equation to numerically simulate the yarn shrinkage process, solves for the amount of yarn shrinkage and the change of the coil angle at different times, and obtains the shrinkage deformation data at the yarn scale.
7. The method according to claim 1, characterized in that, In S600, the step of constructing a fabric-scale shrinkage prediction model based on the physical structure parameters and shrinkage deformation data includes: S601, based on different fabric structures, determines the interaction between yarn shrinkage deformation data and fabric loops; S602, based on the principle of energy conservation, combined with interaction relationships and yarn shrinkage deformation data, establishes a fabric-scale shrinkage prediction model; S603, compare the shrinkage rate predicted by the shrinkage rate prediction model with the actual measured shrinkage rate, adjust the parameters of the shrinkage rate prediction model according to the comparison results, until the comparison results are lower than the deviation threshold, and obtain the adjusted shrinkage rate prediction model.
8. The method according to claim 1, characterized in that, In S700, the step of inputting the historical shrinkage rate data into the shrinkage rate prediction model for iterative training to obtain the trained shrinkage rate prediction model includes: S701 divides the historical shrinkage rate dataset into a training set and a validation set according to a set ratio; S702, Obtain the hyperparameters of the shrinkage rate prediction model, and optimize the hyperparameters of the shrinkage rate prediction model through grid search. The range of grid search includes the range of tree depth and the range of the number of trees. S703: Stop iterative training when the error of the shrinkage rate prediction model on the validation set is lower than the set error threshold, and obtain the trained shrinkage rate prediction model.
9. A shrinkage performance prediction system for knitted fabrics, characterized in that, include, At least one processor; At least one memory for storing at least one program; When the at least one program is executed by the at least one processor, the at least one processor performs the method as described in any one of claims 1 to 8.