Dynamic process testing method for air permeability and moisture permeability of nonwoven fabrics used in mask fabric production

Abstract

This invention relates to the field of textile material testing technology, specifically disclosing a dynamic process testing method for the air permeability and moisture permeability of nonwoven fabrics used in the production of facial mask fabrics. The method simulates the stretching, die-cutting, and facial surface bonding processes during production using a simulation platform to obtain a high-fidelity dataset of pore structure deformation. A dynamic coupling environment is constructed by combining parameters of breathing, sweating, and facial micro-movements, and performance data is simultaneously collected in real-time on a bionic skin interface to obtain a time-series record of air permeability and moisture permeability performance. Subsequently, the influence of breathing on fabric undulations is analyzed, sweating stages are divided and performance fluctuation ranges are determined, the interference of facial movements on moisture permeability is assessed, and the contribution weight of each factor is quantified. Finally, through threshold comparison and parameter iterative optimization, a full-cycle performance spectrum and comfort evaluation report are generated. This invention achieves accurate detection and mechanism diagnosis of the dynamic performance of facial mask substrates under simulated real-world usage scenarios, providing an effective solution for optimizing the comfort of facial mask fabric-related products.

Description

Technical Field

[0001] This invention relates to the field of textile material testing technology, specifically a dynamic process testing method for the air permeability and moisture permeability of nonwoven fabrics used in the production of face mask fabrics. Background Technology

[0002] Nonwoven fabrics, as the base material for face masks, directly determine the user's comfort and skincare effects during application, holding a crucial position in the cosmetic materials field. Face mask application typically lasts 15 to 25 minutes, during which skin temperature, humidity, respiratory airflow, and subtle facial movements continuously change. These dynamic interactions in real-world scenarios significantly impact the fabric's breathability and moisture permeability. Current testing methods generally involve cutting and laying the sample flat, fixing it on an instrument, and measuring air and moisture permeability under constant temperature, air pressure, or humidity differences. This method ignores the real-world changes throughout the entire process from production to actual application. During production, nonwoven fabrics undergo significant stretching and die-cutting; during application, they are subjected to skin curvature and negative pressure adsorption. These processes alter the shape, size, and connectivity of the fabric's internal pores. Flattened samples cannot retain these deformation characteristics, leading to a significant difference between the pore structure measured during testing and that during actual use.

[0003] A deeper problem is that existing tests can only obtain values ​​at a single moment in a stable environment without dynamic interference. They cannot reflect the gradual increase in skin perspiration, the alternation of positive and negative pressure caused by breathing, and the local compression and relaxation caused by facial expressions throughout the entire application process. These factors act simultaneously and influence each other. The gradual seepage of sweat will cause the pores to be filled with liquid, the periodic changes in the direction and intensity of breathing airflow will cause the fabric to undulate slightly, and facial micro-movements will further change the tightness of the fit. All three factors together determine the actual breathability and moisture permeability of the fabric at different times. If these dynamic factors cannot be coupled together for synchronous observation, it is difficult to know the differences in the fabric's performance at different stages of actual application, and it is also impossible to determine which manufacturing process can maintain better comfort throughout the entire use cycle. Summary of the Invention

[0004] The purpose of this invention is to provide a dynamic process testing method for the air permeability and moisture permeability of nonwoven fabrics used in the production of face mask fabrics, which realizes accurate detection, mechanism diagnosis and comfort prediction of the full-cycle dynamic process of air permeability and moisture permeability of face mask substrate under simulated real use environment.

[0005] The objective of this invention can be achieved through the following technical solutions:

[0006] This application provides a dynamic process testing method for the air permeability and moisture permeability of nonwoven fabrics used in the production of facial mask fabrics, including the following steps:

[0007] S1. By constructing a simulation platform, the pore structure change data of nonwoven fabric under stretching, die-cutting and skin curvature bonding is obtained for the full cycle change of mask substrate from production and molding process to application and use state, and an initial deformation feature dataset is generated.

[0008] S2. Based on the initial deformation feature dataset, a dynamic environment simulation module is used to construct a multi-factor coupled environment in a real-world usage scenario by combining the periodic parameters of changes in respiratory airflow, gradual increase in sweating, and minor facial movements, and outputting a dynamic interaction impact dataset.

[0009] S3. For the dynamic interaction impact dataset, simulate the application state on the bionic skin interaction interface, and collect the instantaneous performance capture data of the contact interface between the fabric and the bionic skin in real time to obtain the time sequence change record of breathability and moisture permeability.

[0010] S4. By recording the time-series changes, analyze the degree of influence of changes in respiratory airflow on the slight undulations of the fabric, and combine the changes in pore structure caused by the gradual increase in perspiration to determine the range of air permeability fluctuations in different time periods.

[0011] S5. Based on the aforementioned air permeability fluctuation range, assess the interference of local fit changes caused by minor facial movements on the moisture permeability, obtain the contribution weight of each dynamic factor to the overall performance, and determine the overall influence trend.

[0012] S6. In response to the overall impact trend, a preset threshold comparison method is adopted. If the air permeability and moisture permeability performance is lower than the preset threshold in a certain time period, the dynamic environment simulation parameters are adjusted and the optimized performance time series data is regenerated.

[0013] S7. By using the optimized performance time-series data, construct a full-cycle performance spectrum of breathability and moisture permeability, determine the comfort distribution of the mask substrate in real-world usage scenarios, and output the final evaluation results.

[0014] The beneficial effects of this invention are as follows:

[0015] By constructing a simulation platform covering the entire production and application chain, this invention solves the fundamental defect of traditional testing that loses the true deformation characteristics due to sample cutting and laying. In a virtual environment, this invention accurately reproduces the entire process of nonwoven fabric from stretching and die-cutting on the production line to negative pressure adsorption for surface bonding. It dynamically acquires the evolution data of the fabric pore structure at each stage, so that the initial input for testing is no longer a static, homogeneous, idealized sample, but a high-fidelity digital twin that includes the deformation of the production process and the deformation of the application bonding. This fundamentally ensures that the fabric pore structure on which the subsequent performance analysis is based is highly consistent with the actual use state.

[0016] By constructing a multi-factor dynamic coupling environment and synchronously collecting data in real time at the bionic skin interface, this invention overcomes the bottleneck of traditional steady-state detection, which cannot reflect the interactive effects of multiple factors such as breathing, sweating, and movement during actual use. This invention not only simulates the periodic changes in respiratory airflow, the gradual increase in sweat secretion, and subtle facial movements, but more importantly, through multi-physics field coupling solution and high-frequency sensing synchronous acquisition technology, it achieves accurate capture and recording of the instantaneous breathability and moisture permeability of the fabric-skin interface under the combined effect of these three factors. This allows the evaluation to leap from obtaining single, static laboratory data to obtaining continuous, dynamic usage process data, and realizes the complete observation of the dynamic performance of the face mask substrate under simulated real complex scenarios.

[0017] By establishing a complete closed loop of data acquisition, mechanism analysis, parameter optimization, and comprehensive evaluation, this invention solves the deep-seated problem that existing technologies cannot guide process optimization and accurate prediction of comfort. This invention goes beyond performance measurement, further quantifying the influence weights of various dynamic factors, identifying performance mutation intervals and dominant factors, and iteratively optimizing simulation parameters based on performance threshold feedback. By constructing intuitive full-cycle performance maps and comfort distribution maps, it outputs quantified comfort levels and comprehensive reports, elevating testing from simple performance verification to mechanism diagnosis and process optimization guidance. This provides precise and efficient R&D tools and data support for developing mask products that maintain better comfort throughout their entire lifecycle. Attached Figure Description

[0018] To better understand and implement this application, the technical solution is described in detail below with reference to the accompanying drawings.

[0019] Figure 1 This is a flowchart illustrating the dynamic process detection method for the air permeability and moisture permeability of nonwoven fabrics used in the production of face mask fabrics provided in Embodiment 1 of this application.

[0020] Figure 2 This is a flowchart illustrating step S3 in the dynamic process detection method for the air permeability and moisture permeability of nonwoven fabrics used in the production of face mask fabric provided in Embodiment 1 of this application.

[0021] Figure 3 This is a flowchart illustrating step S4 in the dynamic process detection method for the air permeability and moisture permeability of nonwoven fabrics used in the production of face mask fabrics provided in Embodiment 1 of this application. Detailed Implementation

[0022] To further illustrate the technical means and effects adopted by the present invention to achieve its intended purpose, exemplary embodiments will be described in detail below, examples of which are illustrated in the accompanying drawings. In the following description, when referring to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application.

[0023] The terminology used in this application is for the purpose of describing particular embodiments only and is not intended to be limiting of the application. The singular forms “a,” “the,” and “the” used herein are also intended to include the plural forms unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used herein refers to and includes any and all possible combinations of one or more of the associated listed items.

[0024] The following detailed description of the specific implementation methods, features, and effects of the present invention, in conjunction with the accompanying drawings and preferred embodiments, is provided in detail.

[0025] Example 1

[0026] Please see Figures 1-3 This embodiment provides a dynamic process testing method for the air permeability and moisture permeability of nonwoven fabrics used in the production of facial mask fabrics, including the following steps:

[0027] S1. By constructing a simulation platform, data on the changes in the pore structure of nonwoven fabric under stretching, die-cutting, and skin curvature bonding are obtained for the entire cycle of changes in the mask substrate from production and molding to application and use, and an initial deformation feature dataset is generated.

[0028] Further, step S1 specifically includes:

[0029] In the simulation platform, the nonwoven fabric production and forming process simulation is run to generate an initial three-dimensional digital model with a random fiber network structure. Then, a specific tensile load on the simulated production line is applied to the initial three-dimensional digital model to calculate the displacement and rearrangement of the fiber network. The pore size, shape and connectivity distribution data of the fabric under tension are obtained to complete the deformation simulation of the production process.

[0030] The process of generating an initial 3D digital model with a random fiber network structure includes: pre-setting the raw material fiber properties of the nonwoven fabric (such as fiber diameter, length, and bending stiffness) and key production process parameters (such as web laying speed, randomness, and the process pressure and energy of hot rolling or hydroentangling) in the simulation platform. Subsequently, the platform calls a process-based random fiber deposition algorithm or a statistical fiber network reconstruction algorithm to generate a 3D digital model that statistically matches the sample from the real production line (such as fiber orientation distribution and areal density) based on the above parameters. This model consists of a large number of interwoven cylindrical or beam elements representing fibers, with their intersections forming adhesion or entanglement, thus clearly defining the basis and data structure for generating the initial 3D digital model and ensuring the physical foundation of the simulation.

[0031] The pore structure data obtained from the stretching simulation is input into the die-cutting path planning module. The system presets the die-cutting boundary based on the shape of the mask and compares the key dimensions of the pores (such as the equivalent diameter) in the boundary area in real time. If it is detected that the key pore size exceeds the preset process allowable range due to stretching, the system automatically optimizes and adjusts the die-cutting path or boundary position to avoid areas with poor structure and outputs the fabric three-dimensional mesh model and pore data after simulating the actual die-cutting process.

[0032] The automatic optimization adjustment internally presets a threshold for the allowable range of pore size (e.g., an equivalent diameter between X and Y micrometers). The adjustment logic is as follows: when the proportion of defective pores exceeding the threshold within the die-cutting planning boundary exceeds a set value (e.g., 5%), adjustment is triggered. The optimization algorithm can employ iterative boundary shrinkage or local path offset: first, it attempts to proportionally shrink the die-cutting contour; if this still does not meet the requirements, it increases the path shrinkage in areas with dense defective pores based on the pore size distribution map, while ensuring that the adjusted contour is smooth and continuous and meets the minimum size requirements of the mask product. This process clarifies the triggering conditions, judgment criteria, and execution algorithm for adjustment, making the automatic optimization operable.

[0033] The die-cut 3D mesh model is mapped onto a parameterized biomimetic facial skin surface (such as the cheek and forehead areas). In the simulation platform, the negative pressure adsorption force and surface normal adhesion constraint are applied to the model to simulate the application process, driving the mesh to deform to adapt to the skin curvature. Through transient dynamic analysis, the mesh node displacements are calculated from the initial application to the tight adhesion process. The pore structure change sequence under the application state is output in time series to capture the dynamic structural evolution during the application process.

[0034] The process of applying negative pressure adsorption force and surface normal adhesion constraint to the model during simulated application includes: the surface normal adhesion constraint is defined as: forcing the mesh nodes of the fabric model to always be located in the normal direction of the corresponding projection point of the skin surface model during iterative calculations, or to maintain a preset, minute interval distance (e.g., 0.01 mm) representing the moisturizing essence film with the surface. The negative pressure adsorption force is applied to the surface of the fabric model on the skin-adhering side in the form of a distributed pressure perpendicular to the skin surface. Its magnitude is set according to clinical or experimental data as a negative pressure value that varies with time or is constant (e.g., -0.5 kPa to -2 kPa). The static or transient dynamic equations under this composite boundary condition are solved by a finite element analysis solver, thereby clearly defining the physical field and mathematical constraints driving mesh deformation and ensuring the reliability of the adhesion simulation results.

[0035] Based on the pore structure change sequence obtained from the application simulation, for each pore element, the displacement vector, area change rate, and orientation angle, and other multi-dimensional deformation characteristics are calculated at each time point relative to the initial die-cut state. The feature data of all pore elements at all time steps are organized according to time series and spatial location to construct a structured initial deformation feature dataset, which provides high-fidelity structural change input for subsequent air permeability and moisture permeability analysis coupled with dynamic environmental loads.

[0036] The calculation of multi-dimensional deformation features, such as displacement vector, area change rate, and orientation angle, relative to the initial die-cut state at each time point includes: extracting the displacement vector by comparing the node coordinates of the center point of each pore unit at each time step with the initial coordinates in the application simulation; calculating the area change rate by projecting each pore unit onto the plane of its fabric layer and calculating the ratio of the projected area at each time step to the initial projected area; defining the orientation angle as the angle between the projection of the displacement vector onto the local cutting plane of the fabric and a reference direction (such as the direction of the fabric machine); and organizing all features according to a unified data structure: storing them in the form of a four-dimensional array or database table, with index dimensions including time step number, unique ID of the pore unit, and feature type (displacement, area change rate, orientation angle), thus clearly defining the conversion rules from simulation results to structured datasets and providing a clearly formatted input for subsequent analysis.

[0037] Specifically, by constructing a full-chain simulation platform covering production to application, the fundamental problem of traditional testing being unable to retain the pore deformation characteristics of the mask fabric under actual use due to the use of statically cut samples has been solved. This has achieved a leap from idealized static samples to digital twins of the actual deformation state, laying a reliable structural evolution foundation for the subsequent accurate evaluation of the mask substrate performance in a real dynamic environment.

[0038] S2. Based on the initial deformation feature dataset, the dynamic environment simulation module is used to construct a multi-factor coupled environment in real-world usage scenarios by combining periodic parameters of respiratory airflow changes, gradual increase in sweating, and subtle facial movements, and outputs a dynamic interaction impact dataset.

[0039] Further, step S2 specifically includes:

[0040] The initial deformation feature dataset generated by S1 is used as structural input and loaded into the dynamic environment simulation module. In this module, based on physiological parameters from the actual application process (15-25 minutes), three dynamic environmental loads are simultaneously coupled and simulated: ① an alternating positive and negative pressure airflow field simulating respiratory rate (e.g., 12-20 breaths / minute); ② a time-varying, gradually increasing liquid water transport and evaporation model simulating sweat secretion rate; ③ a periodic local pressure / displacement fluctuation model simulating subtle changes in facial expressions. Through multiphysics coupling, preliminary multi-factor environmental data synchronized with dynamic structural changes (including airflow velocity distribution at each time step, pore liquid water saturation, local contact pressure, etc.) are calculated.

[0041] The synchronous coupling simulation of three dynamic environmental loads requires the establishment of a unified finite element or finite volume computational domain. This domain includes the nonwoven fabric deformation structure, obtained by interpolation from the S1 dataset, the adjacent air domain, and the bionic skin boundary. Coupling is achieved through the following methods: ① The respiratory airflow field is modeled as a time-varying pressure or velocity condition applied to the inlet boundary of the air domain, with a waveform that is sinusoidal or a curve fitted based on respiratory physiological data; ② The sweating model is implemented by defining a liquid water saturation source term that varies with time and space within the porous medium (fabric) domain, with its release rate following a preset incremental curve; ③ Facial micro-movements are applied by directly assigning periodic displacement boundary conditions to the bionic skin boundary nodes. The three physical fields (fluid, porous medium mass transfer, and solid mechanics) are jointly solved at each time step in the same transient solver through a fully coupled or sequentially strongly coupled iterative approach. This ensures that the bidirectional interaction between airflow, sweat transport, and structural deformation is calculated in real time, outputting truly synchronous preliminary multi-factor environmental data.

[0042] Physiological parameters during actual application include: respiratory parameters: typical values ​​such as frequency of 12, 16, 20 breaths / minute, with corresponding airflow pressure amplitude ranges typically from ±10 Pa to ±50 Pa (corresponding to calm breathing to deep breathing), and waveforms can be selected as sine waves or trapezoidal waves containing inspiratory and expiratory plateaus; sweating parameters: onset time (e.g., 5 minutes after application), secretion rate curve (e.g., linear or exponential growth from 0 to 2-10 g / (m²·min)), and physicochemical properties of sweat components (e.g., surface tension, viscosity); facial micro-movement parameters: defining the cycle of typical movements (e.g., raising eyebrows, pursing lips) (e.g., 5-15 seconds per cycle), the area of ​​action (mapped through facial muscle groups), displacement amplitude (e.g., 0.1-0.5 mm), and pressure fluctuation amplitude (e.g., 0.1-0.3 kPa).

[0043] Extract local pressure / deformation fluctuation signals caused by facial micro-movements from preliminary environmental data, and apply time series analysis (such as spectrum analysis) to quantify their periodicity and key influencing factors (such as fluctuation amplitude and dominant frequency). Compare the key influencing factors with preset physiological movement threshold ranges. If the key factors (such as the instantaneous pressure peak caused by local compression) exceed the threshold, dynamically weight and adjust the airflow and sweat transmission parameters of the corresponding area and time period according to the degree of exceedance and spatial location (for example, reduce the local breathability coefficient at the moment of pressure or enhance sweat diffusion during relaxation period), so as to more realistically reflect the mutual inhibition or enhancement effect of dynamic interaction and generate corrected coupled environmental data.

[0044] The process involves dynamically weighting and adjusting airflow and sweat transmission parameters for specific regions and time periods. This includes: triggering adjustments when the peak facial pressure in a region exceeds a threshold P_th during a specific time period; the algorithm executes as follows: calculating a base adjustment factor α (α>1) based on the excess ratio (e.g., (P_peak - P_th) / P_th); for airflow simulation, instantaneously dividing the permeability (or Darcy coefficient) of the grid cell in that region by α; for sweat transmission, multiplying the capillary pressure or evaporation coefficient of the grid cell in that region by α; using a half-width Hanning window as the adjustment time decay function to ensure smooth activation and de-escalation of the adjustment effect within seconds before and after the abnormal pressure peak; and determining the specific numerical mapping relationship of the weighting coefficient (i.e., α) through a pre-set lookup table, which establishes a correspondence between the pressure excess range and the parameter correction magnitude, thereby transforming qualitative adjustments into deterministic numerical calculations.

[0045] The corrected coupled environment data undergoes an overall consistency check, focusing on whether there are any non-physiological conflicts or distortions between the respiratory airflow fluctuation cycle and the facial movement pressure fluctuation cycle. If a mismatch is found, a secondary calibration is performed using preset time series smoothing and phase alignment tools to ensure that the changes in multiple factors conform to the true physiological synchronization logic in the time domain. The calibrated multi-dimensional dynamic interaction parameters (airflow, humidity, pressure field and their spatiotemporal evolution) that are strictly aligned with the structural deformation time series are integrated and structured as a dynamic interaction influence dataset, providing realistic boundary conditions and environmental loads for subsequent transient calculations of air permeability and moisture permeability.

[0046] The calibration process employs a pre-defined time-series smoothing and phase alignment tool. This tool refers to a signal processing function library integrated into the post-processing unit of the simulation module. When the difference between the respiratory cycle T_breath and the dominant facial movement cycle T_motion exceeds the physiological tolerance (e.g., |T_breath - T_motion| > 0.5 seconds), the calibration procedure is initiated: the pressure fluctuation signal is smoothed using a Savitzky-Golay filter to eliminate cycle misjudgments caused by high-frequency noise. Subsequently, phase alignment is performed: the phase difference Δφ between the two signals' dominant frequency components (obtained via FFT) is calculated. The alignment operation is achieved by shifting the time sequence of one of the signals (the shift amount is Δφ / 2π * T), with the facial movement signal being selected for adjustment, as breathing is considered the dominant rhythm. The entire calibration process can be automatically completed by calling relevant functions (savgol_filter, fft, correlation_lags) from the scipy.signal library, and a cycle comparison report before and after calibration is output in the log.

[0047] Specifically, by constructing a dynamic environment simulation module with multi-physics coupling, the bottleneck problem of traditional detection methods being unable to reflect the real-time interaction of multiple factors such as breathing, sweating, and facial movements during the actual application of the mask is solved due to testing under stable and single conditions. This provides highly realistic boundary conditions and environmental loads for subsequent transient performance analysis, achieving a key breakthrough from static stable environment testing to dynamic coupled environment simulation.

[0048] S3. For the dynamic interaction impact dataset, simulate the application state on the bionic skin interaction interface, and collect instantaneous performance capture data of the contact interface between the fabric and the bionic skin in real time to obtain the time sequence change record of breathability and moisture permeability.

[0049] Furthermore, step S3 specifically includes:

[0050] S31. Load the dynamic interaction dataset (containing time-varying airflow, sweat, and pressure parameters) generated in S2 into a simulation platform with a bionic skin interface. On this platform, drive the bionic skin to perform periodic facial micro-movements corresponding to the dataset, while injecting simulated breathing airflow and gradual sweat into the system. Under the condition that the fabric and the bionic skin are in simulated negative pressure contact, synchronously and in real time collect multi-physics field data of the contact interface: including airflow velocity and pressure difference at specific locations (used to calculate instantaneous air permeability), humidity gradient and moisture flux on both sides of the interface (used to calculate instantaneous moisture permeability), and generate the original high-frequency time series of air permeability and moisture permeability performance.

[0051] Synchronous, real-time acquisition of multi-physics field data at the contact interface includes: a micro-array sensing unit integrated with the bionic skin substrate; this unit includes a group of micro-pressure differential sensors arranged at preset positions on the bionic skin (such as below the nostrils and at the high point of the cheek) to measure the local air pressure difference on both sides of the fabric; a distributed pair of micro temperature and humidity sensors (one pair of sensors located below the surface of the bionic skin and near the outer surface of the fabric) to measure the absolute humidity on both sides of the interface; and a micro-displacement actuator integrated in the bionic skin to reproduce facial micro-movements in the S2 data; all sensors are controlled and read through a unified data acquisition card (DAQ), which is driven by a master clock signal to ensure that the timestamps of multiple physical quantities such as air pressure difference, humidity gradient, and motion displacement are perfectly aligned at each sampling moment.

[0052] Generate the original high-frequency time sequence of breathability and moisture permeability, including: high frequency means that the sampling frequency of the data acquisition card is uniformly set to no less than 100 Hz to adapt to the dynamic changes of breathing (about 0.3 Hz) and facial movements (the highest possible number of Hz). The instantaneous air permeability Q(t) is calculated using a simplified form of Darcy's law: Q(t) = K⋅A⋅ΔP(t) / d, where ΔP(t) is the measured pressure difference at time t, A is the nominal area of ​​the test area, d is the fabric thickness (provided by the S1 model), and K is the permeability coefficient related to the fabric structure (obtainable through prior calibration). The instantaneous moisture permeability WVT(t) is calculated using Fick's diffusion law: WVT(t) = D⋅ΔC(t) / d, where ΔC(t) is the water vapor concentration difference at time t calculated by the humidity sensor, and D is the effective diffusion coefficient of water vapor in the fabric (obtained through calibration). Through the above formulas, the original voltage signal sequence is converted in real time into a sequence of air permeability (unit: L / (m²·s)) and moisture permeability (unit: g / (m²·s)).

[0053] S32. The original performance time series is filtered and denoised. A sliding time window analysis (covering several breathing or movement cycles) is used to calculate the variation amplitude (standard deviation or peak difference) of air permeability and moisture permeability within each window, generating air permeability variation amplitude sequences and moisture permeability variation amplitude sequences respectively. Based on a preset physiological rationality threshold, abnormal fluctuations are automatically identified and marked: if the variation amplitude exceeds the threshold for multiple consecutive windows (e.g., a sudden drop in air permeability due to intense facial expression compression), the time period is marked as a performance mutation segment; otherwise, it is marked as a performance stable segment. This outputs segmented marked sequences for air permeability and moisture permeability performance.

[0054] A sliding time window analysis (covering several respiratory or movement cycles) is employed to calculate the variation amplitude of air permeability and moisture permeability within each window. This includes: applying zero-phase digital filtering (e.g., a low-pass Butterworth filter) to the original performance time series to eliminate high-frequency noise. Subsequently, the system dynamically sets the length of the sliding time window based on the real-time monitored major respiratory cycles, ensuring that each window covers an integer number (e.g., 3) of respiratory or facial movement cycles; the window slides along the time axis with a fixed step size (e.g., 1 second); within each window, the variation amplitude of the air permeability and moisture permeability sequences is calculated separately. This amplitude value is typically quantified using the standard deviation of the data within that window to characterize the dispersion of performance; the calculated amplitude values ​​(i.e., standard deviations) for each window are arranged in chronological order, thereby generating air permeability variation amplitude sequences and moisture permeability variation amplitude sequences, respectively. These amplitude sequences provide direct quantitative basis for subsequent automatic segmentation based on preset thresholds.

[0055] S33. Perform time alignment and correlation analysis on the segmented marker sequences of breathability and moisture permeability; accurately identify the time-overlapping intervals of breathability and moisture permeability mutations, i.e., the overlapping mutation intervals. These intervals usually correspond to the strong interference moments caused by the combined effects of facial movements, breathing, and sweating; combine the dynamic environmental data in S2 to deeply analyze the dominant influencing factors of each overlapping mutation interval (e.g., whether it is mainly due to sweat clogging the pores or compression causing excessive tightness), integrate all time sequences, segmented markers, and correlation analysis results to generate a dynamic change map of breathability and moisture permeability performance throughout the entire application cycle (15-25 minutes) of the mask. This map not only contains continuous performance values ​​but also reveals key mutation moments and their causes, comprehensively reflecting the true performance of the fabric during dynamic use.

[0056] Generate a dynamic change map of breathability and moisture permeability throughout the entire application cycle (15-25 minutes) of the face mask. This includes aligning and overlaying the segmented marker sequences of breathability and moisture permeability obtained in S32 on a unified time axis, accurately identifying the intervals where the marked abrupt change segments completely overlap in time, i.e., overlapping abrupt change intervals. For each overlapping abrupt change interval, retrieve the corresponding airflow, sweat saturation, and pressure data from the dynamic interaction influence dataset in S2. By comparing the intensity and phase relationship of the changes in each environmental parameter, determine and label the dominant influencing factors of the performance abrupt change in that interval. Integrate the high-frequency original performance time-series curves (after smoothing), the background color markings of the stable and abrupt performance segments, the highlighted markers of the overlapping abrupt change intervals, and the data table containing explanations of the dominant influencing factors into a chart with time as the horizontal axis and performance value as the vertical axis, and attach key statistical indicators to form a comprehensive map that fully displays the dynamic evolution of performance, key events, and their attributions.

[0057] Specifically, by replicating the real application environment on a high-precision biomimetic skin interaction platform and simultaneously acquiring multi-physics field data, the core problem of traditional testing methods—which can only perform static, single-point measurements and cannot obtain continuous dynamic performance data throughout the entire lifecycle of mask use—was solved. This enabled in-depth analysis from static performance values ​​to the mechanism of performance evolution throughout the entire process, providing direct and comprehensive data for accurately evaluating the dynamic comfort of fabrics in actual use.

[0058] S4. By recording time-series changes, analyze the impact of changes in respiratory airflow on the slight undulations of the fabric, and combine this with the changes in pore structure caused by the gradual increase in perspiration to determine the range of breathability fluctuations in different time periods.

[0059] Furthermore, step S4 specifically includes:

[0060] S41. Based on the synchronous monitoring data acquired in real time during the S3 acquisition process, a continuous microscopic image sequence of the fabric surface and time-series data of the breathing airflow sensor are obtained. Through image analysis technology, the minute fluctuation amplitude of the fabric caused by the alternation of positive and negative breathing pressure is calculated frame by frame to generate fluctuation amplitude time-series data. This sequence is precisely aligned with the breathing airflow data in the time domain. The pairing data of the peak airflow intensity and the corresponding fabric fluctuation amplitude in each breathing cycle are extracted. Through linear regression analysis, the intensity coefficient of the effect of breathing airflow on fabric fluctuation is calculated, thereby first quantifying the degree of direct influence of the periodic mechanical action of breathing on the fabric structural state.

[0061] The process involves using image analysis technology to calculate the minute fluctuations in the fabric caused by the alternating positive and negative pressure of breathing, frame by frame, generating temporal data of these fluctuations. This includes: continuously capturing a sequence of surface images of the fabric during the test (15-25 minutes) at a frame rate of at least 200 frames per second; and performing frame-by-frame analysis of the sequence using digital image correlation (DIC) technology. Using the image in the initial application state as a reference, the grayscale correlation of each speckle sub-region in each subsequent frame is calculated to accurately track and calculate the minute displacement field of each point on the fabric surface in the direction perpendicular to the skin (normal). The fluctuation amplitude corresponding to each frame is quantified as the root mean square value of the normal displacement of all points within the entire observation area of ​​that frame. The fluctuation amplitude values ​​of all frames are arranged in chronological order, thus generating high-resolution temporal data of fluctuations synchronized with the respiratory cycle.

[0062] Linear regression analysis was used to calculate the intensity coefficient of the effect of breathing airflow on fabric undulation. This involved extracting a unique peak airflow intensity Fmax,i and its corresponding peak fabric undulation amplitude Amax,i from the time-domain aligned data for each complete breathing cycle (from one airflow intensity trough to the next), forming a pair of paired data points (Fmax,i,Amax,i), where i=1,2,…,N represents the cycle number. Subsequently, a univariate linear regression was performed on this pair of data sequences using the least squares method, with the model A=k⋅F+b. The slope k obtained from the fitting is the intensity coefficient of the effect of breathing airflow on fabric undulation, which physically represents the change in fabric undulation amplitude caused by a unit change in airflow pressure (unit: mm / Pa). Simultaneously, the coefficient of determination of this regression was calculated and reported to assess the strength of the linear relationship between breathing airflow undulation and the fabric's mechanical response, providing a confidence basis for subsequent S44 judgments.

[0063] S42. Simultaneously acquire the cumulative sweat volume time series on the same time axis. Based on the characteristics of sweat secretion simulating the real skin moisturization process, analyze the increasing trend and slope change of the sequence, identify key turning points (such as sweat starting to penetrate the fabric, forming a continuous liquid film, etc.). Based on this, divide the entire cycle of mask application (such as 15-25 minutes) into several sweat volume increasing intervals with different moisturization characteristics (such as: initial dry state period, sweat penetration period, and moisture saturation period), thereby establishing a staged time framework dominated by changes in the liquid environment for breathability performance analysis.

[0064] Based on the characteristics of sweat secretion simulating the real skin wetting process, the increasing trend and slope changes of the sequence were analyzed to identify key inflection points, including: An automatic identification algorithm based on the mathematical characteristics of the sweat volume sequence was used to process the cumulative sweat volume time series S(t) as follows: First, its first derivative (i.e., the instantaneous sweating rate R(t)) and second derivative were calculated; the inflection point where sweat begins to penetrate the fabric was determined by identifying the first significant upward inflection point on the R(t) curve. Specifically, the algorithm was to find the moment when R(t) first exceeded twice the standard deviation of its initial average (e.g., the average during the dry period); the inflection point of forming a continuous liquid film corresponded to the inflection point where the sweating rate transitioned from a rapid increase to a plateau or a slow growth phase. This was achieved by finding the moment when the second derivative of S(t) changed from a significant positive value to a value close to zero or negative. This could be achieved by finding the value of R(t). The moment when the absolute value of the slope of the sequence after moving average is lower than a certain threshold (such as 10% of the maximum slope); these criteria based on derivatives and statistical thresholds provide programmable and repeatable objective standards for interval division.

[0065] S43. Record the temporal changes in breathability performance and segment the data according to the increasing sweat volume intervals defined in S42. Within each independent interval, extract all instantaneous test values ​​of breathability performance within that time period to form a set of breathability performance values ​​for that interval. Calculate the maximum and minimum values ​​of breathability in each set, and the difference is the range of breathability performance fluctuation for that interval. Arrange the fluctuation ranges of all intervals in chronological order to obtain a preliminary sequence of breathability performance fluctuation intervals related to the sweating process.

[0066] This time period is the start and end time range corresponding to each increasing interval of sweat volume, which is divided by step S42 based on the cumulative sweat volume time sequence;

[0067] S44. Integrate and correlate the respiration intensity coefficient obtained in S41 with the permeability fluctuation ranges obtained in S43 for each interval. The focus is on determining whether the fluctuations in permeability are primarily driven by the periodic modulation of respiration fluctuations, the pore blockage effect caused by increasing sweat, or a combination of both, in different sweating intervals (i.e., different pore moisture states). For example, in the initial dry state, mechanical fluctuations of respiration may be the main cause of the fluctuations; while in the saturated moisture state, the overall decrease and low-amplitude fluctuations in performance may be mainly caused by liquid water occupying the pores. Finally, output the permeability fluctuation interval sequence with attribution of dominant factors as the final analysis result, fully revealing the intrinsic mechanism of dynamic performance changes during application.

[0068] Specifically, by quantifying the impact of breathing on fabric fluctuations, intelligently dividing the sweating stage and analyzing the breathability fluctuations at each stage, and finally correlating and judging the dominant factors of performance fluctuations under different humid conditions, the problem of traditional methods being unable to reveal the intrinsic mechanism of dynamic performance changes has been solved, achieving a fundamental leap from monitoring performance phenomena to elucidating their physical causes.

[0069] S5. Based on the fluctuation range of breathability, assess the interference of local fit changes caused by minor facial movements on moisture permeability, obtain the contribution weight of each dynamic factor to the overall performance, and determine the overall influence trend.

[0070] Furthermore, step S5 specifically includes:

[0071] The baseline data under the stable performance period is extracted from the time-series change records of breathability and moisture permeability, and the change in local fit at each moment is calculated based on the facial micro-movement sequence. The change in local fit is compared with the fluctuation range of breathability performance to identify the moment when the change in local fit exceeds the fluctuation range of breathability performance as the key interference point.

[0072] The identification of moments when localized changes in fit exceed the range of breathability fluctuations includes: pre-setting a minimum duration threshold, such as 0.5 seconds; during identification, comparing the absolute value of the localized fit change calculated at each moment with the upper limit of the breathability fluctuation range corresponding to the sweating zone at that moment; if the absolute value of the localized fit change is greater than the upper limit of the current breathability fluctuation range, and this state is maintained continuously for a time that reaches or exceeds the preset minimum duration threshold, then the starting moment of this continuous excess time is identified and recorded as a key interference point. This rule ensures that the identified interference points have significant deviation and reasonable persistence.

[0073] For each key interference point, the actual measured value of moisture permeability, the change in local fit, the intensity of breathing airflow, and the amount of sweat accumulation at that moment are extracted simultaneously to form a multi-dimensional data set;

[0074] The multi-dimensional data set is composed of structured data tables, which are built on a unified high-precision time reference. For each identified key interference point, the following data extraction operations are performed simultaneously at the precise time: the measured value of moisture permeability at that time is extracted from the time-series record of moisture permeability performance; the change value at that time is extracted from the calculation result of the change in local fit; the airflow intensity value at that time is extracted from the time-series data of the breathing airflow sensor; and the cumulative value at that time is extracted from the time-series sequence of cumulative sweating. All extracted values ​​are marked with the same timestamp and arranged in the order of interference points. Each row of data contains the above four synchronously acquired values, thus forming a multi-dimensional data set with clear rows and columns and strict time alignment.

[0075] Using the actual measured value of breathability as the dependent variable and the changes in local fit, respiratory airflow intensity, and cumulative sweating as independent variables, a multiple linear regression analysis was performed on the multidimensional dataset to obtain the contribution weight of each dynamic factor to the change in breathability. Based on the weight results, the influence of each dynamic factor was ranked.

[0076] The contribution weights of each dynamic factor to the change in breathability were obtained, including: performing multiple linear regression analysis based on the aforementioned multidimensional dataset; in the analysis, the measured value of breathability was used as the predicted variable, and the change in local fit, respiratory airflow intensity, and cumulative sweating were used as predictor variables; the contribution weights of each dynamic factor specifically refer to the standardized regression coefficients of each predictor variable calculated through this regression analysis. These coefficients have eliminated the influence of the variables' own dimensions, and their absolute values ​​directly represent the relative importance of the corresponding dynamic factor to the change in breathability. Based on the absolute values ​​of the standardized regression coefficients of each factor, they were ranked from highest to lowest degree of influence.

[0077] Based on the contribution weight of each dynamic factor and its changing trend throughout the entire application cycle of the mask, the overall change direction of moisture permeability under the combined effect of multiple dynamic factors is analyzed through weighted calculation, and the evaluation results including the ranking of the contribution weights of each factor and the overall influence trend are output.

[0078] The study analyzes the overall change direction of breathability under the combined influence of multiple dynamic factors through weighted calculations. This includes: using standardized regression coefficients obtained from regression analysis as fixed weights; and multiplying the real-time values ​​of three dynamic factors—the change in local fit, the intensity of respiratory airflow, and the amount of sweat accumulation—at each moment along the entire time axis of mask application by their respective standardized regression coefficient weights. The three products are then summed to calculate the comprehensive impact value of the dynamic factors at each moment, forming a curve that changes over time. By analyzing the overall correlation between this comprehensive impact curve and the actual trend curve of breathability change, qualitative and quantitative descriptions are provided. For example, how the upward or downward trend of the comprehensive impact value is related to the change in breathability, the overall direction of influence under the combined influence of multiple dynamic factors is determined, and the results are output in the form of descriptive conclusions combined with key correlation indicators.

[0079] Specifically, by identifying significant fit interference caused by facial movements and using multiple regression and weighted analysis, the contribution weights of facial movements, breathing, and sweating to moisture permeability were quantified. This solved the problem of analyzing the relative influence of each factor under the coupling effect of multiple dynamic factors, achieving a breakthrough from qualitative description to quantitative attribution, and enabling accurate judgment of the overall trend of moisture permeability and its dominant causes.

[0080] S6. In response to the overall impact trend, a preset threshold comparison method is adopted. If the air permeability and moisture permeability performance is lower than the preset threshold in a certain period of time, the dynamic environment simulation parameters are adjusted and the optimized performance time series data is regenerated.

[0081] Furthermore, step S6 specifically includes:

[0082] Obtain the performance time series data corresponding to the current comprehensive influence trend, and divide the performance time series data into multiple continuous time periods; compare the average air permeability and average moisture permeability of each time period with the corresponding preset performance thresholds to determine whether there are time periods where the air permeability or moisture permeability is lower than the preset thresholds.

[0083] The performance time series data is divided into multiple consecutive time periods based on a preset fixed time length or a dynamic cycle based on the characteristics of the performance data itself. One method is to divide the entire application time of the mask (e.g., 20 minutes) into 20 consecutive time periods at fixed time intervals (e.g., every minute). Another method is for the system to adaptively segment the data based on the characteristic points (e.g., peaks, troughs, or trend inflection points) of the breathability or moisture permeability performance time series curve itself. Regardless of the method used, the specific segmentation rules (e.g., fixed duration values ​​or adaptive algorithms) must be clearly set before operation and used as preset parameters of the system to ensure that the segmentation results of each analysis period are consistent.

[0084] If there is a time period in which the breathability or moisture permeability is lower than the preset threshold, the dynamic environment simulation parameters corresponding to that time period are extracted. The parameters include the breathing airflow intensity change curve, the cumulative sweating curve, and the amplitude and frequency data of facial micro-movements during that time period.

[0085] Specifically, dynamic environment simulation parameters (respiration, sweating, and movement data) are extracted from the dynamic environment simulation parameters corresponding to the time period. These parameters are generated in step S2 and are typically stored in a time-indexed data structure (such as a data array with timestamps or a list of time-value pairs). When a time period below a threshold (with a clear start and end time) is identified, all data points whose timestamps fall within this interval are precisely extracted from the aforementioned time-indexed parameter data structure based on these start and end time points. For respiratory airflow intensity, a set of continuous change curve data points within the time period is extracted. For sweating amount, the data segment of its cumulative curve within the time period is extracted. For facial movements, the amplitude and frequency records of all movement events within the time period are extracted.

[0086] Based on the position of the time period below the preset threshold in the entire application cycle and the degree to which the performance is below the threshold, the adjustment amount of respiratory airflow intensity, sweating rate and facial movement amplitude is calculated by linear interpolation method. The calculated adjustment amount is superimposed on the parameter curve of the corresponding time period in the original dynamic environment simulation parameters to form the updated dynamic environment simulation parameters.

[0087] The system employs linear interpolation to calculate adjustments for respiratory airflow intensity, sweating rate, and facial movement amplitude. This includes: calculating the performance deviation based on the percentage of the average breathability or moisture permeability within the current time period that is below a preset threshold; calculating the cycle position coefficient based on the position of the center point of that time period within the entire application cycle; and storing a pre-stored reference adjustment coefficient lookup table indexed by typical performance deviations and cycle positions, containing corresponding suggested adjustment coefficients. For any parameter among respiratory airflow intensity, sweating rate, or facial movement amplitude, the system uses a bilinear interpolation algorithm to perform a two-dimensional lookup and calculation in the lookup table based on the calculated actual performance deviation and cycle position coefficient, thereby obtaining a customized basic adjustment coefficient for that parameter. Finally, the adjustment amount for this parameter is the product of its original average value within the current time period and the basic adjustment coefficient.

[0088] Using the updated dynamic environment simulation parameters, the environmental simulation starting from step S2, the performance acquisition in step S3, and the analysis process from steps S4 to S5 are rerun to obtain new performance time series data and the corresponding comprehensive impact trend. The original data is replaced with the new performance time series data to generate the optimized comprehensive impact trend.

[0089] Specifically, by establishing an automatic feedback and iterative optimization mechanism based on performance thresholds, the problem of test results not being able to directly guide process or material improvements has been solved. It automatically identifies periods when performance is substandard, traces back and quantifies and adjusts the corresponding environmental simulation parameters such as breathing, sweating and facial movements, and generates optimized performance data through re-simulation. This achieves a leap from passively detecting performance to actively optimizing performance prediction, providing a closed-loop digital tuning tool for improving the comfort design of mask fabrics.

[0090] S7. By using the optimized performance time-series data, construct a full-cycle performance spectrum of breathability and moisture permeability, determine the comfort distribution of the mask substrate in real-world usage scenarios, and output the final evaluation results.

[0091] Furthermore, step S7 specifically includes:

[0092] Based on the optimized performance time-series data, curves showing the changes in breathability and moisture permeability over time throughout the entire mask application cycle were plotted to generate a performance spectrum covering the entire cycle. The maximum, minimum, average, and standard deviation of breathability and moisture permeability were extracted from the spectrum to quantify and determine the fluctuation range and stable interval of performance.

[0093] The performance graph covering the entire lifecycle is implemented by inputting optimized time-series data (time-value pairs) of air permeability and moisture permeability into the graph generation engine. The engine automatically draws two continuous curves with different labels according to a preset format (e.g., time axis as the horizontal axis and performance values ​​as the dual vertical axes). It then calls built-in statistical functions to calculate the entire time series array: the maximum and minimum values ​​are obtained by traversing the array to find the extreme values; the average value is obtained by summing all data points and dividing by the total number of data points; and the standard deviation is obtained by calculating the root mean square of the sum of the squares of the differences between each data point and the average value. The quantification of the stable interval refers to identifying all continuous time periods within the range of plus or minus one standard deviation of the average value and highlighting them on the graph as light-colored bands distinct from the background color, thus visually distinguishing between fluctuation and stable phases.

[0094] The performance spectrum is mapped to environmental parameters of real-world usage scenarios to analyze the performance of the mask substrate under different sweating stages, different breathing intensities, and different facial movements, thereby obtaining its environmental adaptability distribution. Based on preset breathability and moisture permeability comfort thresholds, the spectrum is divided into comfortable areas that meet the comfort standards and uncomfortable areas that are below the thresholds, generating a classified comfort space-time distribution map.

[0095] The process involves associating each time point in the performance graph with the sweating stage (e.g., initial dry state) defined in step S2, as well as the breathing intensity level (e.g., calm, deep breathing) and facial movement state (e.g., still, slight movement) at that moment; this association is based on a unified time axis. When dividing comfort and discomfort zones, two preset comfort thresholds are used: one for air permeability and the other for moisture permeability. The division rule is as follows: on the performance graph, for any given time, only when both the air permeability and moisture permeability values ​​are simultaneously higher than or equal to their respective comfort thresholds is that moment classified as a comfortable zone; if either performance value is lower than its corresponding threshold, that moment is classified as a discomfort zone. Subsequently, different background colors (e.g., green for comfortable zones, red for discomfort zones) are used to fill these two types of zones on the time axis graph, thus generating an intuitive comfort space-time distribution map, where space refers to different segments on the time axis.

[0096] Based on the proportion, duration, and occurrence stage of comfortable and uncomfortable areas in the comfort distribution map, and combined with the performance fluctuation range, a weighted scoring model is used to calculate the overall comfort score of the mask substrate. According to the preset level range to which the overall score belongs, the final overall comfort level is determined, and a comprehensive evaluation report including performance map, comfort distribution, level assessment, and improvement suggestions is generated.

[0097] The system employs a weighted scoring model to calculate the overall comfort score of the mask substrate. This model quantifies the score based on three pre-defined weighted dimensions: first, it calculates the percentage of the total comfort zone duration relative to the total duration of the entire testing period, serving as the base score; second, it calculates the average duration of all discomfort zones and compares it to a pre-defined maximum tolerable average duration, calculating a deduction adjustment score; third, it calculates the coefficient of variation of the breathability data throughout the testing period to measure performance volatility, comparing it to a pre-defined maximum tolerable coefficient of variation to obtain a stability score. The system then multiplies the first score by 50%, the second adjustment score by 30%, and the third stability score by 20%. Finally, the weighted values ​​of these three items are summed to obtain the overall comfort score, which falls within a pre-defined range, thus completing the objective determination of comfort level from data.

[0098] Specifically, by transforming optimized dynamic performance data into a visualized full-cycle map and automatically dividing comfortable and uncomfortable zones based on preset comfort thresholds, the system finally outputs a quantitative comfort level and a comprehensive evaluation report by combining a weighted scoring model. This solves the problem that traditional testing methods cannot intuitively and quantitatively evaluate the overall comfort of the face mask throughout its use. It achieves a leap from discrete performance data to intelligent judgment of comprehensive comfort, providing an objective and comprehensive basis for product performance optimization and user experience evaluation.

[0099] Example 2

[0100] This embodiment provides another method for dynamically detecting the air permeability and moisture permeability of nonwoven fabrics used in the production of facial mask fabrics. This embodiment focuses on multi-scale direct observation and data-driven modeling. Its core lies in using high-precision in-situ multimodal sensing technology to simultaneously capture multi-physics field and cross-scale dynamic response data of the fabric during use, and then using this data to construct a data-driven surrogate model to achieve performance prediction, mechanism inversion, and optimization. Specific steps include:

[0101] Multi-scale sample characterization and initial state benchmark establishment: Non-woven fabric mask samples produced from actual production lines are taken, and digital models of their real three-dimensional fiber network and pore structure are obtained using microcomputer tomography or high-resolution three-dimensional scanning, which serve as the physical benchmark for analysis; At the same time, standardized low-stress tensile tests are performed on the samples using a material testing machine, and their stress-strain curves and macroscopic thickness changes are recorded to establish the constitutive relationship of the material.

[0102] Constructing a multimodal dynamic test chamber and synchronous data acquisition system: A dynamic test chamber integrating environmental simulation and multi-scale observation functions was built. The core components of this test chamber include: a multi-axis bionic robotic arm with a bionic skin module at its end, capable of programmatically reproducing complex facial contouring, negative pressure adsorption, and periodic micro-movements (such as the movement trajectories corresponding to smiling and frowning); an environmental simulation subsystem that precisely controls the temperature and humidity within the chamber and includes a programmable breathing simulator (generating periodic airflow) and a micro-liquid pumping system (simulating gradual sweat secretion at specific points on the bionic skin); a multimodal synchronous sensing array integrating distributed micro-pressure sensors and temperature and humidity sensors on the bionic skin surface and above the fabric; simultaneously, a high-speed confocal microscope or optical coherence tomography system is integrated for real-time, in-situ observation of the microscopic pore morphology and liquid (simulating sweat) seepage process in the fabric application area; and a unified time-series acquisition system where all sensors, actuators, and observation equipment are synchronously controlled by a central clock, ensuring strict alignment of mechanical movements, environmental parameters, physical signals (pressure, temperature, and humidity) with microscopic images at every moment.

[0103] Dynamic coupling testing and synchronous acquisition of full information flow were performed: the sample was installed in the test chamber, and a bionic robotic arm drove it to complete the application; during the test, breathing simulation, gradient injection of sweat, and a preset facial movement sequence were executed simultaneously; the system synchronously acquired and generated four types of strictly time-aligned data streams:

[0104] Macroscopic mechanical data stream: force / torque and displacement data of each joint of the bionic robotic arm;

[0105] Interface physics data stream: pressure distribution, temperature gradient, and absolute humidity difference at the fabric-skin interface;

[0106] Environmental parameter data stream: cabin airflow velocity, temperature and humidity, and perspiration injection rate;

[0107] Microstructure image stream: High temporal resolution microscopic images of key observation areas, recording the dynamic processes of pore opening and closing and liquid front movement;

[0108] Feature engineering and machine learning modeling based on dense data streams: Simultaneously acquired full information streams are fused and processed. From the image stream of each time segment, microscopic features such as porosity, pore orientation, and liquid coverage ratio are extracted using computer vision algorithms; macroscopic features such as pressure mean / variance and humidity change rate are extracted from the physical field data stream. These cross-scale features and corresponding environmental parameters (action type, breathing phase, perspiration) are used as input features, and synchronously calculated instantaneous air permeability (converted through pressure difference and calibration model) and moisture permeability (converted through humidity gradient and calibration model) are used as output labels, forming a high-dimensional dynamic process feature dataset. Using this dataset, a machine learning model (such as a gradient boosting decision tree or a temporal convolutional neural network) is trained. This model can learn the complex mapping relationship from multi-scale observation features and environmental parameters to air permeability and moisture permeability performance, becoming a data-driven surrogate model.

[0109] Performance extrapolation and sensitivity analysis of dominant factors based on surrogate models: Virtual experiments are conducted using trained data-driven surrogate models. By systematically changing input features (such as simulating different facial movement amplitudes, respiratory frequency combinations, and sweating curves), the model can quickly predict the corresponding performance change curves. Using model interpretation techniques (such as SHAP value analysis), the contribution and sensitivity of each dynamic factor (specific movement, respiratory intensity changes, sweating stage) to changes in breathability or moisture permeability in any simulated scenario can be quantified, thereby identifying key influencing factors and their interactions.

[0110] Data-driven parameter optimization and verification closed loop: Define a comfort objective function (e.g., maximizing the percentage of time the total breathability exceeds a certain threshold), combine a surrogate model and optimization algorithm (e.g., Bayesian optimization), and automatically search for the optimal combination of environmental simulation parameters (optimal movement pattern, breathing-perspiration coordination curve) in a virtual environment to optimize the objective function. Then, re-execute the optimal parameter combination in a physical test chamber to collect real performance data and verify the optimization effect. The measured data can be further added to the training set to iteratively optimize the surrogate model, forming a hybrid enhanced closed loop of virtual optimization and physical verification.

[0111] Generate a dynamic comfort profile and diagnostic report based on the measured process: Based on the final measured data stream under optimal parameters, not only is a performance time-series graph plotted, but also a microscopic image sequence is further integrated to generate a dynamic comfort profile. This profile uses the time axis as the main line, synchronously displaying macroscopic performance curves, interface pressure cloud map animations, and close-ups of microscopic pore states at key time points. The system automatically marks the periods of performance decline and directly associates them with synchronous microscopic images (such as pore blockage and excessive compression), providing an intuitive diagnosis of physical mechanisms. The final report combines the analysis conclusions of the data-driven model with the dynamic profile, providing quantitative comfort scores and suggestions for improvement based on microscopic mechanisms.

[0112] This embodiment directly captures cross-scale data under real dynamic interaction through highly integrated multimodal synchronous observation, avoiding the uncertainties of complex simulation; it realizes fast and low-cost performance extrapolation and factor analysis by using data-driven models; and it forms a hybrid enhancement closed loop that efficiently combines physical measurement and virtual optimization, providing another efficient, intuitive and mechanistically explainable R&D path for product development.

[0113] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some modifications or alterations to the above-disclosed technical content to create equivalent embodiments without departing from the scope of the present invention. Any simple modifications, equivalent changes and alterations made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the scope of the present invention.

Claims

1. A dynamic process testing method for the air permeability and moisture permeability of nonwoven fabrics used in the production of facial mask fabrics, characterized in that, Includes the following steps: S1. By constructing a simulation platform, the pore structure change data of nonwoven fabric under stretching, die-cutting and skin curvature bonding is obtained for the full cycle change of mask substrate from production and molding process to application and use state, and an initial deformation feature dataset is generated. S2. Based on the initial deformation feature dataset, a dynamic environment simulation module is used to construct a multi-factor coupled environment in a real-world usage scenario by combining the periodic parameters of changes in respiratory airflow, gradual increase in sweating, and minor facial movements, and outputting a dynamic interaction impact dataset. S3. For the dynamic interaction impact dataset, simulate the application state on the bionic skin interaction interface, and collect the instantaneous performance capture data of the contact interface between the fabric and the bionic skin in real time to obtain the time sequence change record of breathability and moisture permeability. S4. By recording the time-series changes, analyze the degree of influence of changes in respiratory airflow on the slight undulations of the fabric, and combine the changes in pore structure caused by the gradual increase in perspiration to determine the range of air permeability fluctuations in different time periods. S5. Based on the aforementioned air permeability fluctuation range, assess the interference of local fit changes caused by minor facial movements on the moisture permeability, obtain the contribution weight of each dynamic factor to the overall performance, and determine the overall influence trend. S6. In response to the overall impact trend, a preset threshold comparison method is adopted. If the air permeability and moisture permeability performance is lower than the preset threshold in a certain time period, the dynamic environment simulation parameters are adjusted and the optimized performance time series data is regenerated. S7. By using the optimized performance time-series data, construct a full-cycle performance spectrum of breathability and moisture permeability, determine the comfort distribution of the mask substrate in real-world usage scenarios, and output the final evaluation results.

2. The method for dynamic process testing of air permeability and moisture permeability of nonwoven fabrics used in mask fabric production according to claim 1, characterized in that, Step S1 specifically includes: In the simulation platform, the nonwoven fabric production and forming process simulation is run to generate an initial three-dimensional digital model with a random fiber network structure. Then, a specific tensile load on the simulated production line is applied to the initial three-dimensional digital model to calculate the displacement and rearrangement of the fiber network and obtain the pore size, shape and connectivity distribution data of the fabric under tension. The pore structure data obtained from the stretching simulation is input into the die-cutting path planning module. The die-cutting boundary is preset according to the shape of the mask, and the key dimensions of the pores in the boundary area are compared in real time. If it is detected that the key pore size exceeds the preset process allowable range due to stretching, the die-cutting path or boundary position is automatically optimized and adjusted, and the fabric three-dimensional mesh model and pore data after simulating the actual die-cutting process are output. The die-cut 3D mesh model is mapped onto a parameterized biomimetic facial skin surface. In the simulation platform, the negative pressure adsorption force and surface normal adhesion constraint are applied to the model to simulate the application, driving the mesh to deform to adapt to the skin curvature. Through transient dynamic analysis, the mesh node displacements are calculated from the initial application to the tight adhesion process, and the pore structure change sequence under the application state is output in time sequence. Based on the pore structure change sequence under the application state, for each pore unit, the displacement vector, area change rate and orientation angle multi-dimensional deformation characteristics are calculated at each time point relative to the initial die-cut state; the feature data of all pore units at all time steps are organized according to time sequence and spatial location to construct a structured initial deformation feature dataset.

3. The method for dynamic process testing of air permeability and moisture permeability of nonwoven fabrics used in mask fabric production according to claim 2, characterized in that, The process of generating an initial three-dimensional digital model with a random fiber network structure includes: based on preset fiber properties and process parameters, calling a random fiber deposition or network reconstruction algorithm to generate a three-dimensional model whose statistical characteristics match those of the real sample.

4. The method for dynamic process testing of air permeability and moisture permeability of nonwoven fabrics used in the production of facial mask fabrics according to claim 1, characterized in that, Step S2 specifically includes: The initial deformation feature dataset is loaded into the dynamic environment simulation module as structural input. In this module, based on the physiological parameters of the actual application process, the respiratory airflow field, liquid water transport and evaporation model, and periodic local pressure / displacement fluctuation model are simultaneously coupled and solved through multi-physics coupling to obtain preliminary multi-factor environmental data. Extract the local pressure / deformation fluctuation signal caused by facial micro-movements from the preliminary environmental data, and use time series analysis to quantify its periodicity and key influencing factors; compare the key influencing factors with the preset physiological movement threshold range. If the key factors exceed the threshold, dynamically weight and adjust the airflow and sweat transmission parameters of the corresponding area and time period according to the degree of exceedance and spatial location to generate corrected coupled environmental data. The overall consistency of the corrected coupled environment data is checked. If there is a non-physiological conflict between the respiratory airflow fluctuation cycle and the facial movement pressure fluctuation cycle, a second calibration is performed using a preset time series smoothing and phase alignment tool. The calibrated multi-dimensional dynamic interaction parameters that are strictly aligned with the structural deformation time series are integrated and the structured output is a dynamic interaction influence dataset.

5. The method for dynamic process testing of air permeability and moisture permeability of nonwoven fabrics used in the production of facial mask fabrics according to claim 4, characterized in that, The dynamic weighted adjustment includes: when the peak pressure of local facial movement is detected to exceed the threshold, an adjustment factor α is calculated based on the excess ratio; for airflow simulation, the air permeability of the corresponding area grid is instantaneously divided by α; for sweat transmission, the capillary pressure or evaporation coefficient of the corresponding area is multiplied by α.

6. The method for dynamic process testing of air permeability and moisture permeability of nonwoven fabrics used in the production of facial mask fabrics according to claim 1, characterized in that, Step S3 specifically includes: S31. Load the dynamic interactive influence dataset into a simulation platform with a bionic skin interactive interface, drive the bionic skin to perform periodic facial micro-movements corresponding to the dataset, and inject simulated breathing airflow and gradual sweat; while the fabric and bionic skin are in simulated negative pressure contact, synchronously and in real time collect multi-physics field data of the contact interface to generate the original high-frequency time series of breathability and moisture permeability. S32. The original performance time series is filtered and denoised. Sliding time window analysis is used to calculate the variation range of air permeability and moisture permeability in each window, and air permeability variation range sequence and moisture permeability variation range sequence are generated respectively. According to the preset physiological rationality threshold, abnormal fluctuations are automatically identified and marked. The time period when the variation range of multiple consecutive windows exceeds the threshold is marked as the performance mutation segment, otherwise it is marked as the performance stable segment, and the segmented marking sequence of air permeability and moisture permeability performance is output. S33. Perform time alignment and correlation analysis on the segmented marker sequences of breathability and moisture permeability to accurately identify the overlapping mutation intervals where breathability mutations and moisture permeability mutations overlap in time; combine dynamic environmental data to analyze the dominant influencing factors of each overlapping mutation interval, integrate all time series, segmented markers and correlation analysis results to generate a dynamic change map of breathability and moisture permeability performance throughout the entire application cycle of the mask.

7. The method for dynamic process testing of air permeability and moisture permeability of nonwoven fabrics used in the production of facial mask fabrics according to claim 1, characterized in that, Step S4 specifically includes: S41. Based on the synchronous monitoring data acquired in real time during S3, a continuous microscopic image sequence of the fabric surface and time-series data from the breathing airflow sensor are obtained. The minute fluctuation amplitude of the fabric caused by the alternation of positive and negative breathing pressure is calculated frame by frame using image analysis technology to generate fluctuation amplitude time-series data. The fluctuation amplitude time-series data and breathing airflow data are precisely aligned in the time domain. The pairing data of the peak airflow intensity and the corresponding fabric fluctuation amplitude in each breathing cycle are extracted. The intensity coefficient of the effect of breathing airflow on fabric fluctuation is calculated through linear regression analysis. S42. Simultaneously acquire the cumulative sweating time series on the same time axis, analyze its increasing trend and slope changes to identify key turning points, and divide the entire cycle of mask application into several sweating increasing intervals with different moisturizing characteristics. S43. Divide the breathability performance time-series change record into data segments according to the defined intervals of increasing sweat volume; within each interval, extract all instantaneous breathability performance test values ​​within that time period to form a numerical set, calculate the maximum and minimum values ​​of breathability in each set, and the difference is the range of breathability performance fluctuation in that interval; arrange the ranges of fluctuation in all intervals in chronological order to obtain a sequence of breathability performance fluctuation intervals associated with the sweating process. S44. Correlation analysis is performed between the respiratory intensity coefficient and the range of air permeability fluctuation in each interval to determine the dominant factors of air permeability fluctuation in different sweating intervals, and output a sequence of air permeability fluctuation intervals with attribution of dominant factors.

8. The method for dynamic process testing of air permeability and moisture permeability of nonwoven fabrics used in the production of facial mask fabrics according to claim 1, characterized in that, Step S5 specifically includes: The baseline data under the stable performance segment was extracted from the time-series change records of breathability and moisture permeability, and the change in local fit at each moment was calculated based on the facial micro-movement sequence. The change in local fit was compared with the fluctuation range of breathability performance, and the moment when the change in local fit exceeded the fluctuation range of breathability performance was identified as the key interference point. For each key interference point, the actual measured value of moisture permeability, the change in local fit, the intensity of breathing airflow, and the amount of sweat accumulation at that moment are extracted simultaneously to form a multi-dimensional data set; Using the actual measured value of breathability as the dependent variable and the changes in local fit, breathing airflow intensity, and cumulative sweating as independent variables, a multiple linear regression analysis was performed on the multidimensional data set to obtain the contribution weight of each dynamic factor to the change in breathability. Based on the weight results, the influence degree of each dynamic factor was ranked. Based on the contribution weight of each dynamic factor and its changing trend throughout the entire application cycle of the mask, the overall change direction of moisture permeability under the combined effect of multiple dynamic factors is analyzed through weighted calculation, and the evaluation results including the ranking of the contribution weights of each factor and the overall influence trend are output.

9. The method for dynamic process testing of air permeability and moisture permeability of nonwoven fabrics used in the production of facial mask fabrics according to claim 1, characterized in that, Step S6 specifically includes: Obtain the performance time series data corresponding to the current comprehensive influence trend and divide it into multiple continuous time periods; compare the average air permeability and average moisture permeability of each time period with the corresponding preset performance thresholds to determine whether there are time periods where the air permeability or moisture permeability is lower than the preset thresholds. If there is a time period below the preset threshold, the dynamic environment simulation parameters corresponding to that time period are extracted. The parameters include the breathing airflow intensity change curve, the cumulative sweating curve, and the amplitude and frequency data of facial micro-movements within that time period. Based on the position of the time period below the preset threshold in the entire application cycle and the degree to which the performance is below the threshold, the adjustment amount of respiratory airflow intensity, sweating rate and facial movement amplitude is calculated by linear interpolation method. The calculated adjustment amount is superimposed on the parameter curve of the corresponding time period in the original dynamic environment simulation parameters to form the updated dynamic environment simulation parameters. Using the updated dynamic environment simulation parameters, the analysis process is rerun to obtain new performance time-series data and corresponding comprehensive impact trends, and an optimized comprehensive impact trend is generated.

10. The method for dynamic process detection of air permeability and moisture permeability of nonwoven fabrics used in the production of facial mask fabrics according to claim 1, characterized in that, Step S7 specifically includes: Based on the optimized performance time-series data, curves showing the changes in breathability and moisture permeability over time throughout the entire mask application cycle were plotted to generate a performance spectrum covering the entire cycle. The maximum, minimum, average, and standard deviation of breathability and moisture permeability were extracted from the spectrum to quantify and determine the fluctuation range and stable interval of performance. The performance graph is mapped to environmental parameters of real-world usage scenarios to analyze the performance of the mask substrate under different sweating stages, different breathing intensities, and different facial movements. Based on preset breathability and moisture permeability comfort thresholds, the graph is divided into comfortable areas that meet the comfort standards and uncomfortable areas that are below the thresholds, generating a classified comfort space-time distribution map. Based on the proportion, duration, and occurrence stage of comfortable and uncomfortable areas in the comfort distribution map, and combined with the performance fluctuation range, a weighted scoring model is used to calculate the overall comfort score of the mask substrate. According to the preset level range to which the overall score belongs, the final overall comfort level is determined, and a comprehensive evaluation report including performance map, comfort distribution, level assessment, and improvement suggestions is generated.

Images