Nonwoven fabric processing process control method

Abstract

The application discloses a non-woven fabric processing process control method, and particularly relates to the field of non-woven fabric processing; an environmental humidity change rate function in a set period is obtained; a humidity impact factor is extracted and a fiber distribution sensitivity coefficient is calculated; a structure abnormality early warning factor is generated in combination with fiber deposition real-time image data; the fiber arrangement path is dynamically adjusted according to the early warning result, so that the fiber deposition tends to be uniform; and the hot-pressing time and temperature distribution are synchronously adjusted to realize hot-pressing compensation; further, the cooling air speed is adjusted according to the dry-wet ratio at the outlet of the cooling section; finally, the fabric structure quality is evaluated through image similarity calculation; the application realizes the whole-process closed-loop control of humidity disturbance-structure response-process adjustment, effectively improves the fiber uniformity and interlayer bonding stability of the non-woven fabric product under complex environment, and is suitable for the intelligent manufacturing scene of high-performance non-woven fabric.

Description

Technical Field

[0001] This invention relates to the field of nonwoven fabric processing, and more specifically to a method for controlling the processing technology of nonwoven fabrics. Background Technology

[0002] Nonwoven fabrics, which can be formed without spinning and weaving processes, are widely used in various industries such as medical, filtration, agriculture, and disposable products, and have advantages such as high forming efficiency, strong flexibility, and diverse structures. However, with the increasing automation of production lines and the growing demand for functional nonwoven fabrics, higher requirements are being placed on their microstructure uniformity, pore distribution, and thermal bonding stability.

[0003] During long-term continuous production, fluctuations in ambient humidity have a particularly significant impact on fiber deposition, electrostatic adsorption, and hot-press bonding processes. Especially in environments with rapid alternations between high and low humidity, incomplete fiber dehumidification and delayed hot-pressing temperature compensation lead to localized fiber concentration, porosity distortion, and a sharp decrease in interlayer separation force. Existing nonwoven fabric control systems mostly rely on single sensor data feedback (such as temperature or humidity), which cannot identify transient structural anomalies under high humidity impacts. Consequently, they cannot adjust web density, hot-pressing intensity, and cooling time in a timely manner, resulting in structural defect bands that severely affect product consistency and mechanical properties. Summary of the Invention

[0004] The purpose of this invention is to provide a method for controlling the processing technology of nonwoven fabrics to overcome the shortcomings of the prior art.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a method for controlling the processing technology of nonwoven fabrics, comprising: S100. Obtain historical relative humidity fluctuation data within a set time period in the target production environment and establish a humidity change rate function R(t); S200. Based on the humidity change rate function R(t), extract the corresponding humidity shock factor group M={m1,m2,...,mn}, where mn represents the fluctuation intensity and duration of the nth humidity change. S300. Calculate the fiber distribution sensitivity coefficient F based on the humidity impact factor group M, where F reflects the risk level of non-uniform fiber deposition. S400. Combine the F with the current real-time monitoring data P of fiber deposition to generate a structural anomaly early warning factor W. S500. Based on the structural anomaly warning factor W, adjust the fiber arrangement path T of the web laying unit so that the fiber distribution migrates to the low W region. S600: Obtain the temperature field distribution B of the hot pressing zone after the fiber arrangement path T is adjusted, and calculate the required hot pressing time compensation Δt. S700, In the hot pressing unit, extend the hot pressing time Δt and dynamically adjust the temperature field B; S800, adjust the cooling fan speed V according to the temperature-humidity ratio at the outlet of the final cooling section; S900: Perform similarity matching between the final fabric fiber image and the preset reference image, and output the quality judgment result.

[0006] Preferably, S100 specifically includes: S101. Multiple environmental parameter acquisition nodes are set up in the nonwoven fabric production line, respectively located in the web laying area, hot pressing area and cooling area, to collect relative humidity data of each key process area. S102. The collected relative humidity data is statistically analyzed in segments according to the set time interval to obtain the average and extreme values ​​of humidity in each area within the target time period. S103. The third-order central difference method is used to calculate the rate of change of relative humidity in each region over time, and the humidity change rate function R(t) is constructed.

[0007] Preferably, S200 specifically includes: S201. The humidity change rate function R(t) is processed by taking the first derivative in the time dimension to obtain the humidity change acceleration function A(t); S202. Set the humidity change detection threshold Rt and the acceleration threshold At. When at any time t, R(t) is greater than Rt and A(t) is greater than At, it is determined to be a humidity shock event. S203. Extract the start time, end time, peak change rate and duration of each humidity shock event to construct a humidity shock unit; S204. Combine all humidity impact units in chronological order to generate humidity impact factor group M.

[0008] Preferably, S300 specifically includes: S301. Using the peak change rate and duration of each humidity impact unit in the humidity impact factor group M as input variables, construct the fiber disturbance influence function. ; S302. Based on historical production data of nonwoven fabrics, establish a fiber distribution anomaly probability model P(F|G) to fit the influence of humidity shock on fiber distribution nonuniformity. S303, Calculate the corresponding humidity impact units for all humidity impact units in the current cycle. The values ​​are then substituted into the anomaly probability model P(F|G) for weighted summation; S304. Output the fiber distribution sensitivity coefficient F for the current time period based on the weighted result.

[0009] Preferably, S400 specifically includes: S401. Obtain real-time image data of fiber deposition at the current netting outlet, and extract the fiber deposition distribution parameter P using grayscale distribution analysis. S402. The obtained fiber distribution sensitivity coefficient F and the fiber deposition distribution parameter P are jointly normalized to form a standardized feature pair. ; S403. Construct a structural anomaly risk inference model based on fuzzy logic judgment rules, in order to and As input, output structural anomaly warning factor W; S404. Classify the structural anomaly level according to the magnitude of the W value.

[0010] Preferably, S500 specifically includes: S501. Divide the netting exit area into multiple spatial grid units, and map the structural anomaly early warning factor W to each unit to form a two-dimensional risk distribution matrix; S502. Based on the spatial gradient information of the W value in the risk distribution matrix, identify target areas where the W value is lower than a preset safety threshold. S503. Combining the current fiber feeding speed and spinneret angle parameters, the optimal offset trajectory of the fiber arrangement path T is calculated using a path planning algorithm. S504, control the web laying unit to execute path T, so that the fiber distribution is concentrated towards the low W region.

[0011] Preferably, S600 specifically includes: S601. Obtain the two-dimensional temperature distribution map of the hot-pressing zone corresponding to the current fiber arrangement path T, and construct the temperature field distribution matrix B; S602. Analyze the local temperature deviation ΔB in the fiber aggregation area based on matrix B, and match ΔB with the historical bonding sufficiency curve; S603. Based on the difference between ΔB and the target bonding temperature Ts, and combined with the current conveying speed Vm, calculate the hot pressing time compensation Δt. S604. Write Δt as a dynamic adjustment parameter into the hot pressing control command to extend the heating time of the hot pressing device in the target path segment.

[0012] Preferably, S700 specifically includes: S701. Synchronously input the hot pressing time compensation value Δt to the hot pressing control device and set the time for the pressure roller to pass through the target path segment; S702. Based on the temperature compensation requirement corresponding to Δt, adjust the target temperature value of each heating unit in the hot pressing zone to construct a dynamic temperature control command. S703. Real-time monitoring of the temperature field distribution matrix B′ in the hot-pressed area, and difference analysis between it and the temperature matrix B before compensation to determine whether the heating response meets the standard. S704. If the temperature of any target point in B′ does not reach the set offset threshold, the heating power is adjusted and Δt is extended through feedback.

[0013] Preferably, S800 specifically includes: S801. At the outlet of the cooling section, the temperature T0 and relative humidity H0 of the cooling airflow are obtained in real time. S802. Calculate the outlet dry-wet ratio R0 of the cooling section based on T0 and H0, and compare it with the target dry-wet ratio Rs to determine the cooling deviation ΔR. S803. Using ΔR as an input variable, and combining the cooling section length L and the fiber layer heat capacity parameter Ct, a linear adjustment model is used to calculate the wind speed regulation value ΔV. S804. Correct the current cooling fan speed V based on ΔV.

[0014] The technical effects and advantages provided by the present invention in the above technical solution are as follows: 1. Compared to existing nonwoven fabric processing methods that rely on single process parameter control, this invention features multi-dimensional information fusion, predictive-control linkage, and image quality self-assessment capabilities. This not only improves process stability in complex environments but also reduces manual intervention and the burden of post-processing inspection. This solution is particularly suitable for processing functional or highly consistent nonwoven fabric products and has broad industrial application value. Attached Figure Description

[0015] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this invention. For those skilled in the art, other drawings can be obtained based on these drawings.

[0016] Figure 1 This is a flowchart of the method of the present invention. Detailed Implementation

[0017] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0018] For examples, please refer to Figure 1As shown, the nonwoven fabric processing technology control method described in this embodiment includes: S100. Obtain historical relative humidity fluctuation data within a set time period in the target production environment and establish a humidity change rate function R(t).

[0019] Step S101: In the nonwoven fabric processing production line, several environmental parameter acquisition nodes are deployed in the web laying area, hot pressing area, and cooling area. Each acquisition node consists of a temperature and humidity sensor, using digital temperature and humidity sensors with a response time of less than 5 seconds and a measurement error of no more than ±2%. The nodes are distributed along the production line at 1-meter intervals, with at least 3 nodes in each area to ensure coverage of the spatial environmental data at each key process point. Each sensor is set to acquire relative humidity data every 30 seconds and transmits it to the central data acquisition unit via a wired network, forming a regional time-series humidity raw dataset.

[0020] Step S102: The relative humidity data collected in step S101 is segmented and statistically analyzed according to a fixed time window Δt of 10 minutes. For the raw humidity data in each time period, the following statistical values ​​are calculated: mean humidity. The calculation method is the arithmetic mean of all relative humidity data within this time period; The humidity extreme values ​​{Hmin, Hmax} represent the maximum and minimum humidity values ​​occurring within the specified time period, respectively. The above statistical operations are performed on each data acquisition node and saved according to node number and process region, forming a humidity feature matrix with the structure {region-node number-time period}.

[0021] Step S103: After obtaining the humidity feature matrix, a humidity change rate function R(t) is constructed based on its time series information to reflect the dynamic change trend of relative humidity in each process area over time. Specifically, the third-order central difference method is used to numerically differentiate the mean humidity of each time series. The mean humidity corresponding to the i-th time period in the time series is set to... If the time step is Δt = 10 minutes, the humidity change rate R(i) at the i-th time point is calculated as follows: Humidity Change Rate Where R(i) is expressed as a percentage per minute (% / min), used to quantify the rate and direction of humidity change at that time point. Forward and backward differencing are used for the first and last points of the sequence to avoid missing data.

[0022] Subsequently, the R(i) values ​​at all time points are fitted to generate the humidity change rate function R(t), which is a continuous function defined on the production cycle time axis, and is expressed as: ,in This is a local weight function constructed based on spline interpolation.

[0023] S200. Based on the humidity change rate function R(t), extract the corresponding humidity impact factor group. , where mn represents the intensity and duration of the nth humidity fluctuation.

[0024] Step S201: Based on the humidity change rate function R(t) constructed in step S103, perform a first-order derivative operation on it in the time dimension to obtain the humidity change acceleration function A(t). This function reflects the changing trend of the humidity change rate itself and can more sensitively capture the steepness of change when abrupt changes occur. Specifically, the central difference method is used to calculate the numerical derivative of R(t), and the time step is set to Δt. Then, at time point... Acceleration of humidity change at the location The calculation method is as follows: A(t) is a time series dataset of the same length as R(t), expressed as a percentage per minute squared. (This parameter is used to identify extreme locations of changing trends and is an important parameter for subsequent identification of humidity shock events.)

[0025] Step S202: To determine abrupt events during humidity changes, two key thresholds need to be set: the humidity change rate threshold Rt and the humidity change acceleration threshold At. The humidity change rate threshold Rt is set to the mean of all R(t) series values ​​plus 1.5 times the standard deviation; the humidity change acceleration threshold At is set to the mean of all A(t) series values ​​plus 2 times the standard deviation. The above statistics adopt a dynamic sliding window method, which is updated every 30 minutes to adapt to real-time changing operating conditions.

[0026] When the following two conditions are met at a certain time point t: R(t) is greater than Rt; A(t) is greater than At; This indicates that a humidity shock event occurred at that time point. This method combines the characteristics of both the first-order rate of change and its acceleration, effectively eliminating misjudgments caused by periodic humidity fluctuations or small disturbances, and ensuring the accuracy of abrupt change identification.

[0027] Step S203: After identifying the time point t where the humidity shock event occurs, further iterate forward and backward through the humidity change rate function R(t) and the humidity change acceleration function A(t) to find: Start time: The point at which the humidity change rate first exceeds the set background fluctuation range; Termination time: The point at which the humidity change rate returns to within the background fluctuation range; Peak rate of change: The maximum value of R(t) within the time interval of the impact event; Duration: End time minus start time.

[0028] The above four parameters are organized into structured data units, defined as humidity impact units, in the following form: ;in, The start time, For the end time, The peak rate of change, For duration (i.e. ).

[0029] Step S204: Arrange the humidity impact units m1 to mn obtained in step S203 in chronological order to construct a humidity impact factor group M, represented as: Each humidity shock unit represents an independent humidity abrupt change process, the intensity of which can be determined by... The duration can be measured by Measurement. The humidity impact factor group M will serve as the input for calculating the fiber distribution sensitivity coefficient in the next stage, quantifying the potential impact of abrupt events on fiber deposition uniformity. The factor group M is arranged in chronological order and includes structured parameters, facilitating matching with the process timeline and enabling feedforward control of the fiber structure response behavior.

[0030] S300. Calculate the fiber distribution sensitivity coefficient F based on the humidity impact factor group M, where F reflects the risk level of non-uniform fiber deposition.

[0031] The implementation method of step S301: Each humidity impact unit extracted in step S204 The two core parameters are the peak rate of change. With duration As input variables, construct the fiber disturbance influence function. This is used to quantify the impact of a single humidity surge on fiber structure disturbance. Fiber disturbance influence function. The definition is as follows: ;in, The unit is percentage per minute (% / min). The unit is minutes. The log function is used to enhance the weight of long-term shock events, avoiding misjudgment of the impact caused by extremely short-term strong abrupt changes and long-term moderate abrupt changes. The larger the value of this function, the more significant the potential impact of the humidity shock on fiber web laying disturbance.

[0032] Step S302: Based on historical nonwoven fabric production data, construct a fiber distribution anomaly probability model P(F|G) to fit the influence of humidity disturbance factors on the risk of non-uniform fiber deposition. The specific operation is as follows: Collect no less than 100 sets of historical production cycle data, and extract the corresponding humidity impact factor group M historical sequence and actual fiber deposition image data; Image processing methods (such as gray-level co-occurrence matrix and variance analysis) are used to extract the non-uniformity index S of fiber distribution as the target label; Using G as the independent variable and S as the dependent variable, a logistic regression model was used for probability fitting to obtain P(F|G), which is the probability of abnormal fiber distribution under the influence of a specific disturbance. After the model is trained, cross-validation is used to ensure that its prediction accuracy is not less than 85%.

[0033] This model is used to predict in real time the structural anomaly risks that disturbance events may cause under current operating conditions, providing a theoretical basis for the F-value.

[0034] Step S303: Within the current production cycle, for each humidity impact unit in the humidity impact factor group M. Calculate its disturbance impact value This value is then substituted into the fiber distribution anomaly probability model P(F|G) established in step S302 to obtain the corresponding fiber non-uniform deposition probability value. Subsequently, the combined impact of each humidity shock on the overall sensitivity was summarized using the following weighting method: This weighting method takes into account both the intensity of the disturbance and the corresponding risk probability, and can more accurately reflect the cumulative risk of multiple humidity abrupt changes on the fiber deposition distribution.

[0035] Step S304: Output the above summary result F as the fiber distribution sensitivity coefficient for the current time period. The value of F ranges from 0 to 1. The higher the value, the more susceptible the fiber distribution is to sudden changes in humidity, and the higher the risk level. The specific level classification is as follows: F less than 0.3: low disturbance, process can maintain current parameters; F between 0.3 and 0.6: moderate disturbance, subsequent process stability needs to be monitored; F greater than 0.6: High disturbance, it is recommended to adjust the mesh laying path and hot pressing parameters to deal with the risk of structural anomalies.

[0036] S400. Combine the F with the current real-time monitoring data P of fiber deposition to generate a structural anomaly early warning factor W.

[0037] Step S401: Vertically install a high-resolution industrial camera at the web-laying exit, setting the acquisition area size to 100 mm × 100 mm and the acquisition frequency to 5 frames per minute. The acquired image is an 8-bit grayscale image, used to represent the fiber distribution density on the nonwoven fabric surface. Image features are extracted using grayscale distribution analysis, specifically including the following processing steps: mean filtering is applied to the image to reduce noise and improve image clarity; The image is divided into 10×10 grid cells, and the mean gray value is calculated for each cell. with standard deviation Calculate the mean and variance of gray levels for the entire image. As the core indicator of fiber deposition distribution parameter P, the formula is: , where n is the number of grids; the higher the value of parameter P, the more uneven the fiber distribution, which can be used as a characterization of real-time structural stability.

[0038] Step S402: Perform joint normalization on the fiber distribution sensitivity coefficient F obtained in step S304 and the fiber deposition distribution parameter P calculated in step S401 to form a normalized feature pair. This is used for subsequent risk reasoning modeling. The normalization method adopts linear minimum-maximum normalization, and the specific processing is as follows: Scaling F within its historical range [F_min, F_max] yields... ; Similarly, scaling P within its historical range [P_min, P_max] yields... ; After processing, both F′ and P′ are between 0 and 1, avoiding interference from inconsistent feature scales on the inference results and improving the stability and generalization ability of the judgment model.

[0039] Step S403: Construct a structural anomaly risk inference model based on fuzzy logic, using standardized feature pairs (F′, P′) as input to output the structural anomaly early warning factor W. The model construction process is as follows: F′ and P′ are divided into three fuzzy levels: low, medium and high, and the corresponding fuzzy membership functions are represented by triangular functions. A fuzzy inference rule base has been established, containing 9 rules, as shown in the example below: If F′ is high and P′ is high, then W is high; If F′ is medium and P′ is high, then W is medium to high; If F′ is low and P′ is low, then W is low; The Mamdani fuzzy inference method is used for rule matching and fuzzy set operations. The centroid method is used to demodel the fuzzy result set, and the declarative value W is output, ranging from 0 to 1, which represents the degree of structural anomaly risk.

[0040] Step S404: Classify the risk level of fiber deposition state according to the numerical range of the structural anomaly warning factor W to guide subsequent process control responses. The specific classification criteria are as follows: W < 0.3: Risk level is low, fiber structure is stable, no adjustment is needed; When W is between 0.3 and 0.6: the risk level is medium, and it is recommended to switch to enhanced monitoring mode; W > 0.6: The risk level is high, and it is recommended to immediately adjust the web laying path and hot pressing temperature for preventive intervention.

[0041] By quantitatively outputting and classifying the structural anomaly warning factor W, feedforward identification and process response triggering for potential fiber distribution anomalies were achieved.

[0042] S500. Based on the structural anomaly warning factor W, adjust the fiber arrangement path T of the web laying unit so that the fiber distribution migrates to the low W region.

[0043] Step S501: First, the mesh laying exit area is divided into a two-dimensional spatial cell array according to an equidistant grid. The size of each cell is set to 20 mm × 20 mm, forming an m-row × n-column spatial grid structure. Each grid cell is numbered G(i,j) to identify its position in the mesh laying plane.

[0044] Combining the structural anomaly warning factor W output in step S404, the W value at the current time point is mapped one-to-one with the fiber image coordinates and mapped to each spatial unit G(i,j) to form a two-dimensional risk distribution matrix W_grid(i,j) corresponding to the netting area, where each matrix element represents the degree of structural anomaly risk at the corresponding location.

[0045] Step S502: In the two-dimensional risk distribution matrix W_grid(i,j), calculate the spatial gradient ΔW of the W values ​​between each adjacent grid. The gradient is calculated as follows: By determining the numerical direction of ΔW(i,j), the decreasing trend of W value is identified, thereby identifying clustered areas with lower risk.

[0046] A safety threshold Ws is set to 0.4. When a certain area grid W_grid(i,j) is less than Ws and its ΔW(i,j) is negative, the grid is determined to be a target area and has priority in receiving fiber arrangement adjustments. All target grids that meet the conditions form the target area set T_target for use in the next step of path planning.

[0047] Step S503: Combining the current real-time process parameters, including: fiber feeding speed V_f (grams per minute); spinneret angle θ_s (degrees); in the fiber outlet coordinate space, taking the current position as the starting point and the region with the lowest W value in T_target as the target point, construct the fiber layout path T using the A* path planning algorithm based on the cost function. The cost function Cost(i,j) is defined as: Where: W_grid(i,j) is the risk value of the current grid; D(i,j) is the Euclidean distance from the starting point to G(i,j); α and β are weighting coefficients, with α set to 0.7 and β set to 0.3 to emphasize the priority of risk avoidance. The path planning is then completed and generated. , representing the grid coordinates that the optimal layout trajectory needs to pass through sequentially.

[0048] Step S504: Convert path T into execution instructions for the web-laying unit, including the nozzle lateral oscillation trajectory and fiber feed position offset. Control parameters are set as follows: based on each target mesh. The center coordinates are used to adjust the movement range of the nozzle lateral drive motor; the outlet position and spinneret angle θ_s of the feeding component are adjusted synchronously to ensure that the fiber stream accurately covers the target trajectory T; the path offset effect is monitored in real time during the execution process, and the W_grid matrix is ​​re-evaluated every 10 seconds to ensure that the path T is continuously aligned with the low-risk area, so as to achieve the goal of fiber migration to the low W area, thereby improving the deposition uniformity and structural stability.

[0049] S600: Obtain the temperature field distribution B of the hot pressing zone after the fiber arrangement path T is adjusted, and calculate the required hot pressing time compensation Δt.

[0050] Step S601: An infrared thermal imaging device is installed above the hot-pressing unit to collect surface temperature distribution data in the horizontal and vertical directions of the hot-pressing area. The infrared image resolution is set to 1 mm × 1 mm, and the acquisition frequency is 10 frames per second. The image is converted into a temperature value matrix to construct a two-dimensional temperature field distribution matrix B, represented as: B(i,j) represents the temperature value at the i-th row and j-th column position in the hot-pressing area, in degrees Celsius.

[0051] Based on each coordinate point in the fiber arrangement path T obtained in step S504, the region corresponding to path T in matrix B is extracted as the analysis object to ensure that the collected temperature data can accurately reflect the heating situation under the actual movement trajectory of the fiber.

[0052] Step S602: Identify fiber aggregation regions in matrix B and calculate the average temperature within these regions. Temperature of the surrounding area The difference is defined as the local temperature deviation ΔB: ;in: The average temperature of the region traversed by path T; To and The average temperature of adjacent control areas. Subsequently, the ΔB value was matched with a pre-established historical adhesion adequacy curve. This curve, based on experimental data of peel strength of different fiber materials under different temperatures and hot-pressing times, is expressed as: Where Y is the bond strength, T is the temperature, and t is the hot-pressing time. By inputting the temperature corresponding to the current ΔB into the function f(T,t), the target bond strength is found. The minimum hot pressing time required under the given conditions provides a reference for subsequent compensation calculations.

[0053] Step S603: Based on the target bonding temperature of the hot-pressing zone (For example, 165 degrees Celsius) and the current average temperature of the region The difference ΔT between them is defined as: Combined with the current conveyor belt speed The unit is millimeters per second. The required hot-pressing time compensation Δt is calculated using a hot-pressing response hysteresis model, as follows: Where γ is an empirical constant representing the linear response coefficient of hot pressing compensation time per unit temperature difference, and Δt represents the additional heating time required for the current path segment, used to ensure that the structural layers can still achieve the target bonding quality under the influence of temperature difference.

[0054] Step S604: The hot pressing time compensation Δt calculated in step S603 is used as a dynamic control parameter and embedded into the hot pressing control command, specifically including: A delayed shutdown command is sent to the heater control terminal of the hot pressing device, extending the heating time Δt in the area corresponding to path T; the cycle time of the pressure rollers is adjusted synchronously to extend the residence time of the fibers in the hot pressing zone by Δt; if Δt exceeds the set safety threshold (e.g., 10 seconds), the conveying speed is reduced in conjunction with the operation. Within a safe range, ensure that the hot-pressed area does not overheat or that hot spots do not spread. This compensation process is performed independently on a path-by-path basis, achieving regional hot-pressing adaptive control and ensuring that the bonding performance in the structural offset area is not lower than the standard range.

[0055] S700, extend the hot pressing time Δt in the hot pressing unit and dynamically adjust the temperature field B.

[0056] Step S701: The hot pressing time compensation value Δt calculated in step S603 is used as a dynamic control parameter and input into the time control module in the hot pressing control device. The passing time of the pressure roller corresponding to the coordinate segment of path T is set as follows: Where: t_base is the standard pressure roller passage time, determined based on the conveying speed and process settings; Δt is the additional compensation time, in seconds. The hot-pressing control device dynamically adjusts the synchronous movement of the pressure rollers and the conveying cycle based on the t_total value, ensuring that the heating time of the fiber within the target path area matches the structural bonding requirements.

[0057] Step S702: To achieve coordinated control of hot pressing time and temperature, the target temperature settings of multiple heating units arranged laterally in the hot pressing zone are adjusted according to the temperature compensation requirement ΔT corresponding to Δt (obtained from step S603). The target temperature T′ of each heating unit is calculated as follows: Where: T_base is the target temperature under normal production conditions, such as 165 degrees Celsius; ΔT local value represents the temperature compensation value required for the corresponding coordinate position of path T in the hot pressing zone, which is calculated by matching Δt after the path point is mapped to the two-dimensional coordinates in the hot pressing zone.

[0058] It should be noted that this temperature compensation ΔT is set only under the premise of keeping T_base constant and achieving the target bonding energy transfer through Δt time compensation. That is: if Δt can fully compensate for the current temperature difference ΔT, then T_base is not adjusted; if the current conveyor belt speed or equipment structure limitations result in Δt being insufficient for compensation, then the heating unit temperature T′ is locally adjusted through ΔT to compensate for the insufficient heat caused by insufficient time.

[0059] Therefore, this temperature control logic is suitable for dynamic adjustment under temperature-time linkage compensation mode, ensuring that the hot pressing zone can still achieve full bonding between fibers under local disturbance scenarios, avoiding the risk of redundant heating or overcompensation due to unmet model application conditions.

[0060] Subsequently, each T′ value is written into the dynamic temperature control command queue, which is then issued and implemented one by one by the temperature control actuator to achieve a regional temperature increase response in the temperature field.

[0061] Step S703: After the temperature command is executed, the infrared thermal imaging device is used to collect a new frame of temperature data of the current hot-pressed zone in real time, and a temperature field distribution matrix is ​​constructed. Then, a one-to-one difference analysis is performed between the original temperature field matrix B before compensation and the matrix B, and the temperature response matrix ΔB′ is calculated: The temperature rise effect of the region corresponding to path T in the ΔB′ matrix is ​​statistically analyzed, and the average and minimum temperature rise values ​​are calculated to determine whether the ΔT set value is met. If the temperature rise of most points in ΔB′ reaches the set target (e.g., more than 90% of ΔT), the hot pressing temperature rise response is judged to be up to standard, and the process proceeds to the next stage.

[0062] Step S704: If any target region in the ΔB′ matrix experiences insufficient temperature rise (i.e., the temperature rise does not reach the preset offset threshold ΔTt), then the thermal pressure feedback correction mechanism is activated. The typical value of the preset offset threshold ΔTt is 5 degrees Celsius, and the specific setting depends on the thermal bonding characteristics of the fiber material. The feedback correction mechanism includes: Increase the output power of heating units corresponding to grids that do not meet the standards, with a maximum increase of no more than 10%; The synchronization is extended again by Δt, with the increment set to 10% to 20% of the initial Δt, depending on the degree to which ΔB′ fails to meet the target; Update the target temperature setpoint T′ and reissue the dynamic control command to enter the next round of temperature rise response monitoring.

[0063] S800, adjust the cooling fan speed V according to the temperature-humidity ratio at the outlet of the final cooling section.

[0064] Step S801: Install a high-precision temperature and humidity acquisition device at the end of the fiber flow direction at the outlet of the cooling section. The acquisition frequency is set to once every 5 seconds to obtain the temperature T0 and relative humidity H0 of the cooling airflow in real time. The temperature T0 is measured in the range of 0 to 80 degrees Celsius, with an accuracy of ±0.2 degrees Celsius. The relative humidity H0 is measured from 0% to 100% with an accuracy of ±2%.

[0065] The data is uploaded to the central control unit via an industrial communication interface for real-time assessment of the cooling section's environmental conditions, providing a basis for subsequent wet-dry ratio analysis.

[0066] Step S802: Based on the temperature T0 and relative humidity H0 obtained in step S801, calculate the wet-dry ratio R0 of the current cooling airflow, which measures the ratio of water vapor content in the air to the mass of dry air. The calculation formula is as follows: Where: P_w is the partial pressure of water vapor calculated from H0 at temperature T0, in kPa; P_total is atmospheric pressure, with a standard value of 101.325 kPa; 0.622 is the molecular mass ratio of water vapor to dry air. The cooling deviation ΔR is obtained by calculating the difference between R0 and the preset target wet-dry ratio Rs. The target wet-dry ratio Rs is a process stability reference value (e.g., set to 0.008), and ΔR reflects the degree to which the current cooling environment deviates from the optimal cooling conditions.

[0067] Step S803: Take the cooling deviation ΔR as an input variable, combined with the following two parameters: The cooling section length L (in meters) represents the effective path length of the fiber layer cooling process; the unit mass heat capacity parameter Ct of the fiber layer (in joules per gram per degree Celsius) is set according to the fiber type and the number of structural layers, for example, the typical value for polypropylene fiber is 2.1 joules per gram per degree Celsius. The wind speed adjustment value ΔV is calculated using a linear adjustment model, with the specific formula as follows: Where: k is the proportional coefficient, set in conjunction with the cooling fan response characteristics experiment; ΔV represents the wind speed correction range.

[0068] Step S804: Based on the sign and magnitude of ΔV, adjust the current cooling section airflow speed V in real time: If ΔV is positive, it indicates that the wet-dry ratio is too high, and the airflow speed needs to be increased to accelerate the removal rate of humid heat; if ΔV is negative, it indicates that cooling is too fast, which may cause the fiber layer structure to shrink, and the airflow speed should be reduced to mitigate the cooling intensity. The correction method is as follows: The wind speed adjustment range is limited to 300 to 1200 meters per hour, and the adjustment step should not exceed 50 meters per minute per hour to prevent excessive disturbance.

[0069] The corrected wind speed value V_new is applied in real time to the cooling fan control unit, driving the variable frequency fan speed adjustment, ultimately achieving dynamic environmental balance in the cooling process and steady-state forming of the fiber layer structure.

[0070] S900: Perform similarity matching between the final fabric fiber image and the preset reference image, and output the quality judgment result.

[0071] An industrial image acquisition device is installed at the fabric output end, covering the entire width of the fabric laterally, with a longitudinal sampling length of 100 mm and an acquisition resolution of no less than 0.1 mm per pixel. At least three images are acquired per production cycle, saved in grayscale format. The sampling time corresponds to the completion time of the hot pressing and cooling processes to ensure the images reflect the final stable fiber structure. After acquisition, the images are uploaded to the image processing unit and used as target sample fabric images for subsequent analysis.

[0072] The acquired images undergo preprocessing, which mainly includes the following: Gray-scale normalization: Normalizes the gray-scale values ​​of an image to the range of 0 to 1, eliminating the interference of lighting differences on texture judgment; Image denoising: Median filtering is used to smooth image boundaries and improve texture stability; Texture feature extraction: The gray-level co-occurrence matrix method is used to extract the statistical features of the image, including four indicators: energy, entropy, contrast and correlation. The above statistical indicators are constructed into a four-dimensional vector, defined as the target image feature vector G1, in the following form: This vector is used to characterize the macroscopic structural features of the current fabric image.

[0073] Based on the current production process parameters (such as raw material type, hot pressing temperature, wind speed level, etc.), the corresponding standard image in the preset reference image database is called, and the reference image feature vector G2 is extracted: The structural similarity algorithm is used to calculate the similarity S between the target image and the reference image. The algorithm uses the reciprocal of the weighted Euclidean distance; the larger the S value, the higher the image similarity. This algorithm can effectively capture differences in texture levels and is suitable for the consistency recognition of fiber structure distribution.

[0074] The similarity value S is compared with a preset quality judgment threshold St to determine whether the current fabric structure meets the standard. The threshold St is determined based on historical experimental data and is generally set between 0.75 and 0.85. The quality judgment logic is as follows: If S is greater than or equal to St, the output result is qualified, indicating that the target image and the reference image have good structural consistency. If S is less than St, the output result is structural deviation risk, marking the fabric area as a potentially defective part for subsequent process traceability or alarm handling.

[0075] The image quality assessment results are synchronously stored in the data record table and can be linked to the process adjustment logic for closed-loop process optimization control.

[0076] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application.

Claims

1. A method for controlling the processing technology of nonwoven fabrics, characterized in that: include: S100. Obtain historical relative humidity fluctuation data within a set time period in the target production environment and establish a humidity change rate function R(t); S200. Based on the humidity change rate function R(t), extract the corresponding humidity shock factor group M={m1,m2,...,mn}, where mn represents the fluctuation intensity and duration of the nth humidity change. S300. Calculate the fiber distribution sensitivity coefficient F based on the humidity impact factor group M, where F reflects the risk level of non-uniform fiber deposition. S400. Combine the F with the current real-time monitoring data P of fiber deposition to generate a structural anomaly early warning factor W. S500. Based on the structural anomaly warning factor W, adjust the fiber arrangement path T of the web laying unit so that the fiber distribution migrates to the low W region. S600: Obtain the temperature field distribution B of the hot pressing zone after the fiber arrangement path T is adjusted, and calculate the required hot pressing time compensation Δt. S700, In the hot pressing unit, extend the hot pressing time Δt and dynamically adjust the temperature field B; S800, adjust the cooling fan speed V according to the dry-wet ratio at the final cooling section outlet; S900: Perform similarity matching between the final fabric fiber image and the preset reference image, and output the quality judgment result; Specifically, S200 includes: S201. The humidity change rate function R(t) is processed by taking the first derivative in the time dimension to obtain the humidity change acceleration function A(t); S202. Set the humidity change detection threshold Rt and the acceleration threshold At. When at any time t, R(t) is greater than Rt and A(t) is greater than At, it is determined to be a humidity shock event. S203. Extract the start time, end time, peak change rate and duration of each humidity shock event to construct a humidity shock unit; S204. Combine all humidity impact units in chronological order to generate humidity impact factor group M; Specifically, S300 includes: S301. Using the peak change rate and duration of each humidity impact unit in the humidity impact factor group M as input variables, construct the fiber disturbance influence function G(mi); S302. Based on historical production data of nonwoven fabrics, establish a fiber distribution anomaly probability model P(F|G) to fit the influence of humidity shock on fiber distribution nonuniformity. S303. Calculate the corresponding G(mi) value for all humidity impact units in the current cycle, and input it into the anomaly probability model P(F|G) for weighted summation. S304. Output the fiber distribution sensitivity coefficient F for the current time period based on the weighted result; Specifically, S400 includes: S401. Obtain real-time image data of fiber deposition at the current netting outlet, and extract the fiber deposition distribution parameter P using grayscale distribution analysis. S402. The obtained fiber distribution sensitivity coefficient F and the fiber deposition distribution parameter P are jointly normalized to form a standardized feature pair. ; S403. Construct a structural anomaly risk inference model based on fuzzy logic judgment rules, in order to and As input, output structural anomaly warning factor W; S404. Classify structural anomalies according to the magnitude of the W value; The S500 specifically includes: S501. Divide the netting exit area into multiple spatial grid units, and map the structural anomaly early warning factor W to each unit to form a two-dimensional risk distribution matrix; S502. Based on the spatial gradient information of the W value in the risk distribution matrix, identify target areas where the W value is lower than a preset safety threshold. S503. Combining the current fiber feeding speed and spinneret angle parameters, the optimal offset trajectory of the fiber arrangement path T is calculated using a path planning algorithm. S504, control the web laying unit to execute path T, so that the fiber distribution is concentrated towards the low W region.

2. The nonwoven fabric processing technology control method according to claim 1, characterized in that: Specifically, S100 includes: S101. Multiple environmental parameter acquisition nodes are set up in the nonwoven fabric production line, respectively located in the web laying area, hot pressing area and cooling area, to collect relative humidity data of each key process area. S102. The collected relative humidity data is statistically analyzed in segments according to the set time interval to obtain the average and extreme values ​​of humidity in each area within the target time period. S103. The third-order central difference method is used to calculate the rate of change of relative humidity in each region over time, and the humidity change rate function R(t) is constructed.

3. The nonwoven fabric processing technology control method according to claim 1, characterized in that: The S600 specifically includes: S601. Obtain the two-dimensional temperature distribution map of the hot-pressing zone corresponding to the current fiber arrangement path T, and construct the temperature field distribution matrix B; S602. Analyze the local temperature deviation ΔB in the fiber aggregation area based on matrix B, and match ΔB with the historical bonding sufficiency curve; S603. Based on the difference between ΔB and the target bonding temperature Ts, and combined with the current conveying speed Vm, calculate the hot pressing time compensation Δt. S604. Write Δt as a dynamic adjustment parameter into the hot pressing control command to extend the heating time of the hot pressing device in the target path segment.

4. The nonwoven fabric processing technology control method according to claim 3, characterized in that: The S700 specifically includes: S701. Synchronously input the hot pressing time compensation value Δt to the hot pressing control device and set the time for the pressure roller to pass through the target path segment; S702. Based on the temperature compensation requirement corresponding to Δt, adjust the target temperature value of each heating unit in the hot pressing zone to construct a dynamic temperature control command. S703. Real-time monitoring of the temperature field distribution matrix B′ in the hot-pressed area, and difference analysis between it and the temperature matrix B before compensation to determine whether the heating response meets the standard. S704. If the temperature of any target point in B′ does not reach the set offset threshold, the heating power is adjusted and Δt is extended through feedback.

5. The nonwoven fabric processing technology control method according to claim 1, characterized in that: The S800 specifically includes: S801. At the outlet of the cooling section, the temperature T0 and relative humidity H0 of the cooling airflow are obtained in real time. S802. Calculate the outlet dry-wet ratio R0 of the cooling section based on T0 and H0, and compare it with the target dry-wet ratio Rs to determine the cooling deviation ΔR. S803. Using ΔR as an input variable, and combining the cooling section length L and the fiber layer heat capacity parameter Ct, a linear adjustment model is used to calculate the wind speed regulation value ΔV. S804. Correct the current cooling fan speed V based on ΔV.

Images