Production control system of microporous filtering membrane based on polytetrafluoroethylene

Abstract

The invention relates to the technical field of microporous filtering membranes, in particular to a polytetrafluoroethylene-based microporous filtering membrane production control system, which comprises an acquisition module, an adjustment module, a correction module, an updating module and an early warning module. According to the method, the non-uniformity of the base band is actively compensated by adjusting the stretching rate, the calendering deviation is pre-corrected before longitudinal extension, then the quality deviation degree is introduced to correct the transverse stretching heating power, and the power correction coefficient and the mixing proportion are reversely optimized through threshold comparison of the adjustment index and the correction index and stability analysis. According to the method, closed-loop setting from back-end performance expression to front-end formula parameters is achieved, grading early warning is triggered based on the double-index deviation degree, and the problems that due to open-loop decoupling of process parameters and lack of feed-forward self-adaptive adjustment, the uniformity of the micropore structure of a finished filter membrane is poor, the air permeability fluctuation is large, and the large-scale production yield is difficult to guarantee are effectively solved.

Description

Technical Field

[0001] This invention relates to the field of microporous filtration membrane technology, and in particular to a production control system for microporous filtration membranes based on polytetrafluoroethylene. Background Technology

[0002] As high-end electronic devices evolve towards miniaturization and high performance, the environmental adaptability requirements for their built-in components are becoming increasingly stringent. Microporous filter membranes, as a key functional layer for gas exchange and foreign object barrier, directly impact the protection level and long-term reliability of these devices. Polytetrafluoroethylene (PTFE) has become the mainstream choice for preparing high-performance filter membranes due to its excellent chemical stability and designable microstructure. However, the molding process of such membranes involves complex coupling of multiple steps and parameters, and there is a non-linear relationship between microstructure and permeability, making product quality extremely sensitive to process fluctuations. Existing production processes largely rely on single-point parameter control and post-production inspection, lacking real-time perception and adaptive adjustment capabilities for dynamic disturbances between processes. This makes it difficult to maintain the consistency and stability of filter membranes in large-scale manufacturing scenarios, hindering further improvements in the integration and service life of high-end electronic components.

[0003] Chinese Patent Publication No. CN105169966A discloses a polytetrafluoroethylene (PTFE) three-dimensional microporous membrane and its preparation method. The method includes: raw material preparation: high-strength PTFE fine powder and water-soluble, colorfast, high-temperature resistant solid ultrafine powder are crushed, mixed, and stirred by a high-speed mixer, and then sieved by a vibrating screen to obtain raw materials for use; blank making: the raw materials prepared in step (1) are molded into hollow blanks by a press of more than 200 tons using a molding method, and the blanks after demolding are placed in an environment of 25℃ (±3℃) for 20 to 24 hours to eliminate the internal stress of the blanks; sintering: the blanks obtained in step (2) are placed in a fully automatic PTFE sintering furnace. In the process, sintering is carried out according to the set heating, constant temperature, cooling and other procedures; turning or rotary cutting: the blank obtained in step (3) is placed in the oven for preheating, so that the internal and external temperatures of the blank are consistent. Then, a special mandrel with trapezoidal teeth on the surface is pressed or pulled into the center hole of the blank. The blank is hoisted on a high-precision CNC lathe or rotary cutting machine by a crane. The blank is turned or rotary cut and rolled up according to the set film thickness by YG type carbide knife; soaking micropores: the film obtained in step (4) is placed in a special cleaning equipment, first soaked in a heated water tank and then cleaned in a room temperature water tank. The film is unwound, soaked, cleaned, dried and rolled up at a speed of 0.5 meters to 1 meter per minute to obtain a polytetrafluoroethylene three-dimensional microporous membrane.

[0004] Therefore, the existing technology has the following problems: the method relies on water-soluble powder to form pores, which easily leads to uneven pore size distribution and difficulty in precise control; the method adopts intermittent processing of molding, sintering and turning, which easily leads to low production efficiency and limited film thickness uniformity; the method relies on high-temperature sintering and water washing processes, which easily leads to changes in film structure and shrinkage deformation, affecting dimensional stability and consistency of permeability. Summary of the Invention

[0005] To address this, the present invention provides a production control system for microporous filter membranes based on polytetrafluoroethylene (PTFE). This system overcomes the problem in the prior art where the lack of a self-evaluation mechanism and reverse optimization path in the process control strategy leads to difficulties in ensuring the consistency of the filter membrane microporous structure and the stability of its permeation performance by constructing an adaptive closed-loop control architecture based on the fusion of multi-level quality characteristics.

[0006] To achieve the above objectives, the present invention provides a production control system for microporous filter membranes based on polytetrafluoroethylene, comprising: The acquisition module is used to acquire the mass ratio of each component in the mixing stage, the baseband thickness at each preset monitoring point in the calendering stage based on the preset roller spacing threshold, the thickness variation coefficient of the filter membrane in the longitudinal stretching stage based on the preset stretching ratio, the fibril surface density, and the membrane air permeability in the transverse stretching stage based on the preset heating power. The adjustment module is used to determine whether the calendering is abnormal based on the degree of deviation of the base strip thickness, and to adjust the preset stretch ratio and the preset roll gap threshold based on the determination result of the calendering abnormality. The correction module is used to correct the preset heating power according to the quality deviation and the preset power correction coefficient, wherein the quality deviation is determined based on the thickness variation coefficient and the fibril areal density obtained again after adjusting the preset stretch ratio and the preset roller spacing threshold. The update module is used to update the preset power correction coefficient and the mass ratio according to the lateral adjustment mass index and the lateral correction mass index, wherein the lateral adjustment mass index is determined based on the membrane air permeability obtained after adjusting the preset stretch ratio and the preset roller spacing threshold, and the lateral correction mass index is determined based on the membrane air permeability obtained after correcting the preset heating power. The early warning module is used to issue early warnings based on the deviation of the lateral adjustment mass index and the lateral correction mass index after updating the preset power correction coefficient and the mass ratio.

[0007] Furthermore, the adjustment module includes: A statistical calculation unit is used to determine the thickness range and the thickness mean based on the baseband thickness of each preset monitoring point, and to calculate the thickness deviation ratio based on the baseband thickness and the thickness mean. An anomaly determination unit is used to determine whether the rolling process is abnormal based on the threshold comparison result of the thickness range, and to determine the dominant direction of the anomaly based on the thickness deviation ratio. The calendering adjustment unit is used to adjust the preset stretching ratio and the preset roll spacing threshold based on the determination result of calendering anomaly, according to the thickness range, the thickness deviation ratio, and the dominant direction of the anomaly.

[0008] Furthermore, the anomaly determination unit includes: The range determination subunit is used to determine rolling anomalies based on the threshold comparison results of the thickness range, and to determine the range deviation degree based on the degree of deviation of the thickness range from its threshold. An orientation identification subunit is used to determine the dominant orientation of the anomaly based on a threshold comparison result of the thickness deviation ratio.

[0009] Furthermore, the rolling adjustment unit includes: A stretching adjustment subunit is used to reduce the preset stretching ratio based on the determination result of the rolling anomaly and the range deviation. A roller pitch adjustment subunit is used to adjust the preset roller pitch threshold according to the abnormal dominant direction, the average thickness, and the baseband thickness.

[0010] Furthermore, the correction module includes: A distance determination unit is used to determine the quality deviation based on the spatial distance between the current longitudinal vector and a preset longitudinal vector, wherein the current longitudinal vector is determined based on the thickness variation coefficient and the fibril surface density obtained after adjusting the preset stretch ratio and the preset roll gap threshold. A longitudinal correction unit is used to correct the preset heating power based on the threshold comparison result of the quality deviation and the preset power correction coefficient.

[0011] Furthermore, the longitudinal correction unit includes: The deviation determination subunit is used to determine longitudinal tension abnormality when the quality deviation is greater than a preset deviation threshold. A longitudinal correction subunit, connected to the deviation determination subunit, is used to increase the preset heating power based on the determination result of the longitudinal tension anomaly, according to a power adjustment amount, wherein the power adjustment amount is determined based on the mass deviation degree and the preset power correction coefficient.

[0012] Furthermore, the update module includes: An index determination unit is used to determine the lateral adjustment quality index and the lateral correction quality index based on the dispersion and central tendency of the membrane air permeability obtained after adjusting the preset stretch ratio and the preset roller spacing threshold and the membrane air permeability obtained after correcting the preset heating power, respectively. The trend determination unit is used to determine the adjustment quality slope and the correction quality slope respectively based on the changes of the lateral adjustment quality index and the lateral correction quality index within a preset monitoring batch. The root cause identification unit is used to determine the root cause of the deviation based on the threshold comparison results of the lateral adjustment quality index, the lateral correction quality index, the adjustment quality slope, and the correction quality slope, respectively. A lateral update unit is used to update the preset power correction coefficient and the quality ratio based on the root cause of the deviation, according to the lateral adjustment quality index, the lateral correction quality index, the adjustment quality slope, and the correction quality slope.

[0013] Furthermore, the root cause identification unit includes: The mixing identification subunit is used to determine that the root cause of the deviation is the mixing stage when the lateral adjustment quality index is greater than a preset adjustment index threshold, the lateral correction quality index is greater than a preset correction index threshold, and the adjustment quality slope and the correction quality slope are both greater than a preset slope threshold. A lateral identification subunit is used to determine that the root cause of the deviation is the lateral stretching stage when the lateral adjustment quality index is greater than a preset adjustment index threshold, the lateral correction quality index is greater than a preset correction index threshold, and the adjustment quality slope is greater than a preset slope threshold or the correction quality slope is greater than a preset slope threshold. The over-adjustment identification subunit is used to determine that the root cause of the deviation is a correction over-adjustment type when the lateral adjustment quality index is less than or equal to a preset adjustment index threshold, the lateral correction quality index is greater than a preset correction index threshold, the adjustment quality slope is less than a preset slope threshold, and the correction quality slope is greater than a preset slope threshold.

[0014] Furthermore, the horizontal update unit includes: A mixing update subunit is used to update the mass ratio according to the mass update amount when the root cause of the deviation is the mixing stage, wherein the mass update amount is determined based on the lateral correction mass index and the preset correction index threshold. A lateral update subunit is used to update the preset power correction coefficient according to the coefficient update amount when the root cause of the deviation is the lateral stretching stage, wherein the coefficient update amount is determined based on the lateral adjustment quality index and the preset adjustment index threshold. An over-adjustment update subunit is used to update the preset power correction coefficient according to the coefficient reverse update amount when the root cause of the deviation is the corrected over-adjustment type, wherein the coefficient reverse update amount is determined based on the lateral correction quality index and the preset correction index threshold.

[0015] Furthermore, the early warning module includes: The fluctuation calculation unit is used to calculate the standard deviation of the horizontal adjustment quality index and the standard deviation of the horizontal correction quality index after updating the preset power correction coefficient and the quality ratio within a preset monitoring period, so as to obtain the adjustment fluctuation value and the correction fluctuation value respectively. The deviation calculation unit is used to determine the adjustment deviation and the correction deviation based on the threshold comparison results of the adjustment fluctuation value and the correction fluctuation value, respectively, according to the degree of deviation of the adjustment mean and the correction mean from their respective thresholds, wherein the adjustment mean and the correction mean are determined based on the horizontal adjustment quality index and the horizontal correction quality index, respectively. An early warning unit is used to issue an early warning based on the threshold comparison results of the adjustment deviation and the correction deviation, respectively.

[0016] Compared with existing technologies, the beneficial effects of this invention are as follows: It utilizes calendering thickness deviation as a criterion for judging the stability of preceding processes. By adjusting the stretching ratio, it actively compensates for baseband unevenness, pre-correcting calendering deviation before longitudinal extension. Furthermore, it introduces a combination of thickness variation coefficient and fibril areal density to construct a quality deviation degree, thereby correcting the transverse stretching heating power and deeply coupling the development degree of the microfibril network with the heat setting process. Based on this, adjustment and correction indices are constructed according to the measured transverse air permeability. Through threshold comparison and stability analysis of these two indices, the power correction coefficient and mixing ratio are optimized in reverse, achieving closed-loop tuning from downstream performance to upstream formulation parameters. Furthermore, it triggers graded early warning based on the degree of deviation of the two indices, effectively overcoming the problems of poor uniformity of the microporous structure of the finished filter membrane, large fluctuations in air permeability, and difficulty in guaranteeing yield in large-scale production caused by open-loop decoupling of process parameters and lack of feedforward adaptive adjustment.

[0017] Furthermore, by statistically analyzing the baseband thickness to calculate the thickness range and mean, the transverse fluctuation amplitude and overall level of the baseband can be quantified. Based on this, a thickness deviation ratio is constructed. This ratio, by comparing the absolute value of the maximum thickness deviation from the mean with the absolute value of the minimum thickness deviation from the mean, effectively reveals the asymmetric characteristics of the baseband thickness distribution. The thickness range is compared with a preset threshold to determine whether the calendering is abnormal. The range, as a core statistical measure characterizing the degree of data dispersion, directly reflects the transverse uniformity of the baseband thickness. When the range exceeds the threshold, it indicates that the baseband thickness consistency has deteriorated to the point requiring intervention. Based on the confirmed calendering anomaly, the preset stretching ratio and preset roll spacing threshold are adjusted specifically according to the dominant direction of the anomaly indicated by the thickness deviation ratio. This proactively corrects the transverse thickness deviation that occurs during calendering within this process, preventing it from being inherited as an initial condition by subsequent processes.

[0018] Furthermore, by using the thickness range as the core statistic describing the dispersion of data, when the range exceeds a preset threshold, it indicates that the difference between the maximum and minimum values ​​of the baseband thickness has exceeded the normal process fluctuation range. This implies that there may be abnormal factors such as uneven roll gap, roll parallelism deviation, or mechanical vibration during the calendering process. Monitoring whether the range exceeds the limit can effectively capture the key characteristic of uncontrolled transverse uniformity in the calendering process. By defining the thickness deviation ratio as the ratio of the absolute value of the maximum thickness deviation from the mean to the absolute value of the minimum thickness deviation from the mean, when this ratio is greater than a preset ratio threshold, it indicates that the maximum thickness deviation is significantly greater than the minimum thickness deviation, that is, there is a significant positive deviation in the baseband thickness on one side, and the direction of the abnormality is determined to be dominated by the maximum value. When the thickness deviation ratio is less than the preset ratio threshold, it indicates that the minimum thickness deviation is significantly greater than the maximum thickness deviation, that is, there is a significant negative deviation in the baseband thickness on the other side, and the direction of the abnormality is determined to be dominated by the minimum value. This can effectively characterize the asymmetric characteristics of the baseband thickness distribution. At the same time, introducing a preset ratio threshold as a judgment benchmark can quantitatively distinguish between random symmetric fluctuations and systematic biases and identify the main source direction of the abnormality.

[0019] Furthermore, the severity of lateral fluctuations in base strip thickness was quantified by the range deviation. A larger range deviation indicates a more significant deterioration in the uniformity of base strip thickness. The stretching ratio, as a core parameter for subsequent longitudinal stretching processes, directly affects the degree of deformation of the base strip in the stretching direction. By introducing a preset sensitivity coefficient to establish a linear relationship between the range deviation and the stretching ratio adjustment, the reduction in the stretching ratio is matched to the severity of base strip thickness fluctuations. This allows for the reduction of longitudinal stretching deformation without altering the base strip's thickness distribution, thereby mitigating the transmission and amplification effects of base strip thickness unevenness on subsequent processes. This proactively compensates for calendering deviations at the process parameter level. By reducing the preset roll gap threshold when the abnormal dominant direction is dominated by the maximum base strip thickness, the extrusion amount on that side is directly reduced by narrowing the calendering roll gap, suppressing the generation of excessively thick base strips at the source. When the abnormal dominant direction is dominated by the minimum base strip thickness, the preset roll gap threshold is increased. By expanding the calender roll gap, the extrusion amount on that side is increased, which makes up for the base strip too thin defect from the source. This makes the adjustment range of the roll gap match the degree of thickness deviation, and realizes the correction of the asymmetry of the base strip thickness distribution.

[0020] Furthermore, the overall deviation between the current longitudinal stretching state and the ideal process state is quantified through the quality deviation degree. When the quality deviation degree exceeds the preset deviation threshold, it indicates that a significant anomaly has occurred in the longitudinal stretching stage. At this point, the preset heating power is increased according to the power adjustment amount. The quality deviation degree reflects the degree of anomaly in the microstructure of the filter membrane after longitudinal stretching, while the preset power correction coefficient establishes a mapping relationship between longitudinal quality anomalies and transverse heating compensation. By multiplying the two and linking them with the current heating power, a linear match between the compensation magnitude and the severity of the anomaly can be achieved. Increasing the heating power in the transverse stretching stage can change the thermal field distribution in the expansion zone, allowing the membrane to obtain a higher thermodynamic driving force during transverse stretching, thereby promoting the further unfolding of the fibrillary network and the homogenization of the microporous structure. This compensates for the microscopic defects caused by insufficient fibrillary development or uneven thickness in the longitudinal stretching stage, realizing proactive intervention of longitudinal stretching anomalies before entering the transverse stretching process.

[0021] Furthermore, by constructing a lateral adjustment quality index and a lateral correction quality index based on the ratio of the standard deviation to the mean of the air permeability width distribution, the quality status before and after compensation is characterized, respectively. Then, by performing linear regression fitting on the two indices within a preset monitoring batch, the adjustment quality slope and correction quality slope are obtained to determine the quality change trend. When both indices exceed the standard and the slopes increase simultaneously, it indicates that the problem originates from a fundamental deviation in the mixing stage, because uneven mixing will simultaneously affect the quality before and after compensation, and the trend continues to worsen. In this case, the quality ratio needs to be updated according to the degree of exceeding the correction index, correcting the formula from the source. When both indices exceed the standard but only one slope increases, it indicates that the problem originates from local process fluctuations in the lateral stretching stage, because these fluctuations can be partially eliminated by heating power compensation, but the compensation effect is unstable. In this case, the power correction coefficient needs to be updated according to the degree of exceeding the adjustment index, optimizing the compensation strategy. When the adjustment index is normal but the correction index exceeds the standard and the correction slope increases, it indicates that the compensation process has over-adjusted, that is, the compensation has introduced new disturbances. At this time, the power correction coefficient needs to be updated in reverse according to the degree of the correction index exceeding the standard, and the compensation intensity is adjusted back. This achieves accurate positioning and targeted correction of the mixing source deviation, the lateral stretching execution deviation, and the compensation over-adjustment deviation.

[0022] Furthermore, by using volatility assessment as a prerequisite, false alarms caused by short-term random fluctuations can be effectively filtered out, ensuring that the degree of deviation from the mean is only assessed when the quality index is in a statistically stable state. Based on this, an alarm is issued when the adjusted or corrected deviation exceeds a preset deviation threshold, accurately identifying a systematic shift in the quality index that has reached a level requiring intervention. This achieves a two-level assessment of the quality status—first stability, then deviation—ensuring both the reliability of the alarm and a timely response to systematic quality drift, avoiding the problems of volatility masking the true shift or frequent false alarms causing alarm failure. Attached Figure Description

[0023] Figure 1 This is a schematic diagram of the production control system for the polytetrafluoroethylene microporous filter membrane in this embodiment; Figure 2 This is the logic diagram for determining rolling anomalies in the range determination subunit of this embodiment; Figure 3 This is the logic diagram for determining longitudinal tensile abnormalities in the deviation determination subunit of this embodiment; Figure 4 This is a logic diagram showing the determination logic of the mixing stage as the root cause of the deviation in the mixing identification subunit of this embodiment. Detailed Implementation

[0024] To make the objectives and advantages of the present invention clearer, the present invention will be further described below with reference to embodiments; it should be understood that the specific embodiments described herein are merely for explaining the present invention and are not intended to limit the present invention.

[0025] Preferred embodiments of the present invention will now be described with reference to the accompanying drawings. Those skilled in the art should understand that these embodiments are merely illustrative of the technical principles of the present invention and are not intended to limit the scope of protection of the present invention.

[0026] Please see Figure 1 The diagram shown is a schematic of the production control system for a polytetrafluoroethylene (PTFE) microporous filter membrane according to this embodiment. This embodiment provides a production control system for a polytetrafluoroethylene (PTFE) microporous filter membrane, including: The acquisition module is used to acquire the mass ratio of each component in the mixing stage, the baseband thickness at each preset monitoring point in the calendering stage based on the preset roller spacing threshold, the thickness variation coefficient of the filter membrane in the longitudinal stretching stage based on the preset stretching ratio, the fibril surface density, and the membrane air permeability in the transverse stretching stage based on the preset heating power. An adjustment module, connected to the acquisition module, is used to determine whether the calendering is abnormal based on the degree of deviation of the base strip thickness, and to adjust the preset stretching ratio and the preset roll spacing threshold based on the determination result of the calendering abnormality. A correction module, which is connected to the acquisition module and the adjustment module respectively, is used to correct the preset heating power according to the quality deviation and the preset power correction coefficient, wherein the quality deviation is determined based on the thickness variation coefficient and the fibril surface density re-acquired after adjusting the preset stretch ratio and the preset roller spacing threshold. An update module, which is connected to the acquisition module, the adjustment module, and the correction module respectively, is used to update the preset power correction coefficient and the mass ratio according to the lateral adjustment mass index and the lateral correction mass index. The lateral adjustment mass index is determined based on the membrane air permeability obtained after adjusting the preset stretch ratio and the preset roller spacing threshold, and the lateral correction mass index is determined based on the membrane air permeability obtained after correcting the preset heating power. An early warning module, connected to the update module, is used to issue an early warning based on the deviation of the lateral adjustment mass index and the lateral correction mass index, which are re-determined after updating the preset power correction coefficient and the mass ratio.

[0027] In this embodiment, the components in the mixing stage include lubricant and polytetrafluoroethylene (PTFE) resin. The production control system based on PTFE microporous filter membranes is applied to the industrial continuous production environment of PTFE microporous filter membranes. It is suitable for the large-scale production of high-end microporous filter membranes for smartphones, smartwatches, wearable devices, and precision electronic components, which have high requirements for filter membrane pore size uniformity and permeability. The microporous membranes need to meet the dual stringent requirements of IP68 waterproof rating and low-frequency sound transmission while maintaining an ultra-thin thickness of 10-15 micrometers. Its micropore size must be precisely controlled within the golden range of 0.1-1 micrometers and have a highly consistent lateral distribution. By real-time acquisition of process parameters for key processes such as mixing, calendering, longitudinal stretching, and transverse stretching, dynamic monitoring and closed-loop control of the entire production process are achieved. The mass ratio refers to the mixing mass ratio of lubricant and polytetrafluoroethylene resin powder in the mixing process. This ratio can be determined by real-time measurement of the feed mass of each material in the mixing equipment using an online weighing system such as a high-precision electronic balance or mass flow meter, and simultaneously calculating and outputting the ratio value. The base strip thickness is obtained by measuring the thickness of the calendered base strip (i.e., the pre-calendered film blank) along the width of the film, from the left edge to the right edge, at multiple preset monitoring points. These preset monitoring points are located every 10 cm along the width direction and can be obtained using non-contact excitation. A light thickness gauge or a capacitive thickness sensor scans and collects thickness data in real time. The thickness variation coefficient represents the degree of fluctuation in the longitudinal thickness of the filter membrane. The calculation formula is the standard deviation of the thickness divided by the average thickness. During the longitudinal stretching stage, a laser thickness measurement system can be used to collect thickness data at multiple points evenly distributed every 10 cm along the longitudinal length of the filter membrane (the stretching direction). The system automatically calculates the ratio of the standard deviation of these thickness data to the average thickness to obtain the variation coefficient. The fibrillation areal density refers to the micro-density per unit area of ​​the longitudinally stretched polytetrafluoroethylene semi-finished membrane. The number of fine fibers is a direct indicator of the degree of development and uniformity of the fibrillary network. Samples can be taken online or offline at the exit of the longitudinal stretching process. After surface treatment, the samples are observed under a scanning electron microscope. Multiple representative microstructure images are acquired at a fixed magnification of 5000-10000x. The images are binarized using image analysis software to identify and count the number of fibers. The fibrillary areal density is calculated by dividing the counted number of fibers by the actual area corresponding to the image. Membrane permeability characterizes the gas permeation performance of the filter membrane along its width after transverse stretching. This can be obtained using the differential pressure method. The membrane sample is sealed in a testing device, with a fixed pressure difference (typically 6.9-7 kPa) applied to one side and atmospheric pressure on the other. The pressure difference is measured in real time using a miniature differential pressure sensor, while a high-precision flow meter measures the gas flow rate through the membrane. The system automatically calculates the permeability (permeability / area / time) and generates permeability distribution data at multiple test points evenly spaced every 10 cm along the width, thus obtaining several membrane permeability values.

[0028] The preset roll gap threshold is the initial gap setting between the two rolls of the calender during the calendering stage. It depends on the target thickness of the final product, the preset stretch ratio, and the compression and springback characteristics of the calendered material. It is calculated using the formula: Preset roll gap threshold = Final product thickness × Longitudinal stretch ratio × Transverse stretch ratio × Compression and springback coefficient. It is usually set between 0.05 mm and 0.5 mm. In this embodiment, it is set to 0.12 mm, which ensures that the thickness of the base strip after calendering accurately reaches the final product thickness requirement after subsequent longitudinal and transverse stretching, ensuring the uniformity of the base strip thickness and reducing the adjustment burden of subsequent processes. The preset stretch ratio refers to the longitudinal elongation multiple of the film set in the longitudinal stretching stage. It depends on the final pore size and porosity requirements of the target product and the polytetrafluoroethylene material. The stretchability of the material is calculated by decomposing the required total stretching area multiple into the contribution ratios of the longitudinal and transverse directions, combined with the limit value of the linear deformation zone in the material's stress-strain curve. It is typically set between 4 and 6 times; in this embodiment, it is set to 5 times, which allows the PTFE particles to be fully stretched to form a uniform fibrillary network. The preset heating power refers to the heater output power set for each independent heating zone during the transverse stretching stage. First, the target temperature value for each heating zone in the transverse stretching is determined according to the process formula. Then, the residence time of the film through the heating zone at the rated speed, as well as the film's unit area mass and specific heat capacity, are measured to calculate the heat absorbed by the film to rise from the initial temperature to the target temperature. Finally, the total heat output required to maintain the target temperature is calculated through heat balance calculations. This includes heat loss from membrane absorption, air convection absorption, and cavity heat dissipation. The total heat value is compared to the heater's factory calibration curve to obtain the corresponding theoretical power value. Finally, a temperature field uniformity verification test is conducted on actual equipment. The theoretical power value is fine-tuned based on the actual temperature distribution measured by multi-point thermocouples to determine the reference power setting value for each heating zone, which is the preset heating power. Based on this, the preset heating power is usually set to automatically adjust within 60% to 80% of the rated power range for each heating zone. In this embodiment, it is set to automatically adjust within 60% to 80% of the rated power range for each heating zone, ensuring uniform temperature field distribution in the expansion zone. The preset power correction coefficient is a mapping coefficient used to convert the quality deviation into a heating power adjustment amount. First... Data from multiple batches was collected from the historical production database, recording the quality deviation, the corresponding adjustment of lateral stretching heating power, and the change in the air permeability of the final product. Linear regression analysis was performed on these data to calculate the regression slope between the quality deviation and the required heating power adjustment, thus initially determining the mapping relationship. Simultaneously, a response surface methodology experiment was designed, setting different longitudinal quality deviation levels on the equipment and applying different power correction coefficients. The response sensitivity under different coefficients was verified by measuring the compensation effect of the air permeability of the final product. The regression results of historical data were comprehensively compared with the experimental verification results, and the value that resulted in the most stable compensation effect and the best response linearity was selected as the preset power correction coefficient. Based on this, the preset power correction coefficient is usually set between 0.5 and 1.The value is between 5 and 1.0 in this embodiment to ensure the accuracy and stability of downstream compensation.

[0029] By using calendering thickness deviation as a criterion for judging the stability of preceding processes, and actively compensating for baseband unevenness by adjusting the stretching ratio, the calendering deviation is pre-corrected before longitudinal stretching. Furthermore, a quality deviation degree is constructed by introducing the combined characteristics of thickness variation coefficient and fibril areal density, thereby correcting the transverse stretching heating power and deeply coupling the development degree of the microfibril network with the heat setting process. Based on this, adjustment and correction indices are constructed according to the measured transverse air permeability. Through threshold comparison and stability analysis of these two indices, the power correction coefficient and mixing ratio are optimized in reverse, achieving closed-loop tuning from downstream performance to upstream formulation parameters. A graded early warning system is triggered based on the degree of deviation of the two indices, effectively overcoming the problems of poor uniformity of the microporous structure of the finished filter membrane, large fluctuations in air permeability, and difficulty in guaranteeing yield in large-scale production caused by open-loop decoupling of process parameters and lack of feedforward adaptive adjustment.

[0030] Specifically, the adjustment module includes: The statistical calculation unit is used to calculate the range and average value of the baseband thickness at each of the preset monitoring points to obtain the thickness range and average value, and to calculate the thickness deviation ratio based on the maximum and minimum values ​​of the baseband thickness and the average value, where H1=|H max -H A | / |H min -H A | where H1 is the thickness deviation ratio, H max H is the maximum baseband thickness at each preset monitoring point. A H is the average baseband thickness of all preset monitoring points. min It is the minimum baseband thickness at each preset monitoring point; An anomaly determination unit, which is connected to the statistical calculation unit, is used to determine whether the rolling process is abnormal based on the threshold comparison result of the thickness range, and to determine the dominant direction of the anomaly based on the thickness deviation ratio. The calendering adjustment unit is connected to the statistical calculation unit and the anomaly determination unit respectively, and is used to adjust the preset stretching ratio and the preset roll spacing threshold according to the determination result of calendering anomaly, the thickness range, the thickness deviation ratio and the dominant direction of the anomaly.

[0031] By statistically analyzing the baseband thickness to calculate the thickness range and mean, the transverse fluctuation amplitude and overall level of the baseband can be quantified. Based on this, a thickness deviation ratio is further constructed. This ratio, by comparing the absolute value of the maximum thickness deviation from the mean with the absolute value of the minimum thickness deviation from the mean, effectively reveals the asymmetric characteristics of the baseband thickness distribution. The thickness range is compared with a preset threshold to determine whether the calendering is abnormal. The range, as a core statistical measure characterizing the degree of data dispersion, directly reflects the transverse uniformity of the baseband thickness. When the range exceeds the threshold, it indicates that the baseband thickness consistency has deteriorated to the point requiring intervention. Based on the confirmed calendering anomaly, the preset stretching ratio and preset roll spacing threshold are adjusted specifically according to the dominant direction of the anomaly indicated by the thickness deviation ratio. This proactively corrects the transverse thickness deviation that occurs during calendering within this process, preventing it from being inherited as an initial condition by subsequent processes.

[0032] Please see Figure 2 As shown, this is the logic diagram for determining rolling anomalies in the range determination subunit of this embodiment. In this embodiment, the anomaly determination unit includes: The range determination subunit is used to determine the rolling abnormality when the thickness range is greater than the preset range threshold, and to determine the range deviation degree based on the relative deviation between the thickness range and the preset range threshold. A direction identification subunit, connected to the range determination subunit, is used to determine that the abnormal dominant direction is dominated by the maximum value of the baseband thickness when the thickness deviation ratio is greater than a preset ratio threshold, and to determine that the abnormal dominant direction is dominated by the minimum value of the baseband thickness when the thickness deviation ratio is less than the preset ratio threshold.

[0033] In this embodiment, the preset ratio threshold is 1. When the thickness deviation ratio is equal to 1, it indicates that the base strip thickness is symmetrically distributed on both sides of the mean. At this time, although there is lateral thickness fluctuation, there is no obvious directional characteristic, and it is only necessary to trigger the conventional calendering parameter adjustment through the range determination subunit. When the thickness deviation ratio is significantly greater than 1, it means that the maximum thickness deviation from the mean is much greater than the minimum thickness deviation. The base strip shows an obvious "thin in the middle, thick at the edges" or local thickness increase pattern. At this time, the direction recognition subunit will determine the abnormal dominant direction as the maximum value dominant, indicating that there may be insufficient pressure in the middle of the calendering roll or the gap increase caused by roll surface wear. Conversely, when the thickness deviation ratio is significantly less than 1, it indicates that the minimum thickness deviation is dominant. The base strip shows a "thick in the middle, thin at the edges" or local thickness decrease pattern. The direction recognition subunit will determine the minimum value dominant accordingly. This usually corresponds to insufficient pressure at both ends of the calendering roll or excessive tension on the edge of the base strip.

[0034] The preset range threshold is a benchmark value used to determine whether the lateral fluctuation of the base strip thickness during the calendering stage exceeds the allowable range. It depends on the mechanical precision of the calendering equipment, the parallelism of the rollers, and the allowable fluctuation range of the target base strip thickness. It is determined by statistically analyzing the normal fluctuation distribution of the base strip thickness range under stable operating conditions and taking the upper limit of the normal fluctuation range. It is usually set between 0.02 mm and 0.05 mm. In this embodiment, it is set to 0.03 mm, which can effectively identify abnormal fluctuations in base strip thickness caused by uneven gaps between calendering rollers or mechanical vibration.

[0035] By using thickness range as the core statistic to describe the degree of data dispersion, when the range exceeds a preset threshold, it indicates that the difference between the maximum and minimum values ​​of the baseband thickness has exceeded the normal process fluctuation range. This implies that there may be abnormal factors such as uneven roll gap, roll parallelism deviation, or mechanical vibration during the calendering process. Monitoring whether the range exceeds the limit can effectively capture the key characteristic of uncontrolled transverse uniformity in the calendering process. The thickness deviation ratio is defined as the ratio of the absolute value of the maximum thickness deviation from the mean to the absolute value of the minimum thickness deviation. When this ratio is greater than a preset ratio threshold, it indicates that the maximum thickness deviation is significantly greater than the minimum thickness deviation, meaning that there is a significant positive deviation in the baseband thickness on one side, and the direction of the anomaly is determined to be dominated by the maximum value. When the thickness deviation ratio is less than the preset ratio threshold, it indicates that the minimum thickness deviation is significantly greater than the maximum thickness deviation, meaning that there is a significant negative deviation in the baseband thickness on the other side, and the direction of the anomaly is determined to be dominated by the minimum value. This effectively characterizes the asymmetric characteristics of the baseband thickness distribution. Simultaneously, introducing a preset ratio threshold as a judgment benchmark can quantitatively distinguish between random symmetrical fluctuations and systematic biases and identify the main source direction of anomalies.

[0036] Specifically, the rolling adjustment unit includes: A stretching adjustment subunit is used to reduce the preset stretching ratio based on the determination result of rolling anomalies and the range deviation, where L'=L×(1-α×J) P ), where L' is the adjusted preset stretch ratio, L is the original preset stretch ratio, α is the preset sensitivity coefficient, and J P It is the deviation from the range; A roll gap adjustment subunit is used to reduce the preset roll gap threshold based on the average thickness and the maximum thickness of the base strip when the abnormal dominant direction is dominated by the maximum value of the base strip thickness, where K'=K-γ×(H max -H A), where K' is the reduced preset roll gap threshold, K is the original preset roll gap threshold, γ is a preset adjustment coefficient, and when the abnormal dominant direction is dominated by the minimum value of the baseband thickness, the preset roll gap threshold is increased according to the average thickness and the minimum value of the baseband thickness, where K''=K+γ×(H A -H min ), where K'' is the increased preset roller spacing threshold.

[0037] The preset sensitivity coefficient is a proportional factor used to convert the range deviation into the stretching ratio adjustment range. First, during the equipment commissioning phase, gradient response experiments are conducted. While keeping other process parameters constant, different range deviation levels are set, and the corresponding stretching ratio adjustments and changes in the coefficient of variation of the filter membrane thickness after adjustment are recorded. Then, the relationship curve between the range deviation and the optimal stretching ratio adjustment is plotted. A linear regression analysis is used to fit the proportional relationship between the two, and finally, the slope of the regression equation is taken as the preset sensitivity coefficient. Based on this, the preset sensitivity coefficient is usually set between 0.2 and 0.5; in this embodiment, it is set to 0.3, which can achieve a linear match between the stretching ratio adjustment range and the severity of calendering anomalies. The adjustment coefficient is a proportional factor used to convert the baseband thickness deviation into the roll gap correction amount. First, a roll gap adjustment step experiment is performed. Different roll gap step values ​​are set under the no-load state of the equipment, and the corresponding baseband thickness changes are measured to establish a transfer function between the roll gap adjustment amount and the baseband thickness change amount. Then, under normal production conditions, multiple sets of baseband thickness deviation data and corresponding optimal roll gap correction amounts are collected. The proportional relationship between the two is determined through regression analysis. Finally, the average value of the regression results of the no-load transfer function and the production data is taken as the preset adjustment coefficient. Based on this, the preset adjustment coefficient is usually set between 0.1 and 0.3. In this embodiment, it is set to 0.2 to ensure that the roll gap correction range matches the thickness deviation degree.

[0038] The severity of lateral fluctuations in baseband thickness was quantified by the range deviation. A larger range deviation indicates a more significant deterioration in baseband thickness uniformity. The stretching ratio, as a core parameter for subsequent longitudinal stretching processes, directly affects the degree of deformation of the baseband in the stretching direction. By introducing a preset sensitivity coefficient to establish a linear relationship between the range deviation and the stretching ratio adjustment, the reduction in the stretching ratio is matched to the severity of baseband thickness fluctuations. This allows for the reduction of longitudinal stretching deformation without altering the baseband's thickness distribution, thereby mitigating the transmission and amplification effects of baseband thickness unevenness on subsequent processes. This proactively compensates for calendering deviations at the process parameter level. By reducing the preset roll gap threshold when the abnormal dominant direction is dominated by the maximum baseband thickness, the extrusion amount on that side is directly reduced by narrowing the calendering roll gap, suppressing excessive baseband thickness at its source. When the abnormal dominant direction is dominated by the minimum base strip thickness, the preset roll gap threshold is increased. By expanding the calender roll gap, the extrusion amount on that side is increased, which makes up for the base strip too thin defect from the source. This makes the adjustment range of the roll gap match the degree of thickness deviation, and realizes the correction of the asymmetry of the base strip thickness distribution.

[0039] Specifically, the correction module includes: A distance determination unit is used to calculate the Mahalanobis distance between the current longitudinal vector and the preset longitudinal vector to obtain the quality deviation, wherein the current longitudinal vector is determined based on the thickness variation coefficient and the fibril surface density obtained after adjusting the preset stretch ratio and the preset roll gap threshold. A longitudinal correction unit, connected to the distance determination unit, is used to correct the preset heating power based on the threshold comparison result of the quality deviation and the preset power correction coefficient.

[0040] Please see Figure 3 As shown, this is the logic diagram for determining longitudinal tension abnormalities in the deviation determination subunit of this embodiment. In this embodiment, the longitudinal correction unit includes: The deviation determination subunit is used to determine longitudinal tension abnormality when the quality deviation is greater than a preset deviation threshold. A longitudinal correction subunit, connected to the deviation determination subunit, is used to increase the preset heating power based on the determination result of the longitudinal tension anomaly, according to the power adjustment amount. The power adjustment amount is determined based on the mass deviation degree and the preset power correction coefficient, where P = k × Z × P0, where P is the power adjustment amount, k is the preset power correction coefficient, Z is the mass deviation degree, and P0 is the preset heating power before correction.

[0041] In this embodiment, the preset longitudinal vector refers to a two-dimensional benchmark vector constructed based on historical best production batch data. Its two dimensions are the thickness variation coefficient and the fibrillation areal density, respectively, which are used to characterize the combination of quality characteristics of the filter membrane when it is in an ideal process state during the longitudinal stretching stage. First, data from multiple batches that have been continuously and stably operating and whose final product air permeability qualification rate is in the top 5% can be extracted from the production database. The average values ​​of the thickness variation coefficient and fibrillation areal density are calculated for each batch. Then, the two-dimensional data of these best batches are statistically analyzed, and the mean value at a 95% confidence level is taken as the coordinate value of the preset longitudinal vector, that is, the preset longitudinal vector = (thickness variation coefficient benchmark value, fibrillation areal density benchmark value). Under the current normal production state, the probability that the measured values ​​of the thickness variation coefficient and fibrillation areal density fall within the confidence ellipse centered on this vector is not less than 95%. At the same time, this vector will be dynamically updated every quarter based on the latest production data to ensure that it can always accurately reflect the optimal quality state under the current process conditions.

[0042] The preset deviation threshold is a critical value for determining whether an abnormality has occurred in the longitudinal stretching process. By collecting data from multiple batches of continuously and stably produced products that are ultimately qualified, the quality deviation of each batch is calculated and its probability distribution curve is plotted. The quantile corresponding to the 95% confidence level is taken as the benchmark value. At the same time, the benchmark value is corrected and determined in combination with the specifications of the air permeability of the final product. Based on this, the preset deviation threshold is usually set between 0.8 and 1.2. In this embodiment, it is set to 1.0 to ensure a balance between the sensitivity and specificity of anomaly identification.

[0043] The quality deviation is used to quantify the overall deviation between the current longitudinal stretching state and the ideal process state. When the quality deviation exceeds a preset deviation threshold, it indicates a significant anomaly in the longitudinal stretching stage. At this point, the preset heating power is increased according to the power adjustment amount. The quality deviation reflects the degree of anomaly in the microstructure of the filter membrane after longitudinal stretching, while the preset power correction coefficient establishes a mapping relationship between longitudinal quality anomalies and transverse heating compensation. By multiplying the two and linking them with the current heating power, a linear match between the compensation magnitude and the severity of the anomaly can be achieved. Increasing the heating power in the transverse stretching stage can change the thermal field distribution in the expansion zone, allowing the membrane to obtain a higher thermodynamic driving force during transverse stretching. This promotes further unfolding of the fibril network and homogenization of the microporous structure, compensating for microscopic defects caused by insufficient fibril development or uneven thickness in the longitudinal stretching stage. This achieves proactive intervention in longitudinal stretching anomalies before entering the transverse stretching process.

[0044] Specifically, the update module includes: An index determination unit is used to calculate the ratio of the adjusted air permeability fluctuation value to the adjusted air permeability mean value to obtain the lateral adjustment quality index, and to calculate the ratio of the corrected air permeability fluctuation value to the corrected air permeability mean value to obtain the lateral corrected quality index. The adjusted air permeability fluctuation value and the adjusted air permeability mean value are determined based on the standard deviation and average value of the membrane air permeability along the width direction after adjusting the preset stretch ratio and the preset roller spacing threshold, respectively. The ratio of the corrected air permeability fluctuation value and the corrected air permeability mean value is determined based on the standard deviation and average value of the membrane air permeability along the width direction after correcting the preset heating power, respectively. A trend determination unit, connected to the index determination unit, is used to perform linear regression fitting on the horizontal adjustment quality index and the horizontal correction quality index within a preset monitoring batch, respectively, to obtain the adjustment quality slope and the correction quality slope, respectively. The root cause identification unit is connected to the index determination unit and the trend determination unit respectively, and is used to determine the root cause of the deviation based on the threshold comparison results of the lateral adjustment quality index, the lateral correction quality index, the adjustment quality slope and the correction quality slope. A lateral update unit, connected to the root cause identification unit, is used to update the preset power correction coefficient and the quality ratio based on the deviation root cause, according to the lateral adjustment quality index, the lateral correction quality index, the adjustment quality slope, and the correction quality slope.

[0045] Please see Figure 4 As shown, this is the determination logic diagram of the mixing identification subunit in this embodiment, which determines that the root cause of the deviation is the mixing stage. In this embodiment, the root cause identification unit includes: The mixing identification subunit is used to determine that the root cause of the deviation is the mixing stage when the lateral adjustment quality index is greater than a preset adjustment index threshold, the lateral correction quality index is greater than a preset correction index threshold, and the adjustment quality slope and the correction quality slope are both greater than a preset slope threshold. A lateral identification subunit is used to determine that the root cause of the deviation is the lateral stretching stage when the lateral adjustment quality index is greater than a preset adjustment index threshold, the lateral correction quality index is greater than a preset correction index threshold, and the adjustment quality slope is greater than a preset slope threshold or the correction quality slope is greater than a preset slope threshold. An over-adjustment identification subunit is used to determine that the root cause of the deviation is a correction over-adjustment type when the lateral adjustment quality index is less than or equal to a preset adjustment index threshold, the lateral correction quality index is greater than a preset correction index threshold, the adjustment quality slope is less than a preset slope threshold, and the correction quality slope is greater than a preset slope threshold. Otherwise, manual intervention is required. Technicians will conduct a comprehensive assessment by combining real-time process parameters with historical production data. If necessary, the production line will be suspended and an anomaly tracing procedure will be initiated. Raw material batches, equipment operation logs, and environmental monitoring records will be checked one by one to determine whether there are complex deviation factors or new anomaly patterns not covered by the model. At the same time, relevant data samples will be fed back to the algorithm optimization module for subsequent iterative updates of identification rules and adaptive adjustments to threshold parameters.

[0046] Specifically, the horizontal update unit includes: A mixing update subunit is used to update the mass ratio according to a mass update amount when the root cause of the deviation is the mixing stage. The mass update amount is determined based on the lateral correction mass index and the preset correction index threshold. ; Where m1 is the quality update amount, r is the preset proportional adjustment step size, and N m It is a horizontally corrected quality index, N m0 It is a preset correction index threshold; A lateral update subunit is used to update the preset power correction coefficient according to a coefficient update amount when the root cause of the deviation is the lateral stretching stage. The coefficient update amount is determined based on the lateral adjustment quality index and the preset adjustment index threshold, where k1 = c × (N a -N a0 ), where k1 is the coefficient update amount, c is the preset coefficient adjustment step size, and N a It is a horizontal adjustment of the quality index, N a0 It is a preset adjustment index threshold; An over-adjustment update subunit is used to update the preset power correction coefficient according to the coefficient reverse update amount when the root cause of the deviation is the corrected over-adjustment type. The coefficient reverse update amount is determined based on the lateral correction quality index and the preset correction index threshold, where k2 = -e×(N) m -N m0 ), where e is the preset reverse update step size.

[0047] The preset monitoring batch size is the sample window size used to calculate the changing trends of the horizontal adjustment quality index and the horizontal correction quality index. It is determined by analyzing the autocorrelation function of the quality index in historical production data to identify its main fluctuation period. The standard deviation of the slope estimate under different window lengths is calculated using the moving range method. The window length that minimizes the coefficient of variation of the slope estimate is selected as the benchmark, while also considering the balance between production cycle time and response timeliness. Based on this, its value is typically set between 5 and 15 batches; in this embodiment, it is set to 10 batches to effectively filter out random fluctuations in single batches. The preset adjustment index threshold is the critical value used to determine whether the horizontal adjustment quality index exceeds the limit. By collecting historical production data on the horizontal adjustment quality index corresponding to qualified batches of final products, a probability distribution curve is plotted. The upper limit of the 95% confidence level is taken as the benchmark. Simultaneously, the benchmark value is corrected based on sensitivity analysis of the adjustment index and the risk of exceeding the air permeability standard. Therefore, its value is typically set between 1.2 and 1.5; in this embodiment, it is set to 1.3, which can promptly trigger anomaly identification when the adjustment index exceeds the normal fluctuation range. The preset correction index threshold is the critical value used to determine whether the horizontal adjustment quality index meets the standard. By analyzing the statistical distribution of the horizontal adjustment quality index in qualified batches of final products, a value that can guarantee above 95% is selected. The product's acceptable quantile is used as the benchmark value. This benchmark value is conservatively adjusted based on the customer's upper limit requirements for air permeability fluctuations. Therefore, its value is typically set between 1.1 and 1.3; in this embodiment, it is set to 1.2. This allows for accurate identification of a state where the quality remains unacceptable after compensation when the correction index exceeds the limit. The preset slope threshold is a critical value used to determine whether there is a continuous deterioration trend in the horizontal adjustment quality index and the horizontal correction quality index. By analyzing the maximum slope distribution of the quality index within the normal fluctuation range in historical data, the quantile corresponding to the 95% confidence level is taken as the benchmark value. This is combined with the average time required from the emergence of a trend to the quality exceeding the limit. The baseline value is corrected and determined. Based on this, its value is usually set between 0.03 and 0.08. In this embodiment, it is set to 0.05, which can realize the forward warning of potential runaway risk. The preset proportional adjustment step size is a proportional factor used to convert the degree of lateral correction of the quality index exceeding the standard into the quality ratio update amount. The change in lateral correction of the quality index corresponding to the change in unit mass ratio is determined by performing a mixing ratio gradient experiment. The gain value of the linear interval of the response curve is taken as the baseline. At the same time, the baseline value is rounded and corrected in combination with the minimum step accuracy of the adjustment mechanism. Based on this, its value is usually set between 0.01 and 0.05. In this embodiment, it is set to 0.02. This ensures that the mixing ratio correction range matches the quality deviation level. The preset coefficient adjustment step size is a proportional factor used to convert the degree of deviation of the lateral adjustment quality index into the power correction coefficient update amount. The change in the lateral adjustment quality index corresponding to a unit coefficient change is determined by performing a power correction coefficient gradient experiment. The gain value of the linear interval of the response curve is taken as a benchmark, and this benchmark value is corrected and determined in conjunction with the minimum adjustable resolution of the control system. Based on this, its value is usually set between 0.05 and 0.15. In this embodiment, it is set to 0.1 to ensure that the coefficient update range matches the quality deviation level. The preset reverse update step size... The length is a proportional factor used to convert the degree of lateral correction of the quality index exceeding the limit into the reverse update amount of the power correction coefficient. By analyzing the correspondence between the optimal correction amplitude after overshooting in historical data and the amount of overshoot in the correction index, the slope value obtained from regression analysis is taken as a benchmark. Simultaneously, this benchmark value is conservatively adjusted based on system stability margin requirements. Therefore, its value is typically set between 0.03 and 0.08; in this embodiment, it is set to 0.05, which enables a corresponding reverse adjustment of the power correction coefficient by 0.005 for every 0.1 overshoot in the correction index, effectively eliminating quality degradation caused by overcompensation and restoring the compensation strategy to a reasonable state.

[0048] By constructing a lateral adjustment quality index and a lateral correction quality index based on the ratio of the standard deviation to the mean of the air permeability width distribution, the quality status before and after compensation is characterized, respectively. Then, by performing linear regression fitting on the two indices within a preset monitoring batch, the adjustment quality slope and correction quality slope are obtained to determine the quality change trend. When both indices exceed the standard and the slopes increase simultaneously, it indicates that the problem originates from a fundamental deviation in the mixing stage, because uneven mixing will affect the quality before and after compensation simultaneously, and the trend continues to worsen. In this case, the quality ratio needs to be updated according to the degree of exceeding the correction index, correcting the formula from the source. When both indices exceed the standard but only one slope increases, it indicates that the problem originates from local process fluctuations in the lateral stretching stage, because these fluctuations can be partially eliminated by heating power compensation, but the compensation effect is unstable. In this case, the power correction coefficient needs to be updated according to the degree of exceeding the adjustment index, optimizing the compensation strategy. When the adjustment index is normal but the correction index exceeds the standard and the correction slope increases, it indicates that the compensation process has over-adjusted, that is, the compensation has introduced new disturbances. At this time, the power correction coefficient needs to be updated in reverse according to the degree of the correction index exceeding the standard, and the compensation intensity is adjusted back. This achieves accurate positioning and targeted correction of the mixing source deviation, the lateral stretching execution deviation, and the compensation over-adjustment deviation.

[0049] Specifically, the early warning module includes: The fluctuation calculation unit is used to calculate the standard deviation of the horizontal adjustment quality index and the standard deviation of the horizontal correction quality index after updating the preset power correction coefficient and the quality ratio within a preset monitoring period, so as to obtain the adjustment fluctuation value and the correction fluctuation value respectively. The deviation calculation unit is used to calculate the relative deviation between the adjustment mean and the preset adjustment index threshold when the adjustment fluctuation value and the correction fluctuation value are both less than the preset fluctuation threshold, so as to obtain the adjustment deviation, and to calculate the relative deviation between the correction mean and the preset correction index threshold, so as to obtain the correction deviation. The adjustment mean and the correction mean are determined based on the average value of the horizontal adjustment quality index and the average value of the horizontal correction quality index for the preset monitoring time, respectively. An early warning unit, which is connected to the fluctuation calculation unit and the deviation calculation unit respectively, is used to issue an early warning when the adjustment deviation is greater than a preset deviation threshold or the correction deviation is greater than a preset deviation threshold.

[0050] The preset monitoring duration is the length of the time window used to calculate the statistical characteristics of the horizontally adjusted quality index and the horizontally corrected quality index. It is determined by analyzing the autocorrelation function of the quality index in historical data to identify its main fluctuation period. The minimum sample size required for mean estimation is calculated using the central limit theorem. A duration that simultaneously includes at least two complete fluctuation periods and meets statistical stability requirements is selected as the benchmark. Based on this, its value is typically set between 4 and 12 hours; in this embodiment, it is set to 8 hours to effectively filter out short-term random fluctuations. The preset fluctuation threshold is the critical value used to determine whether the horizontally adjusted quality index and the horizontally corrected quality index are in a stable state. It is calculated by collecting the standard deviation data of the two quality indices during a period of continuous stable production and qualified final products, plotting their probability distribution curves, and taking the quantile corresponding to the 90% confidence level as the benchmark value. The benchmark value is adjusted and determined by combining the standard deviation and the sensitivity analysis of process fluctuation risk. Based on this, its value is usually set between 0.08 and 0.12. In this embodiment, it is set to 0.1, which can be judged as a stable state when the fluctuation of both indices is within the normal range. The preset deviation threshold is the critical value used to determine whether the adjustment deviation or correction deviation triggers the warning. By establishing a correlation model between the quality index and the air permeability, the maximum index deviation range allowed to ensure that the air permeability pass rate is above 95% is calculated. At the same time, the range is verified and corrected by combining the correspondence between the quality index exceeding the standard and the actual product non-compliance in multiple batches of historical data. Based on this, its value is usually set between 0.1 and 0.2. In this embodiment, it is set to 0.15, which can ensure that intervention is carried out before the quality index deviates significantly from the target value but before it causes the batch of products to be non-compliant.

[0051] By using volatility assessment as a prerequisite, false alarms caused by short-term random fluctuations can be effectively filtered out, ensuring that the degree of deviation from the mean is only assessed when the quality index is in a statistically stable state. Based on this, an alarm is issued when the adjusted or corrected deviation exceeds a preset deviation threshold, accurately identifying a systematic shift in the quality index that has reached a level requiring intervention. This achieves a two-stage assessment of the quality status—first stability, then deviation—ensuring both the reliability of the alarm and a timely response to systematic quality drift, avoiding the problems of volatility masking the true shift or frequent false alarms causing alarm failure.

[0052] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A production control system for microporous filter membranes based on polytetrafluoroethylene, characterized in that, include: The acquisition module is used to acquire the mass ratio of each component in the mixing stage, the baseband thickness at each preset monitoring point in the calendering stage based on the preset roller spacing threshold, the thickness variation coefficient of the filter membrane in the longitudinal stretching stage based on the preset stretching ratio, the fibril surface density, and the membrane air permeability in the transverse stretching stage based on the preset heating power. The adjustment module is used to determine whether the calendering is abnormal based on the degree of deviation of the base strip thickness, and to adjust the preset stretch ratio and the preset roll gap threshold based on the determination result of the calendering abnormality. The correction module is used to correct the preset heating power according to the quality deviation and the preset power correction coefficient, wherein the quality deviation is determined based on the thickness variation coefficient and the fibril surface density obtained after adjusting the preset stretch ratio and the preset roller spacing threshold. The update module is used to update the preset power correction coefficient and the mass ratio according to the lateral adjustment mass index and the lateral correction mass index, wherein the lateral adjustment mass index is determined based on the membrane air permeability obtained after adjusting the preset stretch ratio and the preset roller spacing threshold, and the lateral correction mass index is determined based on the membrane air permeability obtained after correcting the preset heating power. The early warning module is used to issue early warnings based on the deviation of the lateral adjustment mass index and the lateral correction mass index after updating the preset power correction coefficient and the mass ratio.

2. The production control system for microporous filtration membranes based on polytetrafluoroethylene according to claim 1, characterized in that, The adjustment module includes: A statistical calculation unit is used to determine the thickness range and the thickness mean based on the baseband thickness of each preset monitoring point, and to calculate the thickness deviation ratio based on the baseband thickness and the thickness mean. An anomaly determination unit is used to determine whether the rolling process is abnormal based on the threshold comparison result of the thickness range, and to determine the dominant direction of the anomaly based on the thickness deviation ratio. The calendering adjustment unit is used to adjust the preset stretching ratio and the preset roll spacing threshold based on the determination result of calendering anomaly, according to the thickness range, the thickness deviation ratio, and the dominant direction of the anomaly.

3. The production control system for microporous filtration membranes based on polytetrafluoroethylene according to claim 2, characterized in that, The anomaly determination unit includes: The range determination subunit is used to determine rolling anomalies based on the threshold comparison results of the thickness range, and to determine the range deviation degree based on the degree of deviation of the thickness range from its threshold. An orientation identification subunit is used to determine the dominant orientation of the anomaly based on a threshold comparison result of the thickness deviation ratio.

4. The production control system for microporous filtration membranes based on polytetrafluoroethylene according to claim 3, characterized in that, The rolling adjustment unit includes: A stretching adjustment subunit is used to reduce the preset stretching ratio based on the determination result of the rolling anomaly and the range deviation. A roller pitch adjustment subunit is used to adjust the preset roller pitch threshold according to the abnormal dominant direction, the average thickness, and the baseband thickness.

5. The production control system for microporous filtration membranes based on polytetrafluoroethylene according to claim 4, characterized in that, The correction module includes: A distance determination unit is used to determine the quality deviation based on the spatial distance between the current longitudinal vector and a preset longitudinal vector, wherein the current longitudinal vector is determined based on the thickness variation coefficient and the fibril surface density obtained after adjusting the preset stretch ratio and the preset roll gap threshold. A longitudinal correction unit is used to correct the preset heating power based on the threshold comparison result of the quality deviation and the preset power correction coefficient.

6. The production control system for microporous filtration membranes based on polytetrafluoroethylene according to claim 5, characterized in that, The longitudinal correction unit includes: The deviation determination subunit is used to determine longitudinal tension abnormality when the quality deviation is greater than a preset deviation threshold. A longitudinal correction subunit, connected to the deviation determination subunit, is used to increase the preset heating power based on the determination result of the longitudinal tension anomaly, according to a power adjustment amount, wherein the power adjustment amount is determined based on the mass deviation degree and the preset power correction coefficient.

7. The production control system for microporous filtration membranes based on polytetrafluoroethylene according to claim 6, characterized in that, The update module includes: An index determination unit is used to determine the lateral adjustment quality index and the lateral correction quality index based on the dispersion and central tendency of the membrane air permeability obtained after adjusting the preset stretch ratio and the preset roller spacing threshold and the membrane air permeability obtained after correcting the preset heating power, respectively. The trend determination unit is used to determine the adjustment quality slope and the correction quality slope respectively based on the changes of the lateral adjustment quality index and the lateral correction quality index within a preset monitoring batch. The root cause identification unit is used to determine the root cause of the deviation based on the threshold comparison results of the lateral adjustment quality index, the lateral correction quality index, the adjustment quality slope, and the correction quality slope, respectively. A lateral update unit is used to update the preset power correction coefficient and the quality ratio based on the root cause of the deviation, according to the lateral adjustment quality index, the lateral correction quality index, the adjustment quality slope, and the correction quality slope.

8. The production control system for microporous filtration membranes based on polytetrafluoroethylene according to claim 7, characterized in that, The root cause identification unit includes: The mixing identification subunit is used to determine that the root cause of the deviation is the mixing stage when the lateral adjustment quality index is greater than a preset adjustment index threshold, the lateral correction quality index is greater than a preset correction index threshold, and the adjustment quality slope and the correction quality slope are both greater than a preset slope threshold. A lateral identification subunit is used to determine that the root cause of the deviation is the lateral stretching stage when the lateral adjustment quality index is greater than a preset adjustment index threshold, the lateral correction quality index is greater than a preset correction index threshold, and the adjustment quality slope is greater than a preset slope threshold or the correction quality slope is greater than a preset slope threshold. The over-adjustment identification subunit is used to determine that the root cause of the deviation is a correction over-adjustment type when the lateral adjustment quality index is less than or equal to a preset adjustment index threshold, the lateral correction quality index is greater than a preset correction index threshold, the adjustment quality slope is less than a preset slope threshold, and the correction quality slope is greater than a preset slope threshold.

9. The production control system for microporous filtration membranes based on polytetrafluoroethylene according to claim 8, characterized in that, The horizontal update unit includes: A mixing update subunit is used to update the mass ratio according to the mass update amount when the root cause of the deviation is the mixing stage, wherein the mass update amount is determined based on the lateral correction mass index and the preset correction index threshold. A lateral update subunit is used to update the preset power correction coefficient according to the coefficient update amount when the root cause of the deviation is the lateral stretching stage, wherein the coefficient update amount is determined based on the lateral adjustment quality index and the preset adjustment index threshold. An over-adjustment update subunit is used to update the preset power correction coefficient according to the coefficient reverse update amount when the root cause of the deviation is the corrected over-adjustment type, wherein the coefficient reverse update amount is determined based on the lateral correction quality index and the preset correction index threshold.

10. The production control system for microporous filtration membranes based on polytetrafluoroethylene according to claim 9, characterized in that, The early warning module includes: The fluctuation calculation unit is used to calculate the standard deviation of the horizontal adjustment quality index and the standard deviation of the horizontal correction quality index after updating the preset power correction coefficient and the quality ratio within a preset monitoring period, so as to obtain the adjustment fluctuation value and the correction fluctuation value respectively. The deviation calculation unit is used to determine the adjustment deviation and the correction deviation based on the threshold comparison results of the adjustment fluctuation value and the correction fluctuation value, respectively, according to the degree of deviation of the adjustment mean and the correction mean from their respective thresholds, wherein the adjustment mean and the correction mean are determined based on the horizontal adjustment quality index and the horizontal correction quality index, respectively. An early warning unit is used to issue an early warning based on the threshold comparison results of the adjustment deviation and the correction deviation, respectively.

Images