Tightening effect real-time monitoring method based on polyester fabric tightening equipment

Abstract

The invention discloses a tightening effect real-time monitoring method based on polyester fabric tightening equipment, and relates to the field of tension adjustment, and the method comprises the following steps: step 1, deploying tightening data acquisition equipment in a specified area of the tightening equipment; step 2, presetting a plurality of adjustment schemes when the tightening equipment performs tightening operation in different sections along the length direction, the width direction or the specific angle direction of the cloth, wherein each scheme comprises a target tightening speed set value and a target tightening amplitude set value in the direction; 3, the tightening data acquisition equipment sequentially or parallelly applies acting force to the polyester fabric at the tightening speed and the target tightening amplitude set in the adjustment scheme in different specified directions; by collecting the actual deformation quantity, the local flatness and the wrinkle generation coefficient of the polyester fabric and combining a preset speed and amplitude scheme and timing state evaluation, key parameters causing abnormal tightening can be accurately recognized and marked, and the marked parameters are subjected to simulation iteration correction.

Description

Technical Field

[0001] The present invention relates to the technical field of tension adjustment, and particularly to a real-time monitoring method for the tensioning effect based on a polyester fabric tensioning device. Background Art

[0002] As a widely used synthetic fiber material, polyester fabric is commonly used in fields such as clothing, home textiles, and industrial products. During the textile process, tensioning the fabric is to ensure its quality and stability. With the continuous improvement of the automation level of the manufacturing industry, tensioning devices are usually integrated into automated production lines to improve production efficiency and reduce manual intervention. The progress of modern sensor technology has made real-time data collection and analysis possible, such as tension sensors, thermal imaging cameras, vision inspection systems, etc. Currently, most tension monitoring devices use a single sensor or only control the tension in a single direction, making it difficult to perceive local stress unevenness in the width or specific angle direction in real time. Parameters such as the tensioning speed and amplitude are usually statically preset and are difficult to dynamically adjust according to the real-time tensioning state, which is not conducive to timely identification of local abnormalities, resulting in defects flowing into the subsequent processes, increasing the defective rate, and the optimization process highly depends on step-by-step trial and error, with a long adjustment cycle and easy resource waste. Summary of the Invention

[0003] (I) Technical Problems to be Solved Aiming at the above-mentioned disadvantages of the prior art, the present invention provides a real-time monitoring method for the tensioning effect based on a polyester fabric tensioning device, which can effectively solve the problems of the prior art.

[0004] (II) Technical Solutions To achieve the above objectives, the present invention is realized through the following technical solutions: The present invention discloses a real-time monitoring method for the tensioning effect based on a polyester fabric tensioning device, including the following steps: Step 1: Deploy tensioning data collection devices in the designated area of the tensioning device; Step 2: Preset several adjustment schemes for the tensioning device when performing tensioning operations in different sections along the length direction, width direction, or specific angle direction of the fabric. Each scheme includes the set value of the target tensioning speed and the set value of the target tensioning amplitude in this orientation; Step 3: The tensioning data collection devices apply forces to the polyester fabric in sequence or in parallel in the designated different orientations at the tensioning speed and target tensioning amplitude set in the adjustment scheme to perform item-by-item tensioning operations; Step 4: At several preset key time periods, based on the information real-time feedback by the tensioning data collection devices, conduct a state evaluation of the current tensioning state of the polyester fabric one by one to obtain a state evaluation data packet including several indicators; Step 5: Based on the status evaluation data packets for several time periods, establish the correlation between the tensioning speed, amplitude parameters in the adjustment plan and the status change values, and mark the plan parameters that cause the status change amplitude to exceed the preset threshold; Step 6: Construct a simulated tensioning model through a deep learning algorithm. Using the historical status evaluation data as training samples, take the marked parameters and their corresponding adjustment plans as inputs, and iteratively correct the tensioning speed or amplitude parameters. When the status change amplitude of the simulated output is lower than the preset threshold, generate the optimized adjustment parameters in the orientation corresponding to the marked parameters; Step 7: Update the optimized adjustment parameters output by the simulated tensioning model to the parameters in the corresponding orientation in the adjustment plan; Step 8: Based on the updated adjustment plan, repeat Steps 2 to 6 for subsequent orientations or unqualified orientations until no marking behavior is triggered in Step 5.

[0005] Furthermore, the acquisition device includes: a tension sensor, a displacement sensor, a speed sensor and an image acquisition device, which are respectively used to obtain the tension, deformation amount, displacement speed and surface status of the polyester cloth during the tensioning process in real time.

[0006] Furthermore, the evaluation indicators in Step 4 include: actual deformation amount, local flatness and wrinkle generation coefficient.

[0007] Furthermore, the status evaluation process in Step 4 includes the following steps: Step 41: Obtain the tension value and instantaneous displacement value of the polyester cloth in a specified time period, and determine the actual deformation amount of the polyester cloth based on the difference between the displacement value and the pre-determined reference displacement; Step 42: Obtain the surface image of the polyester cloth in the same time period, divide the image into multiple grid units according to the equal area division principle, statistically analyze the pixel gray values in each grid unit, obtain the standard deviation of the gray value, and use the obtained standard deviation as the local flatness index of the corresponding area; Step 43: Based on the same surface image, use the pre-trained edge detection algorithm to extract the wrinkle contour on the surface of the polyester cloth, statistically analyze the area ratio of the extracted contour area in the whole image, and use this ratio as the wrinkle generation coefficient; Step 44: Package the actual deformation amount, the local flatness index of each grid unit, and the wrinkle generation coefficient, and attach the corresponding key time period identifier and sensor position identifier to form the status evaluation data packet for this time period.

[0008] Furthermore, the calculation formula for the status evaluation data packet in Step 4 is: ; In the formula, Represents the comprehensive score of real-time tension status, Represents the weight coefficient of tension balance, represents the tension distribution sensitivity coefficient, Represents the difference between the maximum and minimum tension values of the fabric. Represents the material standard tension threshold, The weight coefficient representing the surface wrinkle degree, Represents the number of machine vision sampling areas, represents the surface wrinkle intensity of the image in the i-th region, J represents the calibration value of the wrinkle-free state, represents the weight coefficient of deformation risk, h represents the percentage coefficient of local strain abnormal increment, Represents the critical value of material yield strain.

[0009] Furthermore, the designated area of the tensioning device includes: a driving mechanism, a tension roller and a clamping point.

[0010] Furthermore, the association relationship establishment process in step 5 is as follows: Step 51: Extracting the actual deformation, local flatness, and wrinkle generation coefficient corresponding to each key period from the state evaluation data packets of the plurality of time periods as state indicators, and calculating the absolute change of each state indicator from the start of the tensioning operation to each key period as a state change value, wherein the state change value includes the deformation change value, the flatness change value, and the wrinkle generation coefficient change value; Step 52: Using a multiple linear regression analysis method, with the tightening speed parameter and the tightening amplitude parameter as independent variables and the state change value as the dependent variable, quantify the direct influence coefficient of the tightening speed parameter and the tightening amplitude parameter on the state change value.

[0011] Furthermore, the marking process in step 5 is as follows: comparing the state change value with a preset threshold value; if the absolute value or the rate of change of any state change value exceeds its corresponding preset threshold value, marking the tensioning speed parameter or tensioning amplitude parameter that causes the exceeding limit, and recording the adjustment scheme to which it belongs; the preset threshold value is set based on historical data or experimental calibration.

[0012] Furthermore, the simulated tension model in step 6 selects samples including marked parameters and corresponding state evaluation vectors in the historical database to construct a training set and a validation set. The input layer is the marked parameters, the hidden layer is the fully connected layer, and the output layer is the predicted state change data. The mean square error is used as the loss function, and training is performed through back propagation until the validation set error converges. For each parameter to be optimized, the current value is input into the trained model, and iterative adjustment is simulated. If the predicted state change data is less than the standard threshold, the parameter is fine-tuned in the gradient direction until the standard threshold is met or the iteration upper limit is reached.

[0013] Furthermore, after the adjustment parameters are updated in step 7, one of the following operations is triggered: interrupt the current tensioning operation in real time and reload the adjustment plan; automatically switch to the updated parameters after completing the current azimuth tensioning operation.

[0014] (III) Beneficial Effects Adopting the technical solution provided by the present invention, compared with the known prior art, it has the following beneficial effects: 1. By arranging a variety of sensors in the key areas of the tensioning device and integrating multiple indicators for real-time collection and evaluation, multi-directional monitoring of the tension state of the polyester fabric is achieved, and the state is evaluated. Hidden dangers such as excessive or too small local tension, wrinkles, and stress concentration can be detected early, the defective rate is greatly reduced, and accurate numerical basis is provided for subsequent process adjustment, thus reflecting superior process controllability and quality stability.

[0015] 2. By introducing a simulation model and iterative optimization mechanism based on deep learning, not only can historical state evaluation data be converted into training samples, but also parameter combinations that cause the state to exceed the boundary can be automatically marked and corrected. Finally, the optimal tensioning speed and amplitude settings for different fabric directions and different sections are generated, thereby reducing the empirical limitations of manual parameter adjustment, significantly improving production efficiency and consistency, while also reducing the trial-and-error cost, and facilitating flexible deployment under various product specifications and production line conditions.

[0016] 3. By presetting multiple tensioning plans and executing them in sequence, and performing state evaluation based on real-time feedback information, fine control under different tensioning states is achieved, the flexibility and adaptability of the production process are improved, and in the face of different types of polyester fabrics and tensioning requirements, the operation plan can be quickly adjusted, thereby effectively responding to market changes and customer needs. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the drawings in the following description are only some embodiments of the present invention, and those of ordinary skill in the art can also obtain other drawings based on these drawings without creative efforts.

[0018] Figure 1 is a schematic flow chart of the present invention; DETAILED DESCRIPTION OF THE EMBODIMENTS To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0019] The present invention will be further described below with reference to the embodiments.

[0020] Embodiment 1 A real-time monitoring method for the tensioning effect of a tensioning device based on polyester fabric, as Figure 1 shown, includes the following steps: Step 1: Deploy tensioning data acquisition devices in the designated area of the tensioning device. The designated area of the tensioning device includes: a driving mechanism, a tension roller, and a clamping point. The acquisition devices include: a tension sensor, a displacement sensor, a speed sensor, and an image acquisition device, which are respectively used to obtain the tension, deformation amount, displacement speed, and surface state of the polyester fabric in real time during the tensioning process. Step 2: Preset several adjustment schemes for the tensioning device when performing tensioning operations in different sections along the length direction, width direction, or specific angle direction of the fabric. Each scheme includes the set value of the target tensioning speed and the set value of the target tensioning amplitude in that orientation. Presetting multi-directional adjustment schemes can plan process parameters in advance, reduce the number of trial and errors, and improve production efficiency.

[0021] Step 3: The tensioning data acquisition devices apply forces to the polyester fabric in sequence or in parallel at different designated orientations with the tensioning speed and target tensioning amplitude set in the adjustment scheme, and perform item-by-item tensioning operations. The item-by-item operation of applying forces according to the scheme supports flexible verification and comparison, and optimizes the debugging process.

[0022] Step 4: At several preset key time periods, based on the information real-time feedback by the tensioning data acquisition devices, conduct a state evaluation on the current tensioning state of the polyester fabric one by one. The evaluation indicators include: actual deformation amount, local flatness, and wrinkle generation coefficient, and obtain a state evaluation data packet including several indicators. The state evaluation at key time periods can timely identify abnormal deformations and flatness problems, and ensure the process quality.

[0023] Step 5: Based on the status evaluation data packets for several time periods, establish the correlation between the tensioning speed, amplitude parameters in the adjustment plan and the status change values. Mark the plan parameters that cause the status change amplitude to exceed the preset threshold. The marking process is as follows: Compare the status change value with the preset threshold. If the absolute value or change rate of any status change value exceeds its corresponding preset threshold, mark the tensioning speed parameter or tensioning amplitude parameter that causes the overlimit, and record the adjustment plan to which it belongs. The preset threshold is set based on historical data or experimental calibration. Through the threshold marking mechanism, the overlimit parameters can be quickly located, providing an accurate basis for subsequent optimization and reducing blind adjustment.

[0024] Step 6: Construct a simulated tensioning model through a deep learning algorithm. Using the historical status evaluation data as training samples, take the marked parameters and their corresponding adjustment plans as inputs, and iteratively correct the tensioning speed or amplitude parameters. When the status change amplitude of the simulated output is lower than the preset threshold, generate the optimized adjustment parameters in the direction corresponding to the marked parameters. The simulated tensioning model selects samples including marked parameters and corresponding status evaluation vectors from the historical database to construct a training set and a validation set. The input layer is the marked parameters, the hidden layer is a fully connected layer, and the output layer is the predicted status change data. Using the mean squared error as the loss function, train through backpropagation and the Adam optimizer until the error of the validation set converges. For each parameter to be optimized, input the current value into the trained model and simulate iterative adjustment. If the predicted status change data is less than the standard threshold, fine-tune the parameter in the gradient direction until the standard threshold is met or the iteration limit is reached. The simulation model based on deep learning uses historical data for iterative optimization, accelerating parameter adjustment, reducing manual intervention, and improving debugging efficiency and accuracy.

[0025] Step 7: Update the optimized adjustment parameters output by the simulated tensioning model to the parameters in the corresponding position in the adjustment plan. After the adjustment parameters are updated, trigger one of the following operations: Interrupt the current tensioning operation in real time and reload the adjustment plan; Automatically switch to the updated parameters after completing the tensioning operation in the current position; Dynamically update the parameters and support real-time or batch switching to achieve fast response and reduce downtime and waste.

[0026] Step 8: Based on the updated adjustment plan, repeat Steps 2 to 6 for subsequent positions or non-compliant positions until no marking behavior is triggered in Step 5.

[0027] In this embodiment, through the combination of real-time multi-sensor data acquisition and deep learning models, dynamic monitoring and intelligent optimization of the polyester fabric tensioning process are realized, effectively improving the stability and consistency of the tensioning quality, reducing human intervention and adjustment time. Compared with traditional methods of single-parameter control or empirical adjustment, it has a higher degree of automation, more accurate state assessment capabilities, and self-adaptive adjustment capabilities, optimizing the real-time monitoring and adjustment efficiency of the tensioning effect, and thus realizing intelligent control of the entire process of textile tensioning.

[0028] Embodiment 2 On other levels, this embodiment also provides another optimization mechanism based on Embodiment 1, specifically a state assessment process for polyester fabric, including the following steps: Step 41: Obtain the tension value and instantaneous displacement value of the polyester fabric during a specified period, and determine the actual deformation of the polyester fabric based on the difference between the displacement value and a pre-determined reference displacement; Step 42: Obtain the surface image of the polyester fabric during the same period, divide the image into multiple grid units according to the principle of equal area division, perform statistical analysis on the pixel gray values within each grid unit, obtain the standard deviation of the gray value, and use the obtained standard deviation as the local flatness index for the corresponding area; Step 43: Based on the same surface image, use a pre-trained edge detection algorithm to extract the wrinkle contour on the surface of the polyester fabric, count the area ratio of the extracted contour area in the entire image, and use this ratio as the wrinkle generation coefficient; Step 44: Package the actual deformation, the local flatness index of each grid unit, and the wrinkle generation coefficient, along with the corresponding key period identifier and sensor position identifier, into a state assessment data packet for this period.

[0029] As a preferred implementation in this embodiment, the calculation formula for the state assessment data packet is: ; In the formula, represents the comprehensive score of the real-time tensioning state, represents the weight coefficient of the tension balance degree, represents the tension distribution sensitivity coefficient, represents the difference between the maximum and minimum tension values on the fabric surface, represents the material standard tension threshold, represents the weight coefficient of the surface wrinkle degree, represents the number of machine vision sampling areas, represents the surface wrinkle intensity of the image in the i-th area, J represents the calibration value for the non-wrinkle state, represents the weight coefficient of the deformation risk, h represents the local strain abnormal increment percentage coefficient, Represents the critical value of the yield strain of the material.

[0030] Compared with the prior art, by obtaining the tension and displacement values in real time to quantify the actual deformation, it provides an accurate physical basis for evaluation. Through the equal-area grid division of the surface image and the analysis of the standard deviation of pixel grayscales, it not only refines the evaluation of local flatness but also can sensitively capture minute surface defects. By using the edge detection algorithm to extract the wrinkle contour and calculate the area ratio, the degree of wrinkle is quantified as a coefficient, providing more intuitive data support, thereby improving the accuracy and stability of the evaluation and making the monitoring of the tightened state more comprehensive and objective.

[0031] Embodiment 3 This embodiment provides a process for establishing an association relationship as follows: Step 51: From the state evaluation data packets of the several time periods, extract the actual deformation, local flatness, and wrinkle generation coefficient corresponding to each key time period as state indicators, and calculate the absolute change amount of each state indicator from the starting moment of the tightening operation to each key time period as the state change value, where the state change value includes the deformation change value, flatness change value, and wrinkle generation coefficient change value; Step 52: Adopt the multiple linear regression analysis method, with the tightening speed parameter and the tightening amplitude parameter as independent variables and the state change value as the dependent variable, to quantify the direct influence coefficient of the tightening speed parameter and the tightening amplitude parameter on the state change value.

[0032] Compared with the prior art, extracting the absolute change amounts of the deformation, flatness, and wrinkle coefficient at each key time period and using multiple linear regression to directly associate the tightening speed and amplitude with the state change amount not only quantifies the influence coefficient of each parameter on the fabric state, enhancing the interpretability of the model, but also efficiently establishes the association relationship with low computational complexity, providing an accurate theoretical basis for subsequent optimization and real-time control, thereby improving the adjustment accuracy and response speed of the tightening process.

[0033] In summary, the present invention can accurately identify and mark the key parameters causing abnormal tightening by collecting the actual deformation, local flatness, and wrinkle generation coefficient of the polyester fabric, combining with the preset speed and amplitude scheme and the timed state evaluation, using the simulated tightening model to perform simulated iterative correction on the marked parameters, realizing the adaptive optimization of the parameters within the threshold, and finally dynamically updating the optimization result to the scheme library for waiting to be applied; By arranging a variety of sensors in the key areas of the tensioning device and integrating multiple indicators for real-time collection and evaluation, multi-faceted monitoring of the tension state of the polyester fabric is achieved, and the state is evaluated. Hidden dangers such as excessive or too little local tension, wrinkles, and stress concentration can be detected early, significantly reducing the defective rate, and providing accurate numerical basis for subsequent process adjustment, thus reflecting superior process controllability and quality stability. By introducing a simulation model based on deep learning and an iterative optimization mechanism, not only can historical state evaluation data be converted into training samples, but also the parameter combinations that cause the state to exceed the limit can be automatically marked and corrected. Finally, the optimal tensioning speed and amplitude settings for different fabric directions and different sections are generated, thus reducing the empirical limitations of manual parameter adjustment, significantly improving production efficiency and consistency, and at the same time reducing the trial-and-error cost, facilitating flexible deployment under various product specifications and production line conditions. By presetting multiple tensioning schemes and executing them in sequence, and performing state evaluation based on real-time feedback information, fine control under different tensioning states is achieved, improving the flexibility and adaptability of the production process, and enabling quick adjustment of the operation plan in the face of different types of polyester fabrics and tensioning requirements, thus effectively responding to market changes and customer needs; This process not only greatly improves the stability and consistency of the tensioning effect, but also reduces manual intervention and experience dependence, has the capabilities of online rapid response, environmental compensation and continuous learning, and significantly improves the intelligent level and production efficiency of the polyester fabric tensioning device.

[0034] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacement of some of the technical features; and these modifications or replacements will not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims

1. A real-time monitoring method for the tensioning effect of a polyester fabric tensioning device, characterized in that, It includes the following steps: Step 1: Deploy a tension data acquisition device in a specified area of the tensioning device; Step 2: Preset several adjustment schemes for the tensioning device when performing tensioning operations in different sections along the length direction, width direction or specific angle direction of the fabric. Each scheme includes a target tensioning speed setting value and a target tensioning amplitude setting value in that orientation; Step 3: The tension data acquisition device sequentially or in parallel applies forces to the polyester fabric at the specified different orientations at the tensioning speed and target tensioning amplitude set in the adjustment scheme to perform item-by-item tensioning operations; Step 4: At a preset number of key time periods, based on the information real-time feedback by the tension data acquisition device, conduct a state assessment for the current tensioning state of the polyester fabric one by one, and obtain a state assessment data packet including several indicators; Step 5: Based on the state assessment data packets of several time periods, establish the correlation between the tensioning speed and amplitude parameters in the adjustment scheme and the state change value, and mark the scheme parameters that cause the state change amplitude to exceed the preset threshold; Step 6: Construct a simulated tensioning model through a deep learning algorithm. Using the historical state assessment data as training samples, take the marked parameters and their corresponding adjustment schemes as inputs, and iteratively correct the tensioning speed or amplitude parameters. When the state change amplitude of the simulated output is lower than the preset threshold, generate optimized adjustment parameters for the orientation corresponding to the marked parameters; Step 7: Update the optimized adjustment parameters output by the simulated tensioning model to the parameters corresponding to the orientation in the adjustment scheme; Step 8: Based on the updated adjustment scheme, repeat Steps 2 to 6 for subsequent orientations or non-compliant orientations until no marking behavior is triggered in Step 5.

2. The real-time monitoring method for the tensioning effect of a polyester cloth tensioning device according to claim 1, characterized in that, The acquisition device includes: a tension sensor, a displacement sensor, a speed sensor and an image acquisition device, which are respectively used to obtain the tension, deformation amount, displacement speed and surface state of the polyester fabric during the tensioning process in real time.

3. A real-time monitoring method for the tensioning effect of a polyester fabric tensioning device according to claim 1, characterized in that, The evaluation indicators in Step 4 include: actual deformation amount, local flatness and wrinkle generation coefficient.

4. The real-time monitoring method for the tensioning effect of a polyester fabric tensioning device according to claim 1, characterized in that The state assessment process in Step 4 includes the following steps: Step 41: Obtain the tension value and instantaneous displacement value of the polyester fabric at a specified time period, and determine the actual deformation amount of the polyester fabric based on the difference between the displacement value and the pre-determined reference displacement; Step 42: Obtain the surface image of the polyester fabric at the same time period, divide the image into multiple grid units according to the equal area division principle, statistically analyze the pixel gray values in each grid unit, obtain the standard deviation of the gray value, and use the obtained standard deviation as the local flatness index of the corresponding area; Step 43: Based on the same surface image, use a pre-trained edge detection algorithm to extract the wrinkle contour on the surface of the polyester fabric, statistically analyze the area ratio of the extracted contour area in the whole image, and use this ratio as the wrinkle generation coefficient; Step 44: Package the actual deformation amount, the local flatness index of each grid unit, and the wrinkle generation coefficient, along with the corresponding key time period identifier and sensor position identifier, to form the state assessment data packet for this time period.

5. A real-time monitoring method for the tensioning effect of a polyester fabric tensioning device according to claim 1, characterized in that, The calculation formula for the state assessment data packet in Step 4 is: ; Wherein, represents the comprehensive score of the real-time tension state, represents the weight coefficient of the tension balance degree, represents the tension distribution sensitivity coefficient, represents the difference between the maximum and minimum tension values of the fabric surface, represents the material standard tension threshold value, represents the weight coefficient of the surface wrinkle degree, represents the number of machine vision sampling areas, represents the surface wrinkle intensity of the image in the i-th area, and J represents the calibration value in the non-wrinkle state, represents the weight coefficient of the deformation risk, and h represents the local strain anomaly increment percentage coefficient, represents the critical value of the material yield strain.

6. The real-time monitoring method for the tensioning effect of a polyester cloth tensioning device according to claim 1, characterized in that, The designated area of the tensioning device includes: a driving mechanism, a tension roller and a clamping point.

7. A real-time monitoring method for the tensioning effect of a polyester fabric tensioning device according to claim 1, characterized in that The process of establishing the association relationship in step 5 is as follows: Step 51: Extracting the actual deformation, local flatness, and wrinkle generation coefficient corresponding to each key period from the state evaluation data packets of the plurality of time periods as state indicators, and calculating the absolute change of each state indicator from the start of the tensioning operation to each key period as a state change value, wherein the state change value includes the deformation change value, the flatness change value, and the wrinkle generation coefficient change value; Step 52: Using a multiple linear regression analysis method, with the tightening speed parameter and the tightening amplitude parameter as independent variables and the state change value as the dependent variable, quantify the direct influence coefficient of the tightening speed parameter and the tightening amplitude parameter on the state change value.

8. A real-time monitoring method for the tensioning effect of a polyester fabric tensioning device according to claim 1, characterized in that The marking process in step 5 is as follows: comparing the state change value with a preset threshold value; if the absolute value or rate of change of any state change value exceeds its corresponding preset threshold value, marking the tensioning speed parameter or tensioning amplitude parameter that causes the exceeding limit and recording the corresponding adjustment scheme; the preset threshold value is set based on historical data or experimental calibration.

9. The real-time monitoring method for the tensioning effect of a polyester cloth tensioning device according to claim 1, characterized in that, The simulated tension model in step 6 selects samples including marked parameters and corresponding state evaluation vectors in the historical database to construct a training set and a validation set. The input layer is the marked parameters, the hidden layer is the fully connected layer, and the output layer is the predicted state change data. The mean square error is used as the loss function, and training is performed through back propagation until the validation set error converges. For each parameter to be optimized, the current value is input into the trained model, and iterative adjustment is simulated. If the predicted state change data is less than the standard threshold, the parameter is fine-tuned in the gradient direction until the standard threshold is met or the iteration upper limit is reached.

10. A real-time monitoring method for the tensioning effect of a polyester cloth tensioning device according to claim 1, characterized in that, After the adjustment parameters in step 7 are updated, one of the following operations is triggered: Interrupt the current tensioning operation in real time and reload the adjustment plan; Automatically switch to the updated parameters after completing the current position tightening operation.

Images