Online detection system for lining cloth production

Abstract

The invention discloses an online detection system for lining cloth production, and relates to the technical field of textiles, and the online detection system for lining cloth production comprises a data management module, a data prediction module, a central processing module and an adjustment control module. By establishing the tension prediction model, the future change trend of the cloth tension is predicted, so that a production manager can predict potential problems and intervene in advance, the quality problem is prevented, the cloth tension change trend is analyzed by using a mode recognition algorithm, a problem mode which causes low efficiency can be recognized more accurately, and the production efficiency is improved. By evaluating and adjusting the key factors influencing the production efficiency, a production manager can accurately control and adjust the operation parameters of the production line, so that the overall production efficiency is improved, cloth damage or defective product generation caused by improper parameter setting can be reduced, and material and energy waste is reduced.

Description

Technical Field

[0001] The invention relates to the technical field of textiles, and in particular to an online detection system for lining cloth production. Background Art

[0002] In the modern textile industry, the use of advanced detection technology to replace manual detection of fabric tension can significantly improve detection efficiency, reduce labor demand, reduce labor intensity, and further ensure the quality of fabrics. Traditionally, the detection of fabric tension mostly relies on manual experience and intuition. This method is not only inefficient, but also the results are easily affected by individual differences, with low reproducibility and unstable detection results. This is particularly unsuitable in large-scale industrial production and cannot meet the high standards and high efficiency requirements of modern textile production. In the textile production process, uneven or inappropriate fabric tension may cause fabric defects, such as uneven stretching, breakage or deformation. If these problems are not discovered and handled in time, a large number of defective products will be produced, resulting in a waste of materials and energy. By installing an online detection system, not only can the uniformity and suitability of fabric tension be monitored in real time, but once an abnormality is detected, measures can be taken quickly to adjust or repair it, thereby effectively reducing the output of defective products.

[0003] When in use, the existing online detection system for lining cloth production is not convenient for real-time acquisition and pre-processing of cloth tension data, and is not convenient for ensuring the accuracy and real-time nature of the data, and is not convenient for predicting future changes in cloth tension, making it difficult for production managers to foresee potential problems and intervene in advance, which easily increases the occurrence of quality problems, and is unable to analyze the cloth tension change trend, making it difficult to accurately identify the problem pattern that leads to inefficiency, and thus making it difficult for production managers to accurately control and adjust the operating parameters of the production line, resulting in a decrease in the overall efficiency of lining cloth production.

[0004] Currently, no effective solution has been proposed for the problems in the related technologies. Summary of the invention

[0005] In view of the deficiencies in the prior art, the present invention proposes an online detection system for lining cloth production, which solves the problem that the existing online detection system for lining cloth production mentioned in the above background technology is not convenient for real-time acquisition and preprocessing of fabric tension data, is not convenient for ensuring the accuracy and real-time nature of the data, and is not convenient for predicting future changes in fabric tension, making it inconvenient for production managers to foresee potential problems and intervene in advance, which easily increases the occurrence of quality problems, and cannot analyze the changing trend of fabric tension, making it inconvenient to accurately identify the problem pattern that leads to low efficiency, thereby making it inconvenient for production managers to accurately control and adjust the operating parameters of the production line, resulting in a decrease in the overall efficiency of lining cloth production.

[0006] To achieve the above objectives, the present invention is implemented through the following technical solutions: An online detection system for lining cloth production, the online detection system for lining cloth production comprises: a data management module, a data prediction module, a central processing module and an adjustment control module; The data management module is used to obtain the real-time data of the fabric tension on the lining fabric production line and pre-process the real-time data of the fabric tension to obtain the fabric tension characteristic data; A data prediction module is used to establish a tension prediction model based on the obtained cloth tension characteristic data, and use the tension prediction model to predict the future cloth tension to obtain the cloth tension change trend; A central processing module is used to analyze the obtained cloth tension change trend by using a pattern recognition algorithm to identify the cloth tension change pattern; The regulating control module is used to evaluate the identified cloth tension variation pattern using a factor evaluation algorithm, obtain key factors affecting the lining cloth production efficiency, and control the operating parameters of the production line based on the key factors; Wherein, the data management module is connected to the central processing module via the data prediction module, and the central processing module is connected to the regulation control module; The data prediction module includes: a model building module and a model prediction module; A model building module is used to build a tension prediction model based on the obtained cloth tension characteristic data; The model prediction module is used to predict the future cloth tension by using the established tension prediction model to obtain the cloth tension change trend; Among them, the model building module and the model prediction module are connected.

[0007] Furthermore, based on the obtained cloth tension characteristic data, a tension prediction model is established including: Convert the obtained cloth tension characteristic data into a tension detection cycle; Substitute the converted tension detection period into the time series method, divide the obtained fabric tension data into periodic indicators and trend indicators, and build and train the tension prediction model through periodic indicators and trend indicators; Statistical error indicators are used as evaluation criteria for the accuracy of the tension prediction model.

[0008] Furthermore, the formula for the statistical error index is: ; In the formula, M Expressed as a statistical error indicator; a It is expressed as the total number of predicted tension values; It is expressed as the actual observed cloth tension value; Expressed as predicted cloth tension value.

[0009] Furthermore, the established tension prediction model is used to predict the future cloth tension, and the cloth tension change trends are obtained, including: Based on the trend indicators, the model optimization algorithm is used to optimize the parameters of the tension prediction model, and the optimal parameter configuration is found. The optimal parameter configuration is applied to the constructed tension prediction model, and the trend indicators are predicted using prediction technology to obtain the fabric tension trend at future moments. The obtained future fabric tension trend is converted into actual production adjustments to construct the final tension prediction model.

[0010] Furthermore, based on the trend indicators, the model optimization algorithm is used to optimize the parameters of the tension prediction model, and the optimal parameter configuration is found. The optimal parameter configuration is applied to the constructed tension prediction model, and the trend indicators are predicted using prediction technology to obtain the fabric tension trend at future moments, including: Initialize the model optimization algorithm and the structural parameters of the tension prediction model; Normalize the fabric tension data and various environmental factor parameters; Update the parameter configuration of the tension prediction model using the model optimization algorithm; Update the optimal parameter configuration and fitness value, and determine whether the set conditions are met. If the conditions are not met, continue to use the model optimization algorithm to update the parameter configuration of the tension prediction model; The tension prediction model is trained, the global optimal parameters found by the model optimization algorithm are applied to the tension prediction model, and the prediction technology is used to predict the trend indicators to obtain the fabric tension trend at future moments.

[0011] Furthermore, the formula for updating the parameter configuration of the tension prediction model using the model optimization algorithm is: ; In the formula, Expressed as t +1 tension prediction model parameter configuration after iteration; Expressed as t Configuration of tension prediction model parameters at the iteration; Expressed as t -1st iteration tension prediction model parameter configuration; f It represents the direction of updating the parameters of the tension prediction model; kIt is expressed as the amplitude of the tension prediction model parameter update; g It is expressed as the magnitude of parameter adjustment from the current worst tension prediction model parameter configuration; Represented as the global worst tension prediction model parameter configuration.

[0012] Furthermore, the obtained cloth tension variation trend is analyzed using a pattern recognition algorithm, and the cloth tension variation patterns are identified to include: Set the parameters of the pattern recognition algorithm and initialize the cloth tension change feature set; The fitness of each feature in the cloth tension change feature set is calculated to evaluate the matching degree with the cloth tension change trend; Among all fabric tension variation characteristics, identifying the best fabric tension variation characteristics and the worst fabric tension variation characteristics; For each cloth tension variation feature, the distance from the optimal cloth tension variation feature and the worst cloth tension variation feature is calculated respectively; Dividing each cloth tension variation feature into an optimal group and a worst group according to the proximity of each cloth tension variation feature to the optimal cloth tension variation feature and the worst cloth tension variation feature; The best group and the worst group are updated, and the updated best group and the worst group are recombined; Enter iterative optimization and repeat the iteration until the current number of iterations equals the maximum number of iterations; After the iteration is completed, the final set of cloth tension change features is analyzed and the cloth tension change pattern is identified.

[0013] Furthermore, the formula for calculating the distance between each cloth tension change characteristic and the optimal cloth tension change characteristic and the worst cloth tension change characteristic is: ; ; In the formula, Expressed as z The distance between a cloth tension variation characteristic and the optimal cloth tension variation characteristic; Expressed as z The distance between the cloth tension variation characteristic and the worst cloth tension variation characteristic; It is represented as the position of the optimal cloth tension variation characteristic; It is represented as the location of the worst cloth tension variation characteristic; Expressed as zThe location of the cloth tension change feature.

[0014] Furthermore, the identified cloth tension variation pattern is evaluated using a factor evaluation algorithm to obtain the key factors affecting the lining cloth production efficiency, and the operating parameters of the production line are controlled based on the key factors, including: Determine the parameters of factors that affect interlining production efficiency and set the initial search range; Set the initial evaluation period value and the number of iterations as the end condition of the evaluation process; Comparing the determined factor parameters affecting the interlining production efficiency with the known factor parameters of the optimized interlining production mode, and updating the factor parameters affecting the interlining production efficiency according to a preset comparison technology; Based on the cross-validation method, the production efficiency and operating parameters of the production line are evaluated after each factor parameter comparison; Based on the evaluation results of production efficiency and operating parameters, the fitness score of each iteration is calculated; The fitness score after each iteration is compared with the fitness score of the previous iteration, and the factor parameter that has the greatest impact on the production efficiency of the lining cloth is selected as the current optimal solution; If the preset maximum number of iterations has been reached, the evaluation process is terminated; Analyze and evaluate the production results after the process is completed, and compare them with the preset production efficiency standards to determine whether the factor parameters have a significant impact on the lining cloth production efficiency. If so, make adjustments. If not, take the current factor parameters as the key factors affecting the lining cloth production efficiency, and adjust the production line operation parameters according to the key factors.

[0015] Furthermore, based on the cross-validation method, evaluating the production efficiency and the operating parameters of the production line after each factor parameter comparison includes the following steps: Initialize the first evaluation of the operating parameters and production efficiency of the lining cloth production line, and divide the parameter data set of factors affecting the lining cloth production efficiency into several subsets. In each evaluation, select one subset as the test set, and the union of the remaining subsets as the training set; The validation model is trained using the adjusted parameters of the factors affecting the interlining production efficiency on the training set, and the prediction performance of the validation model on the production efficiency and the operating parameters of the production line is tested on the test set; According to the actual performance of the parameters of factors affecting the production efficiency of interlining cloth on the test set, the generalization error of the verification model is calculated; The evaluation is repeated until each subset has been used as a test set, and the average of the generalization errors calculated in all evaluations is obtained as the overall evaluation of the production efficiency and production line operation parameters after each factor parameter comparison.

[0016] The beneficial effects of the present invention are: 1. The present invention acquires and pre-processes fabric tension data from the production line in real time to ensure the accuracy and real-time nature of the data, thereby helping to timely identify and adjust problem points in production. By establishing a tension prediction model, the future trend of fabric tension changes is predicted, so that production managers can foresee potential problems and intervene in advance to prevent the occurrence of quality problems. By using a pattern recognition algorithm to analyze the trend of fabric tension changes, the problem pattern that leads to low efficiency can be more accurately identified, and the problem can be quickly located. By evaluating and adjusting the key factors that affect production efficiency, production managers can accurately control and adjust the operating parameters of the production line to improve overall production efficiency, thereby reducing fabric damage or defective products caused by improper parameter settings, and reducing waste of materials and energy.

[0017] 2. The present invention constructs a tension prediction model based on the collected cloth tension feature data, which can accurately map the changing law of cloth tension. Its effectiveness lies in the ability to convert real-time data into useful information, providing a solid foundation for prediction, and using the established tension prediction model to predict future cloth tension, provide future change trends, and help predict problems that may be encountered in production, so as to take measures in advance. By predicting future cloth tension trends, production parameters can be adjusted in time to reduce cloth defects caused by tension mismatch, so as to prevent errors before problems occur, thereby reducing material waste and costs.

[0018] 3. By setting and initializing the parameters of the pattern recognition algorithm and constructing a set of fabric tension change characteristics, the system can effectively identify and distinguish different fabric tension change patterns, improve the accuracy and efficiency of the response, and by calculating the fitness of each fabric tension change characteristic, the system can evaluate the matching degree of these characteristics with the fabric tension change trend, ensure that the fabric maintains the best tension state during the production process, and reduce the generation of defects. By classifying the fabric tension change characteristics into the best group and the worst group, and constantly updating and recombining these groups, the dynamic adjustment of production parameters is achieved. By identifying and applying the best fabric tension change characteristics, fabric defects caused by improper tension, such as overstretching or breakage, can be reduced, thereby reducing the defective rate and material waste. At the same time, the consistency and high standards of fabric quality are ensured to meet the market demand for high-quality lining fabrics.

[0019] 4. The present invention determines the key factor parameters that affect the production efficiency of the lining cloth and sets the initial search range, so that the factor evaluation is more targeted and efficient, and the key adjustment points can be quickly identified. Through iterative evaluation cycles and cross-validation methods, the system continuously compares and updates the factor parameters that affect the production efficiency. This dynamic adjustment mechanism ensures that the production line always operates in the optimal state to adapt to the ever-changing production needs and environment. Through the calculated fitness score, the factor parameters that have the greatest impact on the production efficiency of the lining cloth are selected for optimization, which helps to achieve continuous improvement of the production process, thereby ultimately achieving the goal of improving production efficiency and product quality. The system reduces waste caused by excessive or improper use of resources by accurately controlling and optimizing production parameters, thereby not only saving costs, but also conforming to sustainable production practices. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0021] Figure 1 is a principle block diagram of an online detection system for lining cloth production according to an embodiment of the present invention; Figure 2 The present invention is a structural block diagram of a data prediction module in an online detection system for lining cloth production according to an embodiment of the present invention.

[0022] In the figure: 1. Data management module; 2. Data prediction module; 201. Model building module; 202. Model prediction module; 3. Central processing module; 4. Regulation and control module. DETAILED DESCRIPTION

[0023] The technical solutions in the embodiments of the present invention will be described clearly and completely below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments.

[0024] In the description of the present invention, unless otherwise specified, the meaning of "plurality" is two or more. In addition, the terms "first", "second", "third", etc. are only used for descriptive purposes and cannot be understood as indicating or implying relative importance.

[0025] According to an embodiment of the present invention, an online detection system for lining cloth production is provided.

[0026] The present invention is further described with reference to the accompanying drawings and specific embodiments. Figure 1-Figure 2 As shown, according to an embodiment of the present invention, the online detection system for lining cloth production comprises: a data management module 1, a data prediction module 2, a central processing module 3 and an adjustment control module 4; Data management module 1, used to obtain the real-time data of fabric tension on the lining fabric production line, and pre-process the real-time data of fabric tension to obtain fabric tension characteristic data; It should be explained that the real-time data of fabric tension mainly refers to the real-time measurement data of the tension and changes that the fabric is subjected to during the production process, including: Tension: The tension value of the fabric at a specific point in time, usually measured in Newtons or kilograms. This is a direct indicator of the stretch state of the fabric.

[0027] Tension change rate: The speed or amplitude of change of fabric tension in a time series, which reflects the stability and volatility of fabric tension.

[0028] Position information: Tension data may involve measuring the tension of the fabric at different locations on the production line to monitor uniformity and consistency throughout the production line.

[0029] Environmental factors: Environmental parameters such as temperature and humidity may also be recorded, as these factors can affect the measured and actual performance of fabric tension.

[0030] Timestamp: Each tension data point will be equipped with an accurate time stamp to ensure the real-time and traceability of the data, which is convenient for subsequent analysis and processing.

[0031] It should be explained that the real-time data of fabric tension on the lining production line is obtained, and the real-time data of fabric tension is preprocessed to obtain fabric tension characteristic data including: Deploy fabric tension data collection nodes and use communication protocols to obtain real-time fabric tension data on the lining production line; In the converted frequency domain data, identify and track the environmental interference waves in each frequency band, and obtain the interference wave direction at the calculation sample point; Select several windows centered on the calculation sample point to perform median filtering, and restore the environmental interference signal at the calculation sample point; The interference signals in the wave field records of each frequency band are restored one by one, and the interference wave field is converted back to the time domain using inverse wavelet transform to obtain the entire wave field information of the interference wave; The identified interference wave field is removed from the original real-time cloth tension data to obtain the denoised effective signal wave field, and the denoised cloth tension feature data is generated based on the denoised effective signal wave field.

[0032] It should be explained that multiple sensors, such as tension sensors, are deployed at key locations of the interlining production line, each of which can monitor and record the tension of the fabric in real time. These sensors send data to the backend management in real time via industrial Ethernet or wireless communication protocols (such as Wi-Fi or Zigbee). The time series data collected from the sensors are converted into frequency domain data using Fourier transform to identify and analyze environmental noise generated by mechanical vibration or electromagnetic interference. A specific frequency band window is selected to perform median filtering on the interference wave. This processing method can effectively remove short-term interference without destroying the signal edge. The wave field records of each frequency band are analyzed using wavelet transform, the interference signals are restored one by one, and the interference wave field is converted back to the time domain through inverse wavelet transform to obtain the complete wave field information of the interference wave. The identified and restored interference wave is removed from the original fabric tension data, and the data is further cleaned using signal processing techniques (such as low-pass filtering) to generate a de-noised effective signal wave field.

[0033] Data prediction module 2 is used to establish a tension prediction model based on the obtained cloth tension characteristic data, and use the tension prediction model to predict the future cloth tension to obtain the cloth tension change trend; It should be explained that the cloth tension change trend refers to a general description of how the cloth tension will change in the future based on the existing data, which is inferred by the prediction model, including: Directionality: Tension trends may indicate whether tension is rising, falling, or remaining stable, which can help predict upcoming production challenges or production parameters that need adjustment.

[0034] Fluctuation: Describes the degree of fluctuation in tension changes, for example, whether there are periodic fluctuations or sudden peaks. This information is critical for optimizing machine settings and preventing potential equipment failures.

[0035] Forecast time frame: Trends are often assigned a forecast time frame, such as short-term (minutes to hours), medium-term (hours to days), or long-term (days to weeks), which helps adjust production on different time scales.

[0036] Accuracy and confidence: Trend forecasts often include an assessment of the forecast’s accuracy, such as a confidence interval or margin of error, which helps production managers assess the reliability of the forecast results and make decisions accordingly.

[0037] The central processing module 3 is used to analyze the obtained cloth tension change trend by using a pattern recognition algorithm to identify the cloth tension change pattern; It should be explained that the cloth tension change pattern refers to the representative tension change behavior and law, including: Periodic patterns: Fabric tension may show certain periodic fluctuations, which are usually related to factors such as the machine's operating cycle, the speed of the fabric passing through the production line, etc. The identification of periodic patterns can help adjust the frequency and speed of machine operation to keep the tension stable.

[0038] Trend pattern: refers to the upward or downward trend of fabric tension over a long period of time. For example, during continuous production, the tension may gradually increase or decrease due to machine wear or other external factors.

[0039] Sudden spikes or dips may occur in fabric tension data. These sudden spikes may be caused by equipment failure, operator error, or changes in raw material quality. Identifying these sudden changes can help prevent a large number of rejects by intervening in a timely manner.

[0040] Abnormal modes: These are tension patterns that do not conform to what is expected under normal production conditions and may be caused by equipment failure, improper operation, or external environmental factors such as temperature and humidity changes. By identifying abnormal modes, adjustments or maintenance can be made in a timely manner to ensure production continuity and product quality.

[0041] The adjustment control module 4 is used to evaluate the identified cloth tension change pattern using a factor evaluation algorithm, obtain key factors affecting the lining cloth production efficiency, and control the operation parameters of the production line based on the key factors; It should be explained that the key factors affecting the production efficiency of interlinings include: Machine settings: including machine operating speed, tension control settings, temperature control, etc. These parameters directly affect the production speed and fabric quality.

[0042] Raw material quality: The uniformity, strength and other physical characteristics of the raw material affect the fabric tension during the production process, which in turn affects the quality and yield of the final product.

[0043] Operating technology: The skill level and operating accuracy of operators are crucial to maintaining stable production efficiency. Misoperation may lead to production stagnation or quality problems.

[0044] Environmental conditions: Environmental factors such as temperature and humidity may affect the tension of the fabric and the operating status of the machine, thereby affecting production efficiency and product quality.

[0045] Maintenance and failure rate: The maintenance status and failure rate of the machine directly affect the stability and downtime of the production line. Timely maintenance can reduce unexpected downtime and improve production efficiency.

[0046] Supply chain efficiency: The timeliness of raw material supply and logistics efficiency are also key factors affecting production efficiency. Supply chain problems may lead to production delays or stoppages.

[0047] Energy and resource consumption: Energy efficiency and material utilization efficiency affect production costs and sustainability. Optimizing energy and resource consumption is an important aspect of improving production efficiency.

[0048] Wherein, the data management module 1 is connected to the central processing module 3 via the data prediction module 2, and the central processing module 3 is connected to the regulation control module 4; The data prediction module 2 includes: a model building module 201 and a model prediction module 202; A model building module 201 is used to build a tension prediction model based on the obtained cloth tension feature data; The model prediction module 202 is used to predict the future cloth tension by using the established tension prediction model to obtain the cloth tension change trend; Among them, the model building module 201 and the model prediction module 202 are connected.

[0049] Preferably, establishing a tension prediction model based on the obtained cloth tension characteristic data includes: Convert the obtained cloth tension characteristic data into a tension detection cycle; Substitute the converted tension detection period into the time series method, divide the obtained fabric tension data into periodic indicators and trend indicators, and build and train the tension prediction model through periodic indicators and trend indicators; It should be explained that cyclical indicators describe recurring patterns or cycles in data, which are usually related to seasonal changes, machine operation cycles, or production cycles. Cyclic indicators focus on identifying regular fluctuations in data, which can help us understand how data changes within a fixed time interval. They include: Cycle amplitude: The average level of data in different cycles, reflecting the typical value of data in the same season or cycle.

[0050] Cycle Fluctuation: The deviation of a data point within each cycle from the cycle mean, showing the magnitude and direction of fluctuations within the cycle.

[0051] It should be explained that trend indicators reflect the long-term changes or trends of data over time, whether these changes are rising, falling or stable. Trend indicators help understand the long-term development direction or growth pattern behind the data. They include: Long-term trend: The overall direction of movement of the data, which can be linear or nonlinear. This trend can be determined by calculating a moving average of the data or using a linear regression model.

[0052] Trend slope: describes the speed and direction of trend change, that is, the slope of the trend line, which reflects the speed and strength of the change.

[0053] The statistical error index (i.e., mean absolute percentage error) was used as the evaluation criterion for the accuracy of the tension prediction model.

[0054] It should be explained that fabric tension feature data, such as the average tension value per minute, are first extracted from the real-time data collected by the sensor. These data are then converted into a time series suitable for analysis, where each data point represents a specific detection cycle (for example, every 10 minutes). Time series methods (such as seasonally decomposed time series - STL) are used to decompose the time series data into periodic indicators and trend indicators. Periodic indicators capture regular changes within a fixed period, while trend indicators show long-term trends. Based on the decomposed periodic and trend indicators, a tension prediction model is constructed using statistical or machine learning methods (such as ARIMA models, random forests, or neural networks). During the training of the tension prediction model, the data set is divided into a training set and a test set to verify the prediction ability of the tension prediction model. The accuracy of the tension prediction model is evaluated by calculating the mean square error (MSE) and mean absolute percentage error (MAPE) between the predicted value and the actual tension value.

[0055] Preferably, the formula for the statistical error index is: ; In the formula, M Expressed as a statistical error indicator; a It is expressed as the total number of predicted tension values; It is expressed as the actual observed cloth tension value; Expressed as predicted cloth tension value.

[0056] Preferably, the established tension prediction model is used to predict the future cloth tension, and the cloth tension change trend is obtained including: Based on the trend index, the model optimization algorithm is used to optimize the parameters of the tension prediction model, and the optimal parameter configuration is found. The optimal parameter configuration is applied to the constructed tension prediction model, and the trend index is predicted using the prediction technology (i.e., the second-order adaptive coefficient method) to obtain the fabric tension trend at the future moment. The obtained future fabric tension trend is converted into actual production adjustments to construct the final tension prediction model.

[0057] It should be explained that it is assumed that the trend index of fabric tension has been obtained through time series analysis. Next, model optimization algorithms such as genetic algorithm (GA), particle swarm optimization (PSO), and dung beetle algorithm can be used to optimize the parameters of the tension prediction model. Through these algorithms, the parameter configuration that minimizes the prediction error can be found in the parameter space. After finding the optimal parameter configuration, the second-order adaptive coefficient method is used to predict the trend index. This method can dynamically adjust the parameters of the prediction model to adapt to the latest changes in the data and future trends, and formulate corresponding production adjustment strategies based on the trend prediction results. For example, if the prediction results show that the fabric tension will increase in the future, the set value of the tension controller can be adjusted in advance to avoid exceeding the safe tension range and ensure the smooth progress of production. Combined with the optimized model parameters and actual production adjustment feedback, the tension prediction model can be further calibrated and refined. Ensure that the application of the model in the actual production environment has high accuracy and reliability.

[0058] Preferably, based on the trend index, the parameters of the tension prediction model are optimized using a model optimization algorithm, and the optimal parameter configuration is found, and the optimal parameter configuration is applied to the constructed tension prediction model, and the trend index is predicted using a prediction technology, and the cloth tension trend at a future moment is obtained, including: Initialize the model optimization algorithm and the structural parameters of the tension prediction model; Normalize the fabric tension data and various environmental factor parameters; Update the parameter configuration of the tension prediction model using the model optimization algorithm; Update the optimal parameter configuration and fitness value, and determine whether the set conditions are met. If the conditions are not met, continue to use the model optimization algorithm to update the parameter configuration of the tension prediction model; The tension prediction model is trained, the global optimal parameters found by the model optimization algorithm are applied to the tension prediction model, and the prediction technology is used to predict the trend indicators to obtain the fabric tension trend at future moments.

[0059] It should be explained that a suitable model optimization algorithm is selected, such as the dung beetle algorithm, genetic algorithm (GA), particle swarm optimization (PSO), ant colony algorithm and other model optimization algorithms, and the structural parameters of the tension prediction model, such as the number of layers in the neural network and the number of neurons in each layer, are initialized, the cloth tension data and related environmental factors (such as temperature and humidity) are normalized so that their values ​​are between 0 and 1 to reduce the influence between different dimensions, and the parameters of the tension prediction model are iteratively optimized using the model optimization algorithm. The parameters are adjusted each iteration to improve the performance of the tension prediction model on the validation set. After each iteration, the fitness value of the tension prediction model (such as prediction error) is recorded and compared to see if there is any improvement. If the fitness value does not reach the predetermined threshold or reaches the upper limit of the number of iterations, the optimization is continued, and the best parameter configuration found by the model optimization algorithm is applied to the tension prediction model, and then the tension prediction model is trained, and the performance of the tension prediction model is verified using historical data. The trained tension prediction model is predicted using the second-order adaptive coefficient method, and special attention is paid to trend indicators to obtain future cloth tension trends.

[0060] Specifically, the model optimization algorithm is the Dung Beetle Algorithm (DBO), which is an optimization algorithm based on the behavior of insect dung beetles. Unlike traditional optimization algorithms (such as genetic algorithms, particle swarm algorithms, and ant colony algorithms), DBO has fast convergence, high precision, and high stability. The algorithm mainly obtains the global optimal weights and thresholds (i.e., global optimal parameters) through the position changes of dung beetles, using five behaviors of rolling, dancing, foraging, stealing, and reproduction, and passes the global optimal weights and thresholds (i.e., global optimal parameters) to the tension prediction model to complete the prediction of the fabric tension trend at future moments.

[0061] Preferably, the formula for updating the parameter configuration of the tension prediction model using the model optimization algorithm is: ; In the formula, Expressed as t +1 tension prediction model parameter configuration after iteration; Expressed as t Configuration of tension prediction model parameters at the iteration; Expressed as t -1st iteration tension prediction model parameter configuration; f It represents the direction of updating the parameters of the tension prediction model; k It is expressed as the amplitude of the tension prediction model parameter update; g It is expressed as the magnitude of parameter adjustment from the current worst tension prediction model parameter configuration; Represented as the global worst tension prediction model parameter configuration.

[0062] Preferably, the obtained cloth tension variation trend is analyzed by using a pattern recognition algorithm, and the cloth tension variation pattern is identified to include: Set the parameters of the pattern recognition algorithm and initialize the cloth tension change feature set; The fitness of each feature in the cloth tension change feature set is calculated to evaluate the matching degree with the cloth tension change trend; Among all fabric tension variation characteristics, identifying the best fabric tension variation characteristics and the worst fabric tension variation characteristics; For each cloth tension variation feature, the distance from the optimal cloth tension variation feature and the worst cloth tension variation feature is calculated respectively; Dividing each cloth tension variation characteristic into an optimal group and a worst group according to the proximity of each cloth tension variation characteristic to the optimal cloth tension variation characteristic and the worst cloth tension variation characteristic; The best group and the worst group are updated, and the updated best group and the worst group are recombined; Enter iterative optimization and repeat the iteration until the current number of iterations equals the maximum number of iterations; After the iteration is completed, the final set of cloth tension change features is analyzed and the cloth tension change pattern is identified.

[0063] It should be explained that genetic algorithms, simulated annealing algorithms, gray wolf algorithms and other algorithms are selected as optimization tools for pattern recognition. Initialize the feature set, which may include statistical parameters such as peak value, average value, standard deviation of tension, calculate how each feature matches the known tension change trend pattern, use fitness function (such as prediction accuracy or error minimization) to evaluate, select the best and worst cloth tension change features based on the fitness score, use Euclidean distance or other distance measurement methods to calculate the similarity or difference between each feature and the best feature and the worst feature, and divide the features into groups close to the best feature and groups close to the worst feature based on the proximity of the distance, update the group members through iteration, and reorganize the groups, continuously optimize the feature set, set the maximum number of iterations as the termination condition, and stop iterating after reaching the number, analyze the optimized feature set, and identify which features best represent the change pattern of cloth tension.

[0064] Specifically, the pattern recognition algorithm Dynamic Double-subgroup Strategy GWO (DDGWO) calculates the distance Dist between the current gray wolf individual and the α wolf during the iteration process. i_bestDist from the worst individual of the time i_worst , if Dist i_best ≤(Dist i_worst ) / 2, the gray wolf individuals are divided into a better subgroup with the contemporary optimal position α wolf as the geometric center; otherwise, they are divided into a worse subgroup with the contemporary worse individual as the center. According to the characteristics of the two subgroups, different nonlinear convergence factors are adopted to balance the local search ability and the global search ability. The traditional gray wolf algorithm coordinates the gray wolf population between global and local searches by adjusting the convergence factor. The convergence factor of the standard gray wolf algorithm decreases linearly, and it is difficult to better coordinate the global and local search capabilities. The pattern recognition algorithm of the present invention adopts different convergence factors for the better subgroup and the worse subgroup. For the better subgroup with higher fitness value, the nonlinear convergence factor value decreases rapidly in the early stage of iteration, so that the gray wolves with higher fitness quickly approach the optimal solution; the nonlinear convergence factor value decreases slowly in the later stage of iteration to avoid local optimality.

[0065] Preferably, the formula for calculating the distance from the optimal cloth tension change characteristic and the worst cloth tension change characteristic for each cloth tension change characteristic is: ; ; In the formula, Expressed as z The distance between the cloth tension change characteristics and the optimal cloth tension change characteristics (i.e., the distance between the gray wolf individual and the optimal individual in the gray wolf group in the dynamic dual-group strategy gray wolf algorithm); Expressed as z The distance between the cloth tension change characteristic and the worst cloth tension change characteristic (i.e., the distance between the gray wolf individual and the worst individual in the gray wolf group in the dynamic twin-group strategy gray wolf algorithm); It is represented as the position of the optimal cloth tension variation characteristic; It is represented as the location of the worst cloth tension variation characteristic; Expressed as z The location of the cloth tension change feature.

[0066] Preferably, the identified cloth tension variation pattern is evaluated using a factor evaluation algorithm to obtain key factors affecting the lining cloth production efficiency, and the operating parameters of the production line are controlled based on the key factors, including: Determine the parameters of factors that affect interlining production efficiency and set the initial search range; Set the initial evaluation period value and the number of iterations as the end condition of the evaluation process; Comparing the determined factor parameters affecting the interlining production efficiency with the known factor parameters of the optimized interlining production mode, and updating the factor parameters affecting the interlining production efficiency according to a preset comparison technology; Based on the cross-validation method, the production efficiency and operating parameters of the production line are evaluated after each factor parameter comparison; Specifically, the cross validation method is k-fold cross validation, which is a machine learning model evaluation method, which refers to dividing the factor parameter data set of the determined factor parameters affecting the lining production efficiency into k subsets, one of which is used as a validation set, and the remaining k-1 subsets are used as training sets.

[0067] Based on the evaluation results of production efficiency and operating parameters, the fitness score of each iteration is calculated; The fitness score after each iteration is compared with the fitness score of the previous iteration, and the factor parameter that has the greatest impact on the production efficiency of the lining cloth is selected as the current optimal solution; If the preset maximum number of iterations has been reached, the evaluation process is terminated; Analyze and evaluate the production results after the process is completed, and compare them with the preset production efficiency standards to determine whether the factor parameters have a significant impact on the lining cloth production efficiency. If so, make adjustments. If not, take the current factor parameters as the key factors affecting the lining cloth production efficiency, and adjust the production line operation parameters according to the key factors.

[0068] It is necessary to explain that factors that affect the production efficiency of interlinings, such as machine speed, tension setting, environmental conditions, etc., are identified. An initial search range is set for these factors, such as tension setting from 10 units to 15 units, each evaluation cycle is set to one week, and the total number of iterations is set to 10 times. After each iteration, the parameters are adjusted according to the production data, and the parameters of the new iteration factors are compared with the historically optimized parameters. Data analysis techniques such as deviation analysis are used to determine which parameter changes may lead to efficiency improvements. The parameter settings after each iteration are tested using cross-validation to ensure the stability and effectiveness of these settings under different production conditions. The fitness score of each iteration is calculated. Based on the performance of production efficiency and operating parameters, the parameters with the highest score are selected as the current best solution. Once the preset maximum number of iterations is reached, the evaluation process is terminated. After the iteration, all data and production results are analyzed and compared with the preset production efficiency standards to determine whether further adjustment of the operating parameters of the production line is required.

[0069] Specifically, the factor evaluation algorithm is an improved bird flock optimization algorithm (IBSO), and the improved bird flock optimization algorithm is used to select or optimize the parameters of the SVM to improve the performance of the model.

[0070] Specifically, the improved bird flock optimization algorithm refers to the improvement and variant of the traditional particle swarm optimization algorithm (PSO). Compared with the traditional bird flock optimization algorithm (BSO), the improved bird flock optimization algorithm (IBSO) adds the self-learning coefficient of the bird flock into the foraging behavior and changes the position update method of the beggars in the flying behavior.

[0071] It should be explained that the improved bird flock optimization algorithm (IBSO) includes the following improvements: 1) Improvement of speed and position update rules: In standard PSO, the movement of particles is based on their best historical position and the best position of the entire group. Improved algorithms may use different rules to adjust the speed and position of particles to improve convergence speed and avoid falling into local optimality.

[0072] 2) Parameter adjustment and control: Improved algorithms may include dynamic adjustment mechanisms for PSO parameters (such as inertia weight, individual best and group best learning factors) to balance exploration (global search) and exploitation (local search) during the search process.

[0073] Preferably, based on the cross-validation method, evaluating the production efficiency and the operating parameters of the production line after each factor parameter comparison includes the following steps: Initialize the first evaluation of the operating parameters and production efficiency of the lining cloth production line, and divide the parameter data set of factors affecting the lining cloth production efficiency into several subsets. In each evaluation, select one subset as the test set, and the union of the remaining subsets as the training set; The validation model (i.e., SVM model) is trained using the adjusted parameters of the factors affecting the interlining production efficiency on the training set, and the prediction performance of the validation model on the production efficiency and the operating parameters of the production line is tested on the test set; According to the actual performance of the parameters of the factors affecting the production efficiency of interlining cloth on the test set, the generalization error of the verification model is calculated; The evaluation is repeated until each subset has been used as a test set, and the average of the generalization errors calculated in all evaluations is obtained as the overall evaluation of the production efficiency and production line operation parameters after each factor parameter comparison.

[0074] It should be explained that the operating parameters and production efficiency data on the interlining production line are first collected. Then, these data are divided into several subsets, usually five or ten subsets, for cross-validation. In each evaluation, one subset is selected as the test set, and the remaining union is used as the training set. The parameters of the factors affecting production efficiency are adjusted using the data on the training set, and these parameters are used to train the verification model. The verification model is tested on the test set, and then the generalization error is calculated based on the difference between the production efficiency predicted by the model and the actual production efficiency. The generalization error is a key indicator for evaluating the performance of the model on new data. A lower generalization error means that the model can effectively predict the performance of unknown data, which is an important measure of the performance of the model until each subset has been used as a test set. Then, the average of the generalization errors in all evaluations is calculated as an overall evaluation of the production efficiency and production line operating parameters after the comparison of the influencing factor parameters.

[0075] In summary, with the help of the above technical scheme of the present invention, the present invention constructs a tension prediction model based on the collected cloth tension feature data, which can accurately map the changing law of cloth tension. The effectiveness lies in the ability to convert real-time data into useful information, providing a solid foundation for prediction, and using the established tension prediction model to predict future cloth tension, provide future change trends, and help predict problems that may be encountered in production, so as to take measures in advance. By predicting future cloth tension trends, production parameters can be adjusted in time to reduce cloth defects caused by tension mismatch, so as to prevent errors before problems occur, thereby reducing material waste and costs. By setting and initializing the parameters of the pattern recognition algorithm and constructing a set of cloth tension change features, the system can effectively identify and distinguish different cloth tension change patterns, improve the accuracy and efficiency of the response, and by calculating the fitness of each cloth tension change feature, the system can evaluate the matching degree of these features with the cloth tension change trend, ensure that the cloth maintains the best tension state during the production process, and reduce the generation of defects. By classifying the cloth tension change features into the best group and the worst group, and continuously updating and recombining these groups, the dynamic adjustment of production parameters is realized. By identifying and applying the best cloth tension change features, cloth defects caused by improper tension, such as overstretching or breakage, can be reduced, thereby reducing the defective rate and material waste. At the same time, the consistency and high standards of cloth quality are ensured to meet the market demand for high-quality lining cloth. By determining the key factor parameters that affect the production efficiency of the lining cloth and setting the initial search range, the factor evaluation is more targeted and efficient, and the key adjustment points can be quickly identified. Through iterative evaluation cycles and cross-validation methods, the system continuously compares and updates the factor parameters that affect the production efficiency. This dynamic adjustment mechanism ensures that the production line always operates in the optimal state to adapt to the ever-changing production needs and environment. Through the calculated fitness score, the parameters of the factors that have the greatest impact on the lining production efficiency are selected for optimization, which helps to achieve continuous improvement of the production process, thereby ultimately achieving the goal of improving production efficiency and product quality. The system reduces waste caused by excessive or improper use of resources by accurately controlling and optimizing production parameters, which not only saves costs but also complies with sustainable production practices.

[0076] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. An online detection system for lining cloth production, characterized in that: The online detection system for lining cloth production includes: a data management module, a data prediction module, a central processing module and an adjustment control module; The data management module is used to obtain the real-time data of the fabric tension on the lining fabric production line and pre-process the real-time data of the fabric tension to obtain the fabric tension characteristic data; A data prediction module is used to establish a tension prediction model based on the obtained cloth tension characteristic data, and use the tension prediction model to predict the future cloth tension to obtain the cloth tension change trend; A central processing module is used to analyze the obtained cloth tension change trend by using a pattern recognition algorithm to identify the cloth tension change pattern; The regulating control module is used to evaluate the identified cloth tension variation pattern using a factor evaluation algorithm, obtain key factors affecting the lining cloth production efficiency, and control the operating parameters of the production line based on the key factors; Wherein, the data management module is connected to the central processing module via the data prediction module, and the central processing module is connected to the regulation control module; The data prediction module includes: a model building module and a model prediction module; The model building module is used to build a tension prediction model based on the obtained cloth tension feature data; The model prediction module is used to predict the future cloth tension by using the established tension prediction model to obtain the cloth tension change trend; Wherein, the model building module is connected to the model prediction module.

2. The online detection system for lining cloth production according to claim 1 is characterized in that: The establishing of a tension prediction model based on the obtained cloth tension characteristic data comprises: Convert the obtained cloth tension characteristic data into a tension detection cycle; Substitute the converted tension detection period into the time series method, divide the obtained fabric tension data into periodic indicators and trend indicators, and build and train the tension prediction model through periodic indicators and trend indicators; Statistical error indicators are used as evaluation criteria for the accuracy of the tension prediction model.

3. The online detection system for lining cloth production according to claim 2 is characterized in that: The formula for the statistical error index is: ； In the formula, M Expressed as a statistical error indicator; a It is expressed as the total number of predicted tension values; It is expressed as the actual observed cloth tension value; Expressed as predicted cloth tension value.

4. The online detection system for lining cloth production according to claim 1 is characterized in that: The method of using the established tension prediction model to predict the future cloth tension and obtain the cloth tension change trend includes: Based on the trend indicators, the model optimization algorithm is used to optimize the parameters of the tension prediction model, and the optimal parameter configuration is found. The optimal parameter configuration is applied to the constructed tension prediction model, and the trend indicators are predicted using prediction technology to obtain the fabric tension trend at future moments. The obtained future fabric tension trend is converted into actual production adjustments to construct the final tension prediction model.

5. The online detection system for lining cloth production according to claim 4 is characterized in that: Based on the trend index, the model optimization algorithm is used to optimize the parameters of the tension prediction model, and the best parameter configuration is found. The best parameter configuration is applied to the constructed tension prediction model, and the trend index is predicted using prediction technology to obtain the fabric tension trend at the future moment, including: Initialize the model optimization algorithm and the structural parameters of the tension prediction model; Normalize the fabric tension data and various environmental factor parameters; Update the parameter configuration of the tension prediction model using the model optimization algorithm; Update the optimal parameter configuration and fitness value, and determine whether the set conditions are met. If the conditions are not met, continue to use the model optimization algorithm to update the parameter configuration of the tension prediction model; The tension prediction model is trained, the global optimal parameters found by the model optimization algorithm are applied to the tension prediction model, and the prediction technology is used to predict the trend indicators to obtain the fabric tension trend at future moments.

6. The online detection system for lining cloth production according to claim 5 is characterized in that: The formula for updating the parameter configuration of the tension prediction model using the model optimization algorithm is: ； In the formula, Expressed as t +1 tension prediction model parameter configuration after iteration; Expressed as t Configuration of tension prediction model parameters at the iteration; Expressed as t -1st iteration tension prediction model parameter configuration; f It represents the direction of updating the parameters of the tension prediction model; k It is expressed as the amplitude of the tension prediction model parameter update; g It is expressed as the magnitude of parameter adjustment from the current worst tension prediction model parameter configuration; Represented as the global worst tension prediction model parameter configuration.

7. The online detection system for lining cloth production according to claim 1 is characterized in that: The method of analyzing the obtained cloth tension variation trend by using a pattern recognition algorithm and identifying the cloth tension variation pattern includes: Set the parameters of the pattern recognition algorithm and initialize the cloth tension change feature set; The fitness of each feature in the cloth tension change feature set is calculated to evaluate the matching degree with the cloth tension change trend; Among all fabric tension variation characteristics, identifying the best fabric tension variation characteristics and the worst fabric tension variation characteristics; For each cloth tension variation feature, the distance from the optimal cloth tension variation feature and the worst cloth tension variation feature is calculated respectively; Dividing each cloth tension variation characteristic into an optimal group and a worst group according to the proximity of each cloth tension variation characteristic to the optimal cloth tension variation characteristic and the worst cloth tension variation characteristic; The best group and the worst group are updated, and the updated best group and the worst group are recombined; Enter iterative optimization and repeat the iteration until the current number of iterations equals the maximum number of iterations; After the iteration is completed, the final set of cloth tension change features is analyzed and the cloth tension change pattern is identified.

8. The online detection system for lining cloth production according to claim 7 is characterized in that: The formula for calculating the distance between each cloth tension variation characteristic and the optimal cloth tension variation characteristic and the worst cloth tension variation characteristic is: ； ； In the formula, Expressed as z The distance between a cloth tension variation characteristic and the optimal cloth tension variation characteristic; Expressed as z The distance between the cloth tension variation characteristic and the worst cloth tension variation characteristic; It is represented as the position of the optimal cloth tension variation characteristic; It is represented as the location of the worst cloth tension variation characteristic; Expressed as z The location of the cloth tension change feature.

9. The online detection system for lining cloth production according to claim 1 is characterized in that: The factor evaluation algorithm is used to evaluate the identified cloth tension change pattern to obtain the key factors affecting the lining cloth production efficiency, and the operating parameters of the production line are controlled based on the key factors, including: Determine the parameters of factors that affect interlining production efficiency and set the initial search range; Set the initial evaluation period value and the number of iterations as the end condition of the evaluation process; Comparing the determined factor parameters affecting the interlining production efficiency with the known factor parameters of the optimized interlining production mode, and updating the factor parameters affecting the interlining production efficiency according to a preset comparison technology; Based on the cross-validation method, the production efficiency and operating parameters of the production line are evaluated after each factor parameter comparison; Based on the evaluation results of production efficiency and operating parameters, the fitness score of each iteration is calculated; The fitness score after each iteration is compared with the fitness score of the previous iteration, and the factor parameter that has the greatest impact on the production efficiency of the lining cloth is selected as the current optimal solution; If the preset maximum number of iterations has been reached, the evaluation process is terminated; Analyze and evaluate the production results after the process is completed, and compare them with the preset production efficiency standards to determine whether the factor parameters have a significant impact on the lining cloth production efficiency. If so, make adjustments. If not, take the current factor parameters as the key factors affecting the lining cloth production efficiency, and adjust the production line operation parameters according to the key factors.

10. An online detection system for lining cloth production according to claim 9, characterized in that: The method of evaluating the production efficiency and the operating parameters of the production line after each factor parameter comparison based on the cross-validation method includes the following steps: Initialize the first evaluation of the operating parameters and production efficiency of the lining cloth production line, and divide the parameter data set of factors affecting the lining cloth production efficiency into several subsets. In each evaluation, select one subset as the test set, and the union of the remaining subsets as the training set; The validation model is trained using the adjusted parameters of the factors affecting the interlining production efficiency on the training set, and the prediction performance of the validation model on the production efficiency and the operating parameters of the production line is tested on the test set; According to the actual performance of the parameters of the factors affecting the production efficiency of interlining cloth on the test set, the generalization error of the verification model is calculated; The evaluation is repeated until each subset has been used as a test set, and the average of the generalization errors calculated in all evaluations is obtained as the overall evaluation of the production efficiency and production line operation parameters after each factor parameter comparison.

Images