Intelligent textile fabric production line automatic control method and system

Abstract

The application provides an intelligent textile fabric production line automatic control method and system. It belongs to the field of automatic control technology. The method comprises the following steps: acquiring multi-dimensional sensing data of a textile fabric production line and textile fabric production whole-process data; based on the whole-process data, using system modeling and simulation technology, deeply analyzing the internal logic and mutual relationship of each link in the textile fabric production process, and constructing a precise textile fabric production topology structure model; through real-time data acquisition and intelligent scheduling optimization, the equipment operation state and production parameters can be adjusted in time according to market demand and actual production conditions, the idle equipment and production waiting time are reduced, and the overall operation efficiency of the production line is improved.

Description

TECHNICAL FIELD

[0001] The application provides an intelligent textile fabric production line automatic control method and system, and belongs to the technical field of automatic control. BACKGROUND

[0002] In the traditional textile fabric production mode, the control of the production line mostly depends on manual experience and established procedures. The adjustment of production parameters often lags behind, and it is difficult to respond quickly according to real-time production conditions and changing market demands. For example, in the fabric printing and dyeing link, the deployment of colors and the control of dyeing time often lead to unstable product quality due to errors in manual judgment; in terms of equipment scheduling, there is a lack of scientific planning, which easily causes equipment to be idle or overused, affecting the overall production efficiency. Moreover, the traditional method is difficult to effectively integrate and analyze the massive data in the production process, and cannot provide accurate support for production decisions. With the development of the textile industry towards intelligence and individualization, the existing production control method has been difficult to meet the market demand, and an innovative intelligent textile fabric production line automatic control method is urgently needed. SUMMARY

[0003] The application provides an intelligent textile fabric production line automatic control method and system to solve the problems mentioned in the background art.

[0004] The application provides an intelligent textile fabric production line automatic control method, which comprises the following steps:

[0005] S1: Obtain multi-dimensional sensing data of the textile fabric production line and textile fabric production whole-process data; based on these whole-process data, use system modeling and simulation technology to deeply analyze the internal logic and mutual relationship of each link in the textile fabric production process, and construct a precise textile fabric production topology structure model;

[0006] S2: Based on a data fusion algorithm, deeply fuse the structure information in the textile fabric production topology structure model with the multi-dimensional sensing data, and construct an intelligent sensing network covering the whole production line; based on the intelligent sensing network, use big data analysis and machine learning algorithms, combine multi-source information, construct a market demand prediction model, and generate a textile fabric market demand prediction data set;

[0007] S3: Use a neural network algorithm to dynamically optimize the production process link parameters in the textile fabric production topology structure model; in the optimization process, through continuous iterative calculation, the optimal production process parameter combination is found, and a process optimization data set is formed;

[0008] S4: continuously collecting real-time production line sensor data; comparing the real-time data with the process optimization data set in real time, using the quality detection model to monitor and evaluate the product quality in the production process in real time; once the quality problem is found, the problem is located and the reason is found, and the quality detection data set is generated;

[0009] S5: considering the product quality reflected by the quality detection data set and the market demand change embodied by the market demand prediction data set, using an intelligent scheduling algorithm to dynamically schedule and optimize the production line equipment in the textile fabric production topology structure model.

[0010] The intelligent textile fabric production line automatic control system provided by the application comprises:

[0011] One or more processors;

[0012] Memory for storing one or more programs,

[0013] Wherein, when the one or more programs are executed by the one or more processors, the one or more processors implement the method of any one of the above.

[0014] The application has the following advantages: through real-time data collection and intelligent scheduling optimization, the running state and production parameters of the equipment can be adjusted in time according to the market demand and actual production situation, the idle time and production waiting time of the equipment are reduced, and the overall operation efficiency of the production line is improved. Based on real-time sensor data and process optimization data set for quality detection, quality problems in the production process can be found and adjusted in time, the stability and consistency of the textile fabric product quality are ensured, and the market competitiveness of the product is improved. Combined with market demand prediction for production process parameter optimization and equipment scheduling optimization, the textile fabric production can better adapt to market changes, quickly respond to market demand, reduce inventory risk, and improve the economic benefit of enterprises. The whole control process relies on advanced data analysis algorithm and intelligent control platform, realizes automatic decision and execution of the production process, reduces manual intervention, reduces labor intensity, and improves the intelligent level of production. BRIEF DESCRIPTION OF DRAWINGS

[0015] Figure 1 The method steps of the application are shown in the following figure:

[0016] Figure 2 The method steps of the application are shown in the following figure: Figure 1 The detailed steps of S4 in the application are shown in the following figure. DETAILED DESCRIPTION

[0017] The preferred embodiments of the application are described below with reference to the accompanying drawings, and it should be understood that the preferred embodiments described herein are only used to illustrate and explain the application, and are not used to limit the application.

[0018] One embodiment of the present application, as shown in Figure 1 includes the following steps:

[0019] S1: Obtain multi-dimensional sensing data of the textile fabric production line and textile fabric production whole-process data; based on these whole-process data, use system modeling and simulation technology to deeply analyze the internal logic and mutual relationship of each link in the textile fabric production process, and construct a precise textile fabric production topology structure model;

[0020] S2: Based on the data fusion algorithm, deeply fuse the structure information in the textile fabric production topology structure model with the multi-dimensional sensing data, and construct an intelligent sensing network covering the whole production line; based on the intelligent sensing network, use big data analysis and machine learning algorithms, combine multi-source information, construct a market demand prediction model, and generate a textile fabric market demand prediction data set;

[0021] S3: Use neural network algorithm to dynamically optimize the production process link parameters in the textile fabric production topology structure model; in the optimization process, through continuous iterative calculation, find the optimal production process parameter combination, and form a process optimization data set;

[0022] S4: Continuously collect real-time production line sensing data; compare and analyze the real-time data with the process optimization data set in real time, use the quality detection model to monitor and evaluate the product quality in the production link in real time; once the quality problem is found, locate the link and reason where the problem occurs, and generate a quality detection data set;

[0023] S5: Considering the product quality reflected by the quality detection data set and the market demand change embodied by the textile fabric market demand prediction data set, use intelligent scheduling algorithm to dynamically schedule and optimize the production line equipment in the textile fabric production topology structure model.

[0024] The working principle of the above technical solution is: obtaining multi-dimensional sensing data of the textile fabric production line and textile fabric production whole-process data covering factors such as raw material characteristics, production process, and finished product specifications; based on these whole-process data, using system modeling and simulation technology, the internal logic and mutual relationship of each link in the textile fabric production process are deeply analyzed, and a precise textile fabric production topology model is constructed; this model can intuitively present the entire production path from raw material input to finished product output, as well as the data flow and influence mechanism between nodes; for example, in the fiber processing link, sensors collect data such as fiber humidity, length, fineness, etc.; in the spinning stage, twist, strength, uniformity, etc. Information of yarn is recorded; in the weaving process, the warp and weft density, tension, etc. Parameters of the fabric are obtained; in the printing and dyeing link, data such as dye concentration, temperature, time, and fabric color fastness are monitored. At the same time, collect the raw material characteristics report provided by the raw material supplier, the standard operation process of the production process, and the requirements of the customer for the finished product specifications, etc. Whole-process data to build a comprehensive and accurate textile fabric production topology model.

[0025] Based on the data fusion algorithm, the structure information in the textile fabric production topology model and the multi-dimensional sensing data are deeply fused to construct an intelligent sensing network covering the entire production line; this network can accurately perceive the running state and environmental changes of the production line in real time; based on the intelligent sensing network, using big data analysis and machine learning algorithms, combined with historical sales data, market trends, seasonal factors, and other multi-source information, a market demand prediction model is constructed, and a textile fabric market demand prediction dataset is generated; this dataset can accurately predict the market demand of different specifications and styles of textile fabrics in the future period of time, providing forward-looking guidance for production decisions; for example, through the analysis of textile fabric sales data in different seasons and different regions in the past few years, combined with the current fashion trend and changes in consumer preferences, the demand for a certain functional sports fabric in the eastern region in the next quarter and the sales trend of a specific color curtain fabric in the northern market are predicted.

[0026] With the textile fabric market demand prediction dataset as the guide, the neural network algorithm is used to dynamically optimize the production process link parameters in the textile fabric production topology structure model. In the optimization process, full consideration is given to production efficiency, product quality, energy consumption and other target factors, and through continuous iterative calculation, the optimal production process parameter combination is found to form the process optimization dataset. The dataset is transmitted to the textile fabric production control platform, and the control platform automatically adjusts the running state of the production equipment according to these parameters to realize precise parameter adjustment of the production process and ensure that the production process is always in the best state. For example, according to market demand prediction, the demand for a high-count pure cotton fabric will increase significantly. Through intelligent optimization algorithm, the speed of spinning equipment, twist, and the density of warp and weft of weaving equipment are optimized and adjusted to improve production efficiency and reduce energy consumption while ensuring product quality.

[0027] Relying on the intelligent sensing network of the textile fabric production line, real-time production line sensing data is continuously collected; these real-time data are compared and analyzed with the process optimization dataset in real time, and quality detection models (such as appearance quality detection models based on image recognition and internal quality detection models based on sensor data) are used to monitor and evaluate the product quality in the production link in real time; once quality problems are found, the system can quickly locate the link and reason for the problem and generate a quality detection dataset; the dataset contains detailed information such as the type, location, and severity of the quality problem, providing a basis for subsequent quality improvement and production adjustment; for example, during the printing and dyeing process, the color of the fabric is monitored in real time by a color sensor, and when the color deviation exceeds the set range, the system immediately records the information of the location and analyzes the reasons such as dye ratio, dyeing time, or equipment failure, and the relevant information is included in the quality detection dataset; considering the product quality reflected by the quality detection dataset and the market demand changes embodied in the textile fabric market demand prediction dataset, intelligent scheduling algorithm is used to dynamically schedule and optimize the production line equipment in the textile fabric production topology structure model. In the optimization process, full consideration is given to factors such as the running state of the equipment, the maintenance cycle, and the production capacity, and the production tasks and operation sequence of the equipment are reasonably arranged to improve the utilization rate and production efficiency of the equipment. The optimized equipment scheduling scheme is arranged to form the equipment scheduling optimization dataset, which is transmitted to the textile fabric production control platform, and the control platform executes the equipment scheduling task according to the dataset to realize intelligent and automated scheduling of the production line equipment. For example, according to the quality detection dataset, it is found that a weaving equipment produces a high rate of defective products, and combined with the market demand prediction dataset, the production task of the equipment is adjusted to the production of products with relatively low quality requirements, and the production task of high-quality products is allocated to equipment with more stable performance to ensure that production orders are completed on time and with quality.

[0028] The effect of the above technical solution is that through intelligent sensing network and data optimization algorithm, the production process can be dynamically adjusted according to real-time market demand and quality data. Intelligent scheduling and process optimization can effectively reduce idle time in the production link, improve equipment utilization, and thus significantly improve production efficiency. By optimizing equipment operating parameters (such as spinning equipment speed, weaving equipment warp and weft density, etc.) in the production link, energy consumption can be minimized to the maximum extent while ensuring fabric quality. This not only reduces production costs, but also has a positive impact on environmental protection.

[0029] Based on multi-dimensional sensing data and image recognition quality detection model, the quality of the fabric can be monitored in real time, and problems can be found and located to specific links. For example, in the printing and dyeing process, by detecting the color change of the fabric in real time, the color consistency of the product can be ensured, and the rework or customer complaints caused by quality problems can be reduced. Combined with market demand prediction data set, the market demand of different specifications and styles of textile fabrics in the future can be accurately predicted, so as to adjust the production strategy and ensure that the product can follow the market trend and improve the market adaptability of the product.

[0030] By optimizing the production process link, the system can reduce unnecessary waste in the production process while maintaining high quality, and improve the utilization rate of equipment through intelligent scheduling, reduce labor costs and machine maintenance costs. The system can respond to quality problems or market demand changes in production in real time and quickly adjust production strategies. Intelligent equipment scheduling can reasonably arrange the production order according to different production tasks, improve the flexibility and responsiveness of the production line.

[0031] By accurately predicting market demand, the produced fabric can meet the requirements of customers for products in terms of style, specification and quality. The accurate quality monitoring system can also ensure that the products received by customers are consistent and stable in quality, and improve the trust of customers in the brand. The whole system combines data modeling, real-time monitoring and machine learning algorithms to provide accurate data support and decision-making basis for management. Through continuous iteration and optimization of production parameters, the scientificity and foresight of the decision can be ensured.

[0032] One embodiment of the present application, the S1, comprises:

[0033] S11, a plurality of types of sensors are arranged at each key link of the textile fabric production line to obtain multi-dimensional sensing data; textile fabric production whole-process data is collected;

[0034] S12, using system modeling and simulation technology, based on the collected whole-process data, the internal logic and mutual relationship of each link in the textile fabric production process are analyzed; a textile fabric production system model comprising a plurality of subsystems is constructed;

[0035] S13, simulate the model by simulation software, observe the data flow and influence mechanism between each subsystem, and analyze the influence of different parameter changes on the production process and product quality;

[0036] S14, according to the results of system modeling and simulation analysis, construct a textile fabric production topology structure model; and give each node and edge in the model corresponding attributes and parameters.

[0037] The working principle of the above technical solution is that a plurality of types of sensors are arranged at each key link of the textile fabric production line to obtain multi-dimensional sensing data; for example, a temperature and humidity sensor is installed in the fiber preparation area to monitor the temperature and humidity of the fiber storage environment in real time, ensuring that the fiber quality is not affected by the environment; a yarn tension sensor and a speed sensor are arranged in the spinning section to accurately master the tension change and equipment speed in the yarn production process; a warp and weft density detection sensor and a fabric tension sensor are provided in the weaving workshop to obtain the warp and weft density and tension of the fabric in a timely manner; a temperature sensor, a color detection sensor and a pH value sensor are installed in the printing and dyeing area to monitor the temperature, color and pH value in the printing and dyeing process in real time. Through these sensors, various data in the textile fabric production process are comprehensively and accurately collected. The textile fabric production whole-process data covering raw material characteristics, production process, product specifications and other elements are collected; a data sharing mechanism is established with raw material suppliers to obtain detailed characteristic reports of raw materials, including information such as the type, length, fineness and strength of the fiber; the standard operation process of the production process is arranged to clearly define the process parameters and operation requirements of each process; the requirements of customers on product specifications, such as the width, thickness, weight, color and the like of the fabric, are collected;

[0038] The internal logic and mutual relationship of each link in the textile fabric production process are analyzed based on the collected whole-process data by using system modeling and simulation technology; a textile fabric production system model including raw material processing, spinning, weaving, printing and dyeing, finishing and other subsystems is constructed;

[0039] The model is simulated by simulation software to observe the data flow and influence mechanism between each subsystem, and analyze the influence of different parameter changes on the production process and product quality; for example, the influence of adjusting the speed of the spinning equipment on the yarn quality and production efficiency is simulated to provide a theoretical basis for optimizing the production process;

[0040] According to the results of system modeling and simulation analysis, a textile fabric production topology structure model is constructed; the model presents the entire production path from raw material input to finished product output in a graphical manner, clearly shows the connection relationship and data flow between each production link; and each node and edge in the model is given corresponding attributes and parameters, such as device name, process parameter, material information and the like, so that the model can accurately reflect the characteristics and laws of the actual production process.

[0041] The effect of the above technical solution is that by arranging various sensors at key links of the production line, the real-time monitoring and accurate collection of various key data in the production process are ensured. The application of temperature and humidity sensors ensures the stability of the fiber storage environment, thereby reducing the quality fluctuations caused by environmental fluctuations.

[0042] By installing yarn tension sensors, speed sensors, fabric tension sensors, etc., accurate monitoring of each production link is achieved, ensuring the stability and consistency of textiles. Through the use of system modeling and simulation technology, the production process can be simulated and adjusted in real time, improving the intelligent level of the production process and reducing the need for manual intervention.

[0043] Simulation analysis helps enterprises identify potential problems and optimize solutions in advance, reducing the trial-and-error costs and time waste in traditional manual adjustment. By simulating and analyzing the impact of different process parameters on product quality and production efficiency, the process settings of each link are optimized, improving the overall production efficiency.

[0044] Through full-process data collection and analysis, the influence of each link on the quality of finished products is controllable, thereby improving the consistency of product quality. By precisely controlling the process parameters of each link, waste of raw materials and energy is effectively reduced, improving production efficiency and resource utilization.

[0045] By providing accurate finished product specifications and meeting customer demands, the market adaptability and customer satisfaction of products are improved, enhancing the market competitiveness of enterprises. By constructing a topological structure model, enterprises can more intuitively understand the relevance of each link, helping management make more scientific and efficient decisions.

[0046] One embodiment of the present application, the S12, comprises:

[0047] The collected full-process data is cleaned and preprocessed, and data mining and feature extraction algorithms are used to extract key features from massive production data;

[0048] Each production link in the textile fabric production process is analyzed to determine the input, output, and interaction between each link and other links, and a production process logic diagram is drawn;

[0049] Based on the results of the analysis of the logical relationship between the production links, the textile fabric production process is divided into multiple relatively independent subsystems, and the boundaries and functions of each subsystem are determined, and the main production process and key equipment within the subsystem are determined;

[0050] For each divided subsystem, a system modeling method is used to construct its internal model; combined with the characteristics of the subsystem and the production data, a modeling tool and algorithm are selected;

[0051] Analyze the coupling relationship between each subsystem, that is, how the output of one subsystem affects the input of other subsystems, and how this influence is transmitted and amplified throughout the production process; use system dynamics to establish a coupling relationship model between subsystems to describe the data flow and interaction mechanism between subsystems;

[0052] Integrate each subsystem model and its coupling relationship model to build a textile fabric production system model containing multiple subsystems.

[0053] The working principle of the above technical solution is: clean and preprocess the collected full-process data, use data mining and feature extraction algorithms to extract key features from massive production data, such as raw material characteristic parameters, fluctuation range of process parameters of each process, product quality indicators, etc. These key features will be important inputs for subsequent modeling, which can more accurately reflect the internal laws of the production process. For example, by extracting the features of the raw fiber length data, analyzing its distribution law, and providing a basis for subsequent optimization of the spinning process;

[0054] Analyze each production link in the textile fabric production process, clarify the input, output and interaction relationship with other links of each link, and draw a production process logic diagram; for example, analyze the influence of the fiber preparation link on the spinning link, determine how factors such as fiber pretreatment method and moisture content affect yarn quality and spinning efficiency; study the connection between the spinning link and the weaving link, and clarify the mechanism of yarn quality indicators (such as yarn strength and twist) on fabric warp and weft density and fabric quality;

[0055] Based on the results of the analysis of the logical relationship of the production link, the textile fabric production process is divided into multiple relatively independent subsystems, such as raw material processing subsystem, spinning subsystem, weaving subsystem, printing and dyeing subsystem, finishing subsystem, etc., and the boundaries and functions of each subsystem are clarified, and the main production process and key equipment within the subsystem are determined; for example, the raw material processing subsystem is mainly responsible for the pretreatment of fiber cleaning, opening, carding, etc., to provide qualified fiber raw materials for subsequent spinning; the spinning subsystem includes processes such as cleaning, carding, drawing, roving, and spinning to process fibers into yarns that meet the requirements;

[0056] For each subsystem, a system modeling approach is used to construct its internal model. Modeling tools and algorithms are selected based on the characteristics of the subsystem and production data, such as physics-based modeling methods and data-driven modeling methods. During model construction, factors such as process parameters, equipment performance, and material characteristics within the subsystem are fully considered to ensure that the model accurately reflects the actual operation of the subsystem. For example, when constructing the model for the spinning subsystem, the impact of process parameters such as spinning equipment rotation speed, spindle speed, and roller pressure on yarn quality and production efficiency is considered, and corresponding mathematical models are established to describe the relationships between these parameters.

[0057] The coupling relationships between the various subsystems are analyzed, that is, how the output of one subsystem affects the input of other subsystems, and how this effect is transmitted and amplified throughout the production process. System dynamics is used to establish coupling relationship models between subsystems to describe the data flow and interaction mechanisms between subsystems. For example, the impact of yarn quality output from the spinning subsystem on the weaving subsystem is studied, and a correlation model is established between yarn quality indicators and quality indicators such as fabric warp and weft density and fabric defect rate. The weaving process is optimized by adjusting spinning process parameters.

[0058] By integrating the models of each subsystem and their coupling relationships, a textile fabric production system model is constructed, which includes multiple subsystems such as raw material processing, spinning, weaving, dyeing and finishing.

[0059] The effects of the above technical solution are as follows: By using data cleaning and feature extraction algorithms, key features can be extracted from massive amounts of production data, improving the ability to analyze patterns in the production process and providing a more accurate basis for subsequent optimization. By clarifying the relationships between various production stages, the operation and coordination of each stage are optimized, reducing variables and uncertainties in the production process.

[0060] By analyzing the input-output relationships in the production process and identifying key factors, various process parameters can be optimized, improving production efficiency and product quality. Drawing a production flow diagram visually illustrates each stage and data flow in the production process, enhancing its visibility and facilitating timely adjustments by management.

[0061] System modeling and simulation can predict and analyze the interactions between different stages in advance, reducing trial-and-error costs and resource waste in production, and lowering production costs. By establishing subsystem models and coupling relationship models, it is possible to better control each production stage and ensure that the output of each stage accurately meets expectations, thereby improving the stability of the entire production process.

[0062] By optimizing the process parameters and equipment operation of each link, unnecessary intervention and adjustment in the production process are reduced, thereby improving the overall production efficiency. Integrating various subsystem models can help management make more scientific and accurate decisions in optimizing the production process, further improving the decision-making efficiency of the enterprise.

[0063] Through precise subsystem modeling and interaction analysis, the quality fluctuation between links can be reduced, and the quality stability of finished products can be improved. Through the application of system modeling and simulation technology, enterprises can achieve continuous innovation in production process and equipment optimization, enhancing the market competitiveness and technological leadership of enterprises.

[0064] In one embodiment of the present application, the S2 comprises:

[0065] S21, based on a data fusion algorithm, deeply fusing the structure information in the textile fabric production topology structure model with multi-dimensional sensing data; preprocessing the deeply fused data; and using the data fusion algorithm to associate and integrate the preprocessed sensing data with the structure information in the model, to build an intelligent sensing network covering the entire production line;

[0066] S22, collecting multi-source information, the multi-source data including historical sales data, market trends and seasonal factors, and integrating the multi-source data with the data collected by the intelligent sensing network;

[0067] S23, based on the integrated data, using big data analysis and machine learning algorithms to build a market demand prediction model; using the built market demand prediction model to predict the market demand of different specifications and styles of textile fabrics in a future period of time; and generating a textile fabric market demand prediction dataset according to the prediction results.

[0068] The working principle of the above technical solution is: based on a data fusion algorithm, the structure information in the textile fabric production topology structure model is deeply fused with multi-dimensional sensing data; the deeply fused data is preprocessed; and the preprocessed sensing data is associated and integrated with the structure information in the model using the data fusion algorithm, to build an intelligent sensing network covering the entire production line; this network can accurately perceive the running state and environmental changes of the production line in real time, providing a comprehensive and accurate data basis for subsequent market demand prediction;

[0069] Collect multi-source information, including historical sales data, market trends, and seasonal factors, and integrate them with the data collected by the intelligent sensing network; for example, obtain the sales data of textile fabrics in the past few years from the enterprise's sales system, including the sales volume and sales of different specifications and styles of fabrics in different regions and time periods; pay attention to industry dynamics and market research reports to understand the development trend and emerging demand of the textile fabric market; consider the impact of seasonal factors on textile fabric sales, such as increased demand for lightweight and breathable fabrics in summer and increased demand for warm fabrics in winter. Uniformly encode and store these multi-source information to establish a comprehensive information database;

[0070] Based on the integrated data, use big data analysis and machine learning algorithms to build a market demand prediction model; use the built market demand prediction model to predict the market demand for different specifications and styles of textile fabrics in the future; generate a textile fabric market demand prediction dataset based on the prediction results, which contains information such as predicted time range, fabric specification, style, and predicted market demand; this dataset can provide forward-looking guidance for production decisions, helping enterprises to reasonably arrange production plans and inventory management, and reduce market risks.

[0071] The effect of the above technical solution is: through the deep integration of data fusion algorithm and multi-dimensional sensing data, the running state and environmental changes of the production line can be accurately perceived in real time, providing a more comprehensive and accurate data basis for subsequent market demand prediction. Through real-time data monitoring of the intelligent sensing network, changes on the production line can be better handled, reducing decision-making errors caused by information lag, and thus reducing the risk in production.

[0072] Integrating multi-source information (such as historical sales data, market trends, and seasonal factors) and combining big data analysis and machine learning algorithms can help improve the accuracy of market demand prediction and provide forward-looking guidance for enterprise production decisions. Based on the demand prediction dataset, enterprises can more accurately plan production and manage inventory, thereby avoiding overproduction or inventory accumulation and improving overall production and supply chain efficiency.

[0073] By incorporating seasonal factors and market trends into the prediction model, enterprises can identify changes in market demand trends in advance, reducing the negative impact of market demand fluctuations. By uniformly encoding and storing multi-source data, a comprehensive information database is established, enhancing the efficiency of information integration and storage and providing a reliable foundation for subsequent data analysis and decision-making.

[0074] The construction of intelligent sensing networks and the application of data fusion technology have improved the automation level of production lines, enabling intelligent monitoring and optimization of the production process, and reducing the need for manual intervention. Precise market demand forecasting enables enterprises to reasonably arrange production and inventory, reducing unnecessary production and inventory accumulation, and reducing resource waste.

[0075] By optimizing production planning and inventory management, and adjusting production strategies in a timely manner based on market demand forecasts, enterprises can better respond to changes in market demand, thereby enhancing their competitiveness in a highly competitive market. By combining historical data, market trends, and seasonal factors, and using big data and machine learning models to forecast market demand, the application value of big data analysis technology in actual production is improved.

[0076] In one embodiment of the present application, the S21 comprises:

[0077] The structural information in the textile fabric production topology model is analyzed, key structural features are extracted, and multi-dimensional sensing data is feature-extracted. Different feature extraction methods are used for different types of sensor data.

[0078] The extracted features are processed for spatio-temporal alignment. In the time dimension, the sensing data collected at different times is interpolated or resampled, and in the spatial dimension, the sensing data is spatially mapped and correlated according to the geographical location of the equipment and the spatial layout of the production links in the production topology model.

[0079] The heterogeneous data is converted and standardized. For structural information, it is converted into a unified data structure, and for sensing data, it is converted into a standard numerical or categorical format according to its type and characteristics.

[0080] The pre-processed structural information and sensing data are preliminarily fused. According to the characteristics of the data and the fusion target, the parameters and weights of the algorithm are adjusted to obtain the best fusion effect.

[0081] Based on the preliminary fusion, deep learning is used for deep fusion of structural information and sensing data. A deep neural network model is constructed, with structural features and sensing features as inputs, through multi-layer neuron nonlinear transformation and feature extraction, to mine deep-level associations and internal laws between data.

[0082] Based on the data obtained through deep fusion and association integration, an intelligent sensing network covering the entire production line is constructed, and the functions and data transmission rules of each node in the network are defined.

[0083] The working principle of the above technical solution is: analyzing the structure information in the textile fabric production topology structure model, extracting key structural features such as the topological order of production links, the connection weight between links, the position and role of equipment in the production process, etc. These structural features can reflect the overall architecture and operation logic of the production system, and feature extraction is performed on multi-dimensional sensor data. Different feature extraction methods are used for different types of sensor data; for example, for temperature and humidity sensor data, statistical features such as mean, variance, maximum, minimum, etc. are extracted to reflect the changes in environmental temperature and humidity; for yarn tension sensor data, features such as tension fluctuation frequency and peak value are extracted to analyze the tension stability in the yarn production process. Through data feature analysis and extraction, more representative and discriminant feature vectors are provided for subsequent data fusion;

[0084] Temporal and spatial alignment processing is performed on the extracted features; in the time dimension, interpolation or resampling processing is performed on the sensor data collected at different times to make them have uniform timestamps, ensuring the synchronization of data in time. For example, if the sampling frequency of some sensors is low and the sampling frequency of other sensors is high, the low-frequency data is converted to the same time resolution as the high-frequency data through interpolation method for comprehensive analysis; in the spatial dimension, according to the geographical position of the equipment in the production topology structure model and the spatial layout of the production links, the sensor data is spatially mapped and associated; for example, the data of the warp and weft density detection sensors and the fabric tension sensors distributed in different positions of the weaving workshop are corresponding to the position of the weaving link in the production topology structure, and the specific production position and area corresponding to the data are clear;

[0085] The heterogeneous data is converted and standardized; for structural information, it is converted into a unified data structure, such as storing the production topology structure in the form of a graph database for efficient query and analysis; for sensor data, according to the data type and characteristics, it is converted into a standard numerical format or classification format; for example, the RGB value of the color detection sensor is converted into a standard color code, and the data of the pH value sensor is converted into a value that meets the chemical standard. At the same time, the data is normalized to map different dimensional data to the same numerical range, eliminating the influence of data dimension on the fusion result and improving the accuracy and stability of data fusion;

[0086] The preprocessed structural information and sensor data are initially fused. Based on the characteristics of the data and the fusion objective, the algorithm parameters and weights are adjusted to obtain the optimal fusion effect. For example, when using the weighted average method, different weights are assigned to the structural information and sensor data according to their importance and reliability, and then a weighted sum is performed to obtain the fusion result. During the initial fusion process, the fusion performance of different algorithms is evaluated through experiments and comparative analysis, and the most suitable data fusion algorithm for this textile fabric production scenario is selected.

[0087] Building upon initial fusion, deep learning is employed to deeply integrate structural information and sensor data. A deep neural network model is constructed, using structural and sensor features as inputs. Through nonlinear transformations and feature extraction of multiple layers of neurons, deep correlations and inherent patterns between data are uncovered. For example, convolutional neural networks (CNNs) are used to fuse and analyze image data (such as equipment operating status images) and sensor data from the production topology, automatically extracting key features and patterns. Alternatively, recurrent neural networks (RNNs) and their variants (such as LSTM and GRU) are used to process time-series sensor data and structural information, capturing dynamic changes and correlations in the data over time. Through deep fusion, more comprehensive, accurate, and intelligently perceptive fused data is generated. Simultaneously, the fused data is integrated with other relevant information in the production topology model, such as equipment maintenance records and historical adjustment information of process parameters, further improving the data foundation of the intelligent sensor network and enabling it to more comprehensively and deeply reflect the operating status and environmental changes of the production line.

[0088] Based on deeply fused and integrated data, an intelligent sensor network covering the entire production line is constructed. The functions and data transmission rules of each node in the network are clearly defined to ensure efficient and accurate data flow. For example, the fused data is allocated to corresponding production node locations, the node's status information is updated in real time, and the information is transmitted to upstream and downstream nodes via the network, enabling real-time monitoring and collaborative control of the production process. Simultaneously, the intelligent sensor network is optimized by employing data compression and encryption technologies to reduce data transmission volume and ensure data security. Network topology optimization algorithms are used to improve network reliability and stability, ensuring continued operation even when some nodes fail. Through continuous optimization, the intelligent sensor network can perceive the production line's operating status and environmental changes in real time and accurately, providing a comprehensive and accurate data foundation for subsequent market demand forecasting.

[0089] The above technical solution achieves the following results: Through deep learning and multi-dimensional data feature extraction, the solution improves the data fusion efficiency at each stage and enhances the comprehensive analysis capabilities of different types of sensor data. By aligning, standardizing, and transforming data in the spatiotemporal dimensions, it reduces processing differences between heterogeneous data types, simplifying subsequent data analysis and model training. Adopting a unified data structure (as shown in the database storage structure information) and standardized processing effectively reduces the computational difficulty caused by different data formats and scales.

[0090] By constructing an intelligent sensor network and optimizing its topology, the solution enhances real-time monitoring and collaborative control of all aspects of the production process. Especially in the event of a failure, the network's high reliability and stability ensure uninterrupted data flow, facilitating rapid response to anomalies and guaranteeing stable production efficiency and quality.

[0091] By collecting and integrating various sensor data in real time during the production process, abnormal changes in key parameters such as yarn tension, temperature, and humidity can be identified promptly, providing real-time feedback for adjustments in the production process. This timely adjustment effectively avoids instability in the production process, improves production efficiency, and ensures product quality. By extracting key structural features (such as the topological sequence of production stages and the connection weights of each stage), the overall architecture and operational logic of the production system can be more clearly understood. This is of great significance for optimizing production processes, improving production line coordination, and increasing equipment utilization.

[0092] By applying data compression and encryption technologies, the solution effectively reduces data transmission volume while ensuring data security. This ensures the efficient and secure transmission of large amounts of sensor data during the production process without affecting real-time performance. Based on deep integration, the constructed intelligent sensor network can provide a more accurate data foundation for predicting future market demand. By analyzing historical data and production status, combined with deep learning models, it is possible to make faster and more accurate predictions of changes in market demand.

[0093] In one embodiment of the present invention, S3 includes:

[0094] S31. Using the market demand forecast dataset for textile fabrics as a guide, determine the objectives for optimizing parameters in the production process; based on the market demand forecast results, analyze the quality requirements and production demands for textile fabrics of different specifications and styles.

[0095] S32. Use neural network algorithms to dynamically optimize the parameters of the production process in the topology model of textile fabric production; use the process parameters in the production topology model as input variables of the neural network and the optimization objective function as output variables;

[0096] S33. By leveraging the self-learning and adaptive capabilities of neural networks, the values ​​of input variables are continuously adjusted. During the optimization process, the optimal combination of production process parameters is found through iterative calculations.

[0097] S34. Organize the optimized production process parameters into a process optimization dataset; and record and classify the process optimization dataset in detail.

[0098] S35. Transmit the process optimization dataset to the textile fabric production control platform; the control platform automatically adjusts the operating status of the production equipment according to the parameters.

[0099] The working principle of the above technical solution is as follows: Guided by a market demand forecast dataset for textile fabrics, the optimization objectives for parameters in the production process are determined; based on the market demand forecast results, the quality requirements and production demands for different specifications and styles of textile fabrics are analyzed; for example, if the market demand forecast shows a significant increase in demand for a certain high-count pure cotton fabric, the optimization objective can be set as improving the production efficiency of this fabric, ensuring its quality stability, and reducing energy consumption. Multiple optimization objectives are quantified and weighted to form a comprehensive optimization objective function.

[0100] A neural network algorithm is used to dynamically optimize the parameters of the production process in the topology model of textile fabric production; the process parameters in the production topology model are used as input variables of the neural network, and the optimization objective function is used as the output variable.

[0101] By leveraging the self-learning and adaptive capabilities of neural networks, the values ​​of input variables are continuously adjusted to optimize the output variables. During the optimization process, multiple target factors such as production efficiency, product quality, and energy consumption are fully considered, and the optimal combination of production process parameters is found through continuous iterative calculations. For example, neural network algorithms can be used to optimize parameters such as the speed and twist of spinning equipment, and the warp and weft density and tension of weaving equipment to achieve the optimization of the production process.

[0102] The optimized production process parameters are compiled into a process optimization dataset. This dataset contains the optimal process parameters for each production stage, such as the optimal speed and twist of spinning equipment, the optimal warp and weft density and tension of weaving equipment, and the optimal temperature and time of dyeing and printing equipment. The process optimization dataset is recorded and classified in detail.

[0103] The process optimization dataset is transmitted to the textile fabric production control platform. The control platform automatically adjusts the operating status of the production equipment based on the parameters, achieving precise parameter tuning of the production process. For example, based on the parameters in the process optimization dataset, the control platform automatically adjusts the motor speed and twist adjustment devices of the spinning equipment to bring the yarn production parameters to optimal values; it also adjusts the warp and weft density control devices and tension adjustment devices of the weaving equipment to ensure that the fabric quality meets requirements. Through precise parameter tuning by the control platform, the production process is always kept in optimal condition, improving production efficiency and product quality.

[0104] The above technical solution achieves the following effects: by applying neural network algorithms to optimize production process parameters, the solution realizes the automation and intelligence of the production process. The neural network can dynamically adjust production process parameters based on market demand forecast data, thereby improving the level of intelligent decision-making in production, reducing manual intervention, and increasing overall production efficiency.

[0105] By optimizing for the quality and production demands of different specifications and styles of textile fabrics, especially in reducing energy consumption, the solution can effectively reduce unnecessary energy consumption. For example, by optimizing parameters such as the speed and twist of spinning equipment, energy waste is reduced while ensuring product quality, thus lowering energy consumption during the production process.

[0106] Through the self-learning and adaptive capabilities of neural network algorithms, the solution can continuously adjust production process parameters to achieve optimal production efficiency and product quality. The optimized process parameters precisely adjust the equipment's operating status, ensuring optimal operation at each stage, thereby guaranteeing high efficiency and high quality in textile fabric production.

[0107] By quantifying and weighting multiple optimization objectives, a comprehensive optimization objective function is generated. This approach not only optimizes the production process but also significantly improves production efficiency. For example, the optimal process parameters for each production stage, such as spinning, weaving, and dyeing, ensure efficient equipment operation, thereby improving the overall efficiency of the production process. Through detailed recording and categorized storage of the process optimization dataset, the approach strengthens the standardization of data management. Optimization parameters for each stage are systematically archived, facilitating subsequent analysis and reuse, while also contributing to process standardization and improving the transparency and traceability of production management.

[0108] By precisely adjusting production process parameters, the solution ensures that quality requirements are consistently met at each stage of production. Each optimized parameter adjustment effectively avoids fluctuations in production, thereby improving product quality stability and ensuring that each batch of fabric meets market demands. The self-learning and adaptive capabilities of neural networks enable the production process to flexibly respond to changes in market demand. When market demand changes, the optimization objectives and parameters are adjusted accordingly, allowing the production process to adapt promptly to changes in demand, thus enhancing the adaptability and flexibility of the production system.

[0109] By transmitting optimized production process parameters to the production control platform, the solution enables precise control of the production process. The control platform adjusts the equipment's operating status based on the optimized parameters, ensuring that each stage of the production process is in optimal condition, thus improving production controllability and stability. Automated production parameter adjustments reduce the need for human intervention, thereby lowering the risk of operational errors. Through precise parameter tuning by the control platform, the adjustment process of production equipment becomes more accurate and reliable, reducing the risk of quality problems or decreased production efficiency caused by improper human operation.

[0110] One embodiment of the present invention, such as Figure 2 As shown, S4 includes:

[0111] S41. An intelligent sensor network based on the textile fabric production line continuously collects real-time production line sensor data and transmits the data to a data acquisition system; the data acquisition system performs preliminary processing and storage on the collected data.

[0112] S42. Perform real-time comparative analysis between real-time data and process optimization dataset; calculate the deviation between real-time data and process optimization data using a data comparison algorithm;

[0113] S43. Use a quality inspection model to monitor and evaluate product quality in the production process in real time; if a quality problem is found, the system will locate the link and cause of the problem and generate a quality inspection dataset.

[0114] The working principle of the above technical solution is as follows: Based on the intelligent sensor network of the textile fabric production line, real-time sensor data of the production line is continuously collected; each sensor in the intelligent sensor network collects data in real time according to the set sampling frequency and transmits the data to the data acquisition system; the data acquisition system performs preliminary processing and storage on the collected data to ensure the integrity and accuracy of the data. For example, the yarn tension sensor collects the tension change data of the yarn in real time during the production process and transmits it to the data acquisition system; the fabric tension sensor monitors the tension of the fabric in real time and provides timely feedback to the data acquisition system.

[0115] Real-time data is compared and analyzed with the process optimization dataset in real time. A data comparison algorithm is used to calculate the deviation between the real-time data and the process optimization data. If the deviation is within a preset range, the production process is considered normal; if the deviation exceeds a set threshold, it indicates a potential quality problem. For example, comparing the real-time collected yarn tension data with the optimal tension value in the process optimization dataset, if the deviation exceeds a certain proportion, it indicates a potential problem in the yarn production process.

[0116] The system employs quality inspection models to monitor and evaluate product quality in real time throughout the production process. These models include an image recognition-based appearance quality inspection model and a sensor data-based intrinsic quality inspection model. The image recognition-based appearance quality inspection model captures images of the textile fabric using cameras and applies image processing technology and pattern recognition algorithms to detect surface defects, stains, color differences, and other issues. The sensor data-based intrinsic quality inspection model uses real-time sensor data, such as yarn tension, fabric tension, and dyeing temperature, to evaluate the fabric's intrinsic quality indicators, such as strength, elasticity, and colorfastness. If a quality problem is detected, the system pinpoints the source and cause of the problem and generates a quality inspection dataset. This dataset contains detailed information such as the type, location, and severity of the quality problem. For example, if image recognition detects a defect on the fabric surface, the system can accurately record the defect's location coordinates, size, and shape, and analyze possible causes, such as equipment malfunction, unreasonable process parameters, or raw material quality issues. The quality inspection dataset is then fed back to the production management department and the quality improvement team, providing a basis for subsequent quality improvements and production adjustments.

[0117] The effects of the above technical solution are as follows: The application of neural network algorithms makes the adjustment of production processes more automated and intelligent. The system can dynamically adjust process parameters based on real-time data, thereby reducing manual intervention and improving the level of intelligent decision-making in the production process. By precisely optimizing the operating parameters of textile equipment, such as yarn tension, fabric tension, and equipment speed, the system can reduce unnecessary energy consumption, especially by reducing energy consumption and optimizing production efficiency without affecting product quality.

[0118] The self-learning and adaptive capabilities of neural networks enable precise adjustments to various process parameters during production based on real-time data. This ensures efficient production and improves product quality stability. By optimizing parameters across multiple production stages, the system ensures equipment operates at its best, enhancing production efficiency. In particular, comprehensive optimization of spinning, weaving, and dyeing processes further improves the efficiency of the entire production process.

[0119] The system meticulously records all process parameters and quality inspection data throughout the production process, providing a reliable basis for subsequent analysis, improvement, and quality traceability. This not only enhances the transparency of production management but also strengthens the ability to monitor potential problems during production. By adjusting process parameters in real time, the system can consistently guarantee the quality of each batch of products throughout the entire production process, avoiding quality fluctuations and ensuring that the fabric quality meets market demands and customer expectations.

[0120] When market demand changes, production process parameters are adjusted accordingly, enhancing the system's adaptability. This allows the production process to respond quickly to market changes and flexibly meet different production needs. Automated production processes reduce human intervention, lowering the risk of quality problems and low production efficiency caused by human error. Precise system control ensures the stability of the production process.

[0121] By integrating image recognition and sensor data into a quality inspection model, the system can monitor and evaluate product quality in real time, promptly identify problems, and pinpoint their causes. This makes quality control more precise, helps to identify and resolve problems in a timely manner, and prevents defective products from entering the market. With the help of an intelligent control platform, the system can precisely adjust the operating status of equipment based on optimized process parameters, ensuring that each link operates at its best, thereby improving the controllability and stability of the production process.

[0122] In one embodiment of the present invention, S41 includes:

[0123] S411. Analyze the temporal and spatial correlations between the data from each sensor, and determine the sampling time order and frequency of different sensors based on the analysis results;

[0124] S412. Based on the set sampling strategy, each sensor collects data in real time. During the collection process, the sensor converts the sensed physical quantity into an electrical signal and the analog signal into a digital signal. The sensor transmits the digital signal to the data acquisition system through a combination of various communication technologies. During the data transmission process, data encryption technology is used to encrypt the transmitted data.

[0125] S413. After receiving the data transmitted from the sensor, the data acquisition system performs preliminary processing.

[0126] S414. Store the pre-processed data into the data storage system; adopt a distributed storage architecture to distribute the data across multiple storage nodes, and classify and store the data according to its characteristics and frequency of use.

[0127] The working principle of the above technical solution is as follows: Analyze the temporal and spatial correlations between the data of each sensor, and determine the sampling time sequence and frequency of different sensors based on the analysis results. For example, in the spinning process, the data of the yarn tension sensor and the speed sensor are closely related. In order to accurately analyze the relationship between yarn tension and equipment speed, these two sensors are set to use synchronous sampling, that is, data is collected at the same time point. In the weaving workshop, the sampling frequency of the warp and weft density detection sensor and the fabric tension sensor can be set differently according to the production speed and quality control requirements of the fabric. The sampling frequency of the warp and weft density detection sensor can be appropriately reduced to reduce the amount of data processing, while the fabric tension sensor needs a higher sampling frequency to monitor the tension changes of the fabric in real time.

[0128] Based on the established sampling strategy, each sensor collects data in real time. During the collection process, the sensor converts the sensed physical quantities into electrical signals and analog signals into digital signals. The sensor then transmits the digital signals to the data acquisition system through a combination of various communication technologies. For example, for sensors that are close to each other and have a large amount of data, wired Ethernet communication is used to ensure high-speed and stable data transmission. For sensors on mobile devices or in areas where wiring is difficult, wireless communication technologies such as Wi-Fi, ZigBee, or dedicated wireless protocols for industrial IoT are used to achieve real-time data transmission. Furthermore, data encryption technology is used to encrypt the transmitted data during the data transmission process.

[0129] After receiving data from the sensors, the data acquisition system performs preliminary processing, including a data integrity check to ensure each data packet contains complete information, such as the sensor number, timestamp, and acquired data. If an incomplete data packet is found, a retransmission request is promptly sent to the sensor to ensure complete data acquisition. Next, the data validity is verified by judging whether the acquired data is reasonable based on pre-defined data ranges and logical rules. For example, for data acquired by temperature sensors, it is checked whether the temperature is within a reasonable range; for yarn tension data, it is determined whether it conforms to the normal tension variation pattern during yarn production. If abnormal data is found, it is marked and processed. Methods such as data smoothing and interpolation can be used to correct abnormal data, or the abnormal data can be stored separately for subsequent analysis and processing.

[0130] The pre-processed data is stored in a data storage system. A distributed storage architecture is employed, distributing data across multiple storage nodes. Data is categorized and stored according to its characteristics and usage frequency. For example, sensor data with high real-time requirements is stored in a cache for rapid access and analysis; historical and backup data are stored in disk arrays or cloud storage to ensure data security and long-term preservation. Simultaneously, a comprehensive data management system is established to index and label the stored data, facilitating subsequent data queries and retrieval. Each sensor data point is assigned a unique identifier, recording information such as the data acquisition time, source, and type, enabling refined data management. Through the data management system, the required data can be quickly located and retrieved, facilitating subsequent data analysis and processing.

[0131] The effects of the above technical solution are as follows: By applying neural network algorithms, the adjustment process of production technology becomes more intelligent and automated. The system can dynamically adjust process parameters based on real-time data, reducing manual intervention and thus improving the efficiency and accuracy of production decisions. Precise optimization of equipment operating parameters, especially yarn tension, fabric tension, and equipment speed, can significantly reduce energy consumption without affecting product quality. This optimization effectively reduces unnecessary energy consumption and improves energy efficiency.

[0132] The self-learning capability of neural networks enables process parameters in production to be automatically and precisely adjusted in real time based on data, ensuring stable operation of the production process and improving product quality consistency. Comprehensive optimization of multiple production stages improves equipment operating efficiency, ensuring that each stage of the production process operates at its optimal state, thereby enhancing overall production efficiency.

[0133] By meticulously recording all production processes and quality data, the system enhances the transparency of production management and ensures the reliability of quality traceability. This provides data support for subsequent analysis and improvement, thereby elevating the level of quality control. Through real-time adjustment of process parameters, the system ensures stable quality for each batch of products, avoiding quality fluctuations and guaranteeing the quality levels required by the market and expected by customers.

[0134] When market demand changes, production process parameters can be quickly adjusted to adapt to new production requirements, making the production process more flexible and adaptable. Automated production processes reduce human intervention, thereby reducing the risk of inefficiencies and quality problems that may result from human error.

[0135] The system integrates image recognition and sensor data to monitor product quality in real time, promptly identifying and locating problems. This makes quality control more precise and helps resolve potential production issues promptly. The intelligent control platform can precisely adjust the operating status of equipment, ensuring efficient and stable operation at every stage of the production process, thereby improving the controllability and stability of the entire production process.

[0136] In one embodiment of the present invention, S414 includes:

[0137] Build a distributed storage architecture and set up a distributed storage cluster, connecting multiple storage nodes through a network to form a unified storage resource pool; and configure a redundancy backup mechanism for the distributed storage system.

[0138] Based on the characteristics and usage frequency of different sensor data during the textile fabric production process, a data classification strategy is formulated; according to the formulated data classification strategy, the pre-processed data is allocated to the corresponding storage locations;

[0139] Assign a unique identifier to each sensor data point and establish a comprehensive data indexing system, creating multi-level indexes based on the different attributes and characteristics of the data.

[0140] The working principle of the above technical solution is as follows: A distributed storage architecture is constructed, and a distributed storage cluster is built, connecting multiple storage nodes through a network to form a unified storage resource pool. During the construction process, storage nodes are rationally configured and optimized to ensure load balancing among nodes and avoid single points of failure and performance bottlenecks. For example, based on data access hotspots, hot data is distributed across high-performance storage nodes to improve data access speed. A redundant backup mechanism is configured for the distributed storage system, such as using multiple replicas or erasure coding technology, to ensure that data remains intact and available even when some storage nodes fail, improving data reliability and availability.

[0141] Based on the characteristics and usage frequency of different sensor data during textile fabric production, a data classification strategy is formulated. For example, considering the real-time requirements of the data, sensor data with extremely high real-time requirements, such as yarn tension and fabric tension, which need to be monitored and fed back in real time during production, are classified into one category. Data with relatively lower real-time requirements but need to be stored for long-term historical analysis and quality traceability, such as production equipment operation logs and historical environmental temperature and humidity data, are classified into another category. From the perspective of data business relevance, sensor data closely related to the same production link or process flow are grouped into one category. For example, sensor data related to yarn quality detection in the spinning process, including yarn thickness and uniformity data, are grouped into one category. This classification strategy can more accurately meet the storage, access, and management needs of different data. According to the established data classification strategy, the pre-processed data is allocated to the corresponding storage locations. For data with extremely high real-time requirements, it is stored in a high-speed cache, which can use an in-memory database or a high-performance solid-state drive (SSD) as the storage medium to achieve fast data read / write and real-time access. For example, during the weaving process, real-time fabric tension data needs to be immediately fed back to the control system for adjustments. Storing this type of data in a cache ensures that the control system can obtain the latest data in a very short time, guaranteeing production stability and product quality. For data with relatively lower real-time requirements but needing long-term storage, it can be stored in disk arrays or cloud storage. Disk arrays offer large storage capacity and relatively high cost-effectiveness, making them suitable for storing large amounts of historical data; cloud storage provides flexible scalability and convenient remote access capabilities, facilitating data sharing and utilization among different departments and branches within an enterprise. For instance, storing the production equipment operation logs from the past few years in cloud storage allows enterprise managers to query and analyze equipment operating status anytime, anywhere via the network, providing a basis for equipment maintenance and upgrade decisions.

[0142] Each sensor data point is assigned a unique identifier, which can include key information such as sensor number, acquisition time, and data type, ensuring the uniqueness and identifiability of the data. For example, for data acquired by a yarn tension sensor at a specific time, its identifier can be designed as "Sensor ID_Acquisition Time_Yarn Tension". This identification method allows for quick and accurate location of the specific data. A comprehensive data indexing system is also established, creating multi-level indexes based on different data attributes and characteristics. Indexes can be created according to dimensions such as sensor type, acquisition time range, and production stage, facilitating subsequent data querying and retrieval. For example, when querying all sensor data from a specific production stage within a certain time period, the indexing system can quickly filter out data that meets the criteria, greatly improving the efficiency of data retrieval.

[0143] The above technical solution achieves the following results: By constructing a distributed storage architecture and redundant backup mechanism, it ensures that data remains available even when some storage nodes fail, improving the system's fault tolerance and data reliability. Through reasonable storage node configuration and optimization, hot data is stored on high-performance nodes, improving data access speed, optimizing system performance, and ensuring rapid feedback of data with extremely high real-time requirements.

[0144] By employing load balancing and distributed storage cluster design, single points of failure and performance bottlenecks are avoided, enhancing system stability and reliability. Through data classification strategies, data is rationally stored based on real-time requirements and business relevance, ensuring that the storage, access, and management needs of different data types are accurately met.

[0145] Cloud storage provides flexible scalability and convenient remote access for data with low real-time requirements but long-term storage, enabling various departments and branches within an enterprise to easily share and utilize data. By establishing a multi-level indexing system, data is indexed based on dimensions such as sensor type and acquisition time, improving the efficiency of data querying and retrieval, and reducing the time cost required to find specific data.

[0146] Storing data with high real-time requirements in a cache enables timely access and feedback of critical data during production, ensuring efficient production processes and stable product quality. By strategically selecting storage media, such as a combination of SSDs and disk arrays, the configuration of storage devices is optimized, reducing the cost of large-scale data storage and improving the overall cost-effectiveness of the storage system.

[0147] Each sensor data point is assigned a unique identifier, and a data indexing system has been established to ensure data identifiability and efficient location capabilities, enhance data traceability, and improve the accuracy of data management. Through a robust storage architecture and data indexing, the system can clearly present the storage status and access paths of all data, ensuring transparency in data management and efficient control over the data.

[0148] In one embodiment of the present invention, step S5 includes:

[0149] S51. Based on the product quality status reflected in the comprehensive quality inspection dataset and the market demand changes reflected in the textile fabric market demand forecast dataset, determine the factors for optimizing production line equipment scheduling; and obtain the product quality status of each piece of equipment based on the quality inspection dataset.

[0150] S52. Use intelligent scheduling algorithms to dynamically optimize the production line equipment in the textile fabric production topology model; compile the optimized equipment scheduling schemes into an equipment scheduling optimization dataset.

[0151] S53. Transmit the equipment scheduling optimization dataset to the textile fabric production control platform; the control platform executes the equipment scheduling task according to the dataset.

[0152] The working principle of the above technical solution is as follows: By combining the product quality data reflected in the comprehensive quality inspection dataset and the market demand changes reflected in the textile fabric market demand forecast dataset, the factors for optimizing production line equipment scheduling are determined; based on the quality inspection dataset, the product quality status of each piece of equipment is obtained, with a focus on equipment with more quality problems; based on the market demand forecast dataset, the market demand trends of different specifications and styles of textile fabrics are understood, and production tasks for the equipment are rationally arranged. Simultaneously, factors such as equipment operating status, maintenance cycles, and production capacity are fully considered to ensure the scientific and rational nature of equipment scheduling.

[0153] Intelligent scheduling algorithms are used to dynamically optimize the scheduling of production line equipment in a textile fabric production topology model. These algorithms can employ multi-objective optimization-based scheduling algorithms, reinforcement learning-based scheduling algorithms, and other approaches. Multi-objective optimization-based algorithms comprehensively consider multiple objective factors such as equipment utilization, production efficiency, and product quality. They establish a multi-objective optimization model and use optimization algorithms to solve for the optimal equipment scheduling scheme. Reinforcement learning-based scheduling algorithms, on the other hand, allow the agent to continuously learn and optimize scheduling strategies through interaction with the environment, achieving adaptive equipment scheduling. For example, using a reinforcement learning-based scheduling algorithm, the agent continuously adjusts the production tasks and running sequences of equipment based on information such as equipment status, production tasks, and market demand to improve equipment utilization and production efficiency. The optimized equipment scheduling schemes are compiled into an equipment scheduling optimization dataset. This dataset includes information such as equipment name, production task, running sequence, start time, and end time. The equipment scheduling optimization dataset is recorded and categorized in detail for subsequent execution and monitoring.

[0154] The equipment scheduling optimization dataset is transmitted to the textile fabric production control platform. The control platform executes equipment scheduling tasks according to this dataset, achieving intelligent and automated scheduling of production line equipment by controlling parameters such as equipment start-up, shutdown, and operating speed. For example, based on the instructions in the equipment scheduling optimization dataset, the control platform assigns production tasks to the corresponding equipment and adjusts the equipment's operating sequence and timing to ensure efficient and orderly production. Simultaneously, it monitors the equipment's operating status in real time, promptly identifying and resolving problems that arise during scheduling to guarantee its smooth implementation.

[0155] The above technical solution achieves the following results: By using intelligent scheduling algorithms, combined with quality inspection data and market demand forecast data, the scheduling of production line equipment is optimized, ensuring timely and efficient operation of equipment and improving production efficiency. Through dynamic scheduling and real-time monitoring of equipment operating status, potential equipment failures are promptly detected, and optimized scheduling is performed based on equipment operating status and maintenance cycles, reducing downtime caused by equipment failures.

[0156] By employing intelligent scheduling algorithms, production tasks and operating sequences of equipment are rationally arranged, avoiding idle or overloaded operation, maximizing equipment utilization, and improving production capacity. Based on quality inspection datasets, equipment with fewer quality issues is prioritized for scheduling, thereby effectively reducing product defect rates and improving overall product quality.

[0157] Intelligent scheduling and rational allocation of equipment resources reduce unnecessary energy consumption and equipment idling, thereby reducing resource waste in the production process and lowering overall production costs. Based on market demand forecasts, the priority of production tasks and equipment scheduling are adjusted to ensure rapid response to changes in market demand, enhancing the production line's adaptability to market changes.

[0158] By combining multi-objective optimization models and reinforcement learning algorithms, equipment scheduling becomes more scientific and rational, comprehensively considering multiple factors such as production efficiency, equipment maintenance, and market demand, thus optimizing the overall production scheduling plan. Through real-time monitoring and recording of equipment scheduling data, every step of the production process can be tracked and managed, improving transparency and controllability and ensuring that production tasks are executed according to plan.

[0159] By employing a reinforcement learning-based scheduling algorithm, the system can dynamically adjust the scheduling of production line equipment according to changes in actual production conditions and market demand, thereby improving the flexibility and responsiveness of the production line. Through dynamic scheduling optimization and real-time monitoring mechanisms, the system ensures the stability of equipment operation during production, reduces production interruptions or equipment failures caused by improper scheduling, and thus improves the overall reliability of the production system.

[0160] In one embodiment of the present invention, S52 includes:

[0161] By deploying various sensors on the textile fabric production line, multi-dimensional data related to equipment scheduling is collected in real time; and the collected multi-dimensional data is preprocessed.

[0162] Initialize the key parameters in the intelligent scheduling algorithm; input the preprocessed production environment data into the preset intelligent scheduling algorithm, and the algorithm starts running according to the initialized parameters to generate a preliminary equipment scheduling plan;

[0163] Based on the evaluation index system, the initially generated equipment scheduling plan is evaluated; based on the evaluation results, the initial scheduling plan is optimized.

[0164] The final equipment scheduling schemes that have been evaluated and optimized are compiled into an equipment scheduling optimization dataset; and the equipment scheduling optimization dataset is stored in a structured database.

[0165] The working principle of the above technical solution is as follows: by deploying various sensors on the textile fabric production line, multi-dimensional data related to equipment scheduling is collected in real time; including but not limited to equipment operating status data, such as equipment speed, temperature, pressure, vibration frequency, etc., and the collected multi-dimensional data is preprocessed.

[0166] The key parameters of the intelligent scheduling algorithm are initialized. Preprocessed production environment data is input into the preset intelligent scheduling algorithm, which then runs according to the initialized parameters, generating a preliminary equipment scheduling plan. During the algorithm's operation, the agent (for reinforcement learning algorithms) or individual individuals (for genetic algorithms) continuously tries different equipment scheduling strategies based on the current state information and algorithm rules, evaluating and improving these strategies through interaction with the environment (reinforcement learning) or evolutionary operations (genetic algorithms). For example, in a reinforcement learning-based algorithm, the agent selects an action to execute based on the current equipment operating status and production task requirements, such as assigning an idle piece of equipment to an urgent order. After executing the action, the environment provides a reward signal, and the agent updates its policy network parameters based on the reward signal to adjust the probability of future action selections. Through multiple such interactions and learning processes, the agent gradually generates a preliminary equipment scheduling plan. In a genetic algorithm, the initial population evolves through selection, crossover, and mutation operations, generating a new generation of individuals, each representing an equipment scheduling plan. After a certain number of iterative evolutions, the individual with the highest fitness is selected as the initial equipment scheduling scheme.

[0167] Based on the evaluation index system, the preliminary equipment scheduling plan is evaluated. The evaluation indicators include: equipment utilization rate (the ratio of actual working time to total available time per unit time), reflecting equipment utilization efficiency; production efficiency (measured by the quantity of qualified textile fabrics produced per unit time), measuring the speed and output of the production process; product quality pass rate (calculated based on quality inspection data, representing the proportion of qualified products to total products, reflecting the production process's ability to control product quality); energy consumption (statistically calculating electricity and gas consumption during production to evaluate energy utilization efficiency); and production cost (covering raw material costs, equipment maintenance costs, and labor costs), reflecting the economics of the production plan. Based on the evaluation results, the preliminary scheduling plan is adjusted. The plan is optimized; if low equipment utilization is found, analyze whether there is idle equipment or uneven load, and improve equipment utilization by adjusting the allocation of production tasks or the running sequence of equipment; if production efficiency does not meet the expected target, check for bottlenecks in the production process, optimize the production process or increase corresponding resource input; if the product quality pass rate is not ideal, investigate factors affecting product quality, such as equipment parameter settings and raw material quality, and adjust and improve relevant factors; if energy consumption is too high, analyze the main energy consumption links and take energy-saving measures, such as optimizing equipment operating parameters and using energy-saving equipment; if production costs exceed the budget, find ways to reduce costs, such as optimizing raw material procurement strategies and reducing equipment maintenance costs. Through multiple evaluations and optimizations, the equipment scheduling plan is made optimal in each evaluation indicator or meets production requirements.

[0168] The final equipment scheduling scheme, after evaluation and optimization, is compiled into an equipment scheduling optimization dataset. This dataset should contain detailed and accurate information, including the name and model of the equipment, the specific requirements of the production task (such as textile fabric specifications, styles, order quantities, delivery dates, etc.), the operating sequence and schedule of the equipment, the key parameter settings during equipment operation (such as speed, temperature, pressure, etc.), and potential risks and countermeasures. The equipment scheduling optimization dataset is stored in a structured database, categorized by dimensions such as production date, production batch, and equipment type, to facilitate subsequent querying, analysis, and execution.

[0169] The above technical solution achieves the following results: By collecting multi-dimensional data related to equipment scheduling in real time and combining it with intelligent scheduling algorithms, the equipment scheduling scheme is optimized, ensuring efficient equipment operation and reducing equipment idle time and production task delays. Through real-time monitoring and analysis of equipment utilization, the allocation or execution sequence of equipment production tasks can be effectively adjusted, thereby reducing equipment idle rates and load imbalances, and improving production efficiency.

[0170] Dynamic scheduling strategies based on reinforcement learning or genetic algorithms enable production systems to flexibly adjust equipment operation modes according to real-time feedback information and changes in production tasks, improving the production line's responsiveness to unforeseen circumstances. By continuously optimizing production processes and resource allocation, production bottlenecks are eliminated, production efficiency is improved, and the number of qualified textile fabrics produced per unit time is increased.

[0171] By analyzing the main energy consumption processes and implementing energy-saving measures, such as optimizing equipment operating parameters and using energy-efficient equipment, energy waste is reduced, thereby lowering energy consumption during production. Intelligent scheduling algorithms can assess equipment status in real time and make adjustments, reducing equipment failures or downtime, ensuring stable equipment operation during production, and improving the reliability of the production line.

[0172] By optimizing equipment operating parameters, adjusting production processes, and modifying equipment scheduling strategies based on quality inspection data, the stability and consistency of product quality were ensured, and the product quality pass rate was improved. Furthermore, by optimizing raw material procurement strategies, reducing equipment maintenance costs, and increasing equipment utilization, production costs were lowered, and production economics were improved.

[0173] By constructing a structured database and recording key information for each equipment scheduling session, the scheduling process becomes more transparent and controllable, facilitating subsequent data analysis, querying, and execution. Through multiple optimizations based on evaluation metrics, the intelligent scheduling algorithm can scientifically and rationally adjust production tasks, ensuring the efficient execution of production plans and avoiding unnecessary production fluctuations.

[0174] One embodiment of the present invention provides an intelligent automated control system for a textile fabric production line, comprising:

[0175] One or more processors;

[0176] Memory, used to store one or more programs.

[0177] Wherein, when the one or more programs are executed by the one or more processors, the one or more processors are made to implement the method described in any one of the above.

[0178] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. An intelligent textile fabric production line automation control method, characterized in that, The method comprises: S1: acquiring multi-dimensional sensing data of the textile fabric production line and textile fabric production whole-process data; based on the whole-process data, using system modeling and simulation technology, the internal logic and mutual relationship of each link in the textile fabric production process are analyzed in depth, and a precise textile fabric production topology structure model is constructed; S2: based on a data fusion algorithm, the structure information in the textile fabric production topology structure model is deeply fused with the multi-dimensional sensing data, an intelligent sensing network covering the whole production line is constructed; based on the intelligent sensing network, using big data analysis and machine learning algorithm, combining multi-source information, a market demand prediction model is constructed, and a textile fabric market demand prediction data set is generated; S3: using neural network algorithm to dynamically optimize the production process link parameters in the textile fabric production topology structure model; in the optimization process, the optimal production process parameter combination is found through continuous iterative calculation, and a process optimization data set is formed; S4: continuously collecting real-time production line sensing data; comparing and analyzing the real-time data with the process optimization data set in real time, using a quality detection model to monitor and evaluate the product quality in the production link in real time; once the quality problem is found, the link and reason of the problem are located, and a quality detection data set is generated; S5: considering the product quality reflected by the quality detection data set and the market demand change embodied by the textile fabric market demand prediction data set, using intelligent scheduling algorithm to dynamically schedule and optimize the production line equipment in the textile fabric production topology structure model.

2. The intelligent textile fabric production line automatic control method according to claim 1, characterized in that, The S1 comprises: S11, arranging multiple types of sensors at each key link of the textile fabric production line to acquire multi-dimensional sensing data; collecting textile fabric production whole-process data; S12, using system modeling and simulation technology, based on the collected whole-process data, analyzing the internal logic and mutual relationship of each link in the textile fabric production process; constructing a textile fabric production system model comprising multiple subsystems; S13, simulating the model by simulation software, observing the data flow and influence mechanism between the subsystems, and analyzing the influence of different parameter changes on the production process and product quality; S14, according to the results of system modeling and simulation analysis, constructing a textile fabric production topology structure model; and giving each node and edge in the model corresponding attributes and parameters.

3. The intelligent textile fabric production line automatic control method according to claim 1, characterized in that, The S2 comprises: S21, based on a data fusion algorithm, deeply fusing the structure information in the textile fabric production topology structure model with the multi-dimensional sensing data; preprocessing the deeply fused data; and using the data fusion algorithm to associate and integrate the preprocessed sensing data with the structure information in the model, constructing an intelligent sensing network covering the whole production line; S22, collecting multi-source information, the multi-source data including historical sales data, market trends and seasonal factors, and integrating the multi-source data with the data collected by the intelligent sensing network; S23, based on the integrated data, using big data analysis and machine learning algorithm to construct market demand prediction model; using the constructed market demand prediction model, the market demand of different specifications, style of textile fabric in the future is predicted; according to the prediction result, the textile fabric market demand prediction data set is generated.

4. The intelligent textile fabric production line automatic control method according to claim 1, characterized in that, The S3 comprises: S31, guided by the textile fabric market demand prediction data set, the target of production process parameter optimization is determined; according to the market demand prediction result, the quality requirement and yield demand of different specifications, style of textile fabric are analyzed; S32, using neural network algorithm to dynamically optimize the production process parameter in the textile fabric production topology structure model; the process parameter in the production topology structure model is used as the input variable of neural network, and the optimization objective function is used as the output variable; S33, through the self learning and self adaptive ability of neural network, the value of input variable is adjusted constantly, and in the optimization process, the optimal production process parameter combination is found through constant iteration calculation; S34, the optimized production process parameter is arranged to form process optimization data set; and the process optimization data set is recorded and stored in detail; S35, the process optimization data set is transmitted to the textile fabric production control platform; the control platform automatically adjusts the running state of production equipment according to the parameters.

5. The intelligent textile fabric production line automatic control method according to claim 1, characterized in that, The S4 comprises: S41, based on the intelligent sensing network of textile fabric production line, the real-time production line sensing data is continuously collected; and the data is transmitted to the data acquisition system; the data acquisition system processes and stores the collected data; S42, real-time data and process optimization data set are compared and analyzed in real time; through data comparison algorithm, the deviation value between real-time data and process optimization data is calculated; S43, using quality detection model to monitor and evaluate the product quality in production process; if quality problem is found, the system locates the problem and the reason of the problem, and generates quality detection data set.

6. The intelligent textile fabric production line automatic control method according to claim 5, characterized in that, The S41 comprises: S411, the time correlation and space correlation between each sensor data are analyzed, and the sampling time sequence and frequency of different sensors are determined based on the analysis result; S412, based on the set sampling strategy, each sensor collects data in real time; in the collection process, the sensor converts the physical quantity into electric signal, and converts the analog signal into digital signal; the sensor transmits the digital signal to the data acquisition system through the combination of various communication technologies; and in the data transmission process, data encryption technology is used to encrypt the transmitted data; S413, after the data acquisition system receives the data transmitted by the sensor, the data is preliminarily processed; S414, the data after preliminary processing is stored in the data storage system; distributed storage architecture is adopted, the data is stored in multiple storage nodes, and the data is stored in categories according to the characteristics and frequency of use.

7. The intelligent textile fabric production line automatic control method according to claim 6, characterized in that, The S414 comprises: A distributed storage architecture is constructed, and a distributed storage cluster is built. Multiple storage nodes are connected through a network to form a unified storage resource pool. A redundant backup mechanism is configured for the distributed storage system. According to the characteristics and frequency of use of different sensor data in the textile fabric production process, a data classification strategy is developed. According to the developed data classification strategy, the preliminarily processed data is allocated to the corresponding storage location. A unique identifier is set for each sensor data, and a perfect data indexing system is established. Multiple indexes are created according to different attributes and characteristics of the data. 8.The intelligent textile fabric production line automatic control method of claim 1, wherein, The S5 comprises: S51, comprehensively reflecting the product quality situation of the quality detection data set and the market demand changes of the textile fabric market demand prediction data set, determining the factors of production line equipment scheduling optimization; according to the quality detection data set, the quality status of products produced by each equipment is obtained; S52, using intelligent scheduling algorithm to dynamically schedule and optimize the production line equipment in the textile fabric production topology structure model; the optimized equipment scheduling scheme is arranged to form an equipment scheduling optimization data set; S53, transmitting the equipment scheduling optimization data set to the textile fabric production control platform; the control platform executes the equipment scheduling task according to the data set.

9. The intelligent textile fabric production line automatic control method according to claim 8, characterized in that, The S52 comprises: Through various sensors deployed on the textile fabric production line, multi-dimensional data related to equipment scheduling is collected in real time; and the collected multi-dimensional data is preprocessed; The key parameters in the intelligent scheduling algorithm are initialized and set. The preprocessed production environment data is input into the preset intelligent scheduling algorithm. The algorithm starts running according to the initialized parameters to generate a preliminary equipment scheduling scheme; Based on the evaluation index system, the preliminary generated equipment scheduling scheme is evaluated; according to the evaluation result, the preliminary scheduling scheme is optimized; The final equipment scheduling scheme after evaluation and optimization is arranged to form an equipment scheduling optimization data set; and a structured database is used to store the equipment scheduling optimization data set.

10. An intelligent textile fabric production line automatic control system, comprising: one or more processors; a memory for storing one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors implement the method of any one of claims 1 to 9.

Images