A dyeing quality prediction and closed-loop control method for continuous printing and dyeing production line

Abstract

The application relates to a dyeing quality prediction and closed-loop control method for a continuous printing and dyeing production line. The method comprises the following steps: collecting parameters of gray cloth, dyeing liquid state, equipment running state, environmental parameters and online color detection data in the running process of the continuous printing and dyeing production line; pre-processing and feature fusion are performed on the multi-source data to construct dyeing process state feature data; a dyeing quality prediction model is established based on a time sequence prediction model to predict color difference, uniformity and quality grade in the dyeing process; a dyeing quality evaluation model is established according to the prediction result, and a process parameter adjustment strategy is generated when a quality deviation is detected; dyeing liquid parameters, temperature parameters and equipment running parameters are dynamically adjusted through a closed-loop control mode, and the prediction model is optimized according to the actual dyeing result. The application has the effects of online prediction of dyeing quality in the continuous printing and dyeing process, abnormal early warning and process self-adaptive adjustment.

Description

Technical Field

[0001] This application relates to the fields of intelligent manufacturing and automated control of textile printing and dyeing, and in particular to a method for predicting and controlling the dyeing quality of continuous printing and dyeing production lines. Background Technology

[0002] With the textile industry moving towards intelligent, automated, and high-quality manufacturing, continuous dyeing and finishing production lines have been widely used in the fabric dyeing and processing field. Continuous dyeing and finishing processes typically include multiple continuous processing steps such as fabric input, dye bath padding, steam fixing, washing, drying, and finishing. The dyeing quality is comprehensively affected by various factors, including fabric characteristics, dye ratios, auxiliary agent conditions, processing temperature, equipment operating parameters, and the production environment. In actual production, even with the same process parameters, quality problems such as large color differences, uneven dyeing, and reduced batch consistency can easily occur due to batch variations in raw materials, fluctuations in equipment conditions, and changes in dye bath properties.

[0003] Currently, dyeing quality inspection in continuous dyeing and printing production lines mainly relies on manual sampling or offline color difference analysis. That is, after the fabric has completed the dyeing process, the dyeing results are obtained through manual sampling or laboratory testing, and then the subsequent production parameters are adjusted based on the test results. This method has problems such as long detection cycle, delayed feedback, and inability to reflect the dynamic changes in the production process in real time. When abnormal dyeing quality is found, a certain number of unqualified products have often been produced, resulting in the waste of dyes, energy, and production time.

[0004] Meanwhile, although some existing dyeing and printing production lines are equipped with online color detection equipment, they are mainly used for collecting single color parameters and lack the ability to comprehensively analyze multi-source data of the dyeing process. They cannot fully combine the state of the dye liquor, the operating status of the equipment, and the changing patterns of process parameters to predict the dyeing quality in advance. In addition, traditional production control methods usually rely on the experience of operators to adjust parameters such as dye liquor concentration, temperature, rolling mill pressure, and operating speed. The degree of automation in the control process is low and it is difficult to adapt to the complex and changing working conditions in continuous production.

[0005] With the development of artificial intelligence, big data analysis, and industrial closed-loop control technology, using multi-source data fusion and machine learning models to predict the quality of the production process and dynamically adjust process parameters based on the prediction results has become an important technical direction for improving the quality stability of dyeing and printing products. However, existing technologies still lack a comprehensive control method that can collect data, predict quality, identify anomalies, and automatically adjust the entire continuous dyeing and printing production process, making it difficult to realize the transformation of dyeing quality from "post-event detection" to "process prediction and proactive control".

[0006] Regarding the aforementioned technologies, the inventors believe that existing continuous dyeing processes suffer from problems such as lagging quality inspection, strong reliance on manual adjustments, and insufficient dyeing consistency. Summary of the Invention

[0007] To address the technical problems in existing continuous dyeing and printing production processes, such as lagging dyeing quality detection, difficulty in timely detection of quality anomalies, reliance on manual experience to adjust production process parameters, and insufficient dyeing consistency, this application provides a dyeing quality prediction and closed-loop control method for continuous dyeing and printing production lines.

[0008] The dyeing quality prediction and closed-loop control method for continuous dyeing production lines provided in this application adopts the following technical solution: Firstly, a method for predicting and controlling dyeing quality in continuous dyeing and printing production lines includes the following steps: S1. Obtain multi-source production data during the operation of the continuous dyeing and printing production line; The multi-source production data includes fabric parameter data, dye liquor status data, equipment operating status data, environmental parameter data, and online color detection data; S2. Preprocess the multi-source production data and perform multi-dimensional feature fusion to construct continuous dyeing process state feature data; S3. Based on the state characteristic data of the continuous dyeing process, establish a dyeing quality prediction model, and use the dyeing quality prediction model to predict the dyeing quality of the target fabric to obtain the dyeing quality prediction result. S4. Based on the dyeing quality prediction results, establish a dyeing quality evaluation model, comprehensively evaluate the dyeing effect, and determine whether the current dyeing status meets the preset quality requirements. S5. When the dyeing quality prediction result deviates from the target quality range, a corresponding process parameter adjustment strategy is generated. S6. Dynamically adjust the dye liquor parameters, temperature parameters, and equipment operating parameters in the continuous dyeing and printing production line according to the process parameter adjustment strategy. S7. Collect the adjusted actual dyeing test data, and optimize the dyeing quality prediction model based on the error between the actual dyeing results and the predicted results to form a closed-loop control of dyeing quality prediction and process parameter adjustment.

[0009] By adopting the above technical solutions, intelligent quality prediction and closed-loop control functions for continuous dyeing and printing production processes are realized. By collecting multi-source production data and performing feature fusion, a dyeing quality prediction model is established to predict the dyeing effect in advance. Based on the prediction results, quality evaluation and anomaly judgment are performed, process parameter adjustment strategies are automatically generated, and dye liquor, temperature, and equipment parameters are dynamically adjusted. At the same time, the model is optimized by combining actual production feedback to improve dyeing quality stability, production efficiency, and process control accuracy.

[0010] Optionally, the multi-source production data is acquired through a multi-source data acquisition module installed on the continuous dyeing and printing production line; The continuous dyeing and printing production line includes a greige fabric input unit, a dye liquor supply unit, a dyeing processing unit, a drying and finishing unit, and a finished product output unit. The fabric parameter data includes fabric material, fiber composition, fabric weight, fabric thickness, moisture content, and initial color parameters. The dye liquor state data includes dye concentration, auxiliary agent concentration, pH value, dye liquor temperature, conductivity, and viscosity. The equipment operating status data includes the fabric guiding speed, rolling mill pressure, dye liquor circulation flow rate, and equipment operating load.

[0011] By adopting the above technical solutions, a multi-dimensional state perception function for the continuous dyeing and printing production process is realized. The multi-source data acquisition module acquires key parameters such as fabric, dye liquor, and equipment operation in real time, and comprehensively grasps the production factors affecting dyeing quality. The fusion of multiple types of data forms complete process state information, providing an accurate and reliable data foundation for subsequent dyeing quality prediction, anomaly analysis, and closed-loop control, thereby improving the accuracy of quality prediction and production stability.

[0012] Optionally, the multi-source data preprocessing includes: The data preprocessing module performs data cleaning, outlier removal, time synchronization, and normalization on the collected data. Wherein: the data cleaning unit is used to remove invalid data; The abnormal data removal unit is used to identify and delete abnormal sampled data; The time synchronization unit is used to synchronize data obtained from different acquisition devices according to the production timeline; The data normalization unit is used to eliminate dimensional differences between different types of data.

[0013] By adopting the above technical solutions, a standardized preprocessing function for multi-source production data is realized. Invalid information is removed through data cleaning, data reliability is improved through outlier removal, time synchronization ensures the temporal consistency of data from different acquisition devices, and normalization eliminates differences in data units. This improves the accuracy and stability of multi-source data fusion analysis and provides a high-quality data foundation for subsequent production status monitoring, process optimization, and intelligent decision-making.

[0014] Optionally, the multidimensional feature fusion includes: The data fusion module integrates fabric features, dye liquor features, equipment operation features, and color detection features. The data fusion module includes a multi-source feature fusion unit, a dyeing process state feature vector generation unit, and a time-series data construction unit. Establish the state feature vector of the staining process: F=[Fp,Fd,Fe,Fc] Where: Fp represents the characteristics of the greige fabric; Fd represents the characteristics of the dye liquor state; Fe represents the characteristics of the equipment operating state; Fc represents the characteristics of color detection. The continuous dyeing process time series data is generated based on the state feature vector.

[0015] By adopting the above technical solution, the fusion modeling function of multi-dimensional production characteristics of continuous dyeing and printing process is realized. The properties of greige fabric, dye liquor state, equipment operating parameters and color detection results are uniformly represented, a comprehensive state feature vector is constructed, and the correlation analysis and collaborative expression between data from different sources are realized. This improves the completeness and accuracy of the state description of the dyeing process and provides a reliable data foundation for subsequent process optimization, quality prediction and intelligent control.

[0016] Optionally, the staining quality prediction model includes an input layer, a feature fusion layer, a temporal feature extraction layer (and a quality prediction output layer); The temporal feature extraction layer adopts a recurrent neural network model, including at least one of the Long Short-Term Memory (LSTM) network model and the Gated Recurrent Unit (GRU) model; The dyeing quality prediction model is used to predict dyeing quality parameters based on production status data over a continuous time period.

[0017] By adopting the above technical solution, the intelligent prediction function of dyeing quality in continuous printing and dyeing process is realized. The input layer receives multi-source production status data, the feature fusion layer realizes multi-dimensional feature association, the dynamic change law in the dyeing process is extracted by LSTM or GRU time series network, and the prediction output layer is used to predict quality parameters, thereby improving the accuracy, real-time performance and stability of dyeing quality prediction, realizing the early identification of abnormal dyeing trends, and providing a basis for process adjustment and quality control.

[0018] Optionally, the dyeing quality prediction results output by the quality prediction output layer include: predicted color difference value ΔE, color uniformity index, color fastness prediction index, and batch consistency index. Based on the above prediction results, staining quality evaluation parameters are formed.

[0019] By adopting the above technical solution, and by outputting key quality parameters such as predicted color difference value, color uniformity, color fastness, and batch consistency, the dyeing effect can be quantitatively analyzed from multiple dimensions such as color accuracy, stability, and product consistency, forming comprehensive quality evaluation parameters. This improves the comprehensiveness and reliability of dyeing quality detection and provides data support for production process optimization and quality control.

[0020] Optionally, the dyeing quality evaluation model includes: a color difference evaluation unit, a uniformity evaluation unit, a color fastness evaluation unit, and a batch consistency evaluation unit; The evaluation indicators are weighted and calculated using a comprehensive quality evaluation model: Q = w1E + w2U + w3S + w4B; Where: Q represents the comprehensive dyeing quality evaluation value; E represents the color difference evaluation index; U represents the color uniformity index; S represents the color fastness index; and B represents the batch consistency index.

[0021] By adopting the above technical solutions, color difference, uniformity, color fastness, and batch consistency indicators are independently analyzed, and a comprehensive evaluation value is formed by using a weighted calculation method. This enables quantitative assessment of dyeing quality, improves the accuracy and comprehensiveness of quality evaluation, and provides a reliable basis for production process optimization, quality traceability, and intelligent control.

[0022] Optionally, the anomaly detection and early warning module performs trend analysis based on the staining quality prediction results; When the predicted color difference value exceeds the preset threshold, the anomaly judgment unit generates a dyeing anomaly judgment result and outputs quality warning information through the warning output unit. The quality warning information includes the anomaly type, anomaly location, and corresponding adjustment suggestions; The closed-loop control module controls the operation of the continuous dyeing and printing production line according to the process parameter adjustment strategy. The process parameter adjustment strategy includes: adjusting the amount of dye added, the amount of auxiliary agent added, and the pH value through the dye liquor parameter adjustment unit; The temperature of the dye liquor, the evaporation temperature, and the drying temperature can be adjusted using the temperature parameter adjustment unit. Adjust the guide speed, rolling mill pressure, and dye liquor circulation speed using the equipment parameter adjustment unit; The feedback optimization includes: acquiring actual dyeing quality detection data through an online quality detection unit; The error between the predicted result and the actual detection result is calculated using the prediction error calculation unit: Ep = |Yr - Yp| Where: Ep represents the prediction error; Yr represents the actual detection result; Yp represents the model prediction result; Based on the prediction error, the parameters of the staining quality prediction model are adjusted through the model parameter update unit, and the model is continuously optimized through the self-learning optimization unit.

[0023] By adopting the above technical solutions, intelligent detection, early warning, and closed-loop optimization control of quality anomalies in continuous dyeing and printing processes can be achieved. By analyzing the dyeing quality trend, abnormal states can be identified in a timely manner and targeted adjustment suggestions can be output. At the same time, the dye liquor, temperature, and equipment operating parameters can be automatically adjusted according to the process parameters. In addition, error feedback and model self-learning optimization can be carried out in combination with actual detection results, thereby improving the stability of the production process, the accuracy of quality control, and the level of intelligence.

[0024] Secondly, a dyeing quality prediction and closed-loop control system for continuous dyeing production lines includes: a multi-source data acquisition module for acquiring fabric parameter data, dye liquor state data, equipment operating status data, environmental parameter data, and color detection data during the operation of the continuous dyeing production line. The data preprocessing module is used to clean, synchronize, and normalize the collected data. The data fusion module is used to construct state characteristic data of the staining process; The staining quality prediction module is used to output staining quality prediction results based on the prediction model; The staining quality evaluation module is used to evaluate staining quality based on the prediction results; The anomaly detection and early warning module is used to identify abnormal staining conditions and output early warning information; The closed-loop control module is used to adjust the dyeing and printing production process parameters based on the prediction results. The feedback optimization module is used to optimize the prediction model based on the actual staining results.

[0025] By adopting the above technical solutions, the dyeing quality prediction and intelligent closed-loop control functions of continuous dyeing production lines are realized. A complete production status model is constructed through multi-source data acquisition, preprocessing and fusion. The prediction module is used to assess the dyeing quality in advance, and the evaluation, anomaly warning and process adjustment modules are combined to realize dynamic optimization of the production process. At the same time, the model prediction accuracy is continuously improved through the feedback optimization mechanism, so as to realize intelligent monitoring and control of dyeing quality.

[0026] Thirdly, an electronic device includes a processor, a memory, and a computer program stored in the memory, wherein when the computer program is executed by the processor, the processor performs a dyeing quality prediction and closed-loop control method for a continuous dyeing production line as described in any one of claims 1 to 9.

[0027] The computer program is stored in a computer-readable storage medium.

[0028] By adopting the above technical solution, the software deployment function of dyeing quality prediction and closed-loop control method is realized. The computer program stored in the memory is executed by the processor to realize intelligent analysis, quality prediction, anomaly identification and process control of continuous dyeing and printing production data. At the same time, the stability, scalability and execution efficiency of the system operation are guaranteed, which makes it easy to promote and apply in different production equipment.

[0029] In summary, this application includes at least one of the following beneficial technical effects: 1. By constructing a multi-source data acquisition system, we can achieve comprehensive perception of the quality status of raw materials, dye liquor, equipment, and products in the continuous dyeing and printing production process, thus solving the problem of single data source in traditional detection methods.

[0030] 2. By establishing a dyeing quality prediction model based on time-series data analysis, we can predict the trend of dyeing quality changes in advance, and transform quality control from post-production inspection to proactive prediction during the production process.

[0031] 3. By establishing a comprehensive quality evaluation model, the dyeing quality is evaluated from multiple dimensions such as color difference, uniformity, color fastness, and batch consistency, thereby improving the accuracy of quality judgment.

[0032] 4. Through the closed-loop control module, the dye liquor parameters, temperature parameters and equipment operating parameters are automatically adjusted according to the prediction results, reducing reliance on manual experience and improving the automation control level of the continuous dyeing and printing production line.

[0033] 5. The prediction model is continuously updated through the feedback optimization module, enabling the system to adapt to different fabric types, dye systems, and production environments, thereby improving long-term operational stability. Attached Figure Description

[0034] Figure 1 This is an overall flowchart of a dyeing quality prediction and closed-loop control method for a continuous dyeing production line according to an embodiment of this application.

[0035] Figure 2 This is a schematic diagram of a multi-source data acquisition architecture for a continuous dyeing and printing production line according to an embodiment of this application.

[0036] Figure 3This is a schematic diagram of a dyeing quality prediction model for a continuous dyeing and printing production line according to an embodiment of this application.

[0037] Figure 4 This is a schematic diagram of a data fusion and feature extraction process for a continuous dyeing and printing production line according to an embodiment of this application.

[0038] Figure 5 This is a schematic diagram of a dyeing quality evaluation model for a continuous dyeing and printing production line according to an embodiment of this application.

[0039] Figure 6 This is a schematic diagram of a closed-loop control process for process parameters in a continuous dyeing and printing production line according to an embodiment of this application.

[0040] Figure 7 This is a schematic diagram of a predictive feedback optimization control system for a continuous dyeing and printing production line according to an embodiment of this application.

[0041] Figure labeling: 1. Continuous dyeing and printing production line; 11. Greige fabric input unit; 12. Dye liquor supply unit; 13. Dyeing processing unit; 14. Drying and finishing unit; 15. Finished product output unit; 2. Multi-source data acquisition module; 21. Greige fabric parameter acquisition unit; 22. Dye liquor status acquisition unit; 23. Equipment operating status acquisition unit; 24. Environmental parameter acquisition unit; 25. Online color detection unit; 3. Data preprocessing module; 31. Data cleaning unit; 32. Abnormal data removal unit; 33. Time synchronization unit; 34. Data normalization unit; 35. Feature extraction unit; 4. Data fusion module; 41. Multi-source feature fusion unit; 42. Dyeing process status feature vector generation unit; 43. Time series data construction unit; 5. Dyeing quality prediction module; 51. Prediction model input layer; 52. Feature fusion layer; 53. Time series feature extraction layer; 5 4. Quality Prediction Output Layer; 5. Deep Learning Prediction Model; 6. Dyeing Quality Evaluation Module; 6. Color Difference Evaluation Unit; 6. Uniformity Evaluation Unit; 6. Color Fastness Evaluation Unit; 6. Batch Consistency Evaluation Unit; 6. Comprehensive Quality Evaluation Model; 7. Anomaly Detection and Early Warning Module; 7. Quality Trend Prediction Unit; 7. Anomaly Judgment Unit; 73. Early Warning Output Unit; 8. Closed-Loop Control Module; 8. Control Strategy Generation Unit; 82. Dye liquor parameter adjustment unit; 83. Temperature parameter adjustment unit; 84. Equipment parameter adjustment unit; 85. Execution Control Unit; 9. Feedback Optimization Module; 91. Online Quality Inspection Unit; 92. Prediction Error Calculation Unit; 93. Model Parameter Update Unit; 94. Self-Learning Optimization Unit; 10. Control Center; 101. Industrial Controller; 102. Data Storage Unit; 103. Human-Machine Interface. Detailed Implementation

[0042] The following is in conjunction with the appendix Figure 1-7 This application will be described in further detail.

[0043] This application discloses a method for predicting and controlling dyeing quality in continuous dyeing production lines. This method addresses problems in existing continuous dyeing production processes, such as lagging dyeing quality detection, difficulty in timely adjustment of process parameter fluctuations, reliance on manual experience for dyeing quality control, and insufficient batch consistency. By constructing a multi-source data acquisition system, a dyeing quality prediction model, and a closed-loop adjustment and control mechanism, it achieves online prediction of dyeing quality, anomaly warning, and automatic optimization of process parameters in continuous dyeing processes. Figure 1 As shown, it includes the following steps: S1. Construct a data acquisition system for continuous dyeing and printing production lines. In this embodiment, the continuous dyeing production line 1 is used to complete the continuous dyeing process of fabrics, including a greige fabric input unit 11, a dye liquor supply unit 12, a dyeing processing unit 13, a drying and finishing unit 14, and a finished product output unit 15 arranged in sequence. The fabric input unit 11 is used to continuously transport the fabric to be dyed, so that the fabric enters the subsequent dyeing process at a preset speed; the dye supply unit 12 is used to store, prepare and transport the dye required for dyeing; the dyeing processing unit 13 is used to make the fabric fully contact the dye to achieve dye adsorption and penetration; the drying and finishing unit 14 is used to dry and finish the dyed fabric; and the finished product output unit 15 is used to output the dyed fabric. In the continuous dyeing process, the dyeing quality is affected by a variety of factors such as the characteristics of raw materials, the state of dye liquor, the operating status of equipment and environmental changes. By setting up a multi-source data acquisition module 2, the key status parameters in the continuous dyeing production process can be collected in real time.

[0044] like Figure 2 As shown, the multi-source data acquisition module 2 includes a fabric parameter acquisition unit 21, a dye liquor state acquisition unit 22, an equipment operation status acquisition unit 23, an environmental parameter acquisition unit 24, and an online color detection unit 25. Specifically, the fabric parameter acquisition unit 21 is used to collect the basic parameters of the fabric entering the continuous dyeing and printing production line, including fabric material, fiber composition, fabric structure, fabric weight, fabric thickness, moisture content, and initial color parameters. Since different batches of fabrics have different fiber structures and adsorption properties, the amount of dye adsorption will change. Therefore, by collecting the fabric parameters, basic data support can be provided for subsequent dyeing quality prediction. The dye liquor state acquisition unit 22 is used to collect the real-time state information of the dye liquor during the dyeing process, including dye concentration, auxiliary agent concentration, pH value, dye liquor temperature, conductivity, and dye liquor viscosity. Among them, dye concentration directly affects the final color depth of the fabric, auxiliary agent concentration affects dye diffusion and color fixation effect, pH value affects the binding efficiency between dye and fiber, and dye liquor temperature affects dye migration speed. Therefore, real-time acquisition of dye liquor state parameters can improve the ability to analyze the changing patterns of the dyeing process. The equipment operation status acquisition unit 23 is used to collect mechanical parameters during the operation of the continuous dyeing and printing equipment, including the guide speed, roller pressure, dye liquor circulation flow rate, equipment operating load and transmission status. Among them, the guide speed determines the contact time between the fabric and the dye liquor, the roller pressure affects the fabric liquid carrying rate, and the dye liquor circulation speed affects the uniformity of the dye liquor. Changes in the above parameters will affect the final dyeing effect. The environmental parameter acquisition unit 24 is used to collect the production environment status, including ambient temperature, ambient humidity and air flow status, in order to eliminate the impact of environmental factors on dyeing quality. The online color detection unit 25 is set at the detection position of the continuous printing and dyeing production line to acquire fabric color detection data in real time, including color space parameters, spectral reflectance data and color difference parameter ΔE. Through the above multi-source data acquisition method, the entire process data perception of "raw materials-dye liquor-equipment-environment-product quality" in the continuous printing and dyeing production process can be realized.

[0045] S2, Multi-source data preprocessing and fusion Because different types of sensing devices collect data with different sampling frequencies, data formats, and data ranges, this embodiment uses a data preprocessing module 3 to standardize the collected data. The data preprocessing module 3 includes a data cleaning unit 31, an abnormal data removal unit 32, a time synchronization unit 33, a data normalization unit 34, and a feature extraction unit 35. First, the data cleaning unit 31 performs integrity checks on the collected data, processing missing, duplicate, and invalid data. Second, the abnormal data removal unit 32 identifies abnormal sampling values ​​generated by instantaneous fluctuations in sensors during the production process. Specifically, a parameter change range can be established based on historical production data. When the real-time collected value exceeds the preset range, the data is judged to be abnormal and corrected or removed. Subsequently, a unified time reference is established through the time synchronization unit 33, and the data obtained by different acquisition modules are matched according to the continuous dyeing and printing production rhythm. For example, the dye liquor temperature, dye concentration, fabric guiding speed, and color detection results collected within a certain time period are associated so that data from different sources correspond to the same production state. The data normalization unit 34 standardizes data of different dimensions, such as temperature, pressure, concentration, and color parameters, so that each parameter can be input into the prediction model for unified calculation. After data preprocessing, the feature extraction unit 35 extracts key feature parameters that affect staining quality. In this embodiment, the following staining process state feature vector is established: F=[Fp,Fd,Fe,Fc] Wherein: Fp represents the fabric characteristics; Fd represents the dye liquor state characteristics; Fe represents the equipment operation characteristics; Fc represents the color detection characteristics. Subsequently, the above different types of features are fused through the data fusion module 4. The data fusion module 4 includes a multi-source feature fusion unit 41, a dyeing process state feature vector generation unit 42, and a time-series data construction unit 43 to form continuous dyeing process state data. In this way, the prediction model can comprehensively analyze the correlation between different production factors and improve the accuracy of dyeing quality prediction.

[0046] like Figure 3 As shown, this embodiment establishes a dyeing quality prediction model 55 based on multi-source time-series data analysis through the dyeing quality prediction module 5. This model is used to predict future dyeing quality trends based on historical data and real-time collected data from the continuous dyeing and printing production process. The dyeing quality prediction model 55 includes a prediction model input layer 51, a feature fusion layer 52, a time-series feature extraction layer 53, and a quality prediction output layer 54. The prediction model input layer 51 receives the continuous dyeing and printing process status feature data processed by the data fusion module 4. In actual operation, the data collected within the continuous time window is constructed as the model input sequence: X(t)={F(tn),F(t-n+1),...,F(t)} Where: X(t) represents the production state sequence within the current time window; F(t) represents the coloring process state feature vector at the corresponding time node; n represents the length of the time window. By inputting continuous time series, the model can not only analyze the current production state, but also learn the pattern of change of each parameter over time during the coloring process. The feature fusion layer 52 is used to perform deep fusion processing on the input multi-dimensional production features. Since the dyeing quality is not determined by a single parameter during continuous dyeing, but is affected by the fabric properties, dye liquor state, equipment operating status and environmental changes, the feature fusion layer 52 establishes the correlation between different parameters. For example, when the dye liquor concentration decreases, the fabric guide speed increases and the dye liquor temperature decreases, the model can comprehensively judge the influence of the above factors on the dyeing effect of the fabric, rather than judging based on a single parameter. Furthermore, the temporal feature extraction layer 53 employs a recurrent neural network model to extract the dynamic change patterns in the continuous dyeing process. In this embodiment, a Long Short-Term Memory (LSTM) network is preferably used as the temporal feature extraction model. The LSTM network includes an input gate, a forget gate, and an output gate. By memorizing and updating historical production data, the model can learn long-term production change trends. The specific calculation process is as follows: Forget gate: f_t=σ(W_f[h_(t-1),x_t]+b_f) Input gate: i_t=σ(W_i[h_(t-1),x_t]+b_i) State update: C_t = f_tC_(t-1) + i_t\tilde{C_t} Output gate: o_t=σ(W_o[h_(t-1),x_t]+b_o) Where: x_t represents the input data at the current time; h_(t-1) represents the state at the previous time; C_t represents the current hidden state; W and b represent the model weight parameters and bias parameters, respectively. Through the above time series analysis process, it is possible to learn the parameter variation patterns in the continuous dyeing and printing production process; The quality prediction output layer 54 is used to output the staining quality prediction results. In this embodiment, the output results include: (1) Predict the color difference value ΔE; (2) Color uniformity prediction results; (3) Colorfastness prediction results; (4) Batch consistency prediction results.

[0047] For example, when the model predicts that the color difference value ΔE of the fabric will exceed the set range based on the current dye concentration, temperature, fabric guiding speed and historical color change trends, it is determined that there is a potential dyeing deviation risk in the current production process. Through the above dyeing quality prediction model 55, the traditional "inspection after production" can be transformed into "prediction in advance of production process", thereby improving the initiative of quality control.

[0048] S4. Establish a comprehensive evaluation model for staining quality. like Figure 5As shown, this embodiment uses a dyeing quality evaluation module 6 to comprehensively evaluate the predicted dyeing quality results. The dyeing quality evaluation module 6 includes a color difference evaluation unit 61, a uniformity evaluation unit 62, a color fastness evaluation unit 63, and a batch consistency evaluation unit 64. Specifically, the color difference evaluation unit 61 is used to determine the deviation between the predicted color parameters and the target color parameters based on online color detection data. Specifically, it determines the degree of color deviation by calculating the predicted color difference value ΔE. When ΔE ≤ ΔE0, it indicates that the color meets the quality requirements; when ΔE > ΔE0, it indicates that there is a risk of color deviation. The uniformity evaluation unit 62 is used to analyze the degree of color consistency between different areas of the fabric. In the continuous dyeing process, due to factors such as uneven dye liquor circulation and equipment pressure changes, the color of the fabric may be inconsistent in the transverse or longitudinal direction. Therefore, the uniformity evaluation is used to determine the dyeing stability.

[0049] The color fastness evaluation unit 63 is used to predict the color fastness performance of fabrics based on dye liquor state parameters, processing temperature, and historical production data. The batch consistency evaluation unit 64 is used to analyze the color stability between different production batches. Furthermore, a dyeing quality evaluation function is established through the comprehensive quality evaluation model 65: Q=w1E+w2U+w3S+w4B, where: Q represents the comprehensive dyeing quality evaluation value, E represents the color difference evaluation index, U represents the color uniformity index, S represents the color fastness evaluation index, B represents the batch consistency index, and w1, w2, w3, and w4 represent the corresponding weight coefficients. The current production status is determined based on the comprehensive quality evaluation value Q. When Q reaches the preset standard range, the current production parameters are maintained; when Q is lower than the set threshold, the abnormal detection and process adjustment process is initiated. Through the comprehensive quality evaluation model 65, the judgment error caused by the evaluation of a single color index is avoided, and the comprehensiveness of the dyeing quality evaluation is improved.

[0050] S5. Staining quality trend prediction and anomaly warning like Figure 7 As shown, this embodiment uses an anomaly detection and early warning module 7 to identify dyeing quality risks in advance. The anomaly detection and early warning module 7 includes a quality trend prediction unit 71, an anomaly judgment unit 72, and an early warning output unit 73. The quality trend prediction unit 71 analyzes the quality change trend in the future production process based on the continuous prediction results output by the dyeing quality prediction model 55. For example, it predicts color difference changes within multiple consecutive time windows: ΔE1, ΔE2, ΔE3... The system determines whether the color difference is continuously increasing. When a quality indicator is detected to be continuously changing in an unfavorable direction, it indicates that there may be a process deviation in the production process. The anomaly judgment unit 72 makes a judgment based on the preset quality standard. If it predicts that ΔE > ΔEmax or Q < Qmin, it judges that there is an abnormal risk in the current dyeing quality. The early warning output unit 73 outputs the corresponding early warning information. The early warning information includes: anomaly type, anomaly occurrence time, affected production area, and suggested parameter adjustments. When it is detected that the dye concentration has decreased, resulting in a lighter color, the system outputs: "Dye concentration is insufficient, it is recommended to increase the dye replenishment ratio." Through early warning, operators or control systems can make adjustments before obvious quality defects occur in the product.

[0051] S6. Closed-loop control of process parameters based on prediction results like Figure 6 As shown, in this embodiment, the closed-loop control module 8 dynamically adjusts the key process parameters in the continuous dyeing production line 1 based on the dyeing quality prediction results and quality evaluation results, so as to realize the automatic optimization control of dyeing quality. The closed-loop control module 8 includes a control strategy generation unit 81, a dye liquor parameter adjustment unit 82, a temperature parameter adjustment unit 83, an equipment parameter adjustment unit 84, and an execution control unit 85. When the dyeing quality prediction module 5 outputs the prediction results, the control strategy generation unit 81 first analyzes the deviation between the predicted quality results and the target quality requirements. When the predicted color difference value ΔE gradually increases and the comprehensive quality evaluation value Q is lower than the preset threshold, the control strategy generation unit 81 determines the influencing factors based on the current production status. If the dye liquor concentration is detected to be lower than the target range, it is determined that the main influencing factor is the change in the state of the dye liquor. If fluctuations in rolling mill pressure are detected, the main influencing factor is determined to be changes in equipment operating status; if abnormal drying temperature is detected, the main influencing factor is determined to be temperature control deviation. Based on different influencing factors, the control strategy generation unit 81 generates corresponding adjustment strategies, as follows: Dye liquor parameter closed-loop adjustment: The dye liquor supply unit 12 is controlled by the dye liquor parameter adjustment unit 82; the specific adjustment parameters include: (1) dye replenishment amount; (2) auxiliary agent addition amount; (3) pH adjustment amount; (4) dye liquor circulation flow rate. When the prediction model determines that the fabric color has a tendency to be too light, the dye liquor parameter adjustment unit 82 increases the dye replenishment ratio to restore the dye liquor concentration to the target range. When the pH value of the dye solution deviates from the set value, an adjuster is automatically added to restore the dye solution environment to a stable state. In this way, the state of the dye solution is kept within a range suitable for the dyeing reaction, thereby improving the dye utilization rate. Closed-loop temperature parameter regulation: The dyeing processing unit 13 and the drying and finishing unit 14 are controlled by the temperature parameter regulation unit 83. The regulation parameters include: dye liquor temperature, steaming temperature, and drying temperature. In the actual dyeing process, temperature changes will affect the dye diffusion rate, fiber adsorption capacity, and color fixation effect. Therefore, when the prediction model detects an abnormal color change trend, the temperature parameter regulation unit 83 will automatically adjust the operating status of the heating equipment according to the prediction results. When the predicted fabric color is too dark, the dyeing temperature can be appropriately reduced or the high temperature treatment time can be shortened. When the predicted color is too light, the dyeing temperature can be appropriately increased to improve the dye binding efficiency. Through dynamic temperature regulation, the dyeing process can be precisely controlled. Closed-loop regulation of equipment operating parameters: The operating parameters of the continuous dyeing and printing production line are adjusted by the equipment parameter adjustment unit 84. The adjustment parameters include: fabric guide speed, roller pressure, dye liquor circulation speed, and equipment operating load. When insufficient contact time between the fabric and the dye liquor is detected, the fabric guide speed is reduced to increase the fabric immersion time. When insufficient dye liquor carrying rate is detected, the roller pressure is increased to allow the fabric to carry more dye liquor. When uneven local concentration of dye liquor is detected, the dye liquor circulation speed is increased to keep the dye liquor uniform. Finally, the adjustment command is sent to the corresponding actuator by the execution control unit 85 to realize automatic control in the continuous dyeing and printing production process.

[0052] S7. Actual Quality Feedback and Model Optimization To improve the long-term adaptability of the staining quality prediction model 55, this embodiment further includes a feedback optimization module 9, such as... Figure 7 As shown, the feedback optimization module 9 includes an online quality inspection unit 91, a prediction error calculation unit 92, a model parameter update unit 93, and a self-learning optimization unit 94. After adjusting the process parameters, the online quality inspection unit 91 re-collects color parameters and quality inspection data from the actual production process. Subsequently, the prediction error calculation unit 92 calculates the error between the model prediction result and the actual inspection result. The calculation formula is as follows: Ep = |Yr - Yp| where: Ep represents the prediction error, Yr represents the actual dyeing quality result obtained from the inspection, and Yp represents the dyeing quality result predicted by the model. When the prediction error is less than the set range, it indicates that the current prediction model can accurately reflect the changing patterns of the production process. When the prediction error exceeds the set range, it indicates that the current model needs further optimization. At this time, the model parameter update unit 93 adjusts the weight parameters in the prediction model. Meanwhile, the self-learning optimization unit 94 continuously trains the model using newly added production data. For example, corresponding data samples are established for different fabric types: cotton fiber fabrics, polyester fabrics, blended fabrics, different dye systems, reactive dyes, disperse dyes, and acid dyes, so that the model can adapt to different production conditions. Through a continuous learning mechanism, the adaptability of the dyeing quality prediction model 55 to complex production environments is improved. First, during the operation of the continuous dyeing and printing production line 1, the multi-source data acquisition module 2 collects real-time data on fabric parameters, dye liquor state parameters, equipment operating parameters, environmental parameters, and online color detection data. Then, the data preprocessing module 3 cleans, synchronizes, and normalizes the collected data, and the data fusion module 4 establishes continuous dyeing and printing process state characteristic data. Furthermore, the fused state data is input into the dyeing quality prediction module 5, which analyzes the dynamic change patterns in the production process through a time-series prediction model to predict future dyeing quality trends. When the prediction results meet the quality requirements, the system maintains the current production state. When the prediction results have a risk of deviation, the dyeing quality evaluation module 6 judges the degree of impact, and the anomaly detection and early warning module 7 outputs quality risk information. Subsequently, the closed-loop control module 8 automatically adjusts the dye liquor parameters, temperature parameters, and equipment operating parameters based on the prediction results, so that the production status is restored to the target range. After the adjustment is completed, the feedback optimization module 9 collects the actual production results and optimizes the model parameters based on the prediction error, so that the system forms a closed-loop intelligent control process of "data acquisition - quality prediction - anomaly judgment - parameter adjustment - effect feedback - model optimization".

[0053] The above are all preferred embodiments of this application, and are not intended to limit the scope of protection of this application. Therefore, all equivalent changes made in accordance with the structure, shape and principle of this application should be covered within the scope of protection of this application.

Claims

1. A method for predicting and controlling dyeing quality in continuous dyeing and printing production lines, characterized in that, Includes the following steps: S1. Obtain multi-source production data during the operation of the continuous dyeing and printing production line; The multi-source production data includes fabric parameter data, dye liquor status data, equipment operating status data, environmental parameter data, and online color detection data; S2. Preprocess the multi-source production data and perform multi-dimensional feature fusion to construct continuous dyeing process state feature data; S3. Based on the state characteristic data of the continuous dyeing process, establish a dyeing quality prediction model, and use the dyeing quality prediction model to predict the dyeing quality of the target fabric to obtain the dyeing quality prediction result. S4. Based on the dyeing quality prediction results, establish a dyeing quality evaluation model, comprehensively evaluate the dyeing effect, and determine whether the current dyeing status meets the preset quality requirements. S5. When the dyeing quality prediction result deviates from the target quality range, a corresponding process parameter adjustment strategy is generated. S6. Dynamically adjust the dye liquor parameters, temperature parameters, and equipment operating parameters in the continuous dyeing and printing production line according to the process parameter adjustment strategy. S7. Collect the adjusted actual dyeing test data, and optimize the dyeing quality prediction model based on the error between the actual dyeing results and the predicted results to form a closed-loop control of dyeing quality prediction and process parameter adjustment.

2. The dyeing quality prediction and closed-loop control method for continuous dyeing and printing production lines according to claim 1, characterized in that, The multi-source production data is acquired through a multi-source data acquisition module (2) installed on the continuous dyeing and printing production line (1); The continuous dyeing and printing production line (1) includes a fabric input unit (11), a dye supply unit (12), a dyeing processing unit (13), a drying and finishing unit (14), and a finished product output unit (15). The fabric parameter data includes fabric material, fiber composition, fabric weight, fabric thickness, moisture content, and initial color parameters. The dye liquor state data includes dye concentration, auxiliary agent concentration, pH value, dye liquor temperature, conductivity, and viscosity. The equipment operating status data includes the fabric guiding speed, rolling mill pressure, dye liquor circulation flow rate, and equipment operating load.

3. The dyeing quality prediction and closed-loop control method for continuous dyeing and printing production lines according to claim 1, characterized in that, The multi-source data preprocessing includes: The data preprocessing module (3) performs data cleaning, outlier removal, time synchronization and normalization on the collected data. Wherein: the data cleaning unit (31) is used to remove invalid data; The abnormal data removal unit (32) is used to identify and delete abnormal sampled data; The time synchronization unit (33) is used to synchronize the data obtained by different acquisition devices according to the production time axis; The data normalization unit (34) is used to eliminate dimensional differences between different types of data.

4. The dyeing quality prediction and closed-loop control method for continuous dyeing and printing production lines according to claim 1, characterized in that, The multidimensional feature fusion includes: The data fusion module (4) fuses the fabric features, dye liquor features, equipment operation features and color detection features; the data fusion module (4) includes a multi-source feature fusion unit (41), a dyeing process state feature vector generation unit (42), and a time-series data construction unit (43). Establish the state feature vector of the staining process: F=[Fp,Fd,Fe,Fc] Where: Fp represents the characteristics of the greige fabric; Fd represents the characteristics of the dye liquor state; Fe represents the characteristics of the equipment operating state; Fc represents the characteristics of color detection. The continuous dyeing process time series data is generated based on the state feature vector.

5. The dyeing quality prediction and closed-loop control method for continuous dyeing and printing production lines according to claim 1, characterized in that, The staining quality prediction model includes an input layer (51), a feature fusion layer (52), a temporal feature extraction layer (53), and a quality prediction output layer (54). The temporal feature extraction layer adopts a recurrent neural network model, including at least one of the Long Short-Term Memory (LSTM) network model and the Gated Recurrent Unit (GRU) model; The dyeing quality prediction model is used to predict dyeing quality parameters based on production status data over a continuous time period.

6. The dyeing quality prediction and closed-loop control method for continuous dyeing and printing production lines according to claim 5, characterized in that, The dyeing quality prediction results output by the quality prediction output layer (54) include: predicted color difference value ΔE, color uniformity index, color fastness prediction index, and batch consistency index. Based on the above prediction results, staining quality evaluation parameters are formed.

7. The dyeing quality prediction and closed-loop control method for continuous dyeing and printing production lines according to claim 1, characterized in that, The dyeing quality evaluation model (65) includes: a color difference evaluation unit (61), a uniformity evaluation unit (62), a color fastness evaluation unit (63), and a batch consistency evaluation unit (64). The evaluation indicators are weighted and calculated using a comprehensive quality evaluation model: Q = w1E + w2U + w3S + w4B; Where: Q represents the comprehensive dyeing quality evaluation value; E represents the color difference evaluation index; U represents the color uniformity index; S represents the color fastness index; and B represents the batch consistency index.

8. The dyeing quality prediction and closed-loop control method for continuous dyeing and printing production lines according to claim 1, characterized in that, The anomaly detection and early warning module (7) performs trend analysis based on the staining quality prediction results; When the predicted color difference value exceeds the preset threshold, the anomaly judgment unit (72) generates a dyeing anomaly judgment result and outputs quality warning information through the warning output unit (73); The quality warning information includes the anomaly type, anomaly location, and corresponding adjustment suggestions; The closed-loop control module (8) controls the operation of the continuous dyeing and printing production line according to the process parameter adjustment strategy. The process parameter adjustment strategy includes: adjusting the amount of dye added, the amount of auxiliary agent added, and the pH value through the dye liquor parameter adjustment unit (82); The temperature of the dye liquor, the evaporation temperature and the drying temperature are adjusted by the temperature parameter adjustment unit (83); Adjust the guide speed, rolling pressure and dye liquor circulation speed through the equipment parameter adjustment unit (84); The feedback optimization includes: acquiring actual staining quality detection data through the online quality detection unit (91); The error between the prediction result and the actual detection result is calculated by the prediction error calculation unit (92): Ep=|Yr-Yp| Where: Ep represents the prediction error; Yr represents the actual detection result; Yp represents the model prediction result; Based on the prediction error, the parameters of the staining quality prediction model are adjusted through the model parameter update unit (93), and the model is continuously optimized through the self-learning optimization unit (94).

9. A dyeing quality prediction and closed-loop control system for continuous dyeing and printing production lines, characterized in that, include: The multi-source data acquisition module (2) is used to collect data on fabric parameters, dye liquor status, equipment operating status, environmental parameters, and color detection during the operation of the continuous dyeing and printing production line. The data preprocessing module (3) is used to clean, synchronize, and normalize the collected data; The data fusion module (4) is used to construct the state feature data of the staining process; The staining quality prediction module (5) is used to output staining quality prediction results based on the prediction model; The staining quality evaluation module (6) is used to evaluate staining quality based on the prediction results; The anomaly detection and early warning module (7) is used to identify abnormal staining states and output early warning information; The closed-loop control module (8) is used to adjust the dyeing and printing production process parameters according to the prediction results; The feedback optimization module (9) is used to optimize the prediction model based on the actual staining results.

10. An electronic device, characterized in that, The system includes a processor, a memory, and a computer program stored in the memory. When the computer program is executed by the processor, it causes the processor to execute the dyeing quality prediction and closed-loop control method for continuous dyeing and printing production lines as described in any one of claims 1 to 9. The computer program is stored in a computer-readable storage medium.

Images