Printing and dyeing production decision and execution control system based on fundamental research

Abstract

According to the printing and dyeing production decision-making and execution control system based on the fundamental research, a scientific basis is provided for production decision-making through the process-quality mapping model established through the fundamental research, a traditional mode depending on artificial experience is changed fundamentally, scientificity and accuracy of process formulation are remarkably improved, and the system is suitable for large-scale popularization and application. Therefore, the one-time success rate and the product quality stability are greatly improved; a dynamic optimal process is generated by adopting a multi-objective optimization algorithm, so that the cost, the energy consumption and the time can be comprehensively optimized on the premise of ensuring the quality, and cost reduction, efficiency improvement and green production are realized; real-time execution and correction are carried out through a model predictive control algorithm, so that the system has look-ahead control and anti-interference capabilities, dynamic changes in the production process are effectively coped with, and accurate reproduction of process execution is ensured; the self-learning mechanism can utilize production data to continuously iteratively optimize the core model, so that the system has the ability of continuous evolution, the dependence on initial basic research is gradually reduced, and the production efficiency is improved for a long time.

Description

Technical Field

[0001] This invention relates to the field of dyeing and printing production technology, and in particular to a dyeing and printing production decision and execution control system based on basic research. Background Technology

[0002] As a crucial component of the textile industry, the dyeing and printing industry is a typical process manufacturing sector. Its production process essentially involves a series of complex physical and chemical reactions between dyes and fibers. For a long time, the production management of my country's dyeing and printing enterprises has heavily relied on the personal experience of skilled workers. From formula determination and process route design to production process control, decisions are largely made based on the experience of veteran workers and manual operation. This traditional model results in a lack of solid scientific basis and reproducible data support for the production process, leading to poor production stability and low first-pass yield, severely hindering the transformation and upgrading of the dyeing and printing industry towards high quality, high efficiency, and green practices.

[0003] With the development of industrial automation technology, some control systems aimed at improving the automation level of the dyeing and printing process have emerged in the market. However, most of these systems have limited functionality, only able to achieve simple program control based on preset formulas and fixed process parameters, or single-point feedback control of the equipment's operating parameters. They cannot cope with disturbances caused by dynamic factors such as batch differences in raw materials, fluctuations in equipment status, and environmental changes in dyeing and printing production. Essentially, existing control systems are a simple replication and solidification of "human experience," rather than based on a deep understanding and modeling of the inherent physicochemical laws of the dyeing and printing process. Therefore, they face the dilemma of not being able to fundamentally solve the problem of the scientific nature of process decision-making, and also lack the ability to adaptively adjust and proactively optimize based on real-time production status.

[0004] Existing technical solutions suffer from several problems: First, there is a disconnect between process decision-making and process control. Decision-making relies on experience, and control depends on fixed procedures, failing to form a closed-loop linkage based on data models. Second, there is a lack of in-depth analysis of the dyeing and printing process mechanisms and data integration, making it impossible to construct predictive models that accurately describe the intrinsic relationship between "process parameters, process state, and final quality," thus lacking theoretical guidance for optimization and correction. Third, control strategies are outdated, mostly relying on "post-event" corrections, failing to achieve "pre-event" prediction and "in-event" dynamic adjustment, making it difficult to ensure that production results consistently meet expectations. These problems directly lead to development bottlenecks for dyeing and printing enterprises, such as processing fees rising slower than costs, large fluctuations in product quality, high energy and material consumption, and heavy reliance on skilled workers.

[0005] Therefore, there is an urgent need in this field for a novel solution that can deeply integrate fundamental theoretical research on the dyeing and printing process with modern intelligent control technology. This solution must be able to transform implicit expert experience into explicit, computable, and optimizable digital models, and on this basis, build an intelligent system capable of autonomous decision-making, precise execution, and continuous learning and evolution from practice. This would fundamentally change the traditional experience-based production model of the dyeing and printing industry, achieving a breakthrough in digital and intelligent transformation. Summary of the Invention

[0006] The purpose of this invention is to provide a decision-making and execution control system for dyeing and printing production based on basic research, so as to solve the problems existing in the prior art.

[0007] To achieve the above objectives, the present invention provides the following solution: This invention provides a decision-making and execution control system for dyeing and printing production based on basic research, comprising: The basic research and digital twin module is used to construct a material feature library, an equipment dynamics model library, and a process-quality mapping model by conducting basic experiments on dyes, auxiliaries, and fiber fabrics, thereby forming a process digital twin of the dyeing and printing process. The intelligent decision-making and optimization module, connected to the basic research and digital twin module, is used to receive order information and call the process-quality mapping model to calculate the dynamic optimal process parameters that meet the constraints of quality, cost, energy consumption and time through a multi-objective optimization algorithm. The adaptive execution and control module, connected to the intelligent decision-making and optimization module, is used to receive the dynamic optimal process parameters and perform real-time control and dynamic correction of the dyeing and printing production process based on the model predictive control algorithm. At the same time, it feeds back the production process data to the intelligent decision-making and optimization module and the basic research and digital twin module to form a closed-loop optimization.

[0008] Preferably, in the basic research and digital twin module, the process-quality mapping model is constructed as follows: Machine learning algorithms are used to generate process parameter vectors. Material feature vector and device parameter vector As input, with quality index vector As output, the trained prediction model is obtained, and the expression of the prediction model is: ; Among them, process parameter vector Includes temperature curves, pH curves, additive concentration sequences; material feature vectors. Including dye molecular weight, diffusion coefficient, fabric porosity; equipment parameter vector. Including parameters such as the heat transfer coefficient and heat exchange efficiency of the dyeing vat; Including dye uptake, color difference, and fastness; This is the random error term.

[0009] Preferably, in the intelligent decision-making and optimization module, the expression of the multi-objective optimization algorithm is: ; in, To predict the staining rate, To predict color difference, This represents the total cost of the process. Total process time Total energy consumption of the process. Weighted by the dyeing rate, For color difference weighting, As a cost weight, As time weight, Energy consumption weighting; The multi-objective optimization algorithm is solved using a multi-objective evolutionary algorithm or a particle swarm optimization algorithm.

[0010] Preferably, it further includes: The self-learning and model update module is used to collect actual process data and measured quality data generated during the production process, and periodically use new data to retrain or incrementally update the process-quality mapping model in order to continuously improve the model prediction accuracy.

[0011] A method for decision-making and execution control in dyeing and printing production based on basic research, comprising: S1. Through basic research and digital twin modules, construct a process digital twin of the dyeing and printing process, and train it to obtain a process-quality mapping model; S2. The intelligent decision-making and optimization module receives production orders and generates dynamic optimal process parameters based on the process-quality mapping model and using a multi-objective optimization algorithm. S3. Through the adaptive execution and control module, based on the model predictive control algorithm, the dynamic optimal process parameters are executed, and the dyeing and printing process is monitored and dynamically corrected in real time. S4. Through the self-learning and model update module, production feedback data is collected to iteratively optimize the process-quality mapping model.

[0012] 6. The dyeing and printing production decision-making and execution control method based on basic research according to claim 5, characterized in that, in step S1, constructing the process-quality mapping model specifically includes: Design experimental schemes covering a wide range of processes, and collect sample data on process parameters, material characteristics, equipment parameters and corresponding quality indicators; The sample data is trained using support vector regression, random forest, or neural network algorithms to obtain the process-quality mapping model.

[0013] Preferably, in step S2, generating the dynamic optimal process parameters specifically includes: Analyze production orders to determine target color, fabric type, and weighting coefficients for multi-objective optimization; Using the process-quality mapping model as a simulator, a multi-objective evolutionary algorithm is used to search in the process parameter space to find the process parameter combination that minimizes the objective function, which is then used as the dynamically optimal process parameter.

[0014] Preferably, step S3 specifically includes: Based on the current production status and the process-quality mapping model, predict the impact of process parameters on quality indicators in the future period. Solve a real-time optimization problem by calculating the sequence of future control commands that makes the predicted output closest to the target trajectory; Implement the first control action in the sequence of control commands.

[0015] Preferably, in step S4, the iterative optimization of the process-quality mapping model specifically includes: After each production run, the actual process data and measured quality data are stored as new samples in the historical database. When the number of new samples reaches a preset threshold, the model update process is triggered, and the process-quality mapping model is retrained using all historical data, or the model parameters are updated online using an incremental learning algorithm.

[0016] The present invention achieves the following beneficial technical effects compared to the prior art: This invention provides a decision-making and execution control system for dyeing and printing production based on basic research. The process-quality mapping model established through basic research provides a scientific basis for production decisions, fundamentally changing the traditional model that relies on human experience. This significantly improves the scientific rigor and accuracy of process formulation, thereby greatly increasing the first-pass yield and product quality stability. A multi-objective optimization algorithm generates a dynamically optimal process, which can comprehensively optimize costs, energy consumption, and time while ensuring quality, achieving cost reduction, efficiency improvement, and green production. Real-time execution and correction through model predictive control algorithms enable the system to have forward-looking control and anti-interference capabilities, effectively responding to dynamic changes in the production process and ensuring accurate reproduction of process execution. A self-learning mechanism can continuously iterate and optimize the core model using production data, enabling the system to continuously evolve, gradually reducing reliance on initial basic research and improving production efficiency in the long term. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 A schematic diagram of a dyeing and printing production decision and execution control system based on basic research provided by this invention; Figure 2 The flowchart of the dyeing and printing production decision-making and execution control method based on basic research provided by this invention. Detailed Implementation

[0019] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0020] The purpose of this invention is to provide an intelligent system that can make autonomous decisions, execute precisely, and continuously learn by deeply integrating the physicochemical mechanisms of the dyeing and printing process with data-driven intelligent technology.

[0021] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0022] Example 1: like Figure 1 As shown, the core of the system of this invention lies in constructing a closed-loop control architecture, which is organically composed of a basic research and digital twin module, an intelligent decision-making and optimization module, an adaptive execution and control module, and a self-learning and model update module.

[0023] In practical implementation, these modules connect to communication protocols (such as OPC UA or Industrial Ethernet) via a data bus to ensure real-time interaction between data and command streams. The system is deployed on an industrial server or cloud platform and integrated with field dyeing equipment (such as dyeing vats, sensors, and actuators) through an IoT gateway.

[0024] Please refer to Figure 2 The basic research-based decision-making and execution control method for dyeing and printing production using the above system is as follows: In the basic research and digital twin module, the primary task is to construct a process digital twin of the dyeing and printing process through systematic experimental analysis. This digital twin is not a simple data model, but a virtual mirror that integrates material properties, equipment dynamics, and process-quality mapping relationships.

[0025] Specifically, establishing a material characteristic library requires detailed physicochemical testing of dyes, auxiliaries, and fiber fabrics. For example, the absorption spectrum of dyes is measured using a UV-Vis spectrophotometer, the diffusion coefficient of dyes is determined using the capillary method, and the porosity of the fabric is calculated using microscopic image analysis techniques.

[0026] These parameters are quantized into material feature vectors. This includes the dye molecular weight (unit: g / mol) and diffusion coefficient (unit: m). 2 Key properties include per second (S / s) and fabric porosity (dimensionless). The equipment dynamics model library models different types of dyeing vats to describe their heat transfer, mass transfer, and fluid behavior.

[0027] For example, the heating process of a dye vat can be described by the following kinetic model: ; in, The temperature inside the dye vat at time t is expressed in °C. Initial temperature (unit: °C). Steam heating power (unit: W). The mass of water in the bath (unit: kg). This is the specific heat capacity (constant) of water. Heat exchange efficiency coefficient (unit: s) related to dyeing vat structure and stirring rate -1 ), The time unit is seconds. The model fits parameters using historical running data. This allows for accurate prediction of temperature change trajectories.

[0028] The process-quality mapping model is the core of this module. It is trained using a machine learning algorithm, and its mathematical expression is: ; here, This is a vector of quality metrics, including dye uptake rate. (Value range 0-1), color difference (Calculated based on CIELAB color space), fastness, etc.; This is a vector of process parameters, including temperature profiles. pH curve , Auxiliary agent concentration sequence; This is a vector of device parameters, including the heat transfer coefficient. Heat exchange efficiency, etc.; The random error term follows a normal distribution with a mean of zero. During model training, support vector regression (SVR) or neural network algorithms are used, with a large amount of historical experimental data as samples, to optimize model parameters by minimizing the prediction error. For example, for the reactive dyeing process, dyeing rate and color difference data are collected under different temperatures, pH values, and auxiliary agent dosages to train a nonlinear regression model that can accurately predict the quality output under given process conditions.

[0029] The intelligent decision-making and optimization module is responsible for generating dynamically optimal process parameters based on production orders. During implementation, this module first parses order information, including the target color (LAB value), fabric type, and production constraints (such as maximum energy consumption limits). Then, it calls the process-quality mapping model as a simulator and combines it with a multi-objective optimization algorithm to calculate the optimal process. The optimization objective function is defined as: ; in, and Predicted output from the process-quality mapping model; The total process cost is calculated using the following formula: ; in, For the first The unit price of a dye / auxiliary agent. No. The amount of each dye / auxiliary agent used. This is the unit price for water. For water consumption, This is the steam cost coefficient. For steam energy consumption, This is the electricity cost coefficient. This refers to electricity consumption, encompassing the costs of dyes, auxiliaries, water, steam, and electricity. This refers to the total process time, including the heating, holding, and cooling stages. Calculate the steam and electricity consumption for total energy consumption; Weighted by the dyeing rate, For color difference weighting, As a cost weight, As time weight, Energy consumption weights are dynamically adjusted based on production strategies. For example, quality is prioritized in high-end fabric production, while cost and efficiency are emphasized in bulk orders. The optimization algorithm employs multi-objective evolutionary algorithms or particle swarm optimization, performing a global search in the process parameter space to output a Pareto optimal solution set. Finally, the operator or automated rules select the most suitable dynamic process parameters.

[0030] The adaptive execution and control module is responsible for translating optimized process parameters into real-time control actions. Based on a model predictive control algorithm, this module executes the following steps in each control cycle (e.g., per second): First, it collects real-time data on temperature, pH, conductivity, and K / S ratios measured by an online spectrophotometer within the dyeing vat via a sensor network. Then, starting from the current state, it uses a process-quality mapping model to predict future timeframes (…). The quality change of ). Then, solve a real-time optimization problem: ; in, Indicates in Time prediction Moment quality indicators For reference trajectory, To control increments (such as changes in the opening degree of steam valves). and These are the weight matrices for output error and control increment, respectively. After solving, only the first control action is implemented, and data is re-acquired and rolled optimization is performed in the next cycle. This proactive control strategy can effectively cope with disturbances such as steam pressure fluctuations and fabric batch differences, ensuring that the process execution always approaches the optimal trajectory.

[0031] The self-learning and model update module enables the system to continuously evolve. After each batch of production, the module collects actual process data and measured quality data, storing them as new samples in the historical database. When the accumulated samples reach a preset threshold (e.g., 100 batches of data), the model update process is triggered. Updates can be performed using batch retraining, i.e., retraining the process-quality mapping model using all historical data; or using incremental learning algorithms to quickly adjust model parameters without losing historical knowledge. For example, for newly introduced dye types, the system can quickly adapt to their characteristics through incremental learning, reducing reliance on initial basic research.

[0032] In specific application scenarios, taking reactive dyeing of cotton fabrics as an example, the system first determines the hydrophilicity of the cotton fabric and the reaction kinetic parameters of the dye through basic research, and establishes a heat transfer model for the dyeing vat. Upon receiving an order, the intelligent decision-making module optimizes dynamic process parameters based on the target deep blue color and cost constraints, including segmented heating curves (e.g., heating from 50℃ to 80℃ at 1.5℃ / min, holding for 30 minutes) and auxiliary agent addition sequences. During execution, the MPC controller monitors the K / S value in real time. If the dyeing rate is found to be lower than expected, it automatically adjusts the steam valve opening to accelerate heating, while simultaneously feeding back data to the decision-making module for further optimization. Through this closed-loop operation, the system achieves a fundamental shift from experience-driven to model-driven approaches, significantly improving the first-pass success rate and resource utilization.

[0033] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0034] It should be noted that the components mentioned in the above embodiments are all general standard parts or components known to those skilled in the art. Their structures and principles can be learned by those skilled in the art through technical manuals or conventional experimental methods.

[0035] This invention has illustrated its principles and implementation methods using specific examples. The descriptions of these embodiments are merely illustrative of the method and its core ideas; furthermore, those skilled in the art will recognize that modifications may be made to the specific implementation methods and application scope based on the principles of this invention. Therefore, the content of this specification should not be construed as limiting the invention.

Claims

1. A decision-making and execution control system for dyeing and printing production based on basic research, characterized in that, include: The basic research and digital twin module is used to construct a material feature library, an equipment dynamics model library, and a process-quality mapping model by conducting basic experiments on dyes, auxiliaries, and fiber fabrics, thereby forming a process digital twin of the dyeing and printing process. The intelligent decision-making and optimization module, connected to the basic research and digital twin module, is used to receive order information and call the process-quality mapping model to calculate the dynamic optimal process parameters that meet the constraints of quality, cost, energy consumption and time through a multi-objective optimization algorithm. The adaptive execution and control module, connected to the intelligent decision-making and optimization module, is used to receive the dynamic optimal process parameters and perform real-time control and dynamic correction of the dyeing and printing production process based on the model predictive control algorithm. At the same time, it feeds back the production process data to the intelligent decision-making and optimization module and the basic research and digital twin module to form a closed-loop optimization.

2. The dyeing and printing production decision-making and execution control system based on basic research as described in claim 1, characterized in that, In the basic research and digital twin module, the process-quality mapping model is constructed as follows: Machine learning algorithms are used to generate process parameter vectors. Material feature vector and device parameter vector As input, with quality index vector As output, the trained prediction model is obtained, and the expression of the prediction model is: ； Among them, process parameter vector Includes temperature curves, pH curves, additive concentration sequences; material feature vectors. Including dye molecular weight, diffusion coefficient, fabric porosity; equipment parameter vector. Including parameters such as the heat transfer coefficient and heat exchange efficiency of the dyeing vat; Including dye uptake, color difference, and fastness; This is the random error term.

3. The dyeing and printing production decision-making and execution control system based on basic research as described in claim 2, characterized in that, In the intelligent decision-making and optimization module, the expression for the multi-objective optimization algorithm is: ； in, To predict the staining rate, To predict color difference, This represents the total cost of the process. Total process time Total energy consumption of the process. Weighted by the dyeing rate, For color difference weighting, As a cost weight, As time weight, Energy consumption weighting; The multi-objective optimization algorithm is solved using a multi-objective evolutionary algorithm or a particle swarm optimization algorithm.

4. The dyeing and printing production decision-making and execution control system based on basic research as described in claim 1, characterized in that, Also includes: The self-learning and model update module is used to collect actual process data and measured quality data generated during the production process, and periodically use new data to retrain or incrementally update the process-quality mapping model in order to continuously improve the model prediction accuracy.

5. A method for decision-making and execution control in dyeing and printing production based on basic research, characterized in that, The dyeing and printing production decision and execution control system based on basic research as described in any one of claims 1-4 includes: S1. Through basic research and digital twin modules, construct a process digital twin of the dyeing and printing process, and train it to obtain a process-quality mapping model; S2. The intelligent decision-making and optimization module receives production orders and generates dynamic optimal process parameters based on the process-quality mapping model and using a multi-objective optimization algorithm. S3. Through the adaptive execution and control module, based on the model predictive control algorithm, the dynamic optimal process parameters are executed, and the dyeing and printing process is monitored and dynamically corrected in real time. S4. Through the self-learning and model update module, production feedback data is collected to iteratively optimize the process-quality mapping model.

6. The dyeing and printing production decision-making and execution control method based on basic research according to claim 5, characterized in that, In step S1, constructing the process-quality mapping model specifically includes: Design experimental schemes covering a wide range of processes, and collect sample data on process parameters, material characteristics, equipment parameters and corresponding quality indicators; The sample data is trained using support vector regression, random forest, or neural network algorithms to obtain the process-quality mapping model.

7. The dyeing and printing production decision-making and execution control method based on basic research according to claim 5, characterized in that, Step S2, generating the dynamic optimal process parameters specifically includes: Analyze production orders to determine target color, fabric type, and weighting coefficients for multi-objective optimization; Using the process-quality mapping model as a simulator, a multi-objective evolutionary algorithm is used to search in the process parameter space to find the process parameter combination that minimizes the objective function, which is then used as the dynamically optimal process parameter.

8. The dyeing and printing production decision-making and execution control method based on basic research according to claim 5, characterized in that, Step S3 specifically includes: Based on the current production status and the process-quality mapping model, predict the impact of process parameters on quality indicators in the future period. Solve a real-time optimization problem by calculating the sequence of future control commands that makes the predicted output closest to the target trajectory; Implement the first control action in the sequence of control commands.

9. The dyeing and printing production decision-making and execution control method based on basic research according to claim 5, characterized in that, Step S4, the iterative optimization of the process-quality mapping model specifically includes: After each production run, the actual process data and measured quality data are stored as new samples in the historical database. When the number of new samples reaches a preset threshold, the model update process is triggered, and the process-quality mapping model is retrained using all historical data, or the model parameters are updated online using an incremental learning algorithm.

Images