Method for establishing digital twin model of textile surface quality under periodic friction and wear

Abstract

The present application relates to a kind of periodic friction wear under textile surface quality digital twin model establishment method, it is obtained by using the internet of things technology to dyeing machine, fabric is configured with multiple sensors and so on mode to obtain the tension and fabric friction wear data suffered in fabric dyeing process;On the basis of obtaining data, establish fabric geometric model, finite element deformation simulation model, fabric dyeing friction wear prediction model and other digital twin virtual model, real-time mapping and simulation are carried out to fabric dyeing process;Finally, establish dyeing fabric tension optimization model, optimize the tension suffered when fabric dyeing, form the tension selection rule suffered by fabric, improve the surface quality of fabric dyeing process.

Description

TECHNICAL FIELD

[0001] The application relates to an optimization method for different fabric digital twin models, in particular to a method for establishing a textile surface quality digital twin model under periodic friction and wear, and belongs to the technical field of fabric dyeing friction. BACKGROUND

[0002] Fabric is formed by interconnecting loop structures, and when subjected to stress, the deformation shows strong complexity. How to accurately characterize and analyze the mesoscopic state and macroscopic deformation of the textile in the dyeing process, establish a high-precision fabric low-tension-deformation control model and a fabric dyeing scratch model, and develop a low-tension control technology is not only the focus of basic problem research of fabric dyeing, but also a key technology that needs to be solved in fabric dyeing processing. However, most of the existing technologies still remain at the level of monitoring data display, and cannot intuitively reflect the actual running condition of the fabric in the dyeing process, and cannot truly reflect the running state of the fabric in the dyeing process. Therefore, there is an urgent need for a technology that can intuitively and truly reflect the friction and wear of the fabric to be dyed under the dyeing running condition.

[0003] Digital twin technology is a technology for virtual simulation and physical reality fusion proposed in recent years, which is a technology for integrating physical reality data, integrating multi-field and multi-scale mapping simulation models, and covering the whole process of processing and product life cycle. The research on the fabric dyeing process involves multiple disciplines, including fabric pretreatment, dyeing cycle control and fabric finishing after dyeing, and other processing links. The fabric dyeing equipment is a coupled system based on mechanical structure and electrical element control, and the fabric dyeing involves complex processes such as stretching deformation and dye immersion, and contains multiple data such as motor speed, fabric tension size and dyeing rate. If the digital twin model technology is applied to the key problem research in the field of textile dyeing processing, it will provide a reference for the research on the friction and wear problems of the textile dyeing process.

[0004] Therefore, in order to solve the above problems, it is necessary to provide an innovative method for establishing a textile surface quality digital twin model under periodic friction and wear, so as to overcome the defects in the prior art. SUMMARY

[0005] The purpose of the present application is to provide a method for establishing a textile surface quality digital twin model under periodic friction and wear, which can truly and intuitively reflect the real-time running state of the fabric in the dyeing cycle process, and solve the problem that the friction and wear of the fabric are difficult to determine in the fabric dyeing process,

[0006] To achieve the above purpose, the technical scheme adopted by the present application is as follows: a method for establishing a textile surface quality digital twin model under periodic friction and wear, comprising the following steps:

[0007] 1), obtaining offline data of the fabric to be dyed before dyeing and the dyeing machine, including geometric parameters, structural characteristics and material properties, and the obtained online data of the fabric during the entire dyeing process, including the circumferential driving force applied on the guide roller of the dyeing machine, the pressure of the fabric on it and the friction torque of the fabric and the surface of the guide roller, the surface wear of the fabric during the dyeing cycle and the dyeing environment data of the fabric to be dyed;

[0008] 2), based on the digital twin model, a physical space model and a virtual space model of the fabric to be dyed and the guide roller of the dyeing machine are established; the virtual space model includes a geometric model, a finite element simulation model, a quantitative model of the friction and tension and speed of the fabric, and a quality prediction model of the textile under the processing parameters;

[0009] 3), the data of the tension, friction and cycle speed of the fabric during the dyeing process in actual production are transmitted to the physical space model, and the virtual space model can directly call the real-time information of the physical space model, and then the dyeing process of the fabric through the guide roller of the dyeing machine is displayed in real time through the human-computer interaction device;

[0010] 4), simulation is carried out in the virtual space model of the fabric-guide roller of the dyeing machine, the fabric tension during the dyeing cycle is real-time controlled, and then the finished fabric is predicted according to the dyeing scratch prediction model;

[0011] 5), the adjustment process of the dyeing cycle speed and tension of the textile under the digital twin model and the optimization process of the periodic friction and wear surface quality are executed, when the entire dyeing process of the fabric is completed, all process data are stored to the physical space model of the fabric-guide roller of the dyeing machine, on the basis of the fabric tension calculation model and the fabric friction and wear prediction model, a group of optimal dyeing cycle speed and fabric tension are obtained through the genetic algorithm, so that the scratch degree of the fabric during the dyeing process is minimized, so that the appropriate fabric tension and dyeing cycle speed during the dyeing process are determined, so as to ensure the surface quality of the finally dyed fabric.

[0012] The method for establishing the digital twin model of the surface quality of the textile under the periodic friction and wear of the application further comprises the following steps in step 1):

[0013] 1.1), before dyeing, the fabric to be dyed and the dyeing machine are detected to obtain the geometric parameters, material properties and structural characteristics of the fabric and the dyeing machine;

[0014] 1.2), data monitoring points are set up on the fabric, a force sensor is used to measure the tensile tension of the fabric at the contact position with the dyeing machine guide roller, a speed measuring sensor device is arranged to obtain the fabric dyeing cycle speed data of the dyeing machine monitoring point, and temperature and pressure sensors are arranged in the dyeing machine cylinder to obtain the dyeing process environment data.

[0015] The method for establishing the periodic friction and wear textile surface quality digital twin model further comprises the following steps:

[0016] 2.1), a physical space model of the fabric-dyeing machine guide roller is established, the entire fabric dyeing system is converted from a physical entity into a Web-identified data model, and the fabric-dyeing machine guide roller information is tracked and called;

[0017] 2.2), a virtual space model of the fabric-dyeing machine guide roller is constructed, including a geometric model, a finite element simulation model, a quantitative model of the friction and the fabric tension and speed of the fabric, and a quality prediction model of the textile under the processing parameters, the fabric abrasion state in the dyeing process is mapped in real time, and the abrasion degree is quantified.

[0018] The method for establishing the periodic friction and wear textile surface quality digital twin model further comprises the following steps:

[0019] 2.2.1), a geometric model of the fabric and the dyeing machine guide roller is constructed according to the data obtained in step 1;

[0020] 2.2.2), a finite element simulation model of the fabric under tension deformation is constructed by using a finite element software;

[0021] 2.2.3), based on a deep neural network, a fabric tension control model and a fabric friction and wear surface quality prediction model are established, a nonlinear mapping relationship between the fabric tensile tension and the fabric friction and wear is constructed, the simulated fabric tensile deformation is used as a training sample set of the fabric friction and wear prediction model by using the finite element simulation model, the samples are standardized, and finally the fabric abrasion prediction model is trained.

[0022] The method for establishing the periodic friction and wear textile surface quality digital twin model further comprises the following steps:

[0023] 2.2.3.1), a script file is written according to the established finite element simulation model by using the parametric analysis function of the finite element software, finite element calculation is performed by changing the contact position coordinates and the dyeing cycle speed, and the tensile tension data of the fabric corresponding to the measurement point is obtained;

[0024] 2.2.3.2), determine the input and output of the fabric surface friction and wear prediction model; the input unit of the fabric friction and wear prediction model is the coordinate parameters of the set contact points and the tensile stress load applied to the fabric to be dyed, expressed as x={x n (1≤n≤N), x n ={s n F n}, s n F represents the coordinates of the contact point in the nth sample. n This represents the tensile stress load of the nth sample; the output unit of the fabric friction and wear prediction model is the friction and wear data of the fabric measurement points, represented as y = {y n (1≤n≤N);

[0025] 2.2.3.3) Construct a deep neural network with L layers, where the number of nodes in the input layer, the i-th hidden layer, and the output layer are r, li, and c, respectively. The activation function is the tanh function, as shown below:

[0026]

[0027] The error function chosen is the squared loss function;

[0028] 2.2.3.4) The backpropagation algorithm is used to calculate the gradient and correct the weight parameters and bias parameters of the deep neural network. When the number of iterations is reached or the error is less than or equal to the preset value, the iteration ends and the trained fabric surface friction and wear prediction model is obtained.

[0029] The method for establishing a digital twin model of textile surface quality under periodic friction and wear according to the present invention is further described in step 4):

[0030] 4.1) The actions and system control commands during the fabric dyeing process are transmitted in real time to the virtual space model of the fabric-dyeing machine guide roller, so that the virtual space model of the fabric-dyeing machine guide roller can simulate the friction and wear during the dyeing cycle and realize the three-dimensional visualization simulation of the fabric dyeing behavior.

[0031] 4.2) Establish a transformation function between the tension on the fabric and the friction and wear of the fabric, and then predict fabric abrasion using a fabric dyeing friction and wear prediction model. The specific steps are as follows:

[0032] 4.2.1), according to formula F f =kF converts the tensile stress F collected by the tension sensor into the frictional force F experienced by the fabric during the dyeing cycle. f, where k is the conversion coefficient, which is determined by the following experiment: collect the tensile tension value F′ at the monitoring point, and determine the friction force by microscopically quantifying the degree of wear on the fabric surface at this moment, thereby obtaining the quotient value k′. Multiple experiments yield multiple quotient values ​​k′, and then the average value is taken as the conversion coefficient k.

[0033] 4.2.2) The converted friction force and contact point coordinates are input into the fabric dyeing friction and wear prediction model to obtain the predicted value of fabric surface abrasion. The predicted value is then compared with other relevant data under the same operating conditions. When the error exceeds the allowable value, the fabric friction and wear prediction model is updated. The update steps are to first add the real-time fabric abrasion data to the training set of the fabric friction and wear prediction model, and then train the fabric friction and wear prediction model to obtain the updated prediction model, so as to ensure that the prediction model can accurately map the dyeing abrasion state.

[0034] The method for establishing a digital twin model of textile surface quality under periodic friction and wear according to the present invention is further described as follows: In step 5), the objective function of the genetic algorithm is defined as:

[0035] minf(NET, s, F)

[0036]

[0037] Where f represents the fabric friction and wear prediction model, NET represents the neural network established in the prediction model, s represents the coordinates of the positioning contact point, S represents the set of points where the positioning contact point is located, and F represents the magnitude of the tension. min and F max This indicates the minimum and maximum tension on the fabric.

[0038] An initial population of p individuals is randomly generated; individuals are decoded one by one, and individuals that do not meet the constraints in the optimization model are removed. Individuals that meet the constraints are used to predict the degree of abrasion on the fabric surface through a deep neural network.

[0039] Individual fitness is defined as

[0040]

[0041] In the above formula, U i (1≤i≤p) represents the chromatic individuals in the population, and Δ is a value predetermined based on the objective function value of each generation.

[0042] Compared with existing technologies, this invention has the following advantages: By applying Internet of Things (IoT) technology to actively collect real-time data during the fabric dyeing process, this invention achieves proactive perception of the fabric dyeing process; by introducing the concept of a digital twin model, a digital twin model of the fabric and the guide roller of the dyeing machine is constructed, and its real-time operating state is simulated, thereby obtaining the friction and wear state during the fabric dyeing process and expanding the dimensions of the physical data; the deep neural network, through its strong adaptive and self-learning capabilities, can effectively model situations where the coupling mechanism between factors is unclear and the exact relationship between input and output is difficult to determine, thus accurately and quickly predicting the friction and wear during the fabric dyeing process, and using a genetic algorithm to obtain the fabric tensile tension that minimizes the degree of fabric dyeing abrasion, forming a rule for selecting the fabric dyeing tension, thereby improving the surface quality of the dyed fabric. [Attached Image Description]

[0043] Figure 1 This is the overall process flow diagram of the present invention.

[0044] Figure 2 This is a flowchart of the digital twin model in step 2) of the present invention.

[0045] Figure 3 This is a flowchart of a simulation staining process using the digital twin model of this invention.

Detailed Implementation Methods

[0046] Please refer to the instruction manual appendix. Figure 1 To be continued Figure 3 As shown, this invention provides a method for establishing a digital twin model of textile surface quality under periodic friction and wear. The basic concept is as follows: First, data from the fabric dyeing process is collected using Internet of Things (IoT) technology. Then, a virtual-physical model of the fabric and dyeing machine is created, constructing three-dimensional models of each component and equipment. This involves building both a physical space model and a virtual space model of the dyeing machine and the fabric in operation, defining the physical space model's attributes, including structural parameters, geometric parameters, and boundary conditions. Next, during the finite element simulation of friction and wear in the fabric dyeing process, the model is meshed and its boundaries are processed. Then, dyeing abrasion is predicted. Based on multi-source data from a database, machine learning is used to train and optimize the simulation model. The simulation results are fed back to each model for further optimization. The tension experienced by the fabric during dyeing is also optimized, forming a tension selection rule for the fabric to improve the surface quality of the dyeing process. Specifically, the method includes the following steps:

[0047] 1) Obtain offline data of the fabric to be dyed and the dyeing machine before dyeing, including geometric parameters, structural characteristics and material properties, as well as online data of the fabric acquired throughout the dyeing process, including the circumferential driving force applied to the guide roller of the dyeing machine, the pressure of the fabric on it and the frictional torque between the fabric and the surface of the guide roller, the wear of the fabric fiber surface during the dyeing cycle and the dyeing environment data of the fabric to be dyed.

[0048] Specifically, in this step, IoT technology is applied to the entire fabric dyeing process. Speed ​​and tension sensors are configured for both the dyeing machine and the fabric, and online data is acquired based on the established monitoring points. The acquisition of online data includes the following steps:

[0049] 1.1) Before dyeing, the fabric to be dyed and the dyeing machine are inspected to obtain the geometric parameters, material properties and structural characteristics of the fabric and the dyeing machine.

[0050] 1.2) Set up data monitoring points on the fabric, use force sensors to measure the tensile tension at the contact point between the fabric and the guide roller of the dyeing machine, and arrange speed measuring sensors to obtain the fabric dyeing cycle speed data at the monitoring points of the dyeing machine. Configure temperature and pressure sensors in the cylinder of the dyeing machine to obtain environmental data of the dyeing process.

[0051] 2) Based on the digital twin model, physical and virtual spatial models of the fabric to be dyed and the guide rollers of the dyeing machine are established. The main tasks of the component digital twin model are as follows: First, the fabric dyeing system is divided into modules, and the main modules are modeled, including the dyeing machine model, sensor module model, and abrasion simulation module model. The dyeing machine model includes modeling the dyeing machine itself and the overall fabric-dyeing machine module. The sensor module model includes modeling pressure and flow sensors. Therefore, the abrasion simulation module model includes simulating three influencing factors: the rotational speed of the dyeing machine guide rollers, the tension on the fabric, and the fabric's opening degree during dyeing within the dyeing machine. Finally, the various modules and other components are connected to complete the construction of the digital twin model of the fabric dyeing system.

[0052] The specific steps of the digital twin model include: (1) Selecting physical entities to establish a three-dimensional visualized physical model, defining the geometric attributes, motion attributes, and functional attributes of the physical entities, and establishing a simulation model and a logical model. (2) Mapping the physical model to the logical model, describing the components, organizational structure, and operating mechanism of the logical model graphically and formally, and feeding back the attributes and behaviors of each element to the physical model through the logical model to achieve optimization of the physical model. (3) Mirroring the relevant attributes of the physical model to the simulation model, training and optimizing the simulation model to achieve visualization of its model, as well as visualization of the twin objects, twin structures, and twin processes of the physical entities, and feeding back the ideal effect of the simulation model to the physical model to achieve matching of the physical model. (4) Using model correlation and evaluation algorithms, verifying the consistency and reliability of the results of the simulation model by the logical model, and verifying the components of the logical model based on the simulation model. (5) After a series of optimizations are performed on the established simulation model, it is verified whether it meets the iterative optimization conditions. Finally, deep learning algorithms and iterative optimization are adopted, and the physical entity and virtual twin are iteratively interacted and optimized based on the twin data to integrate the digital twin model.

[0053] The virtual space model includes a geometric model, a finite element simulation model, a quantitative model of the relationship between friction on the fabric and fabric tension and velocity, and a quality prediction model for textiles under processing parameters. The specific establishment process is as follows:

[0054] 2.1) Establish a physical space model of the fabric-dyeing machine guide roller, and transform the entire fabric dyeing system from a physical entity into a Web-recognizable data model for tracking and calling the fabric-dyeing machine guide roller information.

[0055] 2.2) Construct a virtual space model of the fabric-dyeing machine guide roller, including a geometric model, a finite element simulation model, a quantitative model of the relationship between the friction on the fabric and the fabric tension and speed, and a quality prediction model of textiles under processing parameters. Real-time mapping of the fabric abrasion state during the dyeing process and quantification of the degree of abrasion.

[0056] Specifically, step 2.2) is as follows:

[0057] 2.2.1) Based on the data obtained in step 1, a geometric model of the fabric and the guide roller of the dyeing machine is constructed.

[0058] 2.2.2) A finite element simulation model of the fabric under tension deformation is constructed using finite element software.

[0059] 2.2.3) Based on deep neural networks, a fabric tension control model and a fabric friction and wear surface quality prediction model are established. A nonlinear mapping relationship between the magnitude of fabric tensile tension and fabric friction and wear is constructed. A finite element simulation model is used to simulate the fabric tensile deformation as the training sample set for the fabric friction and wear prediction model. The samples are standardized, and finally, the fabric abrasion prediction model is trained. The training method for the above-mentioned fabric abrasion prediction model includes the following steps:

[0060] 2.2.3.1) Using the parametric analysis function of the finite element software, a script file is written based on the established finite element simulation model. By changing the contact position coordinates and dyeing cycle speed, finite element calculations are performed to obtain the tensile tension number of the fabric corresponding to the measurement point.

[0061] 2.2.3.2), determine the input and output of the fabric surface friction and wear prediction model; the input unit of the fabric friction and wear prediction model is the coordinate parameters of the set contact points and the tensile stress load applied to the fabric to be dyed, expressed as x={x n (1≤n≤N), x n ={s n F n}, s n F represents the coordinates of the contact point in the nth sample. n This represents the tensile stress load of the nth sample; the output unit of the fabric friction and wear prediction model is the friction and wear data of the fabric measurement points, represented as y = {y n (1≤n≤N).

[0062] 2.2.3.3) Construct a deep neural network with L layers, where the number of nodes in the input layer, the i-th hidden layer, and the output layer are r, li, and c, respectively. The activation function is the tanh function, as shown below:

[0063]

[0064] The error function chosen is the squared loss function.

[0065] 2.2.3.4) The backpropagation algorithm is used to calculate the gradient and correct the weight parameters and bias parameters of the deep neural network. When the number of iterations is reached or the error is less than or equal to the preset value, the iteration ends and the trained fabric surface friction and wear prediction model is obtained.

[0066] 3) The data of tension, friction and circulation speed of the fabric during the actual production dyeing process are transmitted to the physical space model. The virtual space model can directly call the real-time information of the physical space model. Then, the dyeing process of the fabric through the guide roller of the dyeing machine is displayed in real time through the human-computer interaction device.

[0067] 4) Simulation is performed in the virtual space model of the fabric-dyeing machine guide roller to adjust the fabric tension in real time during the dyeing cycle, and then the dyeing abrasion is predicted for the finished fabric after dyeing based on the dyeing quality prediction model.

[0068] Specifically, this step is as follows:

[0069] 4.1) The actions and system control commands during the fabric dyeing process are transmitted in real time to the virtual space model of the fabric-dyeing machine guide roller, so that the virtual space model of the fabric-dyeing machine guide roller can simulate the friction and wear during the dyeing cycle and realize the three-dimensional visualization simulation of the fabric dyeing behavior.

[0070] 4.2) Establish a transformation function between the tension on the fabric and the friction and wear of the fabric, and then predict fabric abrasion using a fabric dyeing friction and wear prediction model. The specific method is as follows:

[0071] 4.2.1), according to formula F f =kF converts the tensile stress F collected by the tension sensor into the frictional force F experienced by the fabric during the dyeing cycle. f , where k is the conversion coefficient, which is determined by the following experiment: collect the tensile tension value F′ at the monitoring point, and then determine the friction force by microscopically quantifying the degree of wear on the fabric surface at this moment, thereby obtaining the quotient value k′. Multiple experiments yield multiple quotient values ​​k′, and then the average value is taken as the conversion coefficient k.

[0072] 4.2.2) The converted friction force and contact point coordinates are input into the fabric dyeing friction and wear prediction model to obtain the predicted value of fabric surface abrasion. The predicted value is then compared with other relevant data under the same operating conditions. When the error exceeds the allowable value, the fabric friction and wear prediction model is updated. The update steps are to first add the real-time fabric abrasion data to the training set of the fabric friction and wear prediction model, and then train the fabric friction and wear prediction model to obtain the updated prediction model, so as to ensure that the prediction model can accurately map the dyeing abrasion state.

[0073] 5) Execute the debugging process of textile dyeing cycle speed and tension and the optimization process of periodic friction and wear surface quality under the digital twin model. After the entire dyeing process of the fabric is completed, all process data are stored in the physical space model of fabric-dyeing machine guide roller. Based on the fabric tension calculation model and the fabric friction and wear prediction model, a set of optimal dyeing cycle speed and fabric tension is obtained through genetic algorithm so that the degree of abrasion on the fabric during the dyeing process is minimized. Thus, the appropriate fabric tension and dyeing cycle speed during the dyeing process are determined to ensure the surface quality of the finally dyed fabric.

[0074] In this step, the objective function of the genetic algorithm is defined as:

[0075] minf(NET, s, F)

[0076]

[0077] Where f represents the fabric friction and wear prediction model, NET represents the neural network established in the prediction model, s represents the coordinates of the positioning contact point, S represents the set of points where the positioning contact point is located, and F represents the magnitude of the tension. min and F max This indicates the minimum and maximum tension on the fabric.

[0078] An initial population of p individuals is randomly generated; individuals are decoded one by one, and individuals that do not meet the constraints in the optimization model are removed. Individuals that meet the constraints are used to predict the degree of abrasion on the fabric surface through a deep neural network.

[0079] Individual fitness is defined as

[0080]

[0081] In the above formula, U i (1≤i≤p) represents the chromatic individuals in the population, and Δ is a value predetermined based on the objective function value of each generation.

[0082] The digital twin model of textile surface quality under periodic friction and wear established using this invention simulates the dyeing process of the fabric as follows: Figure 3 As shown: First, a virtual scene of the fabric dyeing process is constructed based on the corresponding structural data. Then, real-time data on the fabric's operating status inside the dyeing machine is collected and processed to obtain data reflecting the physical equipment status and scene feature data reflecting the fabric's operating status. Next, the time required for full dyeing of the fabric is predicted using the status data, and a corresponding digital twin model is created in the virtual dyeing scene based on the scene feature data. This digital twin model evolves as the scene feature data changes. Finally, it is determined whether the fabric has been fully dyed within the specified number of cycles, and a new digital twin model is created in the virtual scene based on the scene feature data. The digital twin model reflects the operating status of the dyed fabric in real-time, intuitively, and realistically.

[0083] The above-described specific embodiments are merely preferred embodiments of this invention and are not intended to limit this invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this invention should be included within the protection scope of this invention.

Claims

1. A method for establishing a digital twin model of the surface quality of textiles under periodic friction and wear, characterized in that: Includes the following steps: 1) Obtain offline data of the fabric to be dyed and the dyeing machine before dyeing, including geometric parameters, structural characteristics and material properties, as well as online data of the fabric acquired throughout the dyeing process, including the circumferential driving force applied to the guide roller of the dyeing machine, the pressure of the fabric on it and the frictional torque between the fabric and the surface of the guide roller, the wear of the fabric fiber surface during the dyeing cycle and the dyeing environment data of the fabric to be dyed. 2) Based on the digital twin model, establish a physical space model and a virtual space model of the fabric to be dyed and the guide roller of the dyeing machine; the virtual space model includes a geometric model, a finite element simulation model, a quantitative model of the relationship between the friction of the fabric and the fabric tension and speed, and a quality prediction model of textiles under the processing parameters. 3) The data of tension, friction and circulation speed of the fabric during the actual production dyeing process are transmitted to the physical space model. The virtual space model can directly call the real-time information of the physical space model. Then, the dyeing process of the fabric through the guide roller of the dyeing machine is displayed in real time through the human-computer interaction device. 4) Simulation is performed in the virtual space model of the fabric-dyeing machine guide roller to adjust the fabric tension in real time during the dyeing cycle, and then the dyeing abrasion is predicted for the finished fabric after dyeing based on the dyeing quality prediction model. 5) Execute the debugging process of textile dyeing cycle speed and tension and the optimization process of periodic friction and wear surface quality under the digital twin model. After the entire dyeing process of the fabric is completed, all process data are stored in the physical space model of fabric-dyeing machine guide roller. Based on the fabric tension calculation model and the fabric friction and wear prediction model, a set of optimal dyeing cycle speed and fabric tension is obtained through genetic algorithm so that the degree of abrasion on the fabric during the dyeing process is minimized. Thus, the appropriate fabric tension and dyeing cycle speed during the dyeing process are determined to ensure the surface quality of the finally dyed fabric.

2. The method for establishing a digital twin model of textile surface quality under periodic friction and wear as described in claim 1, characterized in that: Step 1) involves the following steps in obtaining online data: 1.1) Before dyeing, the fabric to be dyed and the dyeing machine are inspected to obtain the geometric parameters, material properties and structural characteristics of the fabric and the dyeing machine. 1.2) Set up data monitoring points on the fabric, use force sensors to measure the tensile tension at the contact point between the fabric and the guide roller of the dyeing machine, and arrange speed measuring sensors to obtain the fabric dyeing cycle speed data at the monitoring points of the dyeing machine. Configure temperature and pressure sensors in the cylinder of the dyeing machine to obtain environmental data of the dyeing process.

3. The method for establishing a digital twin model of textile surface quality under periodic friction and wear as described in claim 1, characterized in that: Step 2) specifically refers to: 2.1) Establish a physical space model of the fabric-dyeing machine guide roller, and transform the entire fabric dyeing system from a physical entity into a Web-recognizable data model for tracking and calling the fabric-dyeing machine guide roller information; 2.2) Construct a virtual space model of the fabric-dyeing machine guide roller, including a geometric model, a finite element simulation model, a quantitative model of the relationship between the friction on the fabric and the fabric tension and speed, and a quality prediction model of textiles under processing parameters. Real-time mapping of the fabric abrasion state during the dyeing process and quantification of the degree of abrasion.

4. The method for establishing a digital twin model of textile surface quality under periodic friction and wear as described in claim 3, characterized in that: Step 2.2) specifically refers to: 2.2.1) Based on the data obtained in step 1, a geometric model of the fabric and the guide roller of the dyeing machine is constructed; 2.2.2) A finite element simulation model of the fabric under tension deformation is constructed using finite element software; 2.2.3) Based on deep neural networks, a fabric tension control model and a fabric friction and wear surface quality prediction model are established. A nonlinear mapping relationship between the magnitude of fabric tensile tension and fabric friction and wear is constructed. The simulated fabric tensile deformation is used as the training sample set for the fabric friction and wear prediction model. The samples are standardized. Finally, the fabric abrasion prediction model is trained.

5. The method for establishing a digital twin model of textile surface quality under periodic friction and wear as described in claim 4, characterized in that: The training method for the fabric abrasion prediction model includes the following steps: 2.2.3.1) Using the parametric analysis function of the finite element software, a script file is written based on the established finite element simulation model. By changing the contact position coordinates and dyeing cycle speed, finite element calculations are performed to obtain the tensile tension data of the fabric corresponding to the measurement point. 2.2.3.2), determine the input and output of the fabric surface friction and wear prediction model; the input unit of the fabric friction and wear prediction model is the coordinate parameters of the set contact points and the tensile stress load applied to the fabric to be dyed, expressed as x={x n (1≤n≤N), x n ={s n F n }, s n F represents the coordinates of the contact point in the nth sample. n This represents the tensile stress load of the nth sample; the output unit of the fabric friction and wear prediction model is the friction and wear data of the fabric measurement points, represented as y = {y n (1≤n≤N); 2.2.3.3) Construct a deep neural network with L layers, where the number of nodes in the input layer, the i-th hidden layer, and the output layer are r, li, and c, respectively. The activation function is the tanh function, as shown below: The error function chosen is the squared loss function; 2.2.3.4) The backpropagation algorithm is used to calculate the gradient and correct the weight parameters and bias parameters of the deep neural network. When the number of iterations is reached or the error is less than or equal to the preset value, the iteration ends and the trained fabric surface friction and wear prediction model is obtained.

6. The method for establishing a digital twin model of textile surface quality under periodic friction and wear as described in claim 1, characterized in that: Step 4) specifically involves: 4.1) The actions and system control commands during the fabric dyeing process are transmitted in real time to the virtual space model of the fabric-dyeing machine guide roller, so that the virtual space model of the fabric-dyeing machine guide roller can simulate the friction and wear during the dyeing cycle and realize the three-dimensional visualization simulation of the fabric dyeing behavior. 4.2) Establish a transformation function between the tension on the fabric and the friction and wear of the fabric, and then predict fabric abrasion using a fabric dyeing friction and wear prediction model. The specific steps are as follows: 4.2.1), according to formula F f =kF converts the tensile stress F collected by the tension sensor into the frictional force F experienced by the fabric during the dyeing cycle. f , where k is the conversion coefficient, which is determined by the following experiment: collect the tensile tension value F′ at the monitoring point, and determine the friction force by microscopically quantifying the degree of wear on the fabric surface at this moment, thereby obtaining the quotient value k′. Multiple experiments yield multiple quotient values ​​k′, and then the average value is taken as the conversion coefficient k. 4.2.2) The converted friction force and contact point coordinates are input into the fabric dyeing friction and wear prediction model to obtain the predicted value of fabric surface abrasion. The predicted value is then compared with other relevant data under the same operating conditions. When the error exceeds the allowable value, the fabric friction and wear prediction model is updated. The update steps are to first add the real-time fabric abrasion data to the training set of the fabric friction and wear prediction model, and then train the fabric friction and wear prediction model to obtain the updated prediction model, so as to ensure that the prediction model can accurately map the dyeing abrasion state.

7. The method for establishing a digital twin model of textile surface quality under periodic friction and wear as described in claim 1, characterized in that: In step 5), the objective function of the genetic algorithm is defined as: minf(NET, s, F) Where f represents the fabric friction and wear prediction model, NET represents the neural network established in the prediction model, s represents the coordinates of the positioning contact point, S represents the set of points where the positioning contact point is located, and F represents the magnitude of the tension. min and F max This indicates the minimum and maximum tension on the fabric. An initial population of p individuals is randomly generated; individuals are decoded one by one, and individuals that do not meet the constraints in the optimization model are removed. Individuals that meet the constraints are used to predict the degree of abrasion on the fabric surface through a deep neural network. Individual fitness is defined as In the above formula, U i (1≤i≤p) represents the chromatic individuals in the population, and Δ is a value predetermined based on the objective function value of each generation.

Images