Intelligent control method and system for toilet paper roll core forming based on data driving

Abstract

The invention relates to a toilet paper roll core forming intelligent control method and system based on data driving, and the method comprises the following steps: S1, obtaining production process parameters and equipment operation state data, and carrying out the preprocessing; s2, constructing a mapping model of process quality and parameter data of production based on SVR; s3, on the basis of the mapping model, performing quality prediction on the technological process, and on the basis of a multi-objective optimization algorithm, obtaining optimization control parameters; s4, pushing the optimized control parameters to an intelligent control system, and realizing closed-loop control by using a fuzzy PID control algorithm; and S5, dynamically feeding back the actual output process data collected in real time to the modeling link to form a complete closed-loop feedback mechanism. According to the method, the environment change and the process disturbance can be dynamically responded in the production process, so that the purposes of improving the product quality, improving the production efficiency and reducing the resource consumption are achieved.

Description

Technical Field

[0001] The present invention relates to the field of intelligent control, and in particular to a data-driven intelligent control method and system for toilet paper roll core forming. Background Art

[0002] With the rapid growth of market demand in the toilet paper industry and the continuous improvement of product quality requirements, the traditional core forming production method faces many challenges, such as low production efficiency, poor product quality stability, and high dependence on operator experience. The traditional control method is mainly based on fixed parameter adjustment, lacking precise control of the relationship between process parameters and finished product quality in the actual production process, resulting in production deviations caused by material fluctuations, equipment performance degradation or environmental changes. It is difficult to correct them in real time, which in turn increases the scrap rate and production costs.

[0003] In recent years, the rapid development of the industrial control field, especially the widespread application of data-driven technology, has provided new solutions to solve the complex process control problems in the toilet paper roll core forming process. Summary of the invention

[0004] In order to solve the above problems, the purpose of the present invention is to provide a data-driven intelligent control method for toilet paper roll core forming, which can dynamically respond to environmental changes and process disturbances during the production process, so as to achieve the purpose of improving product quality, improving production efficiency and reducing resource consumption.

[0005] To achieve the above object, the present invention adopts the following technical solutions: The data-driven intelligent control method for toilet paper roll core forming includes the following steps: S1: Obtain production process parameters and equipment operation status data, and pre-process them; S2: Based on the preprocessed data, a data-driven modeling method is used to build a mapping model between production process quality and parameter data based on SVR; S3: Based on the mapping model, the quality of the process is predicted, and the optimization control parameters are obtained based on the multi-objective optimization algorithm; S4: Push the optimized control parameters to the intelligent control system, compare the real-time collected data with the modeling prediction results, and use the fuzzy PID control algorithm to dynamically adjust the tension, speed, and humidity influencing variables to achieve closed-loop control; S5: The actual output process data collected in real time is dynamically fed back to the modeling link to form a complete closed-loop feedback mechanism.

[0006] Furthermore, the production process parameters and equipment operation status are obtained, specifically: a strain gauge tension sensor is used, installed on both sides of the contact between the winding drum and the paper, to obtain the tension data of the paper during the winding process; a Hall effect encoder is used, installed at the end of the core drum shaft, to measure the speed data of the servo motor; a digital humidity sensor is used, arranged in the environmental humidity influence area and on one side of the paper surface, to obtain the paper moisture content and environmental humidity data; a thin film pressure sensor is used, arranged in the clamping assembly of the winding drum, to monitor the clamping force data in real time; a three-axis vibration accelerometer sensor is arranged on the rotating part of the equipment to monitor the vibration amplitude and frequency data; and a temperature sensor is used to monitor the equipment temperature data in real time.

[0007] Further preprocessing is performed as follows: low-pass filtering is used to smooth the high-frequency noise in the tension and pressure data; Kalman filtering is used to optimize the dynamic changes of humidity and speed data; time series linear interpolation is used to fill missing values ​​in speed and tension data; mean filling is used to fill the vibration and temperature data with the mean of the historical window; finally, all data are normalized.

[0008] Furthermore, a mapping model between production process quality and parameter data is constructed based on SVR, specifically: The input feature vector X is the preprocessed data, X=[T,v,H e ,P,Vib,T e ,H w ]; Where, T is the tension, v is the speed, H e is the ambient humidity, P is the pressure, Vib is the vibration, T e is the device temperature, H w is the moisture content of paper; The output variable is Y=[Y1,Y2,Y3]; Among them, Y1 is the thickness uniformity, Y2 is the roundness of the core, and Y3 is the tensile strength. The mapping model between process quality and parameter data of production based on SVR is: ; Among them, X i is the input feature vector of the i-th sample; is the composite kernel function; is the Lagrange multiplier, which represents the input feature vector X in the model i The corresponding Y k The weight of k (X) is the k-th output Y k The prediction model of k For the corresponding Y kThe bias of; m is the number of samples in the training data; The objective optimization function is: ; Where C is the regularization coefficient; is the model complexity term; For the corresponding Y k Model parameters of represents the Lagrange multiplier; K is the total number of quality indicators; denote the slack variables in the positive and negative directions respectively; And set a dynamic error tolerance in the model to handle the error range of different sample inputs:

[0009] in, is the tolerance range of the i-th training sample; is the global basic tolerance parameter; σ i The local variance of the input parameters for the ith training sample.

[0010] Furthermore, the composite kernel function Including RBF kernel , polynomial kernel and the linear kernel , as follows: ; ; ; ; Where c is the bias term, d is the order of the polynomial, is the width of the RBF kernel; λ1, λ2, λ3 are weight coefficients.

[0011] Further, the training of the mapping model is as follows: Initialize weight coefficients λ1, λ2, λ3 and kernel parameters; and initially adjust regularization coefficient C and global base tolerance parameters ; The optimization problem of support vector regression is solved using the Lagrangian dual form: ; Among them, X j is the input feature vector of the jth sample; is the Lagrange multiplier of the jth sample; Solve the dual problem using the Sequential Minimal Optimization (SMO) algorithm and determine the non-zero Lagrange multipliers The corresponding input feature vector Xi ; is the true value of the i-th sample regarding the k-th quality indicator; Combine the obtained input feature vector to construct each output Y k The mapping model.

[0012] Furthermore, S3 is specifically: Through the trained mapping model f k (X) , predicting quality indicators in real time and calculating each input process parameter X l Mapping Model f k (X) Sensitivity S k,l : ; Through the sensitivity analysis results, the sensitivity S of all process parameters k,l Sort by absolute value, determine the process parameters that have the greatest impact on quality indicators, and prioritize the optimization of key variables; Set the multi-objective optimization objective function F(X): ; in, is the expected value of the kth target quality indicator; is the weight of the kth target; R(X) is the regularization term; λ is the weight coefficient of the regularization term; According to the multi-objective optimization objective function, NSGA-II is used to generate the Pareto solution set and obtain the optimized solution parameter X opt .

[0013] Further, S4 is specifically: The parameter X to be optimized opt Push to the intelligent control system as the initial setting value; Calculate the input control error, including the error and error rate of change : ; ; in, is the actual expected value, t represents time; According to e(t) and Δe(t), dynamically update PID parameters: ; ; ; in, They are the proportional, integral and differential coefficients dynamically adjusted by the fuzzy controller; is the corresponding adaptively adjusted learning rate; is the adjustment value of the kth target; is the time interval; Fuzzy PID control output: ; in, Indicates the time index; express The error corresponding to the moment; u(t) is the change in the adjustment variable: ; in, for X l The new process parameters after adjustment.

[0014] Furthermore, S5 is specifically: Collect the current actual process parameters X real and the actual quality index Y real With the model prediction value Comparison, calculation error : ; Collect new real-time production data , merged into the training set : ; Among them, D old For historical data sets; Update SVR model: ; Among them, m new is the number of training set data after merging; argmin means finding the parameter that minimizes the function; Regularly adjust the fuzzy rule base and PID weights, update the optimization solution, and form a complete self-learning and feedback closed loop.

[0015] A data-driven intelligent control system for toilet paper roll core forming includes a processor, a memory, and a computer program stored in the memory. When the processor executes the computer program, it specifically executes the steps in the data-driven intelligent control method for toilet paper roll core forming as described above.

[0016] The present invention has the following beneficial effects: 1. The present invention can dynamically respond to environmental changes and process disturbances during the production process, thereby achieving the purpose of improving product quality, increasing production efficiency and reducing resource consumption; 2. By strengthening multi-objective prediction, introducing composite kernels and dynamically adjusting error tolerance, the model's nonlinear modeling and generalization capabilities will be significantly improved, and the complex relationship between various process parameters and quality indicators will be easier to analyze and capture; 3. The present invention is based on the fuzzy PID algorithm and dynamically adjusts the optimization parameters. It can respond to changes in equipment operation or external environment in real time, ensure the stability and reliability of the process, compare the quality index data collected in real time with the model prediction value, calculate the error and optimize the parameter setting, truly realizing end-to-end closed-loop control. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 The figure is a flow chart of the method of the present invention. DETAILED DESCRIPTION

[0018] The present invention is further described in detail below with reference to the accompanying drawings and specific embodiments: refer to Figure 1 In this embodiment, a data-driven intelligent control method for toilet paper roll core forming is provided, comprising the following steps: S1: Obtain production process parameters and equipment operation status data, and pre-process them; S2: Based on the preprocessed data, a data-driven modeling method is used to build a mapping model between production process quality and parameter data based on SVR; S3: Based on the mapping model, the quality of the process is predicted, and the optimization control parameters are obtained based on the multi-objective optimization algorithm; S4: Push the optimized control parameters to the intelligent control system, compare the real-time collected data with the modeling prediction results, and use the fuzzy PID control algorithm to dynamically adjust the tension, speed, and humidity influencing variables to achieve closed-loop control; S5: The actual output process data collected in real time is dynamically fed back to the modeling link to form a complete closed-loop feedback mechanism.

[0019] In this embodiment, the production process parameters and the equipment operation status are obtained, specifically: a strain gauge tension sensor is used, which is installed on both sides of the winding roller in contact with the paper to obtain the tension data of the paper during the winding process; a Hall effect encoder is used, which is installed at the end of the core roller shaft to measure the speed data of the servo motor; a digital humidity sensor (capacitive type) is used, which is arranged in the environmental humidity influence area and on one side of the paper surface to obtain the paper moisture content and environmental humidity data; a thin film pressure sensor is used, which is arranged in the clamping assembly of the winding roller to monitor the clamping force data in real time; a three-axis vibration accelerometer sensor is arranged on the rotating part of the equipment to monitor the vibration amplitude and frequency data; and a temperature sensor is used to monitor the equipment temperature data in real time.

[0020] In this embodiment, the preprocessing is as follows: low-pass filtering is used to smooth the high-frequency noise in the tension and pressure data; Kalman filtering is used to optimize the dynamic changes of humidity and speed data; time series linear interpolation is used to fill the missing values ​​of speed and tension data; mean filling is used to fill the vibration and temperature data with the mean of the historical window; finally, each data is normalized.

[0021] In this embodiment, a mapping model between production process quality and parameter data is constructed based on SVR, specifically: The input feature vector X is the preprocessed data, X=[T,v,H e ,P,Vib,T e ,H w ]; Where, T is the tension, v is the speed, H e is the ambient humidity, P is the pressure, Vib is the vibration, T e is the device temperature, H w is the moisture content of paper; The output variable is Y=[Y1,Y2,Y3]; Among them, Y1 is thickness uniformity; Y2 is the roundness of the core; Y3 is the tensile strength; The mapping model between process quality and parameter data of production based on SVR is: ; Among them, X i is the input feature vector of the i-th sample; is the composite kernel function; is the Lagrange multiplier, which represents the input feature vector X in the model i The corresponding Y k The weight of k (X) is the k-th output Y k The prediction model of k For the corresponding Yk The bias of; m is the number of samples in the training data; The objective optimization function is:

[0022] Where C is the regularization coefficient; is the model complexity term; For the corresponding Y k Model parameters of represents the Lagrange multiplier; K is the total number of quality indicators; denote the slack variables in the positive and negative directions respectively; And set a dynamic error tolerance in the model to handle the error range of different sample inputs:

[0023] in, is the tolerance range of the i-th training sample; is the global basic tolerance parameter; The local variance of the input parameters for the ith training sample.

[0024] In this embodiment, the composite kernel function Including RBF kernel , polynomial kernel and the linear kernel , as follows: ; ; ; ; Where c is the bias term, d is the order of the polynomial, is the width of the RBF kernel; λ1, λ2, λ3 are weight coefficients.

[0025] In this embodiment, the training of the mapping model is as follows: Initialize weight coefficients λ1, λ2, λ3 and kernel parameters (such as γ of RBF, d, c of polynomial kernel); and initially adjust regularization coefficient C and global basic tolerance parameter ; The optimization problem of support vector regression is solved using the Lagrangian dual form: ; Among them, X j is the input feature vector of the jth sample; is the Lagrange multiplier of the jth sample; Solve the dual problem using the Sequential Minimal Optimization (SMO) algorithm and determine the non-zero Lagrange multipliers The corresponding input feature vector X i ; is the true value of the i-th sample regarding the k-th quality indicator; Combine the obtained input feature vector to construct each output Y k The mapping model.

[0026] In this embodiment, S3 is specifically: Through the trained mapping model f k (X) , predicting quality indicators in real time and calculating each input process parameter X l Mapping Model f k (X) Sensitivity S k,l : ; Through the sensitivity analysis results, the sensitivity S of all process parameters k,l Sort by absolute value, determine the process parameters that have the greatest impact on quality indicators, and prioritize the optimization of key variables; Set the multi-objective optimization objective function F(X): ; in, is the expected value of the kth target quality index; R(X) is the regularization term; λ is the weight coefficient of the regularization term; According to the multi-objective optimization objective function, NSGA-II is used to generate the Pareto solution set and obtain the optimized solution parameter X opt .

[0027] In this embodiment, S4 is specifically: The parameter X to be optimized opt Push to the intelligent control system as the initial setting value; Calculate the input control error, including the error and error rate of change : ; ; in, is the actual expected value, t represents time; According to e(t) and Δe(t), dynamically update PID parameters: ; ; ; in, They are the proportional, integral and differential coefficients dynamically adjusted by the fuzzy controller; is the corresponding adaptively adjusted learning rate; is the adjustment value of the kth target; is the time interval; Fuzzy PID control output: ; in, Indicates the time index; express The error corresponding to the moment; u(t) is the change in the adjustment variable (such as tension, speed or humidity): ; in, for X l Adjusted new process parameters .

[0028] In this embodiment, S5 is specifically: Collect the current actual process parameters X real and the actual quality index Y real With the model prediction value Comparison, calculation error : ; Collect new real-time production data , merged into the training set : ; Among them, D old For historical data sets; Update SVR model: ; Among them, m new is the number of training set data after merging; argmin means finding the parameter that minimizes the function; Regularly adjust the fuzzy rule base and PID weights, update the optimization solution, and form a complete self-learning and feedback closed loop.

[0029] A data-driven intelligent control system for toilet paper roll core forming includes a processor, a memory, and a computer program stored in the memory. When the processor executes the computer program, it specifically executes the steps in the data-driven intelligent control method for toilet paper roll core forming as described above.

[0030] It will be appreciated by those skilled in the art that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0031] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0032] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0033] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0034] The above is only a preferred embodiment of the present invention, and does not limit the present invention in other forms. Any technician familiar with the profession may use the above disclosed technical content to change or modify it into an equivalent embodiment with equivalent changes. However, any simple modification, equivalent change and modification made to the above embodiment according to the technical essence of the present invention without departing from the technical solution of the present invention still belongs to the protection scope of the technical solution of the present invention.

Claims

1. A data-driven intelligent control method for toilet paper roll core forming, characterized in that: The following steps are included S1: Obtain production process parameters and equipment operation status data, and pre-process them; S2: Based on the preprocessed data, a data-driven modeling method is used to build a mapping model between production process quality and parameter data based on SVR; S3: Based on the mapping model, the quality of the process is predicted, and the optimization control parameters are obtained based on the multi-objective optimization algorithm; S4: Push the optimized control parameters to the intelligent control system, compare the real-time collected data with the modeling prediction results, and use the fuzzy PID control algorithm to dynamically adjust the tension, speed, and humidity influencing variables to achieve closed-loop control; S5: The actual output process data collected in real time is dynamically fed back to the modeling link to form a complete closed-loop feedback mechanism.

2. The data-driven intelligent control method for toilet paper roll core forming according to claim 1, characterized in that: The method for obtaining the production process parameters and the equipment operation status is as follows: a strain gauge tension sensor is installed on both sides of the winding drum in contact with the paper to obtain the tension data of the paper during the winding process; a Hall effect encoder is installed on the end of the winding core drum shaft to measure the rotation speed data of the servo motor; a digital humidity sensor is arranged in the environmental humidity influence area and on one side of the paper surface to obtain the paper moisture content and environmental humidity data; a thin film pressure sensor is arranged in the clamping component of the winding drum to monitor the clamping force data in real time; a three-axis vibration accelerometer sensor is arranged on the rotating part of the equipment to monitor the vibration amplitude and frequency data; and a temperature sensor is used to monitor the equipment temperature data in real time.

3. The data-driven intelligent control method for toilet paper roll core forming according to claim 2 is characterized in that: The preprocessing is specifically as follows: low-pass filtering is used to smooth the high-frequency noise in the tension and pressure data; Kalman filtering is used to optimize the dynamic changes of humidity and speed data; time series linear interpolation is used to fill missing values ​​in the speed and tension data; mean filling is used to fill the vibration and temperature data with the mean of the historical window; and finally, each data is normalized.

4. The data-driven intelligent control method for toilet paper roll core forming according to claim 1, characterized in that: The mapping model between process quality and parameter data of production based on SVR is specifically as follows: The input feature vector X is the preprocessed data, X=[T,v,H e ,P,Vib,T e ,H w ]; Where, T is the tension, v is the speed, H e is the ambient humidity, P is the pressure, Vib is the vibration, T e is the device temperature, H w is the moisture content of paper; The output variable is Y=[Y1,Y2,Y3]; Among them, Y1 is the thickness uniformity, Y2 is the roundness of the core, and Y3 is the tensile strength. The mapping model between process quality and parameter data of production based on SVR is: ； Among them, X i is the input feature vector of the i-th sample; is the composite kernel function; is the Lagrange multiplier, which represents the input feature vector X in the model i The corresponding Y k The weight of k (X) is the k-th output Y k The prediction model of k For the corresponding Y k The bias of; m is the number of samples in the training data; The objective optimization function is: ； Where C is the regularization coefficient; is the model complexity term; For the corresponding Y k Model parameters of represents the Lagrange multiplier; K is the total number of quality indicators; denote the slack variables in the positive and negative directions respectively; And set a dynamic error tolerance in the model to handle the error range of different sample inputs: ； in, is the tolerance range of the i-th training sample; is the global basic tolerance parameter; The local variance of the input parameters for the ith training sample.

5. The data-driven intelligent control method for toilet paper roll core forming according to claim 4 is characterized in that: The composite kernel function Including RBF kernel , polynomial kernel and the linear kernel , as follows: ； ; ； ； Where c is the bias term, d is the order of the polynomial, is the width of the RBF kernel; λ1, λ2, λ3 are weight coefficients.

6. The data-driven intelligent control method for toilet paper roll core forming according to claim 5, characterized in that: The training of the mapping model is as follows: Initialize weight coefficients λ1, λ2, λ3 and kernel parameters; and initially adjust regularization coefficient C and global base tolerance parameters ; The optimization problem of support vector regression is solved using the Lagrangian dual form: ； Among them, X j is the input feature vector of the jth sample; is the Lagrange multiplier of the jth sample; Solve the dual problem using the Sequential Minimal Optimization (SMO) algorithm and determine the non-zero Lagrange multipliers The corresponding input feature vector X i ; is the true value of the i-th sample regarding the k-th quality indicator; Combine the obtained input feature vector to construct each output Y k The mapping model.

7. The data-driven intelligent control method for toilet paper roll core forming according to claim 4, characterized in that: The S3 is specifically: Through the trained mapping model f k (X) , predicting quality indicators in real time and calculating each input process parameter X l Mapping Model f k (X) Sensitivity S k,l : ； Through the sensitivity analysis results, the sensitivity S of all process parameters k,l Sort by absolute value, determine the process parameters that have the greatest impact on quality indicators, and prioritize the optimization of key variables; Set the multi-objective optimization objective function F(X): ； in, is the expected value of the kth target quality index; R(X) is the regularization term; λ is the weight coefficient of the regularization term; According to the multi-objective optimization objective function, NSGA-II is used to generate the Pareto solution set and obtain the optimized solution parameter X opt .

8. The data-driven intelligent control method for toilet paper roll core forming according to claim 7, characterized in that: The S4 is specifically: The parameter X to be optimized opt Push to the intelligent control system as the initial setting value; Calculate the input control error, including the error and error rate of change : ； ； in, is the actual expected value, t represents time; According to e(t) and Δe(t), dynamically update PID parameters: ； ； ； in, They are the proportional, integral and differential coefficients dynamically adjusted by the fuzzy controller; is the corresponding adaptively adjusted learning rate; is the adjustment value of the kth target; is the time interval; Fuzzy PID control output: ； in, Indicates the time index; express The error corresponding to the moment; u(t) is the change in the adjustment variable: ； in, for X l The new process parameters after adjustment.

9. The data-driven intelligent control method for toilet paper roll core forming according to claim 8, characterized in that: The S5 is specifically: Collect the current actual process parameters X real and the actual quality index Y real With the model prediction value Compare and calculate the error : ； Collect new real-time production data , merged into the training set : ； Among them, D old For historical data sets; Update SVR model: ； Among them, m new is the number of training set data after merging; argmin means finding the parameter that minimizes the function; Regularly adjust the fuzzy rule base and PID weights, update the optimization solution, and form a complete self-learning and feedback closed loop.

10. A data-driven intelligent control system for toilet paper roll core forming, characterized in that: It includes a processor, a memory and a computer program stored in the memory. When the processor executes the computer program, it specifically performs the steps in the data-driven intelligent control method for toilet paper roll core forming as described in any one of claims 1 to 9.

Images