Intelligent scheduling and optimal control system of polyester casting stretch shrink film production line

By using an intelligent scheduling and optimization control system, the parameters of each process in the polyester casting stretch shrink film production line are dynamically adjusted in real time, which solves the problem of inconsistent production quality in the existing technology and realizes high-quality shrink film production.

CN118963301AInactive Publication Date: 2024-11-15JIANGSU GUANGHUI PACKAGE

Patent Information

Application Number
CN202411440965.2
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-10-16
Publication Date
2024-11-15
Estimated Expiration
Not applicable · inactive patent

AI Technical Summary

Technical Problem

The control parameter settings of existing polyester casting stretch shrink film production lines lack mathematical and statistical guidance, resulting in inconsistent production quality and difficulty in achieving precise control.

Method used

An intelligent scheduling and optimization control system is adopted, including a feature parameter setting module, a training data collection module, a quality prediction model training module, a parameter weight analysis module, and a control strategy optimization module. Through machine learning models, the parameters of each process are dynamically adjusted in real time to determine the dominant control parameters and provide control suggestions.

Benefits of technology

It enables intelligent analysis of multiple quality indicators and parameters, automatically calculates priority adjustment weights, and dynamically adjusts process parameters in real time, thereby improving the stability and consistency of shrink film production quality.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118963301A_ABST
    Figure CN118963301A_ABST
Patent Text Reader

Abstract

The invention discloses an intelligent scheduling and optimal control system of a polyester casting stretching shrink film production line, and relates to the technical field of intelligent regulation and control of production lines. A training data collection module is set to collect a training sample feature vector set and a quality parameter label vector set in advance, a quality prediction model training module is set to train a quality prediction model, and a parameter weight analysis module is set to obtain the influence weight of each control parameter on each quality parameter. The control strategy optimization module is set to provide control suggestions for various control parameters of various production procedures in the actual production process of the shrink film, and dynamically adjust the parameters of the procedures in real time, so that the product quality of the shrink film is kept within an ideal target range, and the production quality of the shrink film is improved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention relates to the technical field of intelligent control of production lines, and in particular to an intelligent dispatching and optimization control system for a polyester cast stretch shrink film production line. Background Art

[0002] The general production process of polyester cast stretch shrink film production line includes preparation of raw and auxiliary materials, melt extrusion (using exhaust twin-screw extruder for melt extrusion), die forming (the melt enters the hanger-type T-die through the melt metering pump, filter and melt pipe, and is cast onto the cooling drum through the die lip to form a cast sheet), cast sheet (the cast sheet is quickly cooled to below the glass transition temperature through the cooling drum to form a transparent thick sheet), transverse stretching (stretching the cast sheet transversely to improve the longitudinal and transverse strength of the polyester film through molecular orientation), traction, winding and slitting; Each production process includes several control parameters that can be automatically or manually controlled. These control parameters determine the quality of the final shrink film. At present, the setting of various control parameters is often based on experience. The parameters set manually lack scientific guidance in mathematical statistics, which leads to uneven production quality of shrink film. Therefore, a method is needed to automatically and intelligently adjust the control parameters in real time to ensure the quality of the shrink film produced. The Chinese patent with publication number CN117369384A discloses an intelligent defect control system for a thin film production line. The key points of its technical solution are that it includes a user unit, an application unit, a processing unit, a data unit and an equipment unit. The equipment unit includes at least a defect meter and a slitting machine; the application unit is used to interact with the user unit and send processing instructions to the processing unit according to the logical function; the data unit is used to obtain defect data information of the defect meter and upload it to the processing unit; the processing unit is used to perform data flow operations and / or alarm operations according to the processing instructions and defect data information, and send them to the slitting machine through the data unit. However, the control system lacks a quantitative control strategy for each control parameter, and it is difficult to achieve the effect of accurately regulating each control parameter; To this end, the present invention proposes an intelligent scheduling and optimization control system for an ester cast stretch shrink film production line. Summary of the invention

[0003] The present invention aims to solve at least one of the technical problems existing in the prior art. To this end, the present invention proposes an intelligent scheduling and optimization control system for a polyester cast stretch shrink film production line, which dynamically adjusts the parameters of each process in real time to keep the product quality of the shrink film within an ideal target range and improve the production quality of the shrink film.

[0004] To achieve the above purpose, an intelligent scheduling and optimization control system for a polyester cast stretch shrink film production line is proposed, which includes a characteristic parameter setting module, a training data collection module, a quality prediction model training module, a parameter weight analysis module, and a control strategy optimization module; wherein each module is electrically connected; The characteristic parameter setting module predetermines the control parameter set of each production process and the quality parameter set of the produced shrink film; The various production processes include a melt extrusion process at an extruder, a mold forming process at a die head, a sheet casting process at a cooling roller, a transverse stretching process at a stretching unit, and a traction and winding process at a winding head; The feature parameter setting module sends the control parameter set and the quality parameter set to the training data collection module; A training data collection module collects a training sample feature vector set and a quality parameter label vector set of a training quality prediction model in advance in a test environment based on a control parameter set and a quality parameter set; The method of collecting the training sample feature vector set and the quality parameter label vector set of the training quality prediction model is: In the test environment, design N shrink film production experiments, where N is the number of shrink film production experiments selected in advance; In each shrink film production experiment, the test personnel manually adjust different control parameters in the control parameter set of different production processes, start all production processes, complete the production of the test shrink film, collect each control parameter of each production process, and the parameter values ​​of each quality parameter in the quality parameter set of the test shrink film finally produced as the quality parameter label; In each shrink film production experiment, all control parameters of all production processes constitute the training sample feature vector of the shrink film production experiment, and the labels of various quality parameters of the corresponding test shrink films constitute the quality parameter label vector of the shrink film production experiment; The training data collection module sends the training sample feature vector set and the quality parameter label vector set to the quality prediction model training module, and sends the training sample feature vector set to the parameter weight analysis module; A quality prediction model training module, which trains a quality prediction model for predicting various quality parameters of shrink film based on a training sample feature vector set and a quality parameter label vector set; The method of training the quality prediction model for predicting various quality parameters of the shrink film is as follows: Step 11: Use a vector-to-vector encoder-decoder architecture to model the quality prediction model for the prediction process of each quality parameter; The parameter control model is modeled as follows: The encoder is an MLP encoder, and the input of the encoder is a training sample feature vector corresponding to each shrink film production experiment; The decoder is an MLP decoder; Step 12: Construct K output heads at the decoder output and jointly learn K prediction tasks; K is the number of quality parameters; Step 13: Set the comprehensive loss function of the quality prediction model; The comprehensive loss function is set as follows: For each quality parameter, a square loss function is used to measure the difference between the quality parameter label and the quality parameter prediction value output by the output head. The square loss functions of each quality parameter are weighted and summed as the comprehensive loss function of the quality prediction model. Step 14: Using each set of training sample feature vectors as input of the quality prediction model, the quality prediction model is trained until the comprehensive loss function of the quality prediction model converges; The quality prediction model training module sends the quality prediction model to the parameter weight analysis module; The parameter weight analysis module disturbs each control parameter in the training sample feature vector set, outputs the quality parameter prediction values ​​before and after the disturbance through the quality prediction model based on the control parameters after the disturbance, obtains the influence weight of each control parameter on each quality parameter based on the quality parameter prediction values ​​before and after the disturbance, and obtains the dominant control parameter of each quality parameter based on the influence weight of each control parameter on each quality parameter; The method for obtaining the influence weight of each control parameter on each quality parameter is: The number of each process is marked as i, and the control parameters in the i-th process are marked as ij; Mark each quality parameter with a number k; Mark each training sample feature vector in the training sample feature vector set as m; Input the mth training sample feature vector into the quality prediction model to obtain the pre-disturbance prediction value vector QPm of each quality parameter output by the quality prediction model; Mark the control parameter ij in the mth training sample feature vector as Zmij; perform random perturbations on the control parameter Zmij, and fix the values ​​of other control parameters in the training sample feature vector to obtain the perturbed sample feature vector; Input the disturbed sample feature vector into the quality prediction model to obtain the output disturbed prediction value vector QLm of each quality parameter; Calculate the predicted difference vector QDm of each quality parameter before and after the disturbance; For the control parameter ij, the gradient tijk of the kth quality parameter to the control parameter is calculated; the calculation formula of the gradient tijk is: ,in is the kth quality parameter in the predicted difference vector QDm, is the random perturbation value of the control parameter ij in the mth disturbed sample feature vector, M is the number of disturbed sample feature vectors; The influence weight of control parameter ij on the kth quality parameter is set as ; The method of obtaining the dominant control parameters of each quality parameter based on the influence weight of each control parameter on each quality parameter is: Constructing an influence weight vector Wk of the k-th quality parameter, where the influence weight vector Wk includes the influence weights of each control parameter on the k-th quality parameter; Normalize the kth quality parameter to obtain the relative contribution weight vector WNk; For the kth quality parameter, select the s control parameters with the largest influence weights from the relative contribution weight vector as the dominant control parameters of the quality parameter; The parameter weight analysis module sends the dominant control parameters affecting the weights and various quality parameters to the control strategy optimization module; The control strategy optimization module provides control suggestions for various control parameters of each production process based on the dominant control parameters of each quality parameter during the actual production process of shrink film; The providing of control suggestions for various control parameters of each production process includes the following steps: Step 21: Preset an expected range for each quality parameter; Step 22: During the actual production process of the shrink film, actual values ​​of various quality parameters of the last shrink film produced are collected; Step 23: Compare the actual value of each quality parameter with the expected range. If the actual value of any quality parameter is not within the expected range, the quality parameter is regarded as the quality parameter to be adjusted and marked as k0. Steps 24 to 28 are executed. Otherwise, no processing is performed. Step 24: Calculate the deviation value DEVk0 of the quality parameter to be adjusted. The calculation formula of the deviation value DEVk0 is: ;in, is the actual value of the quality parameter to be adjusted, K0_min and K0_max are the left and right boundary values ​​of the expected range of the quality parameter to be adjusted respectively; Step 25: Mark the numbers of the leading control parameters of the k0th quality parameter as ij0; The relative contribution weight corresponding to the dominant control parameter ij0 in the relative contribution weight vector WNk0 is marked as Wk0ij0; Calculate the relative gradient gij0 of the dominant control parameter ij0. The calculation formula for the relative gradient gij0 is: ; Step 26: Mark the parameter value of the dominant control parameter ij0 as Qij0_old; According to the gradient descent rule, modify the value of the dominant control parameter ij0 to Qij0_new; The calculation formula of Qij0_new is: ; in, is the preset adaptive attenuation value; Step 27: Collect the actual values ​​of various control parameters in each production process in the current production stage, and replace the value of the dominant control parameter with the modified value of the dominant control parameter to form a control sample feature vector; Step 28: Input the control sample feature vector into the quality prediction model to obtain the predicted value of the k0th quality parameter. If the predicted value is still not within the corresponding expected range, repeat steps 24 to 27 until the predicted value is within the corresponding expected range. Step 29: Use the modified value Qij0_new of the dominant control parameter ij0 as a control suggestion for the dominant control parameter.

[0005] An intelligent scheduling and optimization control method for a polyester cast stretch shrink film production line is proposed, which includes the following steps: Step 1: predetermine a set of control parameters for each production process and a set of quality parameters for the shrink film produced; Step 2: In advance, in a test environment, based on a control parameter set and a quality parameter set, a training sample feature vector set and a quality parameter label vector set for training a quality prediction model are collected; Step 3: Based on the training sample feature vector set and the quality parameter label vector set, a quality prediction model for predicting various quality parameters of the shrink film is trained; Step 4: disturb each control parameter in the training sample feature vector set, output the quality parameter prediction values ​​before and after the disturbance through the quality prediction model based on the control parameters after the disturbance, obtain the influence weight of each control parameter on each quality parameter based on the quality parameter prediction values ​​before and after the disturbance, and obtain the dominant control parameter of each quality parameter based on the influence weight of each control parameter on each quality parameter; Step 5: In the actual production process of shrink film, based on the dominant control parameters of each quality parameter, provide control suggestions for each control parameter of each production process.

[0006] An electronic device is proposed, comprising: a processor and a memory, wherein the memory stores a computer program that can be called by the processor; The processor executes the above-mentioned intelligent scheduling and optimization control method of the polyester cast stretch shrink film production line by calling the computer program stored in the memory.

[0007] A computer-readable storage medium is provided, on which a rewritable computer program is stored; When the computer program is run on a computer device, the computer device executes the above-mentioned intelligent scheduling and optimization control method for a polyester cast stretch shrink film production line.

[0008] Compared with the prior art, the present invention has the following beneficial effects: The present invention predetermines the control parameter set of each production process and the quality parameter set of the produced shrink film by setting a characteristic parameter setting module, predetermines the training sample feature vector set and the quality parameter label vector set of the training quality prediction model based on the control parameter set and the quality parameter set in a test environment, sets a quality prediction model training module to train a quality prediction model for predicting various quality parameters of the shrink film based on the training sample feature vector set and the quality parameter label vector set, sets a parameter weight analysis module to perturb various control parameters in the training sample feature vector set, and outputs the quality prediction model before and after the perturbation based on the control parameters after the perturbation. The predicted values ​​of quality parameters before and after disturbance are used to obtain the influence weights of various control parameters on various quality parameters, and the dominant control parameters of various quality parameters are obtained based on the influence weights of various control parameters on various quality parameters. The control strategy optimization module is set to provide control suggestions for various control parameters of each production process based on the dominant control parameters of each quality parameter in the actual production process of shrink film. By combining intelligent model modeling and optimization control theory, intelligent analysis of multiple quality indicators and multiple parameters is realized, the priority adjustment weights of various control parameters are automatically calculated, and the parameters of each process are dynamically adjusted in real time to keep the product quality of the shrink film within the ideal target range and improve the production quality of the shrink film. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] Figure 1 This is a module connection relationship diagram of the intelligent scheduling and optimization control system of the polyester cast stretch shrink film production line in Example 1 of the present invention; Figure 2It is a flow chart of the intelligent scheduling and optimization control method of the polyester cast stretch shrink film production line in Example 2 of the present invention; Figure 3 This is a schematic diagram of the structure of an electronic device in Embodiment 3 of the present invention; Figure 4 This is a schematic diagram of the computer-readable storage medium structure in Example 4 of the present invention. DETAILED DESCRIPTION

[0010] The technical solution of the present invention will be clearly and completely described below in conjunction with the embodiments. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0011] Example 1

[0012] like Figure 1 As shown, the intelligent scheduling and optimization control system of the polyester cast stretch shrink film production line includes a characteristic parameter setting module, a training data collection module, a quality prediction model training module, a parameter weight analysis module and a control strategy optimization module; wherein each module is electrically connected; The characteristic parameter setting module is mainly used to predetermine the control parameter set of each production process and the quality parameter set of the produced shrink film; In a preferred embodiment, the various production processes include a melt extrusion process at an extruder, a mold forming process at a die head, a sheet casting process at a cooling roller, a transverse stretching process at a stretching unit, and a traction and winding process at a winding head; Further, the control parameter set is a machine parameter that can be adjusted manually or automatically in each process; Specifically, the specific parameter types of the control parameter set in each process are: For the melt extrusion process, the set of control parameters includes screw speed, barrel heating section temperature, melt pressure and melt temperature; For the mold forming process, the set of control parameters includes mold core temperature, extrusion flow rate, and mold body width; For the casting process, the control parameter set includes roller surface temperature, compact air flow temperature, compact air flow rate, initial solidification film thickness and initial solidification film temperature; For the transverse stretching process, the control parameter set includes the MDO / TDO stretching ratio, the stretching roller temperature, the stretching roller linear speed and the winding tension; For the traction and winding process, the control parameter set includes winding tension, winding speed, and thermal shrinkage rate; Furthermore, the set of quality parameters of the shrink film is a set of quantitative parameters that can reflect the quality of the shrink film produced by the polyester cast stretching process; Specifically, the quality parameter set includes film thickness, heat shrinkage, tensile strength and transparency of the shrink film; It should be noted that the above-mentioned various control parameters or quality parameters can be measured by various sensors or detection instruments; for example, the temperature can be measured by a temperature sensor, the flow rate can be measured by a flow sensor, and the tensile strength can be measured by a tension sensor; The feature parameter setting module sends the control parameter set and the quality parameter set to the training data collection module; Furthermore, the training data collection module is mainly used to collect a training sample feature vector set and a quality parameter label vector set of a training quality prediction model based on a control parameter set and a quality parameter set in advance in a test environment; Specifically, the method of collecting the training sample feature vector set and the quality parameter label vector set of the training quality prediction model is: In the test environment, design N shrink film production experiments, where N is the number of shrink film production experiments selected in advance; In each shrink film production experiment, the test personnel manually adjust different control parameters in the control parameter set of different production processes, start all production processes, complete the production of the test shrink film, collect each control parameter of each production process, and the parameter values ​​of each quality parameter in the quality parameter set of the test shrink film finally produced as the quality parameter label; In each shrink film production experiment, all control parameters of all production processes constitute the training sample feature vector of the shrink film production experiment, and the labels of various quality parameters of the corresponding test shrink films constitute the quality parameter label vector of the shrink film production experiment; The training data collection module sends the training sample feature vector set and the quality parameter label vector set to the quality prediction model training module, and sends the training sample feature vector set to the parameter weight analysis module; Furthermore, the quality prediction model training module is mainly used to train a quality prediction model for predicting various quality parameters of the shrink film based on a training sample feature vector set and a quality parameter label vector set; The method of training the quality prediction model for predicting various quality parameters of the shrink film based on the training sample feature vector set and the quality parameter label vector set is: Step 11: Use a vector-to-vector encoder-decoder architecture to model the quality prediction model for the prediction process of each quality parameter; The parameter control model is modeled as follows: The encoder is an MLP encoder, and the input of the encoder is a training sample feature vector corresponding to each shrink film production experiment; The decoder is an MLP decoder; Step 12: construct K output heads at the output end of the decoder, and jointly learn K prediction tasks; K is the number of quality parameters; specifically, each of the output heads corresponds to a quality parameter, such as film thickness, thermal shrinkage rate, etc., and each output head is used to output parameter values ​​of various quality parameters expected for the produced shrink film; Step 13: Set the comprehensive loss function of the quality prediction model; Specifically, the comprehensive loss function is set as follows: For each quality parameter, a square loss function is used to measure the difference between the quality parameter label and the quality parameter prediction value output by the output head. The square loss functions of each quality parameter are weighted and summed as the comprehensive loss function of the quality prediction model. Step 14: Using each set of training sample feature vectors as input of the quality prediction model, the quality prediction model is trained until the comprehensive loss function of the quality prediction model converges; The quality prediction model training module sends the quality prediction model to the parameter weight analysis module; The parameter weight analysis module is mainly used to perturb each control parameter in the training sample feature vector set, and output the quality parameter prediction values ​​before and after the perturbation through the quality prediction model based on the control parameters after the perturbation, and obtain the influence weights of each control parameter on each quality parameter based on the quality parameter prediction values ​​before and after the perturbation, and obtain the dominant control parameters of each quality parameter based on the influence weights of each control parameter on each quality parameter; Specifically, the method of obtaining the influence weight of each control parameter on each quality parameter is: The number of each process is marked as i, and the control parameters in the i-th process are marked as ij; Mark each quality parameter with a number k; Mark each training sample feature vector in the training sample feature vector set as m; Input the mth training sample feature vector into the quality prediction model to obtain the pre-disturbance prediction value vector QPm of each quality parameter output by the quality prediction model; Mark the control parameter ij in the mth training sample feature vector as Zmij; perform random perturbations on the control parameter Zmij, and fix the values ​​of other control parameters in the training sample feature vector to obtain the perturbed sample feature vector; Input the disturbed sample feature vector into the quality prediction model to obtain the output disturbed prediction value vector QLm of each quality parameter; Calculate the predicted difference vector QDm of each quality parameter before and after the disturbance; For the control parameter ij, the gradient tijk of the kth quality parameter to the control parameter is calculated; the calculation formula of the gradient tijk is: ,in is the kth quality parameter in the predicted difference vector QDm, is the random perturbation value of the control parameter ij in the mth disturbed sample feature vector, M is the number of disturbed sample feature vectors; The influence weight of control parameter ij on the kth quality parameter is set as ; Furthermore, the method of obtaining the dominant control parameters of each quality parameter based on the influence weight of each control parameter on each quality parameter is: Constructing an influence weight vector Wk of the k-th quality parameter, where the influence weight vector Wk includes the influence weights of each control parameter on the k-th quality parameter; Normalize the kth quality parameter to obtain the relative contribution weight vector WNk; For the kth quality parameter, select the s control parameters with the largest influence weights from the relative contribution weight vector as the dominant control parameters of the quality parameter; The parameter weight analysis module sends the dominant control parameters affecting the weights and various quality parameters to the control strategy optimization module; Among them, the control strategy optimization module is mainly used to provide control suggestions for various control parameters of each production process based on the dominant control parameters of each quality parameter during the actual production process of shrink film; The providing of control suggestions for various control parameters of each production process includes the following steps: Step 21: Preset the expected range for each quality parameter; for example, the film thickness can be set to [118μm, 122μm], the thermal shrinkage rate to [14%, 16%], etc. For some indicators such as tensile strength and transparency, it is sufficient to set a single-side target lower limit, such as tensile strength>60MPa, transparency>90%; Step 22: During the actual production process of the shrink film, actual values ​​of various quality parameters of the last shrink film produced are collected; Step 23: Compare the actual value of each quality parameter with the expected range. If the actual value of any quality parameter is not within the expected range, the quality parameter is regarded as the quality parameter to be adjusted and marked as k0. Steps 24 to 28 are executed. Otherwise, no processing is performed. Step 24: Calculate the deviation value DEVk0 of the quality parameter to be adjusted. The calculation formula of the deviation value DEVk0 is: ;in, is the actual value of the quality parameter to be adjusted, K0_min and K0_max are the left and right boundary values ​​of the expected range of the quality parameter to be adjusted respectively; Step 25: Mark the numbers of the leading control parameters of the k0th quality parameter as ij0; The relative contribution weight corresponding to the dominant control parameter ij0 in the relative contribution weight vector WNk0 is marked as Wk0ij0; Calculate the relative gradient gij0 of the dominant control parameter ij0. The calculation formula for the relative gradient gij0 is: ; Step 26: Mark the parameter value of the dominant control parameter ij0 as Qij0_old; According to the gradient descent rule, modify the value of the dominant control parameter ij0 to Qij0_new; The calculation formula of Qij0_new is: ; in, is the preset adaptive attenuation value; Step 27: Collect the actual values ​​of various control parameters in each production process in the current production stage, and replace the value of the dominant control parameter with the modified value of the dominant control parameter to form a control sample feature vector; Step 28: Input the control sample feature vector into the quality prediction model to obtain the predicted value of the k0th quality parameter. If the predicted value is still not within the corresponding expected range, repeat steps 24 to 27 until the predicted value is within the corresponding expected range. Step 29: Use the modified value Qij0_new of the dominant control parameter ij0 as a control suggestion for the dominant control parameter.

[0013] Example 2

[0014] like Figure 2 As shown, the intelligent scheduling and optimization control method of the polyester cast stretch shrink film production line includes the following steps: Step 1: predetermine a set of control parameters for each production process and a set of quality parameters for the shrink film produced; Step 2: In advance, in a test environment, based on a control parameter set and a quality parameter set, a training sample feature vector set and a quality parameter label vector set for training a quality prediction model are collected; Step 3: Based on the training sample feature vector set and the quality parameter label vector set, a quality prediction model for predicting various quality parameters of the shrink film is trained; Step 4: disturb each control parameter in the training sample feature vector set, output the quality parameter prediction values ​​before and after the disturbance through the quality prediction model based on the control parameters after the disturbance, obtain the influence weight of each control parameter on each quality parameter based on the quality parameter prediction values ​​before and after the disturbance, and obtain the dominant control parameter of each quality parameter based on the influence weight of each control parameter on each quality parameter; Step 5: In the actual production process of shrink film, based on the dominant control parameters of each quality parameter, provide control suggestions for each control parameter of each production process.

[0015] Example 3

[0016] Figure 3 Schematic diagram of the structure of an electronic device provided by an embodiment of the present application. Figure 3 As shown, according to another aspect of the present application, an electronic device 100 is also provided. The electronic device 100 may include one or more processors and one or more memories. The memories store computer readable codes, and when the computer readable codes are executed by one or more processors, the intelligent scheduling and optimization control method of the polyester cast stretch shrink film production line as described above can be executed.

[0017] The method or device according to the embodiment of the present application can also be used by Figure 3 The electronic device architecture shown in FIG. Figure 3As shown, the electronic device 100 may include a bus 101, one or more CPUs 102, a ROM 103, a RAM 104, a communication port 105 connected to a network, an input / output component 106, a hard disk 107, etc. The storage device in the electronic device 100, such as the ROM 103 or the hard disk 107, may store the intelligent scheduling and optimization control method for the polyester cast stretch shrink film production line provided in the present application. The intelligent scheduling and optimization control of the polyester cast stretch shrink film production line may, for example, include the following steps: Step 1: predetermine the control parameter set of each production process and the quality parameter set of the shrink film produced; Step 2: pre-collect the training sample feature vector set and the quality parameter label vector set of the training quality prediction model based on the control parameter set and the quality parameter set in the test environment; Step 3: train the quality prediction model for predicting various quality parameters of the shrink film based on the training sample feature vector set and the quality parameter label vector set; Step 4: perturb each control parameter in the training sample feature vector set, and output the quality parameter prediction value before and after the perturbation through the quality prediction model based on the control parameter after the perturbation, obtain the influence weight of each control parameter on each quality parameter based on the quality parameter prediction value before and after the perturbation, and obtain the dominant control parameter of each quality parameter based on the influence weight of each control parameter on each quality parameter; Step 5: in the actual production process of the shrink film, provide control suggestions for each control parameter of each production process based on the dominant control parameter of each quality parameter.

[0018] Furthermore, the electronic device 100 may further include a user interface 108. Figure 3 The architecture shown is only exemplary and can be omitted according to actual needs when implementing different devices. Figure 3 One or more components of an electronic device are shown.

[0019] Example 4

[0020] Figure 4 Schematic diagram of a computer-readable storage medium structure provided by an embodiment of the present application. Figure 4 As shown, it is a computer-readable storage medium 200 according to an embodiment of the present application. Computer-readable instructions are stored on the computer-readable storage medium 200. When the computer-readable instructions are executed by the processor, the intelligent scheduling and optimization control method of the polyester cast stretch shrink film production line according to the embodiment of the present application described with reference to the above figures can be executed. The computer-readable storage medium 200 includes, but is not limited to, for example, volatile memory and / or non-volatile memory. Volatile memory may include, for example, random access memory (RAM) and cache memory (cache), etc. Non-volatile memory may include, for example, read-only memory (ROM), hard disk, flash memory, etc.

[0021] In addition, according to the implementation of the present application, the process described above with reference to the flowchart can be implemented as a computer software program. For example, the present application provides a non-transitory machine-readable storage medium, the non-transitory machine-readable storage medium stores machine-readable instructions, the machine-readable instructions can be run by a processor to execute instructions corresponding to the method steps provided by the present application, and when the computer program is executed by a central processing unit (CPU), the above functions defined in the method of the present application are executed.

[0022] The methods, apparatuses, and devices of the present application may be implemented in many ways. For example, the methods, apparatuses, and devices of the present application may be implemented by software, hardware, firmware, or any combination of software, hardware, and firmware. The above order of steps for the method is for illustration only, and the steps of the method of the present application are not limited to the order specifically described above unless otherwise specified. In addition, in some embodiments, the present application may also be implemented as programs recorded in a recording medium, which include machine-readable instructions for implementing the method according to the present application. Therefore, the present application also covers recording media storing programs for executing the method according to the present application.

[0023] In addition, the parts of the above-mentioned technical solutions provided in the embodiments of the present application that are consistent with the implementation principles of the corresponding technical solutions in the prior art are not described in detail to avoid excessive redundancy.

[0024] The specific implementation modes as described above further describe the purpose, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific implementation mode of the present invention and is not intended to limit the present invention. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

[0025] The above preset parameters or preset thresholds are all set by technicians in this field according to actual conditions or obtained through large-scale data simulation.

[0026] The above embodiments are only used to illustrate the technical method of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical method of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical method of the present invention.

Claims

1. Intelligent scheduling and optimization control system for polyester cast stretch shrink film production line, characterized in that: It includes a characteristic parameter setting module, a training data collection module, a quality prediction model training module, a parameter weight analysis module and a control strategy optimization module; wherein each module is electrically connected; A characteristic parameter setting module predetermines a control parameter set of each production process and a quality parameter set of the produced shrink film, and sends the control parameter set and the quality parameter set to a training data collection module; A training data collection module collects a training sample feature vector set and a quality parameter label vector set for training a quality prediction model in advance in a test environment based on a control parameter set and a quality parameter set, and sends the training sample feature vector set and the quality parameter label vector set to a quality prediction model training module, and sends the training sample feature vector set to a parameter weight analysis module; The quality prediction model training module trains a quality prediction model for predicting various quality parameters of the shrink film based on a training sample feature vector set and a quality parameter label vector set, and sends the quality prediction model to a parameter weight analysis module; The parameter weight analysis module disturbs each control parameter in the training sample feature vector set, outputs the quality parameter prediction values ​​before and after the disturbance through the quality prediction model based on the control parameters after the disturbance, obtains the influence weight of each control parameter on each quality parameter based on the quality parameter prediction values ​​before and after the disturbance, obtains the dominant control parameter of each quality parameter based on the influence weight of each control parameter on each quality parameter, and sends the influence weight and the dominant control parameter of each quality parameter to the control strategy optimization module; The control strategy optimization module provides control suggestions for various control parameters of each production process based on the dominant control parameters of each quality parameter during the actual production process of shrink film; The method for obtaining the influence weight of each control parameter on each quality parameter is: The number of each process is marked as i, and the control parameters in the i-th process are marked as ij; Mark each quality parameter with a number k; Mark each training sample feature vector in the training sample feature vector set as m; Input the mth training sample feature vector into the quality prediction model to obtain the pre-disturbance prediction value vector QPm of each quality parameter output by the quality prediction model; Mark the control parameter ij in the mth training sample feature vector as Zmij; perform random perturbations on the control parameter Zmij, and fix the values ​​of other control parameters in the training sample feature vector to obtain the perturbed sample feature vector; Input the disturbed sample feature vector into the quality prediction model to obtain the output disturbed prediction value vector QLm of each quality parameter; Calculate the predicted difference vector QDm of each quality parameter before and after the disturbance; For the control parameter ij, the gradient tijk of the kth quality parameter to the control parameter is calculated; the calculation formula of the gradient tijk is: ,in is the kth quality parameter in the predicted difference vector QDm, is the random perturbation value of the control parameter ij in the mth disturbed sample feature vector, M is the number of disturbed sample feature vectors; The influence weight of control parameter ij on the kth quality parameter is set as .

2. The intelligent dispatching and optimization control system for the polyester cast stretch shrink film production line according to claim 1, characterized in that: The various production processes include a melt extrusion process at an extruder, a mold forming process at a die head, a sheet casting process at a cooling roller, a transverse stretching process at a stretching unit, and a traction and winding process at a winding head.

3. The intelligent dispatching and optimization control system for the polyester cast stretch shrink film production line according to claim 2, characterized in that: The method of collecting the training sample feature vector set and the quality parameter label vector set of the training quality prediction model is: In the test environment, design N shrink film production experiments, where N is the number of shrink film production experiments selected in advance; In each shrink film production experiment, the test personnel manually adjust different control parameters in the control parameter set of different production processes, start all production processes, complete the production of the test shrink film, collect each control parameter of each production process, and the parameter values ​​of each quality parameter in the quality parameter set of the test shrink film finally produced as the quality parameter label; In each shrink film production experiment, all control parameters of all production processes constitute the training sample feature vector of the shrink film production experiment, and the corresponding quality parameter labels of the test shrink film constitute the quality parameter label vector of the shrink film production experiment.

4. The intelligent dispatching and optimization control system for the polyester cast stretch shrink film production line according to claim 3, characterized in that: The method of training the quality prediction model for predicting various quality parameters of the shrink film is as follows: Step 11: Use a vector-to-vector encoder-decoder architecture to model the quality prediction model for the prediction process of each quality parameter; Step 12: Construct K output heads at the decoder output and jointly learn K prediction tasks; K is the number of quality parameters; Step 13: Set the comprehensive loss function of the quality prediction model; The comprehensive loss function is set as follows: For each quality parameter, a square loss function is used to measure the difference between the quality parameter label and the quality parameter prediction value output by the output head. The square loss functions of each quality parameter are weighted and summed as the comprehensive loss function of the quality prediction model. Step 14: Use the feature vector of each group of training samples as the input of the quality prediction model, and train the quality prediction model until the comprehensive loss function of the quality prediction model converges.

5. The intelligent dispatching and optimization control system for the polyester cast stretch shrink film production line according to claim 4, characterized in that: The parameter control model is modeled as follows: The encoder is an MLP encoder, and the input of the encoder is a training sample feature vector corresponding to each shrink film production experiment; The decoder is a MLP decoder.

6. The intelligent dispatching and optimization control system for the polyester cast stretch shrink film production line according to claim 5, characterized in that: The method of obtaining the dominant control parameters of each quality parameter based on the influence weight of each control parameter on each quality parameter is: Constructing an influence weight vector Wk of the k-th quality parameter, where the influence weight vector Wk includes the influence weights of each control parameter on the k-th quality parameter; Normalize the kth quality parameter to obtain the relative contribution weight vector WNk; For the kth quality parameter, s control parameters with the largest influence weights are selected from the relative contribution weight vector as the dominant control parameters of the quality parameter.

7. The intelligent dispatching and optimization control system for the polyester cast stretch shrink film production line according to claim 6, characterized in that: The providing of control suggestions for various control parameters of each production process includes the following steps: Step 21: Preset an expected range for each quality parameter; Step 22: During the actual production process of the shrink film, actual values ​​of various quality parameters of the last shrink film produced are collected; Step 23: Compare the actual value of each quality parameter with the expected range. If the actual value of any quality parameter is not within the expected range, the quality parameter is regarded as the quality parameter to be adjusted and marked as k0. Steps 24 to 28 are executed. Otherwise, no processing is performed. Step 24: Calculate the deviation value DEVk0 of the quality parameter to be adjusted. The calculation formula of the deviation value DEVk0 is: ;in, is the actual value of the quality parameter to be adjusted, K0_min and K0_max are the left and right boundary values ​​of the expected range of the quality parameter to be adjusted respectively; Step 25: Mark the numbers of the leading control parameters of the k0th quality parameter as ij0; The relative contribution weight corresponding to the dominant control parameter ij0 in the relative contribution weight vector WNk0 is marked as Wk0ij0; Calculate the relative gradient gij0 of the dominant control parameter ij0. The calculation formula for the relative gradient gij0 is: ; Step 26: Mark the parameter value of the dominant control parameter ij0 as Qij0_old; According to the gradient descent rule, modify the value of the dominant control parameter ij0 to Qij0_new; The calculation formula of Qij0_new is: ; in, is the preset adaptive attenuation value; Step 27: Collect the actual values ​​of various control parameters in each production process in the current production stage, and replace the value of the dominant control parameter with the modified value of the dominant control parameter to form a control sample feature vector; Step 28: Input the control sample feature vector into the quality prediction model to obtain the predicted value of the k0th quality parameter. If the predicted value is still not within the corresponding expected range, repeat steps 24 to 27 until the predicted value is within the corresponding expected range. Step 29: Use the modified value Qij0_new of the dominant control parameter ij0 as a control suggestion for the dominant control parameter.

8. An intelligent scheduling and optimization control method for a polyester cast stretch shrink film production line, which is implemented based on the intelligent scheduling and optimization control system for a polyester cast stretch shrink film production line according to any one of claims 1 to 7, characterized in that: The following steps are involved: Step 1: predetermine a set of control parameters for each production process and a set of quality parameters for the shrink film produced; Step 2: In advance, in a test environment, based on a control parameter set and a quality parameter set, a training sample feature vector set and a quality parameter label vector set for training a quality prediction model are collected; Step 3: Based on the training sample feature vector set and the quality parameter label vector set, a quality prediction model for predicting various quality parameters of the shrink film is trained; Step 4: disturb each control parameter in the training sample feature vector set, output the quality parameter prediction values ​​before and after the disturbance through the quality prediction model based on the control parameters after the disturbance, obtain the influence weight of each control parameter on each quality parameter based on the quality parameter prediction values ​​before and after the disturbance, and obtain the dominant control parameter of each quality parameter based on the influence weight of each control parameter on each quality parameter; Step 5: In the actual production process of shrink film, based on the dominant control parameters of each quality parameter, provide control suggestions for each control parameter of each production process.

9. An electronic device, characterized in that: include: processor and memory, wherein The memory stores a computer program that can be called by the processor; The processor executes the intelligent scheduling and optimization control method for the polyester cast stretch shrink film production line according to claim 8 in the background by calling the computer program stored in the memory.

10. A computer-readable storage medium, characterized in that: A rewritable computer program is stored thereon; When the computer program is run on a computer device, the computer device executes the intelligent scheduling and optimization control method for a polyester cast stretch shrink film production line according to claim 8 in the background.

Citation Information

Patent Citations

  • Intelligent defect control system for film production line

    CN117369384A

  • Injection moulding process parameter optimization regulation and control method based on Gaussian prediction

    CN104527009A

  • Supervised learning method and device, label prediction method and device, electronic equipment and storage medium

    CN110781934A

  • Blow molding technological parameter online regulating and control method

    CN111002564A

  • Hydrocracking product quality automatic control method and device and memory

    CN111849544A

Cited By

  • Operation control method and system for die-cutting machine

    CN119458499A

  • Synchronous two-way stretch film production control method and system based on process optimization

    CN120595707A

  • A method and system for controlling the production of synchronous biaxially stretched films based on process optimization

    CN120595707B

  • Intelligent optimization method and system for multi-process production energy efficiency of lithium battery diaphragm

    CN121209433A

  • Optimization control method for technological parameters of PTFE (Polytetrafluoroethylene) fiber membrane production line based on Internet of Things

    CN122151560A