Apparatus and method for designing multilayer film

Abstract

Disclosed are an apparatus and a method for designing a multilayer film. The apparatus for designing a multilayer film may model a multilayer film to be designed in a monolayer structure having a preset thickness; may collect physical characteristic data regarding a film corresponding to the monolayer structure; may read a value prestored in a storage space accessible by the apparatus for designing a multilayer film; may obtain a feature setting mode for designating a different feature setting method according to the value that has been read; may perform a feature setup on the basis of multiple physical indices selected from the physical characteristic data according to the feature setting mode; may select at least one from multiple supervised learning models capable of regression analysis as a machine learning model; may predict the falling weight strength of the multilayer film by using the machine learning model trained by using features as independent variables and using the falling weight strength of the multilayer film as a target variable; and may generate design data regarding the multilayer film by combining prediction values regarding other characteristics and prediction values of the falling weight strength such that multifilm design requirements are satisfied.

Description

Design device and method for multilayer films

[0001] Cross-citation with related applications

[0002] This application claims the benefit of priority to Korean Patent Application No. 10-2023-0129783, filed September 26, 2023, and Korean Patent Application No. 10-2024-0130232, filed September 25, 2024, the entire contents of which are incorporated herein by reference.

[0003] The disclosure relates to a device and method for designing a multilayer film.

[0004] Polymer films are thin, flat structures made from polymer materials. Polymers can have a molecular structure in the form of long chains of repeating monomers. The thickness of polymer films can vary, reaching up to millimeters (mm). Polymer films offer numerous advantages, including flexibility, lightness, cost-effectiveness, and processability. Specifically, polymer films are highly flexible and can be manufactured in a variety of shapes and sizes, allowing them to be applied to a variety of product designs. Furthermore, their lightweight nature makes them suitable for products requiring portability, reducing transportation and handling costs. Furthermore, they can be manufactured using relatively inexpensive raw materials and are capable of mass production, making them cost-effective. Furthermore, polymer films can be applied to various manufacturing and processing techniques, such as thermal processing, extrusion, and bonding, allowing for easy adjustment of product shape, size, and thickness. Furthermore, polymer films possess high durability and strength, and some exhibit excellent optical properties, excellent water resistance, and, if desired, breathability.

[0005] A multilayer film can be composed of two or more layers of polymers or other materials. This multilayer structure can enhance the performance of the overall film by combining the unique properties of each layer. For example, one layer may be used to enhance durability and strength, another layer may provide water resistance or moisture repellency, and yet another layer may provide resistance to specific chemicals. By combining layers with different properties, multilayer films can adapt to a wider range of applications and specific requirements than single-layer films.

[0006] The task to be solved is to provide a design device and method for a multilayer film that can predict the drop strength of a multilayer film without information about the single-layer films constituting the multilayer film and design the multilayer film according to given design requirements.

[0007] According to one embodiment, a design device for a multilayer film is a design device for a multilayer film that executes a program code loaded on one or more memory devices through one or more processors to predict the properties of the multilayer film and design the multilayer film based on the predicted properties, wherein the program code is executed to model the multilayer film to be designed as a single-layer structure having a preset thickness, collect property data for a film corresponding to the single-layer structure, read a value stored in advance in a storage space accessible to the design device for the multilayer film, and obtain a feature setting mode for designating a different feature setting method according to the read value, perform feature setting based on a plurality of physical indices selected from the property data according to the feature setting mode, select at least one machine learning model from among a plurality of supervised learning models capable of regression analysis, and predict the drop strength of the multilayer film using the machine learning model learned with the feature as an independent variable and the drop strength of the multilayer film as a target variable, and make predictions about other characteristics so as to satisfy design requirements of the multilayer film. Design data for the multilayer film can be generated by combining the predicted values ​​of the values ​​and the above drop strength.

[0008] In some embodiments, when the feature setting mode includes a first feature setting mode, performing the feature setting may include setting at least some of a yield stress, a yield strain, a necking stress, a necking strain, a breaking stress, and a breaking strain, and the thickness as features, selected from stress-strain data obtained through experiments on a multilayer film having the single layer structure.

[0009] In some embodiments, selecting the machine learning model comprises selecting any one of a Partial Least Squares (PLS) model, a Lasso regression model, and a random forest model as the machine learning model, wherein the machine learning model selected from the PLS model, the Lasso regression model, and the random forest model may be trained to receive at least some of the yield stress, the yield strain, the necking point stress, the necking point strain, the fracture stress, and the fracture strain, and a vector generated from the thickness, and to predict the drop strength of the multilayer film.

[0010] In some embodiments, when the feature setting mode includes a second feature setting mode, performing the feature setting may include setting the thickness as a feature and a strain energy calculated from at least some of a yield stress, a yield strain, a necking stress, a necking strain, a breaking stress, and a breaking strain selected from stress-strain data obtained through an experiment on a multilayer film having the single layer structure.

[0011] In some embodiments, selecting the machine learning model includes selecting any one of a Partial Least Squares (PLS) model, a Lasso regression model, and a random forest model as the machine learning model, and the machine learning model selected from the PLS model, the Lasso regression model, and the random forest model may be trained to receive a vector generated from the strain energy and the thickness as input and to predict the drop strength of the multilayer film.

[0012] In some embodiments, the strain energy is a first strain energy (Wyield ) and performing the feature setting, the first strain energy (W yield ) may include setting the feature by calculating it according to the following mathematical formula:

[0013]

[0014] Here, σ y is the yield stress, and ε y is the yield strain, and σ br is the above-mentioned fracture stress, d is the joint diameter, t is the thickness of the measurement sample, and a is a constant determined experimentally.

[0015] In some embodiments, the strain energy is a second strain energy (W neck ) and performing the feature setting, the second strain energy (W neck ) may include setting the feature by calculating it according to the following mathematical formula:

[0016]

[0017] Here, σ n is the stress at the necking point, and ε n is the strain at the necking point, and ε y is the yield strain, and σ br is the above-mentioned fracture stress, and σ y is the yield stress, d is the joint diameter, t is the thickness of the measurement sample, and b is a constant determined experimentally.

[0018] In some embodiments, the strain energy is a third strain energy (W str-hard ) and performing the feature setting, the third strain energy (W str-hard ) may include setting the feature by calculating it according to the following mathematical formula:

[0019]

[0020] Here, εbr is the above fracture strain, and ε n is the strain at the necking point, and σ br is the above-mentioned fracture stress, and σ n is the necking point stress, L is the final stretching propagation length, t is the thickness of the measurement sample, and c is a constant determined experimentally.

[0021] In some embodiments, the strain energy is a fourth strain energy (W elast ) and performing the feature setting, the fourth strain energy (W elast ) may include setting the feature by calculating it according to the following mathematical formula:

[0022]

[0023] Here, σ y is the yield stress, and ε y is the yield strain, d is the diameter of the joint, D is the diameter of the fixed disk, L is the final stretching propagation length, t is the thickness of the measurement sample, and e is a constant determined experimentally.

[0024] In some embodiments, generating design data for the multilayer film may include combining a predicted value for at least one of stiffness, strength, strain, tear strength, high-velocity drop strength, sealing initiation temperature (SIT), haze, moisture permeability, and air permeability for the multilayer film with a predicted value of the drop strength, and generating the design data based on the combination of predicted values ​​if the multilayer film designed based on the combination of predicted values ​​satisfies the design requirement, and redesigning the multilayer film by changing the combination to another combination if the multilayer film designed based on the combination of predicted values ​​does not satisfy the design requirement.

[0025] According to one embodiment, a method for designing a multilayer film is provided, which is performed by a computing device including one or more processors and one or more memory devices, for predicting properties of a multilayer film and designing the multilayer film based on the predicted properties, the method comprising: modeling the multilayer film to be designed as a single-layer structure having a preset thickness; collecting property data for a film corresponding to the single-layer structure; reading a value stored in advance in a storage space accessible to a design device of the multilayer film, and acquiring a feature setting mode for designating a different feature setting method according to the read value; performing feature setting based on a plurality of physical indices selected from the property data according to the feature setting mode; selecting at least one machine learning model from a plurality of supervised learning models capable of regression analysis; predicting the drop strength of the multilayer film using the machine learning model learned by using the feature as an independent variable and the drop strength of the multilayer film as a target variable; And, to meet the design requirements of the multilayer film, the method may include a step of generating design data for the multilayer film by combining predicted values ​​for other characteristics and predicted values ​​of the drop strength.

[0026] In some embodiments, when the feature setting mode includes a first feature setting mode, the step of performing the feature setting may include a step of setting at least some of the yield stress, yield strain, necking stress, necking strain, breaking stress and breaking strain, and the thickness as features, selected from stress-strain data obtained through experiments on the multilayer film of the single layer structure.

[0027] In some embodiments, the step of selecting the machine learning model includes the step of selecting any one of a Partial Least Squares (PLS) model, a Lasso regression model, and a random forest model as the machine learning model, and the machine learning model selected from the PLS model, the Lasso regression model, and the random forest model may be trained to receive at least some of the yield stress, the yield strain, the necking point stress, the necking point strain, the fracture stress, and the fracture strain, and a vector generated from the thickness, and to predict the drop strength of the multilayer film.

[0028] In some embodiments, when the feature setting mode includes a second feature setting mode, the step of performing the feature setting may include setting the thickness as a feature and a strain energy calculated from at least some of the yield stress, yield strain, necking stress, necking strain, breaking stress and breaking strain selected from stress-strain data obtained through an experiment on a multilayer film having the single layer structure.

[0029] In some embodiments, the step of selecting the machine learning model includes the step of selecting any one of a Partial Least Squares (PLS) model, a Lasso regression model, and a random forest model as the machine learning model, and the machine learning model selected from the PLS model, the Lasso regression model, and the random forest model may be trained to receive a vector generated from the strain energy and the thickness as input and to predict the drop strength of the multilayer film.

[0030] In some embodiments, the strain energy is a first strain energy (W yield ) and the step of performing the feature setting comprises: the first strain energy (W yield ) may include a step of calculating and setting the feature according to the following mathematical formula:

[0031]

[0032] Here, σ y is the yield stress, and ε y is the yield strain, and σ br is the above-mentioned fracture stress, d is the joint diameter, t is the thickness of the measurement sample, and a is a constant determined experimentally.

[0033] In some embodiments, the strain energy is a second strain energy (W neck ), and the step of performing the feature setting comprises: the second strain energy (W neck ) may include a step of calculating and setting the feature according to the following mathematical formula:

[0034]

[0035] Here, σ n is the stress at the necking point, and ε n is the strain at the necking point, and ε y is the yield strain, and σ br is the above-mentioned fracture stress, and σ y is the yield stress, d is the joint diameter, t is the thickness of the measurement sample, and b is a constant determined experimentally.

[0036] In some embodiments, the strain energy is a third strain energy (W str-hard ) and the step of performing the feature setting comprises: the third strain energy (W str-hard ) may include a step of calculating and setting the feature according to the following mathematical formula:

[0037]

[0038] Here, ε br is the above fracture strain, and ε n is the strain at the necking point, and σ br is the above-mentioned fracture stress, and σ n is the necking point stress, L is the final stretching propagation length, t is the thickness of the measurement sample, and c is a constant determined experimentally.

[0039] In some embodiments, the strain energy is a fourth strain energy (W elast ), and the step of performing the feature setting comprises the fourth strain energy (W elast ) may include a step of calculating and setting the feature according to the following mathematical formula:

[0040]

[0041] Here, σ y is the yield stress, and ε y is the yield strain, d is the diameter of the joint, D is the diameter of the fixed disk, L is the final stretching propagation length, t is the thickness of the measurement sample, and e is a constant determined experimentally.

[0042] In some embodiments, the step of generating design data for the multilayer film may include: combining a predicted value for at least one of stiffness, strength, strain, tear strength, high-velocity drop strength, sealing initiation temperature (SIT), haze, moisture permeability, and air permeability for the multilayer film with a predicted value of the drop strength; generating the design data based on the combination of predicted values ​​when the multilayer film designed based on the combination of predicted values ​​satisfies the design requirement; and redesigning the multilayer film by changing the combination to another combination when the multilayer film designed based on the combination of predicted values ​​does not satisfy the design requirement.

[0043] According to embodiments, it is possible to predict the drop strength of a multilayer film without information about the single-layer films that constitute the multilayer film, and to design a multilayer film that can meet given design requirements by taking into account the prediction results regarding other properties.

[0044] Figure 1 is a block diagram illustrating a design device for a multilayer film according to one embodiment.

[0045] FIG. 2 is a diagram illustrating first stress strain data for a multilayer film according to one embodiment.

[0046] FIG. 3 is a diagram illustrating second stress strain data for a multilayer film according to one embodiment.

[0047] Figures 4 and 5 are drawings for explaining the fall strength measurement environment data according to one embodiment.

[0048] Figure 6 is a flowchart illustrating a design method of a multilayer film according to one embodiment.

[0049] FIG. 7 is a drawing for explaining a computing device according to one embodiment.

[0050] Below, with reference to the attached drawings, embodiments of the present invention are described in detail so that those skilled in the art can easily implement the present invention. However, the present invention may be implemented in various different forms and is not limited to the embodiments described herein. In addition, in the drawings, parts irrelevant to the description are omitted for clarity of description, and similar parts are designated with similar reference numerals throughout the specification.

[0051] Throughout the specification and claims, whenever a part is referred to as "comprising" a component, this does not exclude other components, but rather includes other components, unless otherwise stated. Terms including ordinal numbers, such as "first," "second," etc., may be used to describe various components, but these components are not limited by these terms. These terms are used solely to distinguish one component from another.

[0052] Terms such as "...part," "...unit," and "module" described in the specification may refer to a unit capable of processing at least one function or operation described in the specification, which may be implemented by hardware or a circuit, software, or a combination of hardware or a circuit and software. In addition, at least some of the configurations or functions of the design device and method for a multilayer film according to the embodiments described below may be implemented by a program or software, and the program or software may be stored in a computer-readable medium.

[0053] FIG. 1 is a block diagram for explaining a design device of a multilayer film according to one embodiment, FIG. 2 is a diagram for explaining first stress strain data for a multilayer film according to one embodiment, FIG. 2 is a diagram for explaining second stress strain data for a multilayer film according to one embodiment, and FIGS. 4 and 5 are diagrams for explaining drop strength measurement environment data according to one embodiment.

[0054] Referring to FIG. 1, a design device (10) for a multilayer film according to one embodiment can execute program codes or instructions loaded in one or more memory devices through one or more processors. For example, the design device (10) for a multilayer film can be implemented as a computing device (50) as described below with reference to FIG. 7. In this case, one or more processors may correspond to a processor (510) of the computing device (50), and one or more memory devices may correspond to a memory (520) of the computing device (50). The program codes or instructions, executed by one or more processors, can predict the drop strength of a multilayer film without information about a single-layer film constituting the multilayer film, and design a multilayer film that meets given design requirements. In this specification, the term "module" is used to logically distinguish these functions performed by the program codes or instructions.

[0055] Drop strength can refer to the strength that a material, in this case a film, can withstand when subjected to a physical impact. Drop strength can be used as an indicator for measuring the extent of damage to a film when, for example, a weight is dropped from a predetermined height. The design device (10) for a multilayer film may include all or at least a portion of a data collection module (110), a feature setting module (120), a model selection module (130), a learning module (140), a prediction module (150), and a design data generation module (160). For example, the learning module (150) may be implemented within the design device (10) for a multilayer film. Alternatively, for example, the learning module (140) may be implemented outside the design device (10) for a multilayer film, and the design device (10) for a multilayer film may receive a machine learning model learned from the learning module (140).

[0056] A multilayer film may include multiple single-layer films. Here, the multiple single-layer films may include some types of single-layer films and other types of single-layer films, and the unique characteristics of the single-layer films may differ depending on the type. In addition, the stacking order of the single-layer films within the multilayer film may also vary. The design device (10) for a multilayer film can predict the drop strength of the multilayer film using a feature-based machine learning model set from the physical property data collected by the data collection module (110) described below. At this time, the design device (10) for a multilayer film can model the multilayer film to be designed as a single-layer structure having a preset thickness, rather than modeling it as a form in which multiple single-layer films are stacked. Here, the preset thickness may be selected as, for example, the thickness of the final multilayer film to be manufactured, i.e., the final product. By modeling the multilayer film as a single-layer structure, the drop strength can be predicted based only on data for the multilayer film in its final product form, without having to perform experiments on each single-layer film or obtain information about each single-layer film.

[0057] In predicting drop strength, the properties of the multilayer film itself, such as the final product, are sometimes particularly important. In contrast, there are cases where the properties of a specific single-layer film within the multilayer film must be fixed. In the latter case, the composition of some single-layer films within the multilayer film must be fixed to meet the product's required properties. For example, when a multilayer film is used for food applications, the properties of the single-layer film responsible for functions such as moisture permeability or oxygen permeability may be considered separately. In such cases, it may be useful to define the features applied to the machine learning model for drop strength prediction from the property data of the single-layer films within the multilayer film. However, in the former case, for example, in agricultural or heavy-duty applications where the mechanical properties of the final film are important, mechanical properties such as stiffness, strength, and tearing may also be important factors in addition to drop strength. Therefore, in such cases, it is effective for the features applied to the machine learning model for predicting the impact strength to be set from the material property data for the multilayer film itself, which is treated as a single layer, and can be more suitable for the analysis purpose while using fewer features than when based on the material property data for the single-layer film forming the multilayer film.

[0058] The data collection module (110) can collect property data for a film corresponding to a single-layer structure. That is, the data collection module (110) can collect property data by assuming the multilayer film itself to be a single layer, regardless of the type of single-layer film forming the multilayer film, the stacking order of the single-layer films, etc.

[0059] In some embodiments, the data collection module (110) may collect stress-strain data (hereinafter, first stress-strain data) for a multilayer film as material property data. The first stress-strain data may represent mechanical properties for a single-layer structure, i.e., a multilayer film treated as a single layer, and, referring to FIG. 2, may be expressed in the form of a stress-strain curve. The stress-strain curve may represent stress generated when a force is applied to the multilayer film, and how the multilayer film is deformed as a result thereof.

[0060] In some embodiments, the first stress strain data may include at least one of a yield point, a yield stress, resilience, strain hardening, failure, a strain hardening modulus, flow energy, strain hardening energy, toughness, a 1% modulus, and a 2% modulus. Here, the yield point represents the point at which a material transitions from an elastic deformation region to an inelastic deformation region, after which the material may not fully return to its original shape even when the applied stress is removed. The yield stress may be a stress value at the yield point, and a higher value may indicate that the material can withstand a greater stress before being deformed. Resilience may be a measure of how well a material can return to its original shape when the stress is removed after being deformed by a certain stress, which is the area under the stress strain curve up to the yield point. Strain hardening refers to the phenomenon in which a material strengthens as strain increases to withstand additional stress after the yield point, while failure can indicate the point at which the material can no longer withstand stress and breaks. The strain hardening coefficient is represented by the slope of the stress-strain curve in the strain-hardening region and can be used as an indicator of how quickly a material strengthens. The flow energy is the area under the stress-strain curve from the yield point to strain hardening, and the strain hardening energy can represent the area under the stress-strain curve from strain hardening to failure. Toughness is the total area under the stress-strain curve and indicates how much energy a material can absorb, and the 1% modulus and 2% modulus can represent the stress at 1% and 2% strain, respectively.The values ​​of 1% modulus and 2% modulus can be used as measurements of strength within a range that does not exceed the elastic range of the material.

[0061] In some embodiments, the data collection module (110) may collect drop strength data for a multilayer film through a test according to "Method A" of ASTM D1709. Here, the test may evaluate the impact strength or toughness of the film using a freely falling dart. Further details regarding ASTM D1709 may be referenced from ASTM standard documents, and thus are not described herein.

[0062] In some embodiments, the data collection module (110) may acquire stress-strain curves as data based on the results of a uniaxial tensile test performed on a multilayer film in the Machine Direction (MD) and the Transverse Direction (TD). Here, MD may be the direction in which the film moves between the rollers of the manufacturing machine during the film manufacturing process, and TD may be the direction orthogonal to the rollers of the manufacturing machine, i.e., perpendicular to MD. Meanwhile, the uniaxial tensile test may be performed in a manner of measuring the resistance when the multilayer film is stretched in the uniaxial direction. In this case, the data collection module (120) may collect data regarding the stress-strain curves through a test according to ASTM D882. Further details regarding ASTM D882 may be referred to the ASTM standard document, and thus, a description thereof will be omitted herein.

[0063] In some other embodiments, the data collection module (110) can collect stress-strain data (hereinafter, referred to as second stress-strain data) obtained through a UTM (Universal Testing Machine) experiment on a multilayer film as material property data. The UTM experiment can be designed to perform various mechanical tests on a material, and can measure the load value when a load corresponding to the force is generated by applying work to a specimen for a certain time and in a certain direction. That is, experiments such as compression, tension, and bending can be performed by applying a force to a specimen and measuring the load value until the specimen deforms. The data collection module (110) can collect the results obtained through such a UTM experiment on a specimen for a multilayer film as stress-strain data.

[0064] In some embodiments, the second stress strain data may include at least one of Yield Stress, Yield Strain, Necking Stress, Necking Strain, Breaking Stress, and Breaking Strain. Here, Yield Stress may represent the stress at the yield point, which is the point where the material begins to permanently deform beyond its elastic limit, i.e., the point where the material transitions from an elastic deformation region to an inelastic deformation region, and Yield Strain may represent the strain corresponding to the Yield Stress. Necking Point Stress represents the stress at the point where the cross-section of the material begins to narrow locally (the necking point), and typically, the load capacity of the material may be rapidly reduced after the necking point. Necking Point Strain may represent the strain at the necking point. Breaking Stress represents the stress at the point where the material finally breaks, at which point the material can no longer support the load. Breaking Strain may represent the strain corresponding to the Breaking Stress.

[0065] Referring to Figure 3 together, exemplary points (P) corresponding to yield stress and yield strain on the stress-strain curve y ), exemplary points corresponding to necking point stress and necking point strain (P n ), and exemplary points corresponding to the breaking stress and breaking strain (P br ) can be determined.

[0066] The feature setting module (120) can perform feature setting based on a plurality of physical indices selected from the physical property data collected by the data collection module (110) according to a plurality of feature setting modes. The plurality of feature setting modes are stored and managed as predetermined values ​​in a storage space accessible to a processor of a design device (10) for a multilayer film, and the design device (10) for a multilayer film can read the values ​​stored in advance in the storage space and determine a feature setting mode for specifying a different feature setting method according to the read values.

[0067] In machine learning, a feature is an input variable used to predict the target to be predicted, i.e., the properties of a multilayer film, and can refer to individual independent variables of data. Features provide information related to the dependent variable or target variable to be predicted within a dataset and can have a significant impact on the performance of a machine learning model (20). The learning module (150) can use features to train the machine learning model (20). At this time, features constitute a portion of the data input to the machine learning model (20), and the machine learning model (20) can learn patterns based on the data and perform predictions. In particular, feature engineering, which selects or transforms features, can be an important process for improving the performance of the machine learning model (20). In other words, selecting appropriate features or appropriately transforming features can have a significant impact on the high performance of the machine learning model (20).

[0068] In some embodiments, when the feature setting mode includes a first feature setting mode, the feature setting module (120) may perform feature setting based on a plurality of physical indices selected from the first stress strain data.

[0069] In some embodiments, the feature setting module (120) may set at least some of the yield point, yield stress, elasticity, strain hardening, fracture, strain hardening coefficient, flow energy, strain hardening energy, toughness, 1% modulus, and 2% modulus, and the thickness of the multilayer film as features. That is, the feature setting module (120) may select only some of the first stress strain data collected by the data collection module (110) and set the thickness of the multilayer film as features. Then, the learning module (140) may proceed with learning of the machine learning model (20) based on the features including only some of the first stress strain data and the thickness of the multilayer film. The corresponding embodiment may be adopted in an environment where computing resources for performing learning and prediction for the machine learning model (20) are limited.

[0070] In some embodiments, the feature setting module (120) may set all of the yield point, yield stress, elasticity, strain hardening, fracture, strain hardening coefficient, flow energy, strain hardening energy, toughness, 1% modulus, and 2% modulus, and the thickness of the multilayer film as features. That is, the feature setting module (120) may set all of the first stress strain data collected by the data collection module (110) and the thickness of the multilayer film as features. Then, the learning module (140) may proceed with learning of the machine learning model (20) based on features including all of the first stress strain data and the thickness of the multilayer film. The corresponding embodiment may be adopted in an environment having sufficient computing resources for performing learning and prediction for the machine learning model (20).

[0071] In some embodiments, the first stress-strain data may include first direction stress-strain data and second direction stress-strain data. The first direction stress-strain data may represent mechanical properties of the multilayer film in a first direction, and the second direction stress-strain data may represent mechanical properties of the multilayer film in a second direction perpendicular to the first direction. Specifically, the first direction may be a machine direction (MD), and the second direction may be a transverse direction (TD). The MD may be a direction in which the film moves between rollers of a manufacturing machine in a film manufacturing process, and the TD may be a direction orthogonal to the rollers of the manufacturing machine, and may be perpendicular to the MD. The feature setting module (120) may set a plurality of physical indices selected from at least one of the first direction stress-strain data and the second direction stress-strain data as features.

[0072] In some embodiments, the feature setting module (120) may set at least some of the yield point, yield stress, elasticity, strain hardening, failure, strain hardening coefficient, flow energy, strain hardening energy, toughness, 1% modulus, and 2% modulus for the first direction, and at least some of the yield point, yield stress, elasticity, strain hardening, failure, strain hardening coefficient, flow energy, strain hardening energy, toughness, 1% modulus, and 2% modulus for the second direction, and the thickness of the multilayer film as features. That is, the feature setting module (120) may set only some of the first stress strain data in the first direction collected by the data collection module (120), only some of the first stress strain data in the second direction selected, and the thickness of the multilayer film as features.

[0073] In some embodiments, the feature setting module (120) may set as features all of the yield point, yield stress, elasticity, strain hardening, failure, strain hardening coefficient, flow energy, strain hardening energy, toughness, 1% modulus, and 2% modulus in the first direction, and all of the yield point, yield stress, elasticity, strain hardening, failure, strain hardening coefficient, flow energy, strain hardening energy, toughness, 1% modulus, and 2% modulus in the second direction, and the thickness of the multilayer film. That is, the feature setting module (120) may set as features all of the first stress strain data in the first direction collected by the data collection module (120), all of the first stress strain data in the second direction, and the thickness of the multilayer film.

[0074] In some other embodiments, when the feature setting mode includes a second feature setting mode, the feature setting module (120) can perform feature setting based on a plurality of physical indices selected from the second stress strain data.

[0075] In some embodiments, the feature setting module (120) may set at least a portion of the yield stress, the yield strain, the necking point stress, the necking point strain, the fracture stress, and the fracture strain selected from the second stress strain data, i.e., the stress-strain data obtained through an experiment on a multilayer film having a single-layer structure, and the thickness of the multilayer film as features. Then, the learning module (140) may proceed with learning the machine learning model (20) based on the features including only a portion of the second stress-strain data selected and the thickness of the multilayer film. The embodiment may be adopted in an environment where computing resources for performing learning and prediction for the machine learning model (20) are limited.

[0076] In some embodiments, the feature setting module (120) may set all of the yield stress, yield strain, necking point stress, necking point strain, fracture stress, and fracture strain selected from the second stress strain data, i.e., the stress-strain data obtained through an experiment on a multilayer film having a single-layer structure, and the thickness of the multilayer film as features. Then, the learning module (140) may proceed with learning of the machine learning model (20) based on the features including all of the second stress-strain data and the thickness of the multilayer film. The embodiment may be adopted in an environment having sufficient computing resources for performing learning and prediction for the machine learning model (20).

[0077] In some embodiments, the feature setting module (120) may set at least some of the yield stress, the yield strain, the necking point stress, the necking point strain, the fracture stress, and the fracture strain for the first direction, and at least some of the yield stress, the yield strain, the necking point stress, the necking point strain, the fracture stress, and the fracture strain for the second direction, and the thickness of the multilayer film as features. Then, the learning module (140) may proceed with learning of the machine learning model (20) based on features including only some of the second stress strain data in the first direction, only some of the second stress strain data in the second direction, and the thickness of the multilayer film. The embodiment may be adopted in an environment where computing resources for performing learning and prediction for the machine learning model (20) are limited.

[0078] In some embodiments, the feature setting module (120) may set all of the yield stress, yield strain, necking point stress, necking point strain, fracture stress, and fracture strain in the first direction, all of the yield stress, yield strain, necking point stress, necking point strain, fracture stress, and fracture strain in the second direction, and the thickness of the multilayer film as features. Then, the learning module (140) may proceed with learning of the machine learning model (20) based on features including all of the second stress strain data in the first direction, all of the second stress strain data in the second direction, and the thickness of the multilayer film. The corresponding embodiment may be adopted in an environment having sufficient computing resources for performing learning and prediction for the machine learning model (20).

[0079] In some other embodiments, when the feature setting mode includes a third feature setting mode, the feature setting module (120) may calculate strain energy from a plurality of physical indices selected from the second stress strain data and set the strain energy as a feature.

[0080] In some embodiments, the feature setting module (120) may set the strain energy calculated from at least some of the yield stress, yield strain, necking point stress, necking point strain, fracture stress, and fracture strain selected from the second stress strain data, i.e., the stress-strain data obtained through an experiment on a multilayer film having a single-layer structure, and the thickness of the multilayer film as features. Then, the learning module (140) may proceed with learning of the machine learning model (20) based on the features including the strain energy calculated from only some of the second stress-strain data selected and the thickness of the multilayer film. The embodiment may be adopted in an environment where computing resources for performing learning and prediction for the machine learning model (20) are limited.

[0081] In some embodiments, the feature setting module (120) may set the strain energy for all of the yield stress, yield strain, necking point stress, necking point strain, fracture stress, and fracture strain selected from the second stress strain data, i.e., the stress-strain data obtained through an experiment on a multilayer film having a single-layer structure, and the thickness of the multilayer film as features. Then, the learning module (140) may proceed with learning the machine learning model (20) based on the features including the strain energy calculated from all of the second stress-strain data and the thickness of the multilayer film. The embodiment may be adopted in an environment having sufficient computing resources for learning and predicting the machine learning model (20).

[0082] In some embodiments, the feature setting module (120) obtains the first strain energy (W) from the second stress strain data. yield ) can be set as a feature by calculating the first strain energy (W yield ) can be calculated according to the following mathematical formula (1).

[0083] Mathematical formula (1)

[0084]

[0085] Here, σ y is the yield stress among the second stress strain data, and ε y is the yield strain among the second stress strain data, and σ among the stress strain data br may be the breaking stress among the second stress strain data. These variables may correspond to experimental result data measured by a method other than the drop test (e.g., UTM test).

[0086] Meanwhile, in order to calculate strain energy from the second stress strain data, variables related to the environmental settings of the drop test according to a predetermined standard may be additionally used for the multilayer film whose drop strength is to be predicted. Here, the standard may represent the ASTM (American Society for Testing and Materials) standard, or may represent an arbitrary standard that is customized in consideration of the learning performance of the machine learning model (20). Specifically, as variables related to the environmental settings during the drop test, d is the diameter of the weight joint, t is the thickness of the measurement sample, and a may be a constant experimentally determined in consideration of the experimental environment or the product testing environment. Referring also to FIG. 4, the joint diameter d may refer to the diameter of the circular area that is joined to the additional film when the weight is dropped from a predetermined height. The weight includes a fixed disk having a diameter D, the area indicated by l is the area where stretching is propagated, and the area indicated by L represents the area (or length) where stretching is ultimately propagated. The variables related to the environmental settings of these drop experiments (d, D, l, L, t, etc.) can be appropriately adjusted to improve the accuracy of the drop strength to be ultimately predicted.

[0087] In some embodiments, the feature setting module (120) obtains a second strain energy (W) from the second stress strain data. neck ) can be set as a feature by calculating the second strain energy (W neck ) can be calculated according to the following mathematical formula (2).

[0088] Mathematical formula (2)

[0089]

[0090] Here, σ n is the necking point stress among the second stress strain data, and ε nis the necking point strain among the second stress strain data, and ε y is the yield strain among the second stress strain data, and σ br is the breaking stress among the second stress strain data, and σ y is the yield stress among the second stress strain data, and these variables may correspond to experimental result data measured in a manner other than the drop test as described above.

[0091] Meanwhile, as variables related to the environmental settings of the drop experiment, d may be the diameter of the drop joint, t may be the thickness of the measurement sample, and b may be a constant experimentally determined considering the experimental environment or product testing environment.

[0092] In some embodiments, the feature setting module (120) derives a third strain energy (W) from the second stress strain data. str-hard ) can be set as a feature by calculating the third strain energy (W str-hard ) can be calculated according to the following mathematical formula (3).

[0093] Mathematical formula (3)

[0094]

[0095] Here, ε br is the fracture strain among the second stress strain data, and ε n is the necking point strain among the second stress strain data, and σ br is the breaking stress among the second stress strain data, and σ n may be the necking point stress among the second stress strain data. These variables may correspond to experimental result data measured by a method other than the drop test as described above.

[0096] Meanwhile, as variables related to the environmental settings of the drop experiment, L is the final stretching propagation length, t is the thickness of the measurement sample, and c can be a constant experimentally determined by considering the experimental environment or product test environment.

[0097] In some embodiments, the feature setting module (120) obtains a fourth strain energy (W) from the second stress strain data. elast ) can be set as a feature by calculating the fourth strain energy (W elast ) can be calculated according to the following mathematical formula (4).

[0098] Mathematical formula (4)

[0099]

[0100] Here, σ y is the yield stress among the second stress strain data, and ε y may be the yield strain among the second stress strain data. These variables may correspond to experimental result data measured in a manner other than the drop test as described above.

[0101] Meanwhile, as variables related to the environmental settings of the drop experiment, d is the diameter of the drop joint, D is the diameter of the fixed disk, L is the final stretching propagation length, t is the thickness of the measurement sample, and e can be a predetermined constant.

[0102] In some embodiments, the final stretching propagation length L may be calculated from experimental data measured by methods other than the drop experiment. Specifically, the final stretching propagation length L may be calculated according to the following mathematical expression (5).

[0103] Mathematical formula (5)

[0104]

[0105] Here, σ br is the breaking stress among the second stress strain data, and σ yis the yield stress among the second stress strain data, and d may be the diameter of the additional joint.

[0106] In some embodiments, the feature setting module (120) sets the first strain energy (W yield ), second strain energy (W neck ), third strain energy (W str-hard ) and the fourth strain energy (W elast ) and the thickness of the multilayer film can be set as features. The learning module (140) can proceed with learning of the machine learning model (20) based on the features set in this way. This embodiment can be adopted in an environment with sufficient computing resources for learning and predicting the machine learning model (20).

[0107] Alternatively, in some other embodiments, the feature setting module (120) may set the first strain energy (W yield ), second strain energy (W neck ), third strain energy (W str-hard ) and the fourth strain energy (W elast ) and the thickness of the multilayer film can be set as features. The learning module (140) can proceed with learning of the machine learning model (20) based on the features set in this way. This embodiment can be adopted in an environment where computing resources for learning and predicting the machine learning model (20) are limited.

[0108] In some embodiments, the feature setting module (120) sets a first strain energy (W) for the first direction yield ), second strain energy (W neck ), third strain energy (W str-hard ) and the fourth strain energy (W elast ) and the first strain energy (W) for the second direction yield ), second strain energy (W neck ), third strain energy (W str-hard ) and the fourth strain energy (Welast ) and the thickness of the multilayer film can all be set as features. The learning module (140) can proceed with learning of the machine learning model (20) based on the features set in this way. This embodiment can be adopted in an environment with sufficient computing resources for learning and predicting the machine learning model (20).

[0109] Alternatively, in some other embodiments, the feature setting module (120) may set a first strain energy (W) for the first direction yield ), second strain energy (W neck ), third strain energy (W str-hard ) and the fourth strain energy (W elast ) and at least some of the first strain energy (W) in the second direction. yield ), second strain energy (W neck ), third strain energy (W str-hard ) and the fourth strain energy (W elast ) and the thickness of the multilayer film can be set as features. The learning module (140) can proceed with learning of the machine learning model (20) based on the features set in this way. This embodiment can be adopted in an environment where computing resources for learning and predicting the machine learning model (20) are limited.

[0110] The model selection module (130) can select a machine learning model to be used to predict the physical properties of the multilayer film, i.e., the impact strength. Specifically, the model selection module (130) can select at least one machine learning model from among a plurality of supervised learning models capable of regression analysis for predicting the impact strength of the multilayer film. Supervised learning is a method of training a model using labeled training data, and in supervised learning, the model learns the relationship between input variables and target variables, thereby being able to predict the value of the target variable for new input data. In some embodiments, the model selection module (130) can select a supervised learning model capable of regression analysis from among Partial Least Squares (PLS), Linear Regression, Polynomial Regression, Ridge Regression, Lasso Regression, Logistic Regression, Decision Trees, Random Forests, Gradient Boosting, and Neural Networks. The PLS model can be used to model the relationship between multiple independent variables and multiple dependent variables (or target variables), and can find independent variables that are determined to be important for predicting the dependent variable. In linear regression, the model learns a linear relationship between input variables and the target variable, and in polynomial regression, the model can learn a complex relationship by extending linear regression to include higher-order terms of input variables. Ridge regression and Lasso regression add regularization to linear regression to prevent overfitting and control the complexity of the model. Logistic regression is a classification algorithm for solving binary classification problems, but it can be used in regression analysis because it predicts the outcome as a probability.Decision trees, random forests, and gradient boosting learn complex nonlinear relationships and can be used for regression problems as well as classification. Neural networks can solve various types of regression problems, and deep learning, in particular, can be very effective for learning complex nonlinear relationships. Of course, the supervised learning models capable of regression analysis that can be selected by the model selection module (130) are not limited to those listed above, and other types of models not listed may be selected based on specific implementation objectives, implementation environments, required performance, etc.

[0111] The learning module (140) can train a machine learning model (20) using the features set by the feature setting module (120) as independent variables and the drop strength of the multilayer film as a target variable. Accordingly, the machine learning model (20) can be trained as a model capable of predicting the drop strength of the multilayer film for new data, i.e., material property data regarding the multilayer film, by learning the relationship between the features and the target variable. In particular, the learning module (140) can use different learning methods depending on the model selected by the model selection module (130).

[0112] The prediction module (150) can predict the impact strength of a multilayer film using the machine learning model (20) learned by the learning module (140). That is, the prediction module (150) can transmit new data, i.e., physical property data regarding the multilayer film, to the machine learning model (20) and obtain a prediction result regarding the impact strength of the multilayer film.

[0113] Specifically, the model selection module (130) can perform model selection according to a plurality of model selection modes. The plurality of model selection modes are stored and managed as predetermined values ​​in a storage space accessible to the processor of the design device (10) of the multilayer film, and the design device (10) of the multilayer film can read the values ​​stored in advance in the storage space and determine a model selection mode for designating a different model selection method according to the read values.

[0114] In some embodiments, when the model selection mode is the first model selection mode, the model selection module (130) may select the PLS model as the machine learning model to be used to predict the drop strength of the multilayer film. As described above, the features set by the feature setting module (120) depending on the feature setting mode may be combined into a form such as a single vector and used as input to the PLS model. The drop strength of the multilayer film is set as the target variable, and the learning module (150) may train the PLS model in a direction that maximizes the correlation between the input features and the target variable. At this time, the learning module (150) optimizes the "n_components" parameter so that the PLS model can properly explain the data, and the optimal "n_components" value may be determined through cross-validation. Through the PLS model trained in this way, the drop strength of the multilayer film can be predicted by inputting features for a new single-layer film.

[0115] In some other embodiments, when the model selection mode is the second model selection mode, the model selection module (130) may select a Lasso regression model as a machine learning model to be used to predict the crash strength of a multilayer film. As described above, features set by the feature setting module (120) depending on the feature setting mode may be combined into a form such as a single vector and used as input to the Lasso regression model. The crash strength of the multilayer film may be set as a target variable, and the learning module (150) may train the Lasso regression model in a way that induces the model to select only important features through L1 regularization (regularization that adds the sum of the absolute values ​​of the regression coefficients to the objective function as a penalty). At this time, the learning module (150) optimizes the "alpha" parameter for adjusting the regularization strength during model training, and the optimal "alpha" value may be determined through cross-validation. Through the Lasso regression model trained in this way, the crash strength of a new single-layer film may be predicted by inputting features for the multilayer film.

[0116] In some other embodiments, when the model selection mode is the third model selection mode, the model selection module (130) may select a random forest model as a machine learning model to be used to predict the crash strength of a multilayer film. As described above, features set by the feature setting module (120) according to the feature setting mode may be combined into a form such as a single vector and used as input to the random forest model. The crash strength of the multilayer film may be set as a target variable, and the learning module (150) may train the random forest model by setting hyperparameters such as "n_estimators," "max_depth," and "min_samples_split." Through the random forest model trained in this way, features for a new single-layer film may be input to predict the crash strength of the multilayer film.

[0117] In some embodiments, a machine learning model selected from among a PLS model, a Lasso regression model, and a random forest model may be trained to receive as input at least some or all of yield stress, yield strain, necking point stress, necking point strain, fracture stress, and fracture strain, and a vector generated from the thickness of the multilayer film, and to predict the drop strength of the multilayer film.

[0118] In some other embodiments, a machine learning model selected from among a PLS model, a Lasso regression model, and a random forest model may be trained to receive strain energy calculated from at least some or all of yield stress, yield strain, necking point stress, necking point strain, fracture stress, and fracture strain, and a vector generated from the thickness of the multilayer film, and to predict the drop strength of the multilayer film.

[0119] The design data generation module (160) can generate design data for the multilayer film by combining predicted values ​​for other characteristics with predicted values ​​for drop weight strength obtained through the prediction module (150) to meet design requirements of the multilayer film. In some embodiments, the other characteristics can include at least one of stiffness, strength, strain, tear strength, high-speed drop weight strength, sealing initiation temperature (SIT), haze, moisture permeability, and air permeability for the multilayer film.

[0120] The design data generation module (160) can first combine predicted values ​​for at least one of stiffness, strength, strain, tear strength, high-speed drop strength, seal initiation temperature, transparency, moisture permeability, and air permeability for the multilayer film with predicted values ​​for the drop strength. The design requirements for the multilayer film to be designed may include requirements for ensuring mechanical properties, thermal properties, optical properties, barrier properties, and chemical properties to secure desired performance. If the multilayer film designed based on the first combination of predicted values ​​satisfies the design requirements, the design data generation module (160) can generate design data based on the first combination. In this case, the design data may include specific predicted values.

[0121] In contrast, if the multilayer film designed based on the first combination of predicted values ​​does not meet the design requirements, the design data generation module (160) can redesign the multilayer film by changing the first combination to another combination. That is, the design data generation module (160) can secondarily combine the predicted value for at least one of the stiffness, strength, strain, tear strength, high-speed drop strength, seal initiation temperature, transparency, moisture permeability, and air permeability of the multilayer film with the predicted value of the drop strength so as to have a different combination from the first combination. At this time, the predicted value of the drop strength in the first combination and the predicted value of the drop strength in the second combination may be different values.

[0122] If the multilayer film designed based on the secondary combination of predicted values ​​satisfies the design requirements, the design data generation module (160) can generate design data based on the secondary combination. Alternatively, if the multilayer film designed based on the secondary combination of predicted values ​​also does not satisfy the design requirements, the design data generation module (160) can repeat the process of redesigning the multilayer film by changing the secondary combination to another combination.

[0123] According to this embodiment, when the properties of the multilayer film itself, such as the final product, are particularly important, the multilayer film is treated as a single layer, and learning and prediction are performed using elements selected from the collected property data as features, thereby predicting the drop strength of the multilayer film with high accuracy and using fewer features than when based on the property data of the single-layer film forming the multilayer film, while better meeting the analysis purpose. In addition, by varying the number of indices selected as features when performance is prioritized and the number of indices selected as features when computing resources are saved, flexible design is possible according to the implementation purpose and environment of the design device for the multilayer film. Furthermore, by considering the prediction results regarding other characteristics of the multilayer film together with the prediction results of the drop strength, it is possible to design a multilayer film that can satisfy given design requirements.

[0124] Figure 6 is a flowchart illustrating a design method of a multilayer film according to one embodiment.

[0125] Referring to FIG. 6, a method for designing a multilayer film according to an embodiment includes a step (S601) of modeling a multilayer film to be designed as a single-layer structure having a preset thickness, a step (S602) of collecting physical property data for a film corresponding to the single-layer structure, a step (S603) of reading a value stored in advance in a storage space accessible to a design device of the multilayer film and obtaining a feature setting mode for designating a different feature setting method according to the read value, a step (S604) of setting a feature from a plurality of physical indices selected from the physical property data according to the feature setting mode, a step (S605) of selecting at least one machine learning model among a plurality of supervised learning models capable of regression analysis, a step (S606) of training a machine learning model using features as independent variables and the drop strength of the multilayer film as a target variable, a step (S607) of predicting the drop strength of the multilayer film using the trained machine learning model, and a step (S608) of combining predicted values ​​for other characteristics and predicted values ​​of the drop strength to satisfy the design requirements of the multilayer film. It may include a step (S608) of generating design data for the film.

[0126] For more specific details regarding the design method of a multilayer film, please refer to the description of the embodiments described in this specification and the disclosed feature configuration can be applied, so a redundant description is omitted here.

[0127] FIG. 7 is a drawing for explaining a computing device according to one embodiment.

[0128] Referring to FIG. 7, the design device and method of a multilayer film according to the embodiments can be implemented using a computing device (50). This computing device (50) can be implemented as various types of electronic devices, servers, or similar devices, and its functions can be implemented through a combination of software and hardware.

[0129] The computing device (50) may include at least one of a processor (510), a memory (530), a user interface input device (540), a user interface output device (550), and a storage device (560) that communicate via a bus (520). The computing device (50) may also include a network interface (570) electrically connected to a network (40). The network interface (570) may transmit or receive signals to or from other entities via the network (40).

[0130] The processor (510) may be implemented as various types of computing devices, such as an MCU (Micro Controller Unit), an AP (Application Processor), a CPU (Central Processing Unit), a GPU (Graphics Processing Unit), an NPU (Neural Processing Unit), a QPU (Quantum Processing Unit), etc. The processor (510) is a semiconductor device that executes instructions stored in a memory (530) or a storage device (560), and may play a core role in the system. The program code and data stored in the memory (530) or the storage device (560) instruct the processor (510) to perform a specific task, thereby enabling the overall operation of the system. The processor (510) may be configured to implement various functions and methods described above with reference to FIGS. 1 to 6.

[0131] The memory (530) and storage device (560) may include various types of volatile or non-volatile storage media for storing and accessing data of the system. For example, the memory (530) may include read-only memory (ROM) (531) and random access memory (RAM) (532). In some embodiments, the memory (530) may be built into the processor (510), in which case the data transfer speed between the memory (530) and the processor (510) may be very fast. In other embodiments, the memory (530) may be located external to the processor (510), in which case the memory (530) may be connected to the processor (510) through various data buses or interfaces. Such connection may be made through various means already known in the art, for example, a Peripheral Component Interconnect Express (PCIe) interface for high-speed data transfer or a memory controller.

[0132] In some embodiments, at least some components or functions of the device and method for designing a multilayer film according to the embodiments may be implemented as a program or software running on a computing device (50), and the program or software may be stored on a computer-readable medium. Specifically, the computer-readable medium according to one embodiment may be a program recorded thereon for causing a computer including a processor (510) that executes a program or instructions stored in a memory (530) or a storage device (560) to execute steps included in the device and method for designing a multilayer film according to the embodiments.

[0133] In some embodiments, at least some of the components or functions of the design devices and methods of multilayer films according to the embodiments may be implemented using hardware or circuitry of the computing device (50), or may be implemented as separate hardware or circuitry that can be electrically connected to the computing device (50).

[0134] According to embodiments, when the properties of the multilayer film itself, such as the final product, are particularly important, the drop strength of the multilayer film can be predicted with high accuracy even without information about the single-layer films that constitute the multilayer film. That is, by treating the multilayer film as a single layer and performing learning and prediction using selected elements from the collected property data as features, the drop strength of the multilayer film can be predicted more effectively for the analysis purpose while using fewer features than when based on the property data of the single-layer film that constitutes the multilayer film. Furthermore, by varying the number of indices selected as features when performance is prioritized and the number of indices selected as features when computing resources are saved, flexible design is possible according to the implementation purpose and environment of the design device and method for the multilayer film. Furthermore, by considering the prediction results regarding other characteristics of the multilayer film together with the prediction results of the drop strength, it is possible to design a multilayer film that can satisfy given design requirements.

[0135] Although the embodiments of the present invention have been described in detail above, the scope of the present invention is not limited thereto, and various modifications and improvements made by a person of ordinary skill in the art to which the present invention pertains using the basic concept of the present invention defined in the following claims also fall within the scope of the present invention.

Claims

1. A design device for a multilayer film that executes program code loaded in one or more memory devices through one or more processors to predict the properties of a multilayer film and design the multilayer film based on the predicted properties. The above program code is executed, The multilayer film to be designed is modeled as a single layer structure with a preset thickness, Collecting physical property data for films corresponding to the above single layer structure, The design device of the above multilayer film reads a value stored in advance in a storage space accessible to it, and obtains a feature setting mode for specifying a different feature setting method according to the read value. According to the above feature setting mode, feature setting is performed based on a plurality of physical indicators selected from the above physical property data, Select at least one machine learning model from among multiple supervised learning models capable of regression analysis, The machine learning model learned with the above features as independent variables and the drop strength of the multilayer film as a target variable is used to predict the drop strength of the multilayer film, In order to meet the design requirements of the multilayer film, design data for the multilayer film is generated by combining the predicted values ​​for other characteristics and the predicted values ​​for the drop strength. Design device for multilayer films.

2. In paragraph 1, If the above feature setting mode includes the first feature setting mode, performing the feature setting is: A design device for a multilayer film, comprising setting at least some of the yield stress, yield strain, necking stress, necking strain, breaking stress and breaking strain as features, and the thickness, selected from stress-strain data obtained through experiments on a multilayer film having the single layer structure.

3. In paragraph 2, Choosing the above machine learning model is, Including selecting one of the PLS (Partial Least Squares) model, the Lasso regression model, and the random forest model as the machine learning model, The machine learning model selected from the above PLS model, Lasso regression model, and random forest model is, A design device for a multilayer film, which receives at least some of the yield stress, the yield strain, the necking point stress, the necking point strain, the breaking stress and the breaking strain, and a vector generated from the thickness, and learns to predict the drop strength of the multilayer film.

4. In paragraph 1, If the above feature setting mode includes a second feature setting mode, performing the feature setting is: A design device for a multilayer film, comprising setting the thickness as a feature and calculating strain energy from at least some of the yield stress, yield strain, necking stress, necking strain, breaking stress and breaking strain selected from stress-strain data obtained through an experiment on a multilayer film having the single layer structure.

5. In paragraph 4, Choosing the above machine learning model is, Including selecting one of the PLS (Partial Least Squares) model, the Lasso regression model, and the random forest model as the machine learning model, The machine learning model selected from the above PLS model, Lasso regression model, and random forest model is, A design device for a multilayer film, which receives a vector generated from the above strain energy and the above thickness and learns to predict the drop strength of the multilayer film.

6. In paragraph 4, The above strain energy is the first strain energy (W yield ) and Performing the above feature settings is as follows: The first strain energy (W) yield ) is set as the above feature by calculating it according to the following mathematical formula, and a design device for a multilayer film: Here, σ y is the yield stress, and ε y is the yield strain, and σ br is the above-mentioned fracture stress, d is the joint diameter, t is the thickness of the measurement sample, and a is a constant determined experimentally.

7. In paragraph 4, The above strain energy is the second strain energy (W neck ) and Performing the above feature settings is as follows: The second strain energy (W) neck ) is set as the above feature by calculating it according to the following mathematical formula, and a design device for a multilayer film: Here, σ n is the stress at the necking point, and ε n is the strain at the necking point, and ε y is the yield strain, and σ br is the above-mentioned fracture stress, and σ y is the yield stress, d is the joint diameter, t is the thickness of the measurement sample, and b is a constant determined experimentally.

8. In paragraph 4, The above strain energy is the third strain energy (W str-hard ) and Performing the above feature settings is as follows: The third strain energy (W) str-hard ) is set as the above feature by calculating it according to the following mathematical formula, and a design device for a multilayer film: Here, ε br is the above fracture strain, and ε n is the strain at the necking point, and σ br is the above-mentioned fracture stress, and σ n is the necking point stress, L is the final stretching propagation length, t is the thickness of the measurement sample, and c is a constant determined experimentally.

9. In paragraph 4, The above strain energy is the fourth strain energy (W elast ) and Performing the above feature settings is as follows: The fourth strain energy (W) elast ) is set as the above feature by calculating it according to the following mathematical formula, and a design device for a multilayer film: Here, σ y is the yield stress, and ε y is the yield strain, d is the diameter of the joint, D is the diameter of the fixed disk, L is the final stretching propagation length, t is the thickness of the measurement sample, and e is a constant determined experimentally.

10. In paragraph 1, Generating design data for the above multilayer film is as follows: Combining the predicted value of the drop strength with the predicted value of at least one of the properties of stiffness, strength, strain, tear strength, high-speed drop strength, sealing initiation temperature (SIT), haze, moisture permeability, and air permeability for the multilayer film, If the multilayer film designed based on the above combination of predicted values ​​satisfies the above design requirements, the design data is generated based on the above combination, A design device for a multilayer film, comprising redesigning the multilayer film by changing the combination to another combination when the multilayer film designed based on the above combination of predicted values ​​does not satisfy the above design requirements.

11. A method for designing a multilayer film, which is performed by a computing device including one or more processors and one or more memory devices, for predicting the properties of a multilayer film and designing the multilayer film based on the predicted properties, A step of modeling the multilayer film to be designed as a single layer structure having a preset thickness; A step of collecting physical property data for a film corresponding to the single layer structure; A step of reading a value pre-stored in a storage space accessible to the design device of the multilayer film, and obtaining a feature setting mode for specifying a different feature setting method according to the read value; A step of performing feature setting based on a plurality of physical indicators selected from the physical property data according to the above feature setting mode; A step of selecting at least one machine learning model from among multiple supervised learning models capable of regression analysis; A step of predicting the fall strength of the multilayer film using the learned machine learning model with the above features as independent variables and the fall strength of the multilayer film as a target variable; and In order to meet the design requirements of the multilayer film, a step of generating design data for the multilayer film by combining predicted values ​​for other characteristics and predicted values ​​of the drop strength is included. Design method of multilayer films.

12. In paragraph 11, If the above feature setting mode includes the first feature setting mode, the step of performing the feature setting is: A method for designing a multilayer film, comprising a step of setting at least some of the yield stress, yield strain, necking stress, necking strain, breaking stress and breaking strain as features, and the thickness, selected from stress-strain data obtained through an experiment on a multilayer film having the single layer structure.

13. In paragraph 12, The steps for selecting the above machine learning model are: A step of selecting one of a PLS (Partial Least Squares) model, a Lasso regression model, and a random forest model as the machine learning model is included. The machine learning model selected from the above PLS model, Lasso regression model, and random forest model is, A design method for a multilayer film, wherein at least some of the yield stress, the yield strain, the necking point stress, the necking point strain, the breaking stress and the breaking strain, and a vector generated from the thickness are input, and the method is learned to predict the drop strength of the multilayer film.

14. In paragraph 11, If the above feature setting mode includes a second feature setting mode, the step of performing the feature setting is: A design method for a multilayer film, comprising a step of setting the thickness as a feature and calculating strain energy from at least some of the yield stress, yield strain, necking stress, necking strain, breaking stress and breaking strain selected from stress-strain data obtained through an experiment on a multilayer film having the single layer structure.

15. In paragraph 14, The steps for selecting the above machine learning model are: A step of selecting one of a PLS (Partial Least Squares) model, a Lasso regression model, and a random forest model as the machine learning model is included. The machine learning model selected from the above PLS model, Lasso regression model, and random forest model is, A design method for a multilayer film, wherein a vector generated from the above strain energy and the above thickness is input and learned to predict the drop strength of the multilayer film.

16. In paragraph 14, The above strain energy is the first strain energy (W yield ) and The steps for performing the above feature settings are: The first strain energy (W) yield ) is set as the feature by calculating it according to the following mathematical formula, and a design method for a multilayer film: Here, σ y is the yield stress, and ε y is the yield strain, and σ br is the above-mentioned fracture stress, d is the joint diameter, t is the thickness of the measurement sample, and a is a constant determined experimentally.

17. In paragraph 14, The above strain energy is the second strain energy (W neck ) and The steps for performing the above feature settings are: The second strain energy (W) neck ) is set as the feature by calculating it according to the following mathematical formula, and a design method for a multilayer film: Here, σ n is the stress at the necking point, and ε n is the strain at the necking point, and ε y is the yield strain, and σ br is the above-mentioned fracture stress, and σ y is the yield stress, d is the joint diameter, t is the thickness of the measurement sample, and b is a constant determined experimentally.

18. In paragraph 14, The above strain energy is the third strain energy (W str-hard ) and The steps for performing the above feature settings are: The third strain energy (W) str-hard ) is set as the feature by calculating it according to the following mathematical formula, and a design method for a multilayer film: Here, ε br is the above fracture strain, and ε n is the strain at the necking point, and σ br is the above-mentioned fracture stress, and σ n is the necking point stress, L is the final stretching propagation length, t is the thickness of the measurement sample, and c is a constant determined experimentally.

19. In paragraph 14, The above strain energy is the fourth strain energy (W elast ) and The steps for performing the above feature settings are: The fourth strain energy (W) elast ) is set as the feature by calculating it according to the following mathematical formula, and a design method for a multilayer film: Here, σ y is the yield stress, and ε y is the yield strain, d is the diameter of the joint, D is the diameter of the fixed disk, L is the final stretching propagation length, t is the thickness of the measurement sample, and e is a constant determined experimentally.

20. In paragraph 11, The step of generating design data for the above multilayer film is as follows: A step of combining a predicted value for at least one of the properties of stiffness, strength, strain, tear strength, high-speed drop strength, sealing initiation temperature (SIT), haze, moisture permeability, and air permeability for the multilayer film with a predicted value of the drop strength; If the multilayer film designed based on the above combination of predicted values ​​satisfies the above design requirements, a step of generating the design data based on the above combination; and A method for designing a multilayer film, comprising the step of redesigning the multilayer film by changing the combination to another combination, if the multilayer film designed based on the above combination of predicted values ​​does not satisfy the above design requirements.

Images