System for design of kirigami pattern, method for design of kirigami pattern using thereof
Patent Information
- Application Number
- KR1020250030390
- Authority / Receiving Office
- KR · KR
- Patent Type
- Applications
- Current Assignee / Owner
- Filing Date
- 2025-03-10
- Publication Date
- 2026-09-17
Smart Images
Figure PAT00005_ABST
Abstract
Description
Technology Field
[0001] The present invention relates to a kirigami pattern design system and a kirigami pattern design method using the same, and more specifically, to a kirigami pattern design system and a kirigami pattern design method using the same that enables the design of kirigami patterns simply and efficiently. Background Technology
[0002] A kirigami pattern can be applied to a piezoelectric film and used in sensors or energy harvesters.
[0003] Kirigami patterns (structures or shapes) allow for vertical deformation through a repetitive cut structure and possess a specific structure or arrangement formed by joints that connect the cut sections (patterns) to each other. Kirigami patterns can take various forms, such as octagons, squares, stars, square fillets, and circles.
[0004] Applying a kirigami pattern can improve the stretchability, response speed, and sensitivity of the sensor. The kirigami structure is structurally more flexible than planar sensors and can enhance mechanical responsiveness by utilizing deformation. Since the kirigami structure has the characteristic of quickly returning to its original state when deformed, the speed at which the sensor detects changes and outputs results is increased.
[0005] However, in designing sensors with applied kirigami patterns, achieving the desired flexibility, response speed, and sensitivity requires complex simulation and optimization processes, and the design process may involve significant trial and error and time. Prior art literature
[0006] Korean Published Patent No. 10-2023-0140535 (Title of Invention: Methods and Mechanisms for Measuring Patterned Substrate Properties During Substrate Manufacturing, Publication Date: Oct. 06, 2023) The problem to be solved
[0007] The objective of the present invention is to solve these conventional problems by providing a kirigami pattern design system that improves the efficiency of sensor design by enabling the easy acquisition of design parameters for a kirigami pattern suitable for sensor performance, and a kirigami pattern design method using the same.
[0008] The problems of the present invention are not limited to those mentioned above, and other problems not mentioned will be clearly understood by a person skilled in the art from the description below. means of solving the problem
[0009] To solve the above-mentioned problem, a kirigami pattern design system according to one aspect of the present invention may be characterized by comprising: a learning model generation unit that generates a learning model by training an artificial neural network with design parameters of a learning kirigami pattern as input values and performance of a learning sensor composed of a kirigami pattern as output values; and a design parameter acquisition unit that acquires design parameters of a kirigami pattern corresponding to input data related to the performance of the sensor based on the learning data of the learning model.
[0010] Design parameters may include one or more of the number of patterns in the kirigami pattern, the number of connecting parts that connect the patterns to each other, and the width of the connecting parts.
[0011] The kirigami pattern has a square shape, the number of patterns ranges from 2 to 6, the number of connecting parts is 2 or 4, and the width of the connecting part can be at least 0.6 mm.
[0012] The learning model generation unit learns one or more of the elasticity, response speed, and sensitivity of the learning sensor, which are measured according to the deformation of the learning sensor caused by an external force, as output values representing the performance of the learning sensor, and the design parameter acquisition unit can acquire design parameters of a kirigami pattern corresponding to the input data based on the learning data of the learning model when data regarding one or more of the elasticity, response speed, and sensitivity required for sensor design and manufacturing is input.
[0013] The system further includes a simulation unit that performs finite element analysis to measure the amount of voltage generated due to stress caused by deformation of the learning sensor to detect displacement, and calculates one or more of elasticity, response speed, and sensitivity representing the performance of the learning sensor based on the detected data. The learning model generation unit performs learning on a plurality of learning models and compares the learning data of the plurality of learning models with the finite element analysis data to select the learning model with the smallest error rate.
[0014] The learning model may include one or more of an Artificial Neural Network (ANN) and Bayesian Ridge Regression models.
[0015] The learning model generation unit further performs the task of generating a learning model by training an artificial neural network with the performance of a learning sensor to which the kirigami complex pattern is applied as an output value, and the design parameters including the number of patterns, the number of connections, and the width of connections of each kirigami center pattern and kirigami surrounding pattern of a learning kirigami complex pattern including a kirigami center pattern and a kirigami surrounding pattern are input values, and the design parameter acquisition unit can acquire design parameters of a kirigami complex pattern corresponding to the input data based on the training data of the learning model when data regarding one or more of elasticity, response speed, and sensitivity related to the performance of the sensor is input.
[0016] A method for designing a kirigami pattern according to another aspect of the present invention may include a learning model generation step of training an artificial neural network with design parameters of a learning kirigami pattern as input values and performance of a learning sensor composed of a kirigami pattern as output values to generate a learning model, and a design parameter acquisition step of acquiring design parameters of a kirigami pattern corresponding to input data related to the performance of the sensor based on the learning data of the learning model.
[0017] In the design parameter acquisition step, the design parameters may include one or more of the number of patterns of the kirigami pattern, the number of connecting parts that connect the patterns to each other, and the width of the connecting parts.
[0018] The learning model generation step includes a process of training one or more of the elasticity, response speed, and sensitivity of the learning sensor, measured according to the deformation of the learning sensor caused by an external force, as output values representing the performance of the learning sensor, and the design parameter acquisition step may include a process of acquiring design parameters of a kirigami pattern corresponding to the input data based on the learning data of the learning model when data regarding one or more of the elasticity, response speed, and sensitivity required for sensor design and manufacturing is input.
[0019] The learning model generation step may include the process of performing learning on multiple learning models, performing finite element analysis to measure the amount of voltage generated due to stress caused by deformation of the learning sensor to detect displacement, calculating one or more of elasticity, response speed, and sensitivity representing the performance of the learning sensor based on the detected data, and comparing the learning data of multiple learning models with the finite element analysis data to select the learning model with the smallest error rate.
[0020] In the learning model generation step, the method further includes a process of generating a learning model by training an artificial neural network with design parameters including the arrangement method of the kirigami center pattern and the surrounding kirigami peripheral pattern constituting the learning kirigami complex pattern, the number of patterns, the number of connections, and the width of the connections of each kirigami center pattern and kirigami peripheral pattern as input values, and the performance of the learning sensor to which the kirigami complex pattern is applied as output values. In the design parameter acquisition step, when data regarding one or more of elasticity, response speed, and sensitivity related to the performance of the sensor is input, the method may include a process of acquiring design parameters of the kirigami complex pattern corresponding to the input data based on the training data of the learning model. Effects of the invention
[0021] According to the kirigami pattern design system and the kirigami pattern design method using the same of the present invention, the following effects are achieved.
[0022] First, by enabling the easy acquisition of design parameters for a Kirigami pattern that matches the sensor performance, the efficiency of sensor design can be increased.
[0023] Second, design parameters of a kirigami pattern that meet one or more of the desired sensor elasticity, response speed, and sensitivity can be obtained without the need to repeat complex simulations and experiments.
[0024] Third, reliable design parameters for the Kirigami pattern can be obtained based on the training data of the learning model with the smallest error rate compared with the finite element analysis data.
[0025] The effects of the present invention are not limited to those mentioned above, and other unmentioned effects will be clearly understood by a person skilled in the art from the description in the claims. Brief explanation of the drawing
[0026] FIG. 1 is a block diagram showing a kirigami pattern design system according to one embodiment of the present invention. FIG. 2 illustrates examples of various forms of a kirigami pattern according to one embodiment of the present invention. FIG. 3 shows the design parameters of a kirigami pattern according to one embodiment of the present invention. Figure 4 illustrates the process of generating a learning model using a Kirigami pattern design system according to one embodiment of the present invention. FIG. 5 illustrates the process of obtaining design parameters using a Kirigami pattern design system according to one embodiment of the present invention. Figure 6 shows the results of finite element analysis of the design parameters of Figure 5. FIGS. 7 and FIGS. 8 show a kirigami composite pattern according to one embodiment of the present invention. FIG. 9 shows various examples of a kirigami composite pattern according to one embodiment of the present invention. Figure 10 is a flowchart illustrating a method for designing a kirigami pattern using the kirigami pattern design system of Figure 1. Specific details for implementing the invention
[0027] The advantages and features of the present invention and the methods for achieving them will become clear by referring to the embodiments described below in detail together with the accompanying drawings. However, the present invention is not limited to the embodiments disclosed below but may be implemented in various different forms. These embodiments are provided merely to ensure that the disclosure of the present invention is complete and to fully inform those skilled in the art of the scope of the invention, and the present invention is defined only by the scope of the claims. Throughout the specification, the same reference numerals refer to the same components.
[0028] The size or shape of components depicted in the drawings attached to this specification may be exaggerated for clarity and convenience of explanation. It should be noted that identical components in each drawing may be depicted with the same reference numeral. Furthermore, detailed descriptions of functions and configurations of known technology that are deemed to unnecessarily obscure the essence of the invention may be omitted.
[0029] The terms used herein are for describing specific embodiments and are not intended to limit the invention. As used herein, the singular form may include the plural form unless the context clearly indicates otherwise. Furthermore, throughout this specification, when a part is described as "comprising" a certain component, it means that it may include additional components unless specifically stated otherwise.
[0030] When it is stated that one component is "connected" or "connected" to another component, it should be understood that while it may be directly connected or connected to that other component, there may also be other components in between. Conversely, when it is stated that one component is "directly connected" or "directly connected" to another component, it should be understood that there are no other components in between. Other expressions used to describe relationships between components should be interpreted in the same way.
[0031] Terms such as top, bottom, upper surface, lower surface, or upper, lower, used in this specification are used to distinguish relative positions among components. For example, while the upper part of a drawing may be designated as the upper part and the lower part as the lower part for convenience, in practice, the upper part may be designated as the lower part and the lower part as the upper part without departing from the scope of the present invention.
[0032] Terms including ordinal numbers, such as "the first," "the second," etc., as described in this specification may be used to describe various components, but said components are not limited by said terms. These terms are used merely to distinguish that each component is a different component and are not bound by the order of manufacture; furthermore, the names may not match between the detailed description of the invention and the claims.
[0033] All terms used herein, including technical or scientific terms, have the same meaning as generally understood by those skilled in the art to which the present invention pertains, unless otherwise defined. Terms such as those defined in commonly used dictionaries should be interpreted as having a meaning consistent with their meaning in the context of the relevant technology, and should not be interpreted in an ideal or overly formal sense unless explicitly defined in this specification.
[0034] The symbols attached to each step are used to identify each step and do not indicate the order of the steps relative to one another; the steps may be performed differently from the specified order unless a specific order is clearly indicated in the context.
[0035] Hereinafter, embodiments of the present invention will be described with reference to the accompanying drawings.
[0036] FIG. 1 is a block diagram of a kirigami pattern design system according to one embodiment of the present invention, and FIG. 2 illustrates examples of various forms of a kirigami pattern according to one embodiment of the present invention.
[0037] Referring to FIGS. 1 and 2, the Kirigami pattern design system (100) may include a learning model generation unit (110), a simulation unit (120), a design parameter acquisition unit (130), and a storage unit (140).
[0038] The kirigami pattern design system (100) according to an embodiment of the present invention allows the user to easily obtain design parameters of a kirigami pattern that match one or more of the elasticity, response speed, and sensitivity of the sensor desired by the user, without the need to repeat complex simulations and experiments.
[0039] Applying a kirigami pattern can increase one or more of the sensor's stretchability, response speed, and sensitivity.
[0040] The kirigami pattern (structure or shape) can have various pattern forms (also called shapes or forms), such as octagons, squares, square fillets, stars, and circles. The shape of such a kirigami pattern can affect the elasticity, response speed, and sensitivity of the sensor.
[0041] Specifically, when an octagonal kirigami pattern is applied to a sensor, multi-directional deformation is possible, which can increase elasticity.
[0042] When a square-shaped kirigami pattern is applied to a sensor, the cut portion is straight, allowing it to easily stretch in a consistent direction and facilitating deformation control.
[0043] When a square fillet-shaped kirigami pattern is applied to a sensor, the rounded corners reduce stress concentration and enable uniform deformation.
[0044] When a star-shaped kirigami pattern is applied to a sensor, the cuts are made in multiple directions, making it highly flexible and allowing it to be easily deformed even by a small force.
[0045] When a circular kirigami pattern is applied to a sensor, uniform elasticity in all directions is possible, making multi-directional deformation easy.
[0046] In addition, non-linear patterns, rather than symmetric patterns, can be applied to Kirigami patterns. While symmetric patterns generally induce uniform deformation, the characteristics of the sensor may differ when non-linear patterns are applied.
[0047] For example, in the case of curved patterns within non-linear patterns, more flexible deformation is possible compared to straight cuts, and torsion and multi-directional deformation can be easily achieved. Furthermore, fractal patterns can maximize sensitivity by responding sensitively to even small forces, and since multi-stage deformation is possible, they may be suitable for sensors that detect various deformation conditions.
[0048] In the following embodiments of the present invention, the kirigami pattern is described as a symmetrical pattern in the shape of a square, but is not limited thereto, and the kirigami pattern design system (100) can be applied to various types of kirigami patterns.
[0049] Kirigami patterns utilize fine cut and fold structures for elasticity, allowing sensors to stretch more. While conventional sensors can easily tear or experience performance degradation, Kirigami patterns are stretchable and can return to their original state even after deformation. Kirigami patterns can be used in wearable sensors, soft robots, e-skin, and the like.
[0050] Kirigami patterns induce stress concentration through a cut structure in terms of sensitivity and increase the surface area of the sensor, enabling the detection of even finer changes. Because they can respond sensitively to very small forces or deformations, more precise detection is possible. This can be utilized in electronic skin sensors that respond to low pressure, biomedical sensors (heartbeat detection), etc.
[0051] The Kirigami pattern design system (100) may include a learning model generation unit (110), a simulation unit (120), a design parameter acquisition unit (130), and a storage unit (140).
[0052] The learning model generation unit (110) generates a learning model by training an artificial neural network using design parameters of a learning kirigami pattern as input values and performance of a learning sensor composed of a learning kirigami pattern as output values. The learning model generation unit (110) can generate a learning model by learning the relationship between the design parameters of the learning kirigami pattern and the performance of the learning sensor composed of the learning kirigami pattern.
[0053] FIG. 3 shows the design parameters of a kirigami pattern according to one embodiment of the present invention.
[0054] Referring to FIG. 3, the design parameters may include one or more of the number of patterns (101a) of the kirigami pattern, the number of joints (101b) connecting the patterns (101a) to each other, and the width of the joints (101c).
[0055] For example, the number of patterns may range from 2 to 6, the number of connections may be 2 or 4, and the width of the connections may range from 0.6 mm to 2.2 mm.
[0056] The number of patterns (101a) is the number of cut patterns of the kirigami pattern, and the number of connecting parts (101b) can be defined as the number of joints connecting the patterns.
[0057] The number of patterns (101a), the number of connecting parts (101b), and the width of the connecting part (101c) may vary depending on the overall size of the pattern.
[0058] The kirigami pattern (structure or shape) can have various pattern forms, such as octagons, squares, stars, square fillets, and circles. A kirigami pattern can be applied to a piezoelectric film (101) and used in a sensor or energy harvester.
[0059] In a piezoelectric film (101) with a kirigami pattern applied, if the cut portion (region) of the kirigami pattern is stretched, a stress focal point may occur at the connection portion (101b).
[0060] As the piezoelectric film (101) stretches and stress increases at the connection part (101b), the change in the voltage signal (charge generation amount) may be large, and a specific peak may appear in the sensor signal depending on the stress concentration point.
[0061] The voltage output of the piezoelectric film (101) can rise in a parabolic shape as the output increases rapidly as the stress increases. When the piezoelectric film (101) is stretched, it shows an upward section in the parabolic graph due to the increase in signal, and can reach a signal peak at the maximum deformation point (maximum stress). Then, when it is released, it shows a downward section in the parabolic graph due to the decrease in signal, and can return to its original state and recover to the initial value.
[0062] The learning model generation unit (110) can train one or more of the elasticity, response speed, and sensitivity of the learning sensor, which are measured according to the deformation of the learning sensor caused by an external force, into output values representing the performance of the learning sensor corresponding to the input values.
[0063] The learning model generation unit (110) can perform learning on a plurality of learning models. The learning models may include, for example, an Artificial Neural Network (ANN) and a Bayesian Ridge Regression model.
[0064] The simulation unit (120) performs finite element analysis (FEA) to measure the amount of voltage generated due to stress caused by deformation of the learning sensor composed of a learning kirigami pattern, detects displacement, and can calculate one or more of elasticity, response speed, and sensitivity representing the performance of the learning sensor based on the detected data.
[0065] For example, experimental samples can be prepared by cutting a PVDF (Polyvinylidene Fluoride) film, and it is possible to verify whether voltage is generated upon mechanical deformation through a compression test.
[0066] In finite element analysis, the stress values that cause plastic deformation in a PVDF film with a kirigami pattern can be determined. Additionally, the model's data set can be configured based on various design parameters such as the number of patterns, the number of connections, and the width of the connections. Furthermore, the tensile length of each model can be selected so that the stress does not exceed the set value.
[0067] The simulation unit (120) can mathematically simulate how stress, displacement, voltage, etc. change inside the sensor when the sensor of a complex structure is deformed by a physical force (e.g., tension, compression, bending, etc.) in finite element analysis. In the case of a piezoelectric sensor, the amount of voltage generated (amount generated) according to deformation can be calculated.
[0068] The simulation unit (120) can perform stress analysis when the sensor is deformed. When the sensor receives an external force, stress is concentrated in specific parts (e.g., connecting parts of the kirigami pattern, cut areas, etc.), and the points of stress concentration and their magnitude can be measured through finite element analysis. The deformation state of the sensor can be quantitatively analyzed through the stress distribution inside the sensor.
[0069] The simulation unit (120) can perform voltage measurement in the sensor due to stress. The piezoelectric sensor generates voltage when subjected to stress. Through finite element analysis, it is possible to quantitatively analyze how much voltage is generated at a specific deformation. Through this, a change pattern of the output voltage signal according to the degree of deformation can be derived.
[0070] The simulation unit (120) can detect the displacement of the sensor. Displacement is a value indicating how much the sensor has moved or deformed from its original position. Through simulation, it is possible to calculate how much the sensor stretches or bends when a specific force is applied. In the case of a piezoelectric sensor, since a larger displacement results in a larger change in voltage, the displacement can be estimated based on the output signal.
[0071] The simulation unit (120) can perform data-based performance evaluation. Performance evaluation is possible based on the voltage signal output by the sensor, and elasticity, response speed, and sensitivity can be calculated.
[0072] Specifically, the simulation unit (120) can measure how much the sensor can stretch (maximum strain, Strain Limit) when calculating elasticity. Through finite element analysis, it can calculate whether the sensor stretches by 100% or up to 200% when a specific force (e.g., 10N) is applied, and the higher the elasticity, the more suitable it is for use as a wearable sensor or a soft robot sensor.
[0073] The simulation unit (120) can measure how quickly the sensor responds to an external stimulus in calculating the response speed. The response speed can be calculated by analyzing how quickly the voltage signal changes after deformation is applied through simulation.
[0074] The simulation unit (120) can measure how small a force the sensor can respond to when calculating sensitivity. For example, if a voltage signal is detected even when a small force of 1 N is applied, it can be said to be a sensor with high sensitivity. The higher the sensitivity, the more advantageous it can be for electronic skin sensors that detect minute pressure or deformation.
[0075] The above-described learning model generation unit (110) can compare the learning data of a plurality of learning models with the above-described finite element analysis data to select the learning model with the smallest error rate.
[0076] Figure 4 illustrates the process of generating a learning model using a Kirigami pattern design system according to one embodiment of the present invention.
[0077] Referring to FIG. 4, the learning model generation unit (110) may include an input layer (112, Input Layer), a hidden layer (114, Hidden Layer), and an output layer (116, Output Layer).
[0078] In the learning process, the input layer (112) receives design parameters of a learning kirigami pattern as input values, including one or more of the number of patterns, the number of joints connecting the patterns to each other, and the width of the joints.
[0079] The hidden layer (114) may include multiple layers, for example, the first layer (114a) may have 64 neurons, the second layer (114b) may have 32 neurons, and the third layer (114c) may have 16 neurons.
[0080] The hidden layer (114) can learn the non-linear relationship between the input design parameters and the performance of the sensor (stretchability, response speed, sensitivity). That is, a process of analyzing the influence of the number of patterns, the number of connections, and the width of the connections on the sensor performance can be performed. For example, in a multi-layer perceptron (MLP) neural network, each neuron can learn the non-linear relationship by applying weights and activation functions. The hidden layer (114) can extract the relationship between the design parameters and strengthen the correlation with the performance of the sensor, such as its stretchability, response speed, and sensitivity. Generally, the non-linearity can be strengthened using a Rectified Linear Unit (ReLU) function. After passing through the hidden layer, features that have learned the optimal relationship between the design parameters and the sensor performance can be extracted.
[0081] The output layer (116) may have a single neuron and may provide a result value that predicts the optimal performance (stretchability, response speed, sensitivity, etc.) of the sensor as learned by the model.
[0082] For example, a neural network can be trained by iterating through the entire dataset 1,000 times. During the training process, optimization is performed in the direction of minimizing the loss function, and training can be completed when the neural network finally converges to a sufficiently low loss value.
[0083] The design parameter acquisition unit (130) acquires design parameters of a Kirigami pattern corresponding to input data related to the performance of the sensor based on the training data of the training model.
[0084] FIG. 5 illustrates the process of obtaining design parameters using a Kirigami pattern design system according to one embodiment of the present invention.
[0085] Referring to FIG. 5 (a), the design parameter acquisition unit (130) receives data for one or more of the elasticity, response speed, and sensitivity required for sensor performance when designing and manufacturing the sensor.
[0086] For example, a target displacement (stretchability) can be input within the range of 2.2 mm to 46.7 mm.
[0087] Referring to FIG. 5(b), the design parameter acquisition unit (130) can acquire design parameters corresponding to input data based on the training data of the above-described training model. Here, the number of acquired patterns ranges from 2 to 6, the number of connection parts is 2 or 4, and the width of the connection part can range from 0.6 mm to 2.2 mm.
[0088] For example, when the first input value is received as the target displacement value, the number of patterns (output value) can be 4, the number of connections is 2, and the width of the connection is 0.66 mm.
[0089] In addition, when the first input value is received as the target displacement value, the number of patterns is 6, the number of connections is 4, and the width of the connection is 1.03 mm as design parameters. In this case, the loss value may be approximately 0.31.
[0090] When a second input value that is nearly twice as large as the first input value described above is received as the target displacement value, the number of patterns is 5, the number of connections is 2, and the width of the connection is 0.73 mm as design parameters.
[0091] In addition, when a second input value is received as the target displacement value, the number of patterns is 4, the number of connections is 2, and the width of the connection is 1.57 mm as design parameters. In this case, the loss value may be approximately 0.27.
[0092] In this way, when multiple design parameters are obtained for a target displacement value, the design parameter suitable for the sensor design and purpose can be selected and used.
[0093] Figure 6 shows the results of finite element analysis of the design parameters of Figure 5.
[0094] Referring to FIG. 6, when a first input value is received as a target displacement value through an embodiment of the present invention as described above, the design parameter obtained through the learning model did not have an error rate of within 10% when compared with the displacement value detected based on the deformation of the sensor with the corresponding design parameter applied in the finite element analysis result.
[0095] Referring to Fig. 6(b), when a second input value that is nearly twice as large as the first input value described above is received as the target displacement value, the design parameter obtained through the learning model has an error rate of less than 10% when compared to the displacement value detected based on the deformation of the sensor with the corresponding design parameter applied in the finite element analysis result.
[0096] In this way, the kirigami pattern design system (100) according to an embodiment of the present invention can easily obtain design parameters of a kirigami pattern suitable for sensor performance, thereby increasing the efficiency of sensor design and maintaining a safety margin when considering the yield strength (20MPa to 50MPa) of the actual PVDF film.
[0097] Meanwhile, multiple kirigami patterns can be combined to generate a kirigami composite pattern, and a customized sensor can be designed.
[0098] FIGS. 7 and FIGS. 8 show a kirigami composite pattern according to one embodiment of the present invention.
[0099] Referring to FIGS. 7 and 8, for Kirigami pattern A (105), for example, the number of patterns is 2, the number of connecting parts is 4, and the width of the connecting part is 2 mm. For Kirigami pattern B (106), the number of patterns is 2, the number of connecting parts is 2, and the width of the connecting part is 0.62 mm.
[0100] Referring to FIG. 7, a kirigami pattern B (106) (kirigami center pattern) is located in the center, and a kirigami pattern A (105) (kirigami peripheral pattern) is arranged around it to create a kirigami composite pattern. In this kirigami composite pattern, the displacement sensitivity was approximately 20.7. Displacement sensitivity indicates how sensitively the sensor structure responds when deformed. In the voltage change rate graph, the voltage change rate showed a gradual appearance.
[0101] Referring to FIG. 8, a kirigami pattern A (105) is positioned in the center, and a kirigami pattern B (106) is arranged around it to create a kirigami composite pattern. In this kirigami composite pattern, the displacement sensitivity was approximately 4.5. In the voltage change rate graph, the voltage change rate appeared larger compared to the voltage change rate shown in FIG. 7.
[0102] In this way, for the design of a sensor having a kirigami complex pattern, a learning model can be generated by training an artificial neural network using the design parameters of the learning kirigami complex pattern as input values and the performance of the learning sensor composed of the kirigami complex pattern as output values.
[0103] Specifically, the learning model generation unit (110) can generate a learning model by training an artificial neural network using design parameters including the number of patterns, the number of connections, and the width of connections of each of the learning kirigami center pattern and the kirigami surrounding pattern of the learning kirigami complex pattern including the kirigami center pattern and the kirigami surrounding pattern as input values, and the performance of the learning sensor to which the kirigami complex pattern is applied as output values.
[0104] In this case, the design parameter acquisition unit (130) can acquire design parameters of a kirigami composite pattern corresponding to input data related to the performance of the sensor based on the training data of the training model. The design parameters of the kirigami composite pattern may include the number of patterns, the number of connecting parts, and the width of connecting parts for each kirigami center pattern and kirigami peripheral pattern.
[0105] FIG. 9 shows various examples of a kirigami composite pattern according to one embodiment of the present invention.
[0106] Referring to FIG. 9, not only the kirigami composite pattern described in FIG. 7 and FIG. 8 but also kirigami composite patterns of various structures can be formed. In FIG. 9, for convenience of explanation, the shape related to the number of patterns, the number of connecting parts, and the width of the connecting parts has been omitted.
[0107] Referring to FIG. 9 (a), a square-shaped first kirigami pattern (151, kirigami center pattern) is positioned in the center, a square-shaped second kirigami pattern (152, kirigami peripheral pattern) is arranged around the first kirigami pattern (151), and a square-shaped third kirigami pattern (153, kirigami peripheral pattern) is arranged around the second kirigami pattern (152) to form a kirigami composite pattern.
[0108] That is, the first kirigami pattern (151) in the center can be formed as the kirigami center pattern, and two or more kirigami peripheral patterns can be layered around the periphery to form a kirigami composite pattern.
[0109] In the case of a composite kirigami pattern in which kirigami patterns of the same shape are arranged in layers, the voltage change of the sensor may have the characteristic of starting most significantly at the central kirigami pattern and gradually spreading to the surrounding kirigami patterns. Additionally, thanks to the hierarchical structure, the durability of the sensor is increased, the response speed is more likely to be controlled more evenly, and overall, the sensitivity of the sensor is likely to vary depending on the location.
[0110] In this way, a kirigami complex pattern can be formed by a hierarchical combination of kirigami patterns in which multiple kirigami peripheral patterns surround a kirigami central pattern. In other examples, not only square-shaped kirigami patterns but also various shapes of kirigami patterns such as the aforementioned octagon, square fillet, star, and circle can be applied.
[0111] Referring to FIG. 9(b), a circular first kirigami pattern (161, kirigami center pattern) may be positioned in the center, and a square second kirigami pattern (163, kirigami peripheral pattern) may be arranged around the first kirigami pattern (161) to surround it. In this way, a kirigami composite pattern may be formed by a combination of different shapes of kirigami patterns.
[0112] A kirigami composite pattern having a combination of different types of kirigami patterns is a form in which multidirectional deformation of the kirigami central pattern and linear deformation of the kirigami peripheral pattern are combined, such that the circular kirigami central pattern can exhibit smooth and uniform voltage changes, and the square-shaped kirigami peripheral pattern can exhibit abrupt voltage changes at a specific deformation limit.
[0113] Referring to FIG. 9 (c), a kirigami composite pattern can be formed by a combination in which a square-shaped first kirigami pattern (171) and a second kirigami pattern (172) are arranged side by side. In this case, in other examples, not only square-shaped kirigami patterns but also various shapes of kirigami patterns such as the aforementioned octagon, square fillet, star, and circle may be applied, and three or more kirigami patterns may be arranged side by side.
[0114] A kirigami composite pattern having a combination of two or more kirigami patterns arranged side by side is likely to have increased overall elasticity because each kirigami pattern can be deformed independently, and the response speed of each kirigami pattern may operate individually. Additionally, sensitivity may vary depending on the spacing between kirigami patterns and the stress distribution. That is, if the distance between the first kirigami pattern (171) and the second kirigami pattern (172) is close, the two kirigami patterns (171, 172) may react simultaneously and increase sensitivity, and if they are placed far apart from each other, they deform individually, so the sensitivity may not be uniform.
[0115] Meanwhile, each component of the Kirigami pattern design system (100) according to an embodiment of the present invention may be provided in a modular form or implemented by one or more processors (not shown).
[0116] The processor may include a memory (not shown) that stores data for an algorithm or a program that reproduces the algorithm for controlling the operation of each component of the Kirigami pattern design system (100). In this case, the memory and the processor may each be implemented as separate chips or as a single chip. The processor may be a general-purpose processor such as a CPU (Central Processing Unit), AP (Application Processor), DSP (Digital Signal Processor), a graphics-dedicated processor such as a GPU (Graphics Processing Unit) or VPU (Vision Processing Unit), or an artificial intelligence-dedicated processor such as an NPU (Neural Processing Unit).
[0117] The processor can be controlled to process input data according to predefined operation rules or artificial intelligence models stored in memory. Alternatively, if the processor is an AI-dedicated processor, the AI-dedicated processor may be designed with a hardware structure specialized for processing a specific AI model. The predefined operation rules or AI models are characterized by being created through learning. Here, being created through learning means that a basic AI model is trained using multiple learning data by a learning algorithm, thereby creating predefined operation rules or AI models configured to perform desired characteristics (or objectives). Such learning may be performed on the device itself where the AI according to the present disclosure is executed, or it may be performed through a separate server and / or system. Examples of learning algorithms include supervised learning, unsupervised learning, semi-supervised learning, or reinforcement learning, but are not limited to the examples described above.
[0118] An artificial intelligence model may be composed of multiple neural network layers. Each of the multiple neural network layers has multiple weight values and performs neural network operations through calculations between the results of previous layers and the multiple weights. The multiple weights possessed by the multiple neural network layers can be optimized based on the learning results of the artificial intelligence model. For example, the multiple weights can be updated so that the loss or cost values obtained by the artificial intelligence model during the learning process are reduced or minimized.
[0119] Artificial neural networks may include deep neural networks (DNNs), such as Convolutional Neural Networks (CNNs), Deep Neural Networks (DNNs), Recurrent Neural Networks (RNNs), Restricted Boltzmann Machines (RBMs), Deep Belief Networks (DBNs), Bidirectional Recurrent Deep Neural Networks (BRDNNs), or Deep Q-Networks, but are not limited to the examples mentioned above.
[0120] The storage unit (140) can store the pattern shape (e.g., triangle, hexagon, grid, etc.), cutting length, number of cuts, cutting angle, pattern arrangement method, structural design variables, etc., as training data required for creating a learning model of the kirigami pattern design system (100).
[0121] Additionally, the storage unit (140) can store information regarding mechanical performance (strain, tensile strength, durability, etc.), electrical characteristics (rate of change in resistance, sensitivity, etc.) and environmental influences (change in performance due to temperature change), as output data related to sensor performance characteristics.
[0122] Additionally, the storage unit (140) can store information such as a dataset for evaluating the performance of a learned model, verification and test data, preprocessed data, model weights and checkpoints, hyperparameters and setting information.
[0123] The storage unit (140) may be implemented as at least one of a non-volatile memory device such as a cache, ROM (Read Only Memory), PROM (Programmable ROM), EPROM (Erasable Programmable ROM), EEPROM (Electrically Erasable Programmable ROM), and flash memory, a volatile memory device such as RAM (Random Access Memory), or a storage medium such as a hard disk drive (HDD) and CD-ROM, but is not limited thereto.
[0124] Figure 10 is a flowchart illustrating a method for designing a kirigami pattern using the kirigami pattern design system of Figure 1.
[0125] Referring to FIG. 10, a learning model generation step is performed to generate a learning model by training an artificial neural network with design parameters of a learning kirigami pattern as input values and performance of a learning sensor composed of a kirigami pattern as output values (S1001).
[0126] Here, the design parameters may include one or more of the number of patterns in the kirigami pattern, the number of connecting parts that connect the patterns to each other, and the width of the connecting parts.
[0127] The learning model generation step (S1001) may include a process of training one or more of the elasticity, response speed, and sensitivity of the learning sensor, measured according to the deformation of the learning sensor caused by an external force, into an output value representing the performance of the learning sensor corresponding to the input value.
[0128] The learning model generation step (S1001) can perform learning on multiple learning models.
[0129] The learning model generation step (S1001) may include the process of performing finite element analysis to measure the amount of voltage generated due to stress caused by deformation of the learning sensor to detect displacement, calculating one or more of elasticity, response speed, and sensitivity representing the performance of the learning sensor based on the detected data, and comparing the learning data of a plurality of learning models with the finite element analysis data to select the learning model with the lowest error rate.
[0130] In another embodiment, when performing learning on a kirigami composite pattern, the learning model generation step (S1001) can generate a learning model by training an artificial neural network using design parameters including the number of patterns, the number of connections, and the width of connections of each kirigami central pattern and kirigami surrounding pattern of a learning kirigami composite pattern including a kirigami central pattern and a kirigami surrounding pattern as input values, and the performance of a learning sensor to which the kirigami composite pattern is applied as output values.
[0131] Next, a design parameter acquisition step is performed to acquire design parameters of a kirigami pattern corresponding to input data related to sensor performance based on the training data of the training model (S1011).
[0132] The design parameter acquisition step (S1011) may include a process of acquiring design parameters of a kirigami pattern corresponding to the input data based on the training data of a training model when data regarding one or more of elasticity, response speed, and sensitivity related to the performance of a sensor to be designed or manufactured is input.
[0133] In another embodiment, when learning of the kirigami composite pattern is performed as described above, the design parameter acquisition step (S1011) can acquire design parameters of the kirigami composite pattern corresponding to the input data based on the training data of the learning model when data regarding one or more of elasticity, response speed, and sensitivity related to the performance of the sensor is input. The design parameters of the kirigami composite pattern may include the number of patterns, the number of connecting parts, and the width of the connecting parts of each kirigami center pattern and kirigami peripheral pattern.
[0134] In this way, the kirigami pattern design system (100) and the method according to the embodiment of the present invention can easily obtain design parameters of a kirigami pattern suitable for sensor performance, thereby increasing the efficiency of sensor design.
[0135] In addition, efficient sensor design and fabrication can be performed by obtaining design parameters of a kirigami pattern that match one or more of the elasticity, response speed, and sensitivity of the sensor desired by the user, without the need to repeat complex simulations and experiments.
[0136] Methods according to the various embodiments disclosed in this document may be provided as included in a computer program product. The computer program product may be traded between a seller and a buyer as a product. The computer program product may be distributed in the form of a device-readable storage medium (e.g., compact disc read-only memory (CD-ROM)), or distributed online (e.g., download or upload) through an application store (e.g., Play Store™) or directly between multiple user devices (e.g., smartphones). In the case of online distribution, at least a portion of the computer program product (e.g., downloadable app) may be temporarily stored or temporarily created on a device-readable storage medium, such as the memory of a manufacturer's server, an application store's server, or a relay server.
[0137] As described above, preferred embodiments of the present invention have been illustrated and described with reference to the drawings; however, the present invention is not limited to the specific embodiments described above. Various modifications are possible by those skilled in the art without departing from the essence of the invention as claimed in the patent claims, and such modifications should not be understood individually from the technical spirit or perspective of the present invention. Explanation of the symbols
[0138] 100: Kirigami Pattern Design System 110: Learning Model Generation Section 112: Input layer 114: Hidden layer 116: Output layer 120: Simulation section 130: Design Parameter Acquisition Section 140: Storage section
Claims
Claim 1 A kirigami pattern design system characterized by comprising: a learning model generation unit that generates a learning model by training an artificial neural network using design parameters of a learning kirigami pattern as input values and performance of a learning sensor formed by said kirigami pattern as output values; and a design parameter acquisition unit that acquires design parameters of said kirigami pattern corresponding to input data related to the performance of the sensor based on the learning data of said learning model. Claim 2 A kirigami pattern design system according to claim 1, wherein the learning model generation unit learns one or more of the elasticity, response speed, and sensitivity of the learning sensor measured according to the deformation of the learning sensor caused by an external force as output values representing the performance of the learning sensor, and the design parameter acquisition unit acquires design parameters of the kirigami pattern corresponding to the input data based on the learning data of the learning model when data regarding one or more of the elasticity, response speed, and sensitivity related to the performance of the sensor is input. Claim 3 A Kirigami pattern design system according to claim 2, further comprising a simulation unit that performs finite element analysis to measure the amount of voltage generated due to stress occurring according to deformation of the learning sensor to detect displacement, and calculates one or more of elasticity, response speed, and sensitivity representing the performance of the learning sensor based on the detected data, and the learning model generation unit performs the learning on a plurality of learning models and compares the learning data of the plurality of learning models with the finite element analysis data to select the learning model with the smallest error rate. Claim 4 In paragraph 3, the learning model is a Kirigami pattern design system comprising one or more of an Artificial Neural Network (ANN) and a Bayesian Ridge Regression model. Claim 5 A method for designing a kirigami pattern using a kirigami pattern design system according to any one of claims 1 to 4, comprising: a learning model generation step of training an artificial neural network to generate a learning model by taking design parameters of a learning kirigami pattern as input values and the performance of a learning sensor formed by said kirigami pattern as output values; and a design parameter acquisition step of acquiring design parameters of said kirigami pattern corresponding to input data related to the performance of the sensor based on the learning data of said learning model. Claim 6 In claim 5, the learning model generation step includes a process of learning one or more of the elasticity, response speed, and sensitivity of the learning sensor, measured according to the deformation of the learning sensor caused by an external force, as output values representing the performance of the learning sensor, and the design parameter acquisition step includes a process of acquiring design parameters of the kirigami pattern corresponding to the input data based on the learning data of the learning model when data regarding one or more of the elasticity, response speed, and sensitivity related to the performance of the sensor is input. Claim 7 In claim 5, the learning model generation step comprises the process of performing the learning on a plurality of learning models, performing finite element analysis to detect displacement by measuring the amount of voltage generated due to stress caused by deformation of the learning sensor, calculating one or more of elasticity, response speed, and sensitivity representing the performance of the learning sensor based on the detected data, and comparing the learning data of the plurality of learning models with the finite element analysis data to select the learning model with the smallest error rate.