Estimation device, design assistance system, estimation method, and program

Abstract

This estimation device comprises: an input unit that receives an input of a target condition related to a physical property of a composition; a generation unit that generates design information of the composition; an estimation unit that estimates the physical property on the basis of the design information; and an output unit that outputs a score indicating the degree of conformity between an estimated value of the physical property and the target condition in association with the design information.

Description

Estimation Device, Design Support System, Estimation Method, and Program

[0008] , ,

[0001] The present disclosure relates to an estimation device, a design support system, an estimation method, and a program.

[0002] There is a technology for supporting the design of compositions. For example, Patent Document 1 discloses a design support device that proposes candidates for design condition information of a polymer that satisfies the required property information of the polymer and predicts the properties of a resin composition based on the design condition information of the resin composition containing the polymer.

[0003] International Publication No. 2023 / 120290

[0004] However, there is room for improving the efficiency of the design of compositions in the prior art. For example, if the result of evaluating the design information is shown together with the design information, the user can easily determine whether the proposed design information is promising.

[0005] One aspect of the present disclosure aims to improve the efficiency of the design of compositions.

[0006] An estimation device according to one aspect of the present disclosure includes an input unit that receives an input of target conditions regarding the physical properties of a composition, a generation unit that generates design information of the composition, an estimation unit that estimates physical properties based on the design information, and an output unit that outputs, in association with the design information, a score indicating the degree of conformity between the estimated value of the physical properties and the target conditions.

[0007] According to one aspect of the present disclosure, the design of compositions can be made more efficient.

[0008] FIG. 1 is a block diagram showing an example of the overall configuration of a design support system. FIG. 2 is a block diagram showing an example of the hardware configuration of a computer. FIG. 3 is a block diagram showing an example of the functional configuration of a design support system. FIG. 4 is a diagram showing an example of a forward estimation input screen. FIG. 5 is a diagram showing an example of a forward estimation output screen. FIG. 6 is a diagram showing an example of a reverse estimation input screen. FIG. 7 is a diagram showing an example of a reverse estimation output screen. FIG. 8 is a diagram showing an example of a reverse estimation detailed screen. FIG. 9 is a flowchart showing an example of a learning process. FIG. 10 is a sequence diagram showing an example of a forward estimation process. FIG. 11 is a sequence diagram showing an example of a reverse estimation process.

[0009] Hereinafter, embodiments of this disclosure will be described with reference to the accompanying drawings. In this specification and the drawings, components having substantially the same functional configuration are denoted by the same reference numerals, and redundant descriptions will be omitted.

[0010] [Embodiment] One embodiment of the present disclosure is an information processing system that supports the design of a composition. Hereinafter, the information processing system according to this embodiment will be referred to as the "design support system". In this embodiment, the composition may be a rubber composition. The rubber composition may be, for example, a special rubber, a general-purpose rubber, or a semi-general-purpose rubber.

[0011] Special rubber is a synthetic rubber with excellent liquid resistance and heat resistance. Examples of special rubber may include hydrogenated nitrile rubber, nitrile rubber, acrylic rubber, fluororubber, silicone rubber, epichlorohydrin rubber, etc.

[0012] General-purpose rubber or semi-general-purpose rubber is synthetic rubber that does not fall under the category of special rubber. General-purpose rubber may include, for example, styrene-butadiene rubber, butadiene rubber, isoprene rubber, etc. Semi-general-purpose rubber may include, for example, ethylene-propylene-diene rubber, chloroprene rubber, etc.

[0013] In this embodiment, the design support system has a function to estimate the physical properties of the rubber composition. The design support system also has a function to search for design information of the rubber composition based on the estimated physical properties. Hereinafter, the process of estimating the physical properties will be referred to as "forward estimation," and the process of searching for design information based on the estimated physical properties will be referred to as "backward estimation."

[0014] The design support system performs forward or backward estimation based on user specifications. By performing forward estimation, the user can obtain estimated values ​​for the physical properties of the rubber composition in the desired design information. Alternatively, by performing backward estimation, the user can obtain design information for a rubber composition whose physical properties meet target conditions. By repeatedly performing forward and backward estimation as desired, the user can proceed with the design of a rubber composition with excellent physical properties.

[0015] In this embodiment, the objective is to streamline the design of compositions. To this end, this embodiment accepts input of target conditions regarding the physical properties of the composition and outputs a score indicating the degree of conformity between the estimated values ​​of the physical properties and the target conditions, in association with the design information of the composition.

[0016] In one respect, according to this embodiment, the design information of the composition and the score that evaluates that design information are output in association, making it easy to determine whether the design information of the composition is promising or not. Therefore, according to this embodiment, the design of the composition can be made more efficient. In another respect, according to this embodiment, since the composition can be designed efficiently, compositions with excellent physical properties can be developed efficiently.

[0017] <Overall Configuration> The overall configuration of the design support system in this embodiment will be explained with reference to Figure 1. Figure 1 is a block diagram showing an example of the overall configuration of the design support system.

[0018] As shown in Figure 1, the design support system 1000 includes a learning device 10, an estimation device 20, and a terminal device 30. The learning device 10, the estimation device 20, and the terminal device 30 are connected via a communication network N such as a LAN (Local Area Network) or the Internet, enabling data communication.

[0019] The learning device 10 is an example of an information processing device that constructs a machine learning model for estimating the physical properties of a rubber composition. Hereinafter, the machine learning model constructed by the learning device 10 will also be referred to as the "physical property estimation model". The learning device 10 may be a computer such as a personal computer, workstation, or server.

[0020] The property prediction model is a machine learning model that has learned the relationship between the design information of the rubber composition and the physical properties of the rubber composition. Any type of machine learning model may be used for the property prediction model. Examples of machine learning models include Gaussian Process Regression (GPR), Partial Least Squares Regression (PLS), neural networks, random forests, or gradient boosting decision trees. Examples of gradient boosting decision trees include LightGBM (Light Gradient Boosting Machine) or XGBoost (eXtreme Gradient Boosting).

[0021] The estimation device 20 is an example of an information processing device that estimates the physical properties of a rubber composition based on a physical property estimation model. The estimation device 20 may be a computer such as a personal computer, workstation, or server. In addition to the physical properties of the composition, the estimation device 20 may also estimate numerical information such as price information of the composition or CO2 emission source unit price.

[0022] The physical properties of the rubber composition may include any physical properties. The physical properties of the rubber composition may include the unvulcanized and vulcanized properties of the rubber composition. The physical properties of the rubber composition may include at least one of the following properties: processability, mechanical properties, heat resistance, liquid resistance, compression set resistance, or cold resistance. Liquid resistance, for example, refers to resistance to various liquids including fuel oil (e.g., gasoline or similar fuels), lubricating oil (e.g., engine oil or similar engine oil), coolant (or similar to automobile antifreeze), automatic transmission fluid, solvents, aqueous solutions, etc.

[0023] Physical properties related to processability may include, for example, Mooney viscosity, Mooney scorch, or vulcanization rate. Physical properties related to mechanical properties may include, for example, hardness, elongation, or tensile strength. Physical properties related to heat resistance may include, for example, rate of change in hardness, rate of change in elongation, rate of change in tensile strength, or rate of change in volume. Physical properties related to liquid resistance may include, for example, rate of change in hardness, rate of change in elongation, rate of change in tensile strength, or rate of change in volume. Physical properties related to resistance to compression set may include, for example, compression set. Physical properties related to cold resistance may include, for example, the embrittlement temperature measured by the TR (Temperature retraction) test (e.g., TR10, TR70), the temperature measured by the Gehman torsion test (e.g., T5, T10), or the temperature measured by the low-temperature impact embrittlement test (e.g., 50% impact embrittlement temperature).

[0024] Terminal device 30 is an example of an information processing device operated by a user of the design support system 1000. Terminal device 30 may be a computer such as a personal computer, smartphone, or tablet terminal.

[0025] The terminal device 30 may display an input screen for inputting information used in forward or backward calculation. The terminal device 30 may also display an output screen for displaying the results of forward or backward calculation.

[0026] The overall configuration of the design support system 1000 shown in Figure 1 is just one example, and various system configurations are possible depending on the application and purpose. For example, one or more of the learning device 10, estimation device 20, and terminal device 30 may be included in multiple units of the design support system 1000. For example, the learning device 10 or estimation device 20 may be implemented by multiple computers, or as a cloud computing service. For example, the learning device 10 and estimation device 20 may be implemented by a single server device that combines the functions of both. For example, the design support system 1000 may be implemented by a standalone computer. The classification of devices such as the learning device 10, estimation device 20, and terminal device 30 shown in Figure 1 is just one example.

[0027] <Hardware Configuration> The hardware configuration of the design support system 1000 will be explained with reference to Figure 2. The learning device 10, estimation device 20, and terminal device 30 included in the design support system 1000 are implemented by, for example, a computer. Figure 2 is a block diagram showing an example of the computer's hardware configuration.

[0028] As shown in Figure 2, the computer 500 includes a CPU (Central Processing Unit) 501, ROM (Read Only Memory) 502, RAM (Random Access Memory) 503, HDD (Hard Disk Drive) 504, input device 505, display device 506, communication interface 507, and external interface 508. The CPU 501, ROM 502, and RAM 503 form what is known as a computer. Each piece of hardware in the computer 500 is interconnected via a bus line 509. The input device 505 and display device 506 may also be used by connecting them to the external interface 508.

[0029] The CPU 501 is an arithmetic unit that controls and implements the functions of the entire computer 500 by reading programs and data from a storage device such as the ROM 502 or HDD 504 onto the RAM 503 and executing processing.

[0030] ROM 502 is an example of a non-volatile semiconductor memory (storage device) that can retain programs and data even when the power is turned off. ROM 502 functions as the main memory, storing various programs and data necessary for the CPU 501 to execute the various programs installed on HDD 504. Specifically, ROM 502 stores boot programs such as BIOS (Basic Input Output System) and EFI (Extensible Firmware Interface) that are executed when the computer 500 starts up, as well as data such as OS (Operating System) settings and network settings.

[0031] RAM 503 is an example of volatile semiconductor memory (storage device) whose programs and data are erased when the power is turned off. RAM 503 can be, for example, DRAM (Dynamic Random Access Memory) or SRAM (Static Random Access Memory). RAM 503 provides a workspace that is expanded when various programs installed on HDD 504 are executed by CPU 501.

[0032] HDD 504 is an example of a non-volatile storage device that stores programs and data. The programs and data stored in HDD 504 include the OS, which is the basic software that controls the entire computer 500, and applications that provide various functions on the OS. Note that the computer 500 may use a storage device that uses flash memory as its storage medium (for example, an SSD: Solid State Drive) instead of HDD 504.

[0033] The input device 505 includes a touch panel used by the user to input various signals, operation keys and buttons, a keyboard and mouse, and a microphone for inputting sound data such as voice.

[0034] The display device 506 consists of a display such as a liquid crystal or organic EL (Electro-Luminescence) that displays a screen, and a speaker that outputs sound data such as audio.

[0035] Communication I / F 507 is an interface that connects to a communication network and allows the computer 500 to perform data communication.

[0036] External I / F 508 is an interface to external devices. External devices include drive devices 510, etc.

[0037] The drive device 510 is a device for setting the recording medium 511. The recording medium 511 here includes media that record information optically, electrically, or magnetically, such as CD-ROMs, flexible disks, and magneto-optical disks. The recording medium 511 may also include semiconductor memory that records information electrically, such as ROMs and flash memory. This allows the computer 500 to read and / or write to the recording medium 511 via the external interface 508.

[0038] The various programs to be installed on the HDD 504 are installed, for example, when the distributed recording medium 511 is set in a drive device 510 connected to an external I / F 508, and the various programs recorded on the recording medium 511 are read by the drive device 510. Alternatively, the various programs to be installed on the HDD 504 may be installed by downloading them via the communication I / F 507 from a network other than the communication network.

[0039] <Functional Configuration> The functional configuration of the design support system 1000 will be explained with reference to Figure 3. Figure 3 is a block diagram showing an example of the functional configuration of the design support system.

[0040] <<Learning Device>> As shown in Figure 3, the learning device 10 comprises a data storage unit 101, a preprocessing unit 110, a model generation unit 120, and a model output unit 130. The learning device 10 functions as the data storage unit 101, preprocessing unit 110, model generation unit 120, and model output unit 130 when a pre-installed learning program is executed.

[0041] For example, the data storage unit 101 is implemented by the HDD 504 shown in Figure 2. For example, the preprocessing unit 110, the model generation unit 120, and the model output unit 130 are implemented by a process in which a program loaded from the HDD 504 shown in Figure 2 onto the RAM 503 is executed by the CPU 501.

[0042] In the data storage unit 101, teacher data is stored in advance. The teacher data is electronic data used to construct a physical property estimation model. The teacher data may include design information of the rubber composition and correct values of the physical properties of the rubber composition.

[0043] The design information of the rubber composition may include the formulation of the rubber composition. The formulation of the rubber composition may include compounding agents, categories, and weight ratios.

[0044] The compounding agent is information indicating the substances contained in the rubber composition. The information indicating the substance may include, for example, substance name, product name, structural formula, SMILES (Simplified Molecular Input Line Entry System) information, ECFP (Extended Connectivity Circular Fingerprints) information, etc. Also, the information indicating the substance may include price information or CO2 emission per unit, etc.

[0045] The category is information indicating the type or role of the compounding agent. The category may include, for example, polymer, filler, plasticizer, antioxidant, processing aid, vulcanization aid, vulcanizing agent, vulcanization accelerator, vulcanization retarder, etc.

[0046] The weight ratio is information indicating the weight of the compounding agent in one unit of the rubber composition. The weight ratio may be expressed, for example, in phr (parts per hundred rubber).

[0047] The correct value of the physical property may be the measured value obtained by measuring the physical property using a sample of the rubber composition. When the correct value of the physical property is the measured value, the teacher data may further include test conditions. The test conditions are information indicating the test environment when the measured value of the physical property is measured. The test conditions may vary depending on the physical property. The test conditions may be test conditions related to an aging test or test conditions related to an unaged test. The test conditions related to the aging test may include, for example, aging medium, test temperature, and test time.

[0048] The data storage unit 101 may store a database having one or more tables for storing teacher data. The teacher data may be stored分散 in a plurality of tables included in the database. For example, the database may have a design information table and a test result table. The design information table may store the design information of the rubber composition. The test result table may store the test results using the rubber composition. The test results may include the measured values of physical properties and the test conditions.

[0049] The preprocessing unit 110 performs preprocessing on the teacher data. The preprocessing unit 110 may perform preprocessing on the teacher data read from the data storage unit 101. The preprocessing unit 110 may format the teacher data into a format suitable for constructing a physical property estimation model.

[0050] For example, the preprocessing unit 110 may combine the teacher data stored in a plurality of tables. For example, the preprocessing unit 110 may convert the categorical variables included in the design information into one-hot representation. For example, the preprocessing unit 110 may calculate the weighted average of numerical data. For example, the preprocessing unit 110 may delete the data related to the compounding agents with low usage rates.

[0051] The model generation unit 120 generates a physical property estimation model. The model generation unit 120 may generate a physical property estimation model based on the teacher data. The model generation unit 120 may generate a physical property estimation model based on the teacher data read from the data storage unit 101. The model generation unit 120 may generate a physical property estimation model based on the preprocessed teacher data by the preprocessing unit 110.

[0052] The model generation unit 120 may generate a learned physical property estimation model by learning the relationship between the design information and test conditions of the rubber composition and its physical properties. The model generation unit 120 may learn the relationship between the design information and test conditions and its physical properties based on supervised learning. The model generation unit 120 may learn the relationship between the design information and test conditions shown in the training data and the correct values ​​of the physical properties. For example, the model generation unit 120 may update the parameters of the physical property estimation model based on backpropagation. For example, the model generation unit 120 may update the parameters of the physical property estimation model so as to minimize the loss based on the output of the physical property estimation model when the design information and test conditions are input, and the correct values ​​shown in the training data.

[0053] The model output unit 130 outputs a trained physical property estimation model. The model output unit 130 may output a physical property estimation model generated by the model generation unit 120. The model output unit 130 may transmit the trained physical property estimation model to the estimation device 20. The trained physical property estimation model may be stored in the model storage unit 201 of the estimation device 20.

[0054] <Estimation Device> As shown in Figure 3, the estimation device 20 comprises a model storage unit 201, an input unit 210, a generation unit 220, an estimation unit 230, a calculation unit 240, a search unit 250, and an output unit 260. The estimation device 20 functions as the model storage unit 201, input unit 210, generation unit 220, estimation unit 230, calculation unit 240, search unit 250, and output unit 260 when a pre-installed estimation program is executed.

[0055] For example, the model storage unit 201 is implemented by the HDD 504 shown in Figure 2. For example, the input unit 210, generation unit 220, estimation unit 230, calculation unit 240, search unit 250, and output unit 260 are implemented by a process in which a program loaded from the HDD 504 shown in Figure 2 onto the RAM 503 is executed by the CPU 501.

[0056] The model storage unit 201 stores a pre-trained material property estimation model. The model storage unit 201 may also store a pre-trained material property estimation model. The material property estimation model stored in the model storage unit 201 may be generated by the learning device 10.

[0057] The input unit 210 accepts input of information to be used for forward or reverse calculation. The input unit 210 may accept input of design information of the rubber composition as information to be used for forward calculation. For example, the input unit 210 may present a list of compounds classified into multiple categories and accept input of the formulation of a rubber composition containing a compound selected from the list of compounds. The input unit 210 may further accept input of test conditions as information to be used for forward calculation. The input unit 210 may accept input of target conditions regarding the physical properties of the rubber composition as information to be used for reverse calculation. The target conditions may include information indicating the range of the physical properties. The information indicating the range may include, for example, information indicating upper and lower limits. The target conditions may also include information indicating test conditions.

[0058] The input unit 210 may output an input screen for inputting information. The input unit 210 may transmit screen data for displaying the input screen to the terminal device 30. The input unit 210 may output an input screen for inputting information used for forward calculation (hereinafter referred to as the "forward calculation input screen"). The input unit 210 may output an input screen for inputting information used for reverse calculation (hereinafter referred to as the "reverse calculation input screen"). The input unit 210 may output either the forward calculation input screen or the reverse calculation input screen based on the user's specification.

[0059] Screen data may be electronic data written in, for example, HTML (HyperText Markup Language). Screen data may also include applications written in, for example, JavaScript (registered trademark).

[0060] The generation unit 220 generates design information for a rubber composition. The generation unit 220 may generate multiple design information sets. The generation unit 220 may generate design information randomly. For example, the generation unit 220 may randomly select one or more compounding agents to be included in the rubber composition and randomly determine the weight ratio of each compounding agent. The generation unit 220 may generate design information to satisfy predetermined constraints. For example, the generation unit 220 may determine the weight ratio of each compounding agent between an upper limit and a lower limit so as to satisfy constraints that define an upper limit and a lower limit for each category of compounding agent.

[0061] The generation unit 220 may generate new design information based on the already generated design information. The generation unit 220 may also generate new design information by modifying a part of the already generated design information. The generation unit 220 may also generate new design information based on design information with a high score. The score may be calculated by the calculation unit 240.

[0062] The estimation unit 230 estimates the physical properties of the rubber composition. The estimation unit 230 may estimate the physical properties of the rubber composition based on the design information of the rubber composition. The estimation unit 230 may estimate the physical properties of the rubber composition based on the design information and test conditions of the rubber composition. The estimation unit 230 may estimate the physical properties of the rubber composition based on a physical property estimation model. The estimation unit 230 may estimate the physical properties of the rubber composition based on a learned physical property estimation model read from the model storage unit 201. For example, the estimation unit 230 may estimate the physical properties of the rubber composition by inputting the design information and test conditions of the rubber composition into the physical property estimation model. The estimation unit 230 may obtain the estimated values ​​of the physical properties output from the physical property estimation model.

[0063] In the forward estimation, the estimation unit 230 may estimate the physical properties of the rubber composition based on the design information received by the input unit 210. The estimation unit 230 may also estimate the physical properties of the rubber composition based on the test conditions received by the input unit 210.

[0064] In the reverse estimation, the estimation unit 230 may estimate the physical properties of the rubber composition based on the design information generated by the generation unit 220. The estimation unit 230 may estimate the physical properties of the rubber composition for each of the multiple design information generated by the generation unit 220. The estimation unit 230 may further estimate the physical properties of the rubber composition based on the test conditions indicated in the target conditions received by the input unit 210.

[0065] The calculation unit 240 calculates a score. The calculation unit 240 may calculate a score indicating the degree of fit between the estimated physical properties and the target conditions. The calculation unit 240 may calculate a score that shows the maximum value when the estimated physical properties satisfy the target conditions. The calculation unit 240 may calculate a score that shows a smaller value as the estimated physical properties deviate from the target conditions when they do not satisfy the target conditions. For example, if the estimated physical properties fall within the range indicated by the target conditions, the calculation unit 240 may calculate a score that shows the maximum value (e.g., 10 or 100). Also, if the estimated physical properties fall outside the range indicated by the target conditions, the calculation unit 240 may calculate a score by subtracting a value corresponding to the difference with the nearest boundary value from the maximum value.

[0066] The calculation unit 240 may calculate a score based on numerical information of the materials. The numerical information of the materials may be pre-stored in a storage device such as an HDD 504. The numerical information of the materials may include, for example, price information of the materials or CO2 emission intensity. The calculation unit 240 may calculate a numerical value converted for the entire blend based on the numerical information of the materials and use these values ​​as the score.

[0067] The calculation unit 240 may calculate multiple scores for each of the multiple target conditions. The calculation unit 240 may also calculate a total score by integrating the multiple scores. The total score may be a statistical value of the multiple scores. For example, the total score may be the mean, median, mode, sum, weighted sum, etc. of the multiple scores.

[0068] The search unit 250 searches for design information for rubber compositions. The search unit 250 may also search for design information to propose to the user. The search unit 250 may search for design information to propose to the user from a plurality of design information generated by the generation unit 220. The search unit 250 may also search for design information based on a score calculated by the calculation unit 240. The search unit 250 may also search for design information with a high score.

[0069] The output unit 260 outputs the result of forward or reverse calculation. The output unit 260 may transmit the result of forward or reverse calculation to the terminal device 30. The output unit 260 may display the result of forward or reverse calculation on the display device 506 of the calculation device 20. The output unit 260 may store the result of forward or reverse calculation in a storage device. The storage device may be, for example, the HDD 504 of the calculation device 20, or an external storage device connected to the calculation device 20.

[0070] The output unit 260 may output estimated values ​​of physical properties as a result of forward estimation. In other words, the output unit 260 may output estimated values ​​of physical properties when the input unit 210 receives input of design information and test conditions. The output unit 260 may also output estimated values ​​of physical properties estimated by the estimation unit 230 based on the design information and test conditions received by the input unit 210.

[0071] The output unit 260 may output design information as a result of inverse calculation. In other words, the output unit 260 may output design information when the input unit 210 receives input of target conditions. The output unit 260 may output design information associated with a score. The output unit 260 may output design information searched by the search unit 250. The output unit 260 may output one or more design information items with high scores. The output unit 260 may output design information items with scores equal to or greater than a predetermined value. The output unit 260 may output a predetermined number of design information items in descending order of scores. The output unit 260 may output multiple design information items in a predetermined order. For example, the output unit 260 may output multiple design information items in descending order of scores.

[0072] The output unit 260 may output an output screen for displaying the results. The output unit 260 may transmit screen data for displaying the output screen to the terminal device 30. The output unit 260 may output an output screen for displaying the results of forward estimation (hereinafter referred to as the "forward estimation calculation screen"). The output unit 260 may output an output screen for displaying the results of inverse estimation (hereinafter referred to as the "inverse estimation calculation screen").

[0073] <User Interface> The user interface of the design support system 1000 will be described with reference to Figures 4 to 8. The user interface of the design support system 1000 may include, as an example, a forward calculation input screen, a forward calculation power screen, a reverse calculation input screen, a reverse calculation power screen, and a reverse calculation details screen. The user interface of the design support system 1000 may also be displayed on the display device 506 of the terminal device 30, as an example. The forward calculation input screen is an example of the first screen. The reverse calculation input screen is an example of the second screen.

[0074] ≪Forward Estimation Input Screen≫ Figure 4 shows an example of a forward estimation input screen. As shown in Figure 4, the forward estimation input screen 600 may have a forward estimation start menu 601, a forward estimation result menu 602, a reverse estimation start menu 603, a reverse estimation result menu 604, a formulation input area 605, a test condition input area 606, a start button 607, a cancel button 608, etc.

[0075] The forward calculation start menu 601, the forward calculation result menu 602, the reverse calculation start menu 603, and the reverse calculation result menu 604 may be displayed in common on the forward calculation input screen, the forward calculation power screen, the reverse calculation input screen, and the reverse calculation power screen.

[0076] The forward calculation start menu 601 is a menu button for starting forward calculation. When the forward calculation start menu 601 is pressed, the terminal device 30 accepts the operation to start forward calculation. In response to the operation to start forward calculation, the terminal device 30 displays the forward calculation input screen 600 (see Figure 4) on the display device 506 of the terminal device 30.

[0077] The forward calculation result menu 602 is a menu button for displaying the results of the forward calculation. When the forward calculation result menu 602 is pressed, the terminal device 30 displays the forward calculation power screen 620 (see Figure 5) on the display device 506 of the terminal device 30.

[0078] The forward estimation calculation screen 620 may be automatically displayed on the display device 506 of the terminal device 30 when the forward estimation is completed. Alternatively, the forward estimation calculation screen 620 may be displayed on the display device 506 of the terminal device 30 when one result is selected from the list of forward estimation results.

[0079] The reverse calculation start menu 603 is a menu button for starting reverse calculation. When the reverse calculation start menu 603 is pressed, the terminal device 30 accepts the operation to start reverse calculation. In response to the operation to start reverse calculation, the terminal device 30 displays the reverse calculation input screen 640 (see Figure 6) on the display device 506 of the terminal device 30.

[0080] The reverse calculation result menu 604 is a menu button for displaying the reverse calculation result. When the reverse calculation result menu 604 is pressed, the terminal device 30 displays the reverse calculation calculation screen 660 (see Figure 7) on the display device 506 of the terminal device 30.

[0081] The inverse calculation power screen 660 may be automatically displayed on the display device 506 of the terminal device 30 when the inverse calculation is completed. Alternatively, the inverse calculation power screen 660 may be displayed on the display device 506 of the terminal device 30 when one result is selected from the list of inverse calculation results.

[0082] The mixture input area 605 may have a mixture input unit 605a and a mixture display unit 605b. The mixture input unit 605a is an area that receives input of design information used for forward estimation. The mixture display unit 605b is an area that displays the design information used for forward estimation.

[0083] The formulation input unit 605a may accept input of category, compounding agent, and weight ratio. The formulation input unit 605a may have a combo box that displays a predetermined category for selection. The formulation input unit 605a may have a combo box that displays a predetermined compounding agent for selection. The formulation input unit 605a may have a text box into which a numerical value to be used as the weight ratio can be entered. The formulation input unit 605a may have an add button that adds the entered category, compounding agent, and weight ratio to the formulation display unit 605b.

[0084] The formulation display section 605b may display a list of compounding agents to be included in the rubber composition. The formulation display section 605b may also display the category, compounding agent name, and weight ratio for each compounding agent.

[0085] The test condition input area 606 may have a test condition input unit 606a and a test condition display unit 606b. The test condition input unit 606a is an area that receives input of test conditions to be used for forward calculation. The test condition display unit 606b is an area that displays the test conditions to be used for forward calculation.

[0086] The test condition input unit 606a may accept input for test temperature and test time. The test condition input unit 606a may have a text box for accepting input for test temperature and test time for aging tests relating to heat resistance. The test condition input unit 606a may have an add button for adding test conditions for aging tests relating to heat resistance. The test condition input unit 606a may have a text box for accepting input for test temperature, test time, and type of test liquid for aging tests relating to liquid resistance. The test condition input unit 606a may have an add button for adding test conditions for aging tests relating to liquid resistance to the test condition display unit 606b.

[0087] The test condition display unit 606b may display multiple test conditions. For each of the multiple test conditions, the test condition display unit 606b may display the category (e.g., heat resistance, liquid resistance), test temperature, test time, and type of test liquid.

[0088] When the start button 607 is pressed, the forward calculation input screen 600 accepts the operation to perform the forward calculation. When the cancel button 608 is pressed, the forward calculation input screen 600 discards the information entered in the formulation input area 605 and the test condition input area 606, and closes the screen.

[0089] ≪Forward Estimation Calculation Screen≫ Figure 5 shows an example of a forward estimation calculation screen. As shown in Figure 5, the forward estimation calculation screen 620 may have a forward estimation start menu 601, a forward estimation result menu 602, a reverse estimation start menu 603, a reverse estimation result menu 604, a combination display area 621, an estimation result display area 622, etc.

[0090] The formulation display area 621 is an area for displaying the formulation used in the forward estimation. The formulation display area 621 may also display the category, compounding agent name, and weight ratio for each compounding agent included in the rubber composition. The formulation displayed in the formulation display area 621 is the same as the formulation displayed in the formulation display section 605b of the forward estimation input screen 600.

[0091] The estimation result display area 622 is an area for displaying the results of forward estimation. The estimation result display area 622 may have one or more physical property display units 623 (623a to 623e). The physical property display unit 623 displays the estimated value for each estimated physical property. For physical properties for which test conditions have been entered, the physical property display unit 623 displays the estimated value of the physical property for each test condition.

[0092] In the example shown in Figure 5, the physical property display unit 623a displays estimated values ​​for physical properties related to processability (e.g., Mooney viscosity). The physical property display unit 623b displays estimated values ​​for physical properties related to mechanical properties (e.g., hardness, elongation). The physical property display unit 623c displays estimated values ​​for physical properties related to heat resistance (e.g., hardness change rate, elongation change rate) under specified test conditions (e.g., test temperature, test time). The physical property display units 623d and 623e display estimated values ​​for multiple liquid resistance properties (e.g., hardness change rate, elongation change rate) under different test conditions (e.g., test temperature, test time, one or more types of test liquid).

[0093] ≪Inverse Estimation Input Screen≫ Figure 6 shows an example of an inverse estimation input screen. As shown in Figure 6, the inverse estimation input screen 640 may have a forward estimation start menu 601, a forward estimation result menu 602, an inverse estimation start menu 603, an inverse estimation result menu 604, a target condition input area 641, a start button 607, a cancel button 608, etc.

[0094] The target condition input area 641 is an area that accepts input of target conditions to be used for inverse calculation. The target condition input area 641 may have one or more target condition input units 642 (642a to 642d). The target condition input unit 642 accepts input of target conditions for each physical property to be calculated. The target condition input unit 642 may also accept input of test conditions for predetermined physical properties.

[0095] In the example shown in Figure 6, the target condition input unit 642a accepts input of target conditions related to material properties related to machinability (e.g., Mooney viscosity). The target conditions for Mooney viscosity may include lower and upper limits. The target condition input unit 642b accepts input of target conditions related to material properties related to mechanical properties (e.g., hardness, elongation). The target conditions for hardness may include lower and upper limits. The target conditions for elongation may include lower and upper limits.

[0096] The target condition input unit 642c accepts input of target conditions related to physical properties concerning heat resistance (e.g., hardness change rate, elongation change rate). The target condition input unit 642c also accepts input of test conditions related to heat resistance. The target conditions for hardness change rate may include lower and upper limits. The target conditions for elongation change rate may include lower and upper limits. The test conditions related to heat resistance may include test temperature and test time.

[0097] The target condition input unit 642d accepts input of target conditions related to physical properties (e.g., hardness change rate, elongation change rate) related to liquid resistance. The target condition input unit 642d also accepts input of test conditions related to liquid resistance. The target conditions related to hardness change rate may include lower and upper limits for the hardness change rate. The target conditions related to elongation change rate may include lower and upper limits for the elongation change rate. The test conditions related to liquid resistance may include test temperature, test time, and type of test liquid.

[0098] When the start button 607 is pressed, the reverse calculation input screen 640 accepts the operation to perform the reverse calculation. When the cancel button 608 is pressed, the reverse calculation input screen 640 discards the information entered in the target condition input area 641 and closes the screen.

[0099] ≪Inverse Estimation Power Screen≫ Figure 7 shows an example of the inverse estimation power screen. As shown in Figure 7, the inverse estimation power screen 660 may have a forward estimation start menu 601, a forward estimation result menu 602, an inverse estimation start menu 603, an inverse estimation result menu 604, one or more blend display areas 661 (661-1 to 661-4), a display order setting unit 662, etc.

[0100] The formulation display area 661 is an area for displaying the formulation proposed to the user. The formulation display area 661 may include a polymer display section 661a, a total score display section 661b, a score display section 661c, and a formulation display section 661d.

[0101] The polymer display unit 661a displays information indicating the polymers included in the formulation (e.g., substance name or product name). The total score display unit 661b displays the total score calculated based on the formulation. The score display unit 661c displays the scores calculated for each target condition. The formulation display unit 661d displays the details of the formulation (compounding agents and weight ratio).

[0102] The formula display area 661 may be displayed in a selectable format. If any of the formula display areas 661 is selected on the inverse calculation power screen 660, the system may transition from the inverse calculation power screen 660 to the inverse calculation details screen.

[0103] The display order setting unit 662 is a screen component that sets the display order of the formulation display area 661. The display order setting unit 662 may also be a combo box that displays a predetermined rule regarding the display order in a selectable format. Initially, the display order setting unit 662 may be set to descending order of total score. In other words, the inverse calculation power screen 660 may arrange the formulation display area 661 in a predetermined direction in descending order of total score. The display order setting unit 662 may also display ascending or descending order of a specific score as an option. As an example, the display order setting unit 662 may display descending order of processability score. In other words, the inverse calculation power screen 660 may arrange the formulation display area 661 in a predetermined direction in descending order of processability score. When the rule regarding the display order is changed in the display order setting unit 662, the inverse calculation power screen 660 may rearrange the formulation display area 661 according to the changed rule.

[0104] ≪Inverse Estimation Details Screen≫ Figure 8 shows an example of the inverse estimation details screen. As shown in Figure 8, the inverse estimation details screen 680 may have a polymer display area 681, a total score display area 682, a formulation display area 683, a score display area 684, an estimated value display area 685, and a recalculation button 686, etc.

[0105] The polymer display area 681 displays information (e.g., substance name or product name) indicating the polymers included in the formulation selected on the inverse power estimation screen 660. The total score display area 682 displays the total score calculated based on the formulation selected on the inverse power estimation screen 660. The formulation display area 683 displays the details of the formulation selected on the inverse power estimation screen 660 (compounding agents and weight ratio). The score display area 684 displays the scores calculated for each target condition for the formulation selected on the inverse power estimation screen 660. The estimated value display area 685 displays the estimated values ​​for each physical property calculated based on the formulation selected on the inverse power estimation screen 660.

[0106] When the recalculation button 686 is pressed, the screen transitions from the reverse calculation details screen 680 to the forward calculation input screen 600. At this time, the formula displayed in the formula display area 683 may be automatically entered into the formula input area 605 of the forward calculation input screen 600. In addition, the test conditions entered in the target condition input area 641 of the reverse calculation input screen 640 may be automatically entered into the test condition input area 606 of the forward calculation input screen 600.

[0107] On the forward estimation input screen 600, which is accessed from the reverse estimation details screen 680, the automatically entered formulas may be edited. In other words, the forward estimation input screen 600 may accept input of new formulas based on the formulas found by reverse estimation. Furthermore, on the forward estimation input screen 600, the automatically entered test conditions may be edited. In other words, the forward estimation input screen 600 may accept input of new test conditions based on the test conditions used in reverse estimation. Note that the new formulas accepted by the forward estimation input screen 600 are an example of the second design information.

[0108] <Processing Procedure> The design support method executed by the design support system 1000 will be explained with reference to Figures 9 to 11. The design support method may include a learning process (see Figure 9), a forward estimation process (see Figure 10), and a reverse estimation process (see Figure 11).

[0109] ≪Learning Process≫ Figure 9 is a sequence diagram showing an example of the learning process. The learning process is the process of constructing a material property prediction model. The learning process may be performed by the learning device 10.

[0110] In step S1, the preprocessing unit 110 of the learning device 10 reads the training data from the data storage unit 101. The preprocessing unit 110 performs preprocessing on the read training data. The preprocessing unit 110 sends the preprocessed training data to the model generation unit 120.

[0111] In step S2, the model generation unit 120 of the learning device 10 receives training data from the preprocessing unit 110. The model generation unit 120 obtains one or more design information and test conditions from the training data. The model generation unit 120 inputs the acquired design information and test conditions into the physical property estimation model. The physical property estimation model estimates the physical properties based on the input design information and test conditions and outputs the estimated values. The model generation unit 120 obtains the estimated values ​​of the physical properties output from the physical property estimation model.

[0112] In step S3, the model generation unit 120 of the learning device 10 obtains the correct values ​​of the physical properties corresponding to the design information obtained from the training data in step S2. The model generation unit 120 calculates the loss based on the estimated values ​​of the physical properties obtained in step S2 and the correct values ​​of the physical properties obtained from the training data. For example, the model generation unit 120 may calculate the error between the estimated values ​​of the physical properties and the correct values ​​of the physical properties.

[0113] In step S4, the model generation unit 120 of the learning device 10 updates the parameters of the material property estimation model based on the loss calculated in step S3. The model generation unit 120 may update the parameters of the material property estimation model based on the backpropagation method so as to minimize the loss based on the estimated values ​​of the material properties and the correct values ​​of the material properties.

[0114] In step S5, the model generation unit 120 of the learning device 10 determines whether or not learning is complete. Whether or not learning is complete can be determined by whether or not the model parameters have converged, or whether or not a predetermined number of iterations has been exceeded, etc.

[0115] If learning is complete (YES), the model generation unit 120 sends the learned physical property estimation model to the model output unit 130 and proceeds to step S6. On the other hand, if learning is not complete (NO), the model generation unit 120 returns to step S2. After that, the model generation unit 120 obtains different design information and test conditions from the training data and executes the processes from step S2 to step S5 again. In this way, the model generation unit 120 repeatedly executes the processes from step S2 to step S5 until it determines in step S5 that learning is complete.

[0116] In step S6, the model output unit 130 of the learning device 10 receives the learned physical property estimation model from the model generation unit 120. The model output unit 130 transmits the learned physical property estimation model to the estimation device 20. The estimation device 20 receives the learned physical property estimation model from the learning device 10. The estimation device 20 stores the learned physical property estimation model in the model storage unit 201.

[0117] <Forward Estimation Process> Figure 10 is a sequence diagram showing an example of the forward estimation process. The forward estimation process is a process that estimates the physical properties of a rubber composition based on the design information. The forward estimation process may be performed by the estimation device 20.

[0118] In step S11, the user of the design support system 1000 performs a forward calculation start operation on the terminal device 30. The forward calculation start operation may be, for example, pressing the forward calculation start menu 601 on the screen displayed on the display device 506 of the terminal device 30. In response to the forward calculation start operation, the terminal device 30 sends a request to the calculation device 20 to acquire the forward calculation input screen 600.

[0119] The estimation device 20 receives a request from the terminal device 30 to acquire the forward estimation input screen 600. The input unit 210 of the estimation device 20 transmits the screen data of the forward estimation input screen 600 to the terminal device 30. The terminal device 30 receives the screen data of the forward estimation input screen 600 from the estimation device 20. Based on the screen data of the forward estimation input screen 600, the terminal device 30 displays the forward estimation input screen 600 on the display device 506.

[0120] In step S12, the user of the design support system 1000 inputs design information and test conditions into the forward calculation input screen 600. The terminal device 30 receives the design information and test conditions entered into the forward calculation input screen 600.

[0121] In step S13, the user of the design support system 1000 performs a forward calculation operation on the terminal device 30. The forward calculation operation may be, for example, pressing the start button 607 on the forward calculation input screen 600. In response to the forward calculation operation, the terminal device 30 sends a forward calculation execution request to the calculation device 20. The forward calculation execution request includes the design information and test conditions received in step S12.

[0122] In step S14, the estimation device 20 receives a request to perform a forward estimation from the terminal device 30. The input unit 210 of the estimation device 20 obtains design information and test conditions from the forward estimation execution request received by the estimation device 20. The input unit 210 sends the obtained design information and test conditions to the estimation unit 230.

[0123] In step S15, the estimation unit 230 of the estimation device 20 receives design information and test conditions from the input unit 210. The estimation unit 230 reads the physical property estimation model from the model storage unit 201. The estimation unit 230 inputs the design information and test conditions into the physical property estimation model. The physical property estimation model estimates the physical properties based on the input design information and test conditions and outputs the estimated values. The estimation unit 230 acquires the estimated values ​​of the physical properties output by the physical property estimation model. The estimation unit 230 sends the estimated values ​​of the physical properties to the output unit 260.

[0124] In step S16, the output unit 260 of the estimation device 20 receives estimated values ​​of physical properties from the estimation unit 230. The output unit 260 generates screen data for the forward estimation calculation screen 620. The output unit 260 embeds the forward estimation results into the screen data of the forward estimation calculation screen 620. The forward estimation results may include, for example, design information and estimated values ​​of physical properties. The output unit 260 transmits the screen data of the forward estimation calculation screen 620 to the terminal device 30.

[0125] In step S17, the terminal device 30 receives screen data of the forward estimation power screen 620 from the estimation device 20. The terminal device 30 receives screen data of the forward estimation power screen 620 from the estimation device 20. Based on the screen data of the forward estimation power screen 620, the terminal device 30 displays the forward estimation power screen 620 on the display device 506.

[0126] Users of the design support system 1000 can obtain forward estimation results by referring to the forward estimation power screen 620 displayed on the display device 506 of the terminal device 30. Users may use the forward estimation results in the design and development of rubber compositions. For example, a user may input new design information or test conditions into the forward estimation input screen 600 and perform forward estimation again. Alternatively, for example, a user may prototype a rubber composition based on the design information for which good estimation values ​​were obtained through forward estimation.

[0127] <<Inverse Estimation Process>> Figure 11 is a sequence diagram showing an example of the inverse estimation process. The inverse estimation process is a process that searches for design information of the rubber composition based on the estimated values ​​of the physical properties. The inverse estimation process may be performed by the estimation device 20.

[0128] In step S21, the user of the design support system 1000 performs an operation to start the reverse calculation on the terminal device 30. The operation to start the reverse calculation may be, for example, by pressing the reverse calculation start menu 603 on the screen displayed on the display device 506 of the terminal device 30. In response to the operation to start the reverse calculation, the terminal device 30 sends a request to the calculation device 20 to acquire the reverse calculation input screen 640.

[0129] The estimation device 20 receives a request from the terminal device 30 to acquire the reverse estimation input screen 640. The input unit 210 of the estimation device 20 transmits the screen data of the reverse estimation input screen 640 to the terminal device 30. The terminal device 30 receives the screen data of the reverse estimation input screen 640 from the estimation device 20. Based on the screen data of the reverse estimation input screen 640, the terminal device 30 displays the reverse estimation input screen 640 on the display device 506.

[0130] In step S22, the user of the design support system 1000 enters the target conditions into the reverse calculation input screen 640. The terminal device 30 receives the target conditions entered into the reverse calculation input screen 640.

[0131] In step S23, the user of the design support system 1000 performs an inverse calculation operation on the terminal device 30. The inverse calculation operation may be, for example, pressing the start button 607 on the inverse calculation input screen 640. In response to the inverse calculation operation, the terminal device 30 sends an inverse calculation execution request to the calculation device 20. The forward calculation execution request includes the target conditions received in step S22.

[0132] In step S24, the estimation device 20 receives a request to perform reverse estimation from the terminal device 30. The input unit 210 of the estimation device 20 obtains the target conditions from the reverse estimation request received by the estimation device 20. The input unit 210 sends the obtained target conditions to the calculation unit 240. The input unit 210 also sends the test conditions included in the target conditions to the estimation unit 230.

[0133] In step S25, the generation unit 220 of the estimation device 20 randomly generates one or more design information. The generation unit 220 sends the generated design information to the estimation unit 230.

[0134] In step S26, the estimation unit 230 of the estimation device 20 receives test conditions from the input unit 210. The estimation unit 230 also receives design information from the generation unit 220. The estimation unit 230 reads the physical property estimation model from the model storage unit 201. The estimation unit 230 inputs the design information and test conditions into the physical property estimation model. The physical property estimation model estimates the physical properties based on the input design information and test conditions and outputs the estimated values. The estimation unit 230 acquires the estimated values ​​of the physical properties output by the physical property estimation model. The estimation unit 230 sends the estimated values ​​of the physical properties to the calculation unit 240.

[0135] In step S27, the calculation unit 240 of the estimation device 20 receives the target conditions from the input unit 210. The calculation unit 240 also receives estimated values ​​of physical properties from the estimation unit 230. Based on the estimated values ​​of physical properties, the calculation unit 240 calculates a score for each target condition. Based on the scores calculated for each target condition, the calculation unit 240 calculates a total score. The calculation unit 240 sends the design information and scores (including the total score) to the search unit 250.

[0136] In step S28, the search unit 250 of the estimation device 20 receives design information and a score from the calculation unit 240. The search unit 250 stores the design information and the score in association with each other in a storage device. The calculation unit 240 may store the design information and the score in association with each other if the total score is equal to or greater than a predetermined threshold.

[0137] In step S29, the search unit 250 of the estimation device 20 determines whether or not to terminate the search. Whether or not to terminate the search can be determined by whether or not a predetermined number of design information has been stored in the storage device, whether or not a predetermined number of iterations has been exceeded, or whether or not design information with a score of a predetermined value or higher has been obtained. If the search is to be terminated (YES), the search unit 250 proceeds to step S30. On the other hand, if the search is not to be terminated (NO), the search unit 250 returns to step S25.

[0138] When the process returns to step S25, the generation unit 220 of the estimation device 20 generates new design information. At this time, the generation unit 220 may generate new design information based on the already generated design information. The generation unit 220 may also generate new design information based on the design information stored in the memory device (in other words, the design information with a high score). For example, the generation unit 220 may generate new design information by randomly changing at least one of some of the compounding agents or weight ratios among the design information stored in the memory device.

[0139] Subsequently, the estimation device 20 repeats the processes from step S25 to step S29 based on the new design information. At this time, in step S28, if there is design information with a higher score than the design information stored in the memory device, the search unit 250 may associate the design information with the score and store it. In this way, the estimation device 20 repeatedly executes the processes from step S25 to step S29 until it determines in step S29 that the search is finished.

[0140] In step S30, the output unit 260 of the estimation device 20 reads the design information and score from the storage device. The output unit 260 generates screen data for the inverse estimation power screen 660. The output unit 260 embeds the inverse estimation result into the screen data of the inverse estimation power screen 660. The inverse estimation result may include, for example, the design information and score read from the storage device. The output unit 260 transmits the screen data of the inverse estimation power screen 660 to the terminal device 30.

[0141] In step S31, the terminal device 30 receives screen data of the inverse calculation power screen 660 from the calculation device 20. The terminal device 30 receives screen data of the inverse calculation power screen 660 from the calculation device 20. Based on the screen data of the inverse calculation power screen 660, the terminal device 30 displays the inverse calculation power screen 660 on the display device 506.

[0142] Users of the design support system 1000 can obtain inverse calculation results by referring to the inverse calculation power screen 660 displayed on the display device 506 of the terminal device 30. Users may use the inverse calculation results for the design and development of rubber compositions. For example, a user may input new target conditions into the inverse calculation input screen 640 and perform the inverse calculation again. Alternatively, for example, a user may prototype a rubber composition based on the design information with a high score among the design information displayed on the inverse calculation power screen 660. Alternatively, for example, a user may input the design information displayed on the inverse calculation power screen 660 into the forward calculation input screen 600 and perform a forward calculation.

[0143] <Effects of the Embodiment> The estimation device 20 according to one embodiment of the present disclosure receives input of target conditions relating to the physical properties of a composition, generates design information of the composition, estimates the physical properties based on the design information, and outputs a score indicating the degree of conformity between the estimated physical property values ​​and the target conditions, in association with the design information.

[0144] In one respect, according to this embodiment, the design information of the composition and the score that evaluates that design information are output in association, making it easy to determine whether the design information of the composition is promising or not. Therefore, according to this embodiment, the design of the composition can be made more efficient. In another respect, according to this embodiment, since the composition can be designed efficiently, compositions with excellent physical properties can be developed efficiently.

[0145] The estimation device 20 may generate multiple design information sets. The estimation device 20 may search for the design information set with the highest score among the multiple design information sets. In one aspect, according to this embodiment, the estimated physical properties of the design information set can be output from among the multiple design information sets in which the estimated values ​​best match the target conditions.

[0146] The estimation device 20 may accept input of multiple target conditions. The estimation device 20 may output multiple scores for each of the multiple target conditions in association with the design information. In one aspect, according to this embodiment, since scores corresponding to the target conditions are output in association with the design information, the design information can be evaluated from various perspectives.

[0147] The score may show a maximum value when the estimated value meets the target conditions, and a smaller value as the estimated value deviates further from the target conditions. In one aspect, according to this embodiment, the degree to which the design information conforms to the target conditions can be easily understood.

[0148] The estimation device 20 may further accept input of design information and test conditions. The estimation device 20 may estimate physical properties based on the input design information and test conditions. When the estimation device 20 accepts input of design information and test conditions, it may output estimated values ​​of physical properties. In one aspect, according to this embodiment, the physical properties of the rubber composition can be estimated according to the test conditions.

[0149] The design information may include the composition of the compounds. The estimation device 20 may present a list of compounds classified into multiple categories and accept input of a composition including a compound selected from the list. In one aspect, according to this embodiment, the user can intuitively input the design information.

[0150] The estimation device 20 may accept input of second design information based on the output design information. In one respect, according to this embodiment, the physical properties of new design information similar to promising design information can be easily estimated.

[0151] The estimation device 20 may output a first screen for inputting target conditions, or a second screen for inputting design information and test conditions, based on the user's specifications. In one aspect, according to this embodiment, the user can selectively perform forward estimation or reverse estimation as needed.

[0152] The composition may include a rubber composition. In one respect, according to this embodiment, the rubber composition can be designed efficiently.

[0153] [Supplement] Each function of the embodiments described above can be realized by one or more processing circuits. Hereinafter, "processing circuit" in this specification includes processors programmed to execute each function by software, such as CPUs (Central Processing Units) or GPUs (Graphics Processing Units) implemented by electronic circuits, as well as devices such as ASICs (Application Specific Integrated Circuits), DSPs (Digital Signal Processors), FPGAs (Field Programmable Gate Arrays), and conventional circuit modules designed to execute each function described above.

[0154] While embodiments of the present disclosure have been described in detail above, the embodiments disclosed herein are illustrative and not restrictive in all respects. The embodiments can be modified and improved in various ways without departing from the scope and spirit of the appended claims. The features described in the above embodiments can be combined in any way that is not inconsistent with other configurations.

[0155] Furthermore, the following forms are possible for disclosure technology.

[0156] (Note 1) An estimation device comprising: an input unit configured to accept input of target conditions relating to the physical properties of a composition; a generation unit configured to generate design information of the composition; an estimation unit configured to estimate the physical properties based on the design information; and an output unit configured to output a score indicating the degree of conformity between the estimated values ​​of the physical properties and the target conditions, in association with the design information.

[0157] (Note 2) The estimation device according to Note 1, wherein the generation unit is configured to generate a plurality of the design information, and further comprises a search unit configured to search for the design information with a high score among the plurality of the design information.

[0158] (Note 3) The estimation device according to Note 1 or 2, wherein the input unit is configured to accept input of a plurality of target conditions, and the output unit is configured to output a plurality of scores relating to each of the plurality of target conditions in association with the design information.

[0159] (Note 4) The estimation device according to any one of Notes 1 to 3, wherein the score shows a maximum value when the estimated value satisfies the target condition, and a smaller value as the estimated value deviates further from the target condition.

[0160] (Note 5) The estimation device according to any one of Notes 1 to 4, wherein the input unit is configured to further accept input of the design information and test conditions, the estimation unit is configured to estimate the physical properties based on the design information and test conditions input to the input unit, and the output unit is configured to output the estimated values ​​of the physical properties when the input unit accepts input of the design information and test conditions.

[0161] (Note 6) The estimation device according to Note 5, wherein the design information includes the formulation of the composition, and the input unit is configured to present a list of compounds classified into multiple categories and to accept input of the formulation including the compound selected from the list.

[0162] (Note 7) The estimation device according to Note 5 or 6, wherein the input unit is configured to receive input of second design information based on the design information output by the output unit.

[0163] (Note 8) The estimation device according to any one of Notes 5 to 7, wherein the input unit is configured to output a first screen for inputting the target conditions, or a second screen for inputting the design information and the test conditions, based on the user's specifications.

[0164] (Note 9) The estimation device according to any one of Notes 1 to 8, wherein the composition comprises a rubber composition.

[0165] (Note 10) A design support system in which a terminal device operated by a user and an estimation device can communicate via a network, wherein the estimation device comprises: an input unit configured to receive target conditions relating to the physical properties of a composition from the terminal device; a generation unit configured to generate design information of the composition; an estimation unit configured to estimate the physical properties based on the design information; and an output unit configured to transmit to the terminal device a score indicating the degree of conformity between the estimated values ​​of the physical properties and the target conditions in association with the design information, and the terminal device comprises a display device configured to display an input screen for inputting the target conditions and an output screen that displays the design information and the score in association.

[0166] (Note 11) The design support system as described in Note 10, wherein the input unit is configured to further accept input of the design information and test conditions, the output unit is configured to output estimated values ​​of the physical properties based on the design information and test conditions input to the input unit when the input unit has accepted input of the design information and test conditions, and the display device is configured to display a first screen for inputting the target conditions or a second screen for inputting the design information and test conditions, based on the user's specification.

[0167] (Note 12) An estimation method comprising: a computer receiving input of target conditions relating to the physical properties of a composition; generating design information for the composition; estimating the physical properties based on the design information; and outputting a score indicating the degree of conformity between the estimated values ​​of the physical properties and the target conditions, in association with the design information.

[0168] (Note 13) A program for causing a computer to perform the following processes: receiving input of target conditions relating to the physical properties of a composition; generating design information for the composition; estimating the physical properties based on the design information; and outputting a score indicating the degree of conformity between the estimated values ​​of the physical properties and the target conditions, in association with the design information.

[0169] This application claims priority to Japanese Patent Application No. 2024-231026, filed with the Japan Patent Office on 26 December 2024, which is incorporated herein by reference to its entire contents.

[0170] 10: Learning device 20: Estimation device 30: Terminal device 101: Data storage unit 110: Preprocessing unit 120: Model generation unit 130: Model output unit 201: Model storage unit 210: Input unit 220: Generation unit 230: Estimation unit 240: Calculation unit 250: Search unit 260: Output unit 1000: Design support system

Claims

1. An estimation device comprising: an input unit configured to accept input of target conditions relating to the physical properties of a composition; a generation unit configured to generate design information of the composition; an estimation unit configured to estimate the physical properties based on the design information; and an output unit configured to output a score indicating the degree of conformity between the estimated values ​​of the physical properties and the target conditions, in association with the design information.

2. The estimation device according to claim 1, wherein the generation unit is configured to generate a plurality of the design information, and further comprises a search unit configured to search for the design information with a high score among the plurality of the design information.

3. The estimation device according to claim 1 or 2, wherein the input unit is configured to accept input of a plurality of target conditions, and the output unit is configured to output a plurality of scores relating to each of the plurality of target conditions in association with the design information.

4. The estimation device according to any one of claims 1 to 3, wherein the score shows a maximum value when the estimated value satisfies the target condition, and shows a smaller value as the estimated value deviates further from the target condition.

5. The estimation device according to any one of claims 1 to 4, wherein the input unit is configured to further accept input of the design information and test conditions, the estimation unit is configured to estimate the physical properties based on the design information and test conditions input to the input unit, and the output unit is configured to output the estimated values ​​of the physical properties when the input unit has accepted input of the design information and test conditions.

6. The estimation device according to claim 5, wherein the design information includes the formulation of the composition, and the input unit is configured to present a list of compounds classified into a plurality of categories and to accept input of the formulation including the compound selected from the list.

7. The estimation device according to claim 5 or 6, wherein the input unit is configured to receive input of second design information based on the design information output by the output unit.

8. The estimation device according to any one of claims 5 to 7, wherein the input unit is configured to output a first screen for inputting the target conditions, or a second screen for inputting the design information and the test conditions, based on the user's specifications.

9. The estimation apparatus according to any one of claims 1 to 8, wherein the composition comprises a rubber composition.

10. A design support system in which a terminal device operated by a user and an estimation device can communicate via a network, wherein the estimation device comprises: an input unit configured to receive target conditions relating to the physical properties of a composition from the terminal device; a generation unit configured to generate design information of the composition; an estimation unit configured to estimate the physical properties based on the design information; and an output unit configured to transmit to the terminal device a score indicating the degree of conformity between the estimated values ​​of the physical properties and the target conditions, in association with the design information, and the terminal device comprises a display device configured to display an input screen for inputting the target conditions and an output screen that displays the design information and the score in association.

11. The design support system according to claim 10, wherein the input unit is configured to further accept input of the design information and test conditions, the output unit is configured to output estimated values ​​of the physical properties based on the design information and test conditions input to the input unit when the input unit has accepted input of the design information and test conditions, and the display device is configured to display a first screen for inputting the target conditions or a second screen for inputting the design information and test conditions, based on the user's specification.

12. An estimation method comprising: a computer receiving input of target conditions relating to the physical properties of a composition; generating design information for the composition; estimating the physical properties based on the design information; and outputting a score indicating the degree of conformity between the estimated values ​​of the physical properties and the target conditions, in association with the design information.

13. A program for causing a computer to perform the following processes: receiving input of target conditions relating to the physical properties of a composition; generating design information for the composition; estimating the physical properties based on the design information; and outputting a score indicating the degree of conformity between the estimated values ​​of the physical properties and the target conditions, in association with the design information.

Images