Characteristic prediction device, characteristic prediction method, and program

Abstract

The relationship between the design value and the characteristic value of the target substance is visualized so as to be easily grasped. The present invention is provided with: a reference design determination unit configured to determine a reference design condition indicating a design condition in a target substance manufactured using two or more material substances; a peripheral design generation unit configured to generate a plurality of peripheral design conditions in which one design value varies with respect to a reference design condition; a characteristic prediction unit configured to predict two or more characteristic values in each of the peripheral design conditions; and a visualization unit configured so that a symbol corresponding to the peripheral design condition is arranged on a plane having the design value and the first characteristic value as axes and represented by a color corresponding to the second characteristic value.

Description

Technical Field

[0001] The present invention relates to a characteristic prediction device, a characteristic prediction method and a program. Background Art

[0002] In the past, material development was based on the experience and intuition of engineers. In recent years, some technologies that use machine learning and other methods to improve the efficiency of material development have been used.

[0003] For example, Patent Document 1 discloses an invention for predicting the physical properties of a composition composed of a plurality of raw materials and displaying the characteristic amount of a specific raw material contained in the composition in association with the predicted value of the physical property.

[0004] [Prior art literature]

[0005] [Patent Literature]

[0006] [Patent Document 1] Japanese Patent Application Publication No. 2022-128962 Summary of the invention

[0007] [Problems to be solved by the invention]

[0008] However, in the prior art, it is difficult to grasp the changing trends of multiple feature quantities or multiple physical property values. For example, in the invention disclosed in Patent Document 1, in order to grasp the changing trends of multiple feature quantities or multiple physical property values, it is necessary to display multiple coordinate systems representing the relationship between feature quantities and physical property values ​​for comparison.

[0009] In view of the above-mentioned technical problems, an object of the present disclosure is to visualize the relationship between the design value and the characteristic value of a target substance so that it can be easily grasped.

[0010]

Methods to solve the problem

[0011] The present disclosure includes the configurations shown below.

[0012] [1] A characteristic prediction device comprising:

[0013] A reference design determination unit configured to determine reference design conditions representing design conditions in a target substance manufactured using two or more material substances;

[0014] a peripheral design generating unit configured to generate a plurality of peripheral design conditions by changing one design value with respect to the reference design condition;

[0015] a characteristic prediction unit configured to predict two or more characteristic values ​​in each of the peripheral design conditions; and

[0016] The visualization unit is configured to arrange symbols corresponding to the peripheral design conditions in a color corresponding to the second characteristic value on a plane having the design value and the first characteristic value as axes.

[0017] [2] A characteristic prediction device comprising:

[0018] A reference design determination unit configured to determine reference design conditions representing design conditions in a target substance manufactured using two or more material substances;

[0019] a peripheral design generating unit configured to generate a plurality of peripheral design conditions by changing two design values ​​with respect to the reference design condition;

[0020] a characteristic prediction unit configured to predict one or more characteristic values ​​in each of the peripheral design conditions; and

[0021] The visualization unit is configured to arrange symbols corresponding to the peripheral design conditions on a plane having the first design value and the second design value as axes, by expressing the symbols in colors corresponding to the characteristic values.

[0022] [3] The characteristic prediction device described in [2] above, wherein:

[0023] The characteristic prediction unit is configured to predict two or more characteristic values.

[0024] The visualization unit is configured to arrange symbols formed by combining regions represented by color expressions corresponding to the respective characteristic values, on coordinates corresponding to the peripheral design conditions.

[0025] [4] The characteristic prediction device described in [3] above, wherein:

[0026] The respective regions of the symbol are formed in the same shape and are arranged in a point-symmetrical manner.

[0027] [5] The characteristic prediction device described in any one of [1] to [4] above, wherein:

[0028] The reference design determination unit is configured to determine, as the reference design condition, the peripheral design condition corresponding to the symbol selected from the symbols displayed by the visualization unit.

[0029] The visualization unit is configured to re-arrange, on the plane, the symbols corresponding to the peripheral design conditions generated based on the reference design conditions when the reference design conditions are re-selected.

[0030] [6] The characteristic prediction device described in any one of [1] to [4] above, wherein:

[0031] The reference design determination unit is configured to determine the reference design condition by editing the peripheral design condition corresponding to the symbol displayed by the visualization unit.

[0032] The visualization unit is configured to re-arrange, on the plane, the symbols corresponding to the peripheral design conditions generated based on the reference design conditions when the reference design conditions are re-selected.

[0033] [7] The characteristic prediction device described in any one of [1] to [6] above, wherein:

[0034] The peripheral design generating unit is configured to generate the peripheral design condition by changing the design value within a range of a predetermined upper limit value and a predetermined lower limit value and within a predetermined variation range.

[0035] [8] The characteristic prediction device described in any one of [1] to [7] above, wherein:

[0036] The characteristic prediction unit is configured to predict the characteristic value using a machine learning model that uses the design value as an explanatory variable and the characteristic value as a target variable.

[0037] [9] A characteristic prediction method, wherein a computer performs the following steps:

[0038] Determining reference design conditions representing design conditions in a target substance manufactured using two or more material substances;

[0039] generating a plurality of peripheral design conditions by changing one design value with respect to the reference design condition;

[0040] predicting two or more characteristic values ​​in each of the peripheral design conditions; and

[0041] On a plane having the design value and the first characteristic value as axes, symbols corresponding to the peripheral design conditions are arranged in a color corresponding to the second characteristic value.

[0042]

[10] A characteristic prediction method, wherein a computer executes the following steps:

[0043] Determining reference design conditions representing design conditions in a target substance manufactured using two or more material substances;

[0044] generating a plurality of peripheral design conditions by varying two design values ​​with respect to the reference design condition;

[0045] predicting one or more characteristic values ​​in each of the peripheral design conditions; and

[0046] On a plane with the first design value and the second design value as axes, symbols corresponding to the peripheral design conditions are arranged in colors corresponding to the characteristic values.

[0047]

[11] A program that causes a computer to execute the following steps:

[0048] Determining reference design conditions representing design conditions in a target substance manufactured using two or more material substances;

[0049] generating a plurality of peripheral design conditions by changing one design value with respect to the reference design condition;

[0050] predicting two or more characteristic values ​​in each of the peripheral design conditions; and

[0051] On a plane having the design value and the first characteristic value as axes, symbols corresponding to the peripheral design conditions are arranged in a color corresponding to the second characteristic value.

[0052]

[12] A program that causes a computer to execute the following steps:

[0053] Determining reference design conditions representing design conditions in a target substance manufactured using two or more material substances;

[0054] generating a plurality of peripheral design conditions by varying two design values ​​with respect to the reference design condition;

[0055] predicting one or more characteristic values ​​in each of the peripheral design conditions; and

[0056] On a plane with the first design value and the second design value as axes, symbols corresponding to the peripheral design conditions are arranged in colors corresponding to the characteristic values.

[0057] [Effects of the invention]

[0058] According to one embodiment of the present disclosure, the relationship between the design value and the characteristic value of a target substance can be easily grasped. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] Figure 1 This is a block diagram showing an example of the overall configuration of the design support system.

[0060] Figure 2 This is a block diagram showing an example of the hardware configuration of a computer.

[0061] Figure 3 This is a block diagram showing an example of the functional structure of the design support system.

[0062] Figure 4 This is a flowchart showing an example of the processing steps of the characteristic prediction method.

[0063] Figure 5 This is a diagram showing an example of reference design data.

[0064] Figure 6 This is a diagram showing an example of a reference design input screen.

[0065] Figure 7 This is a flowchart showing an example of the procedure of the peripheral design generation process.

[0066] Figure 8 It is a diagram showing a first example of peripheral design conditions.

[0067] Fig. 9 It is a diagram showing a second example of the peripheral design conditions.

[0068] Fig.10 It is a diagram showing a first example of characteristic prediction results.

[0069] Fig.11 It is a diagram showing a second example of the characteristic prediction result.

[0070] Fig.12 It is a diagram showing the third example of the characteristic prediction result.

[0071] Fig.13 It is a diagram showing a first example of the characteristic prediction screen.

[0072] Fig.14 It is a diagram showing a second example of the characteristic prediction screen.

[0073] Fig.15 It is a diagram showing a third example of the characteristic prediction screen.

[0074] Fig.16 (A) is a diagram showing an example of symbols showing two predicted values. Fig.16 (B) is a diagram showing an example of symbols showing two predicted values. Fig.16 (C) is a diagram showing an example of symbols showing three predicted values. Fig.16 (D) is a diagram showing an example of symbols showing four predicted values.

[0075] Fig.17 It is a diagram showing the fourth example of the characteristic prediction result.

[0076] Fig.18 This is a diagram showing an example of a reference design editing screen. DETAILED DESCRIPTION

[0077] Hereinafter, embodiments of the present invention will be described with reference to the accompanying drawings. In addition, in this specification and the drawings, components having substantially the same functional configuration are denoted by the same reference numerals to omit repeated descriptions.

[0078] [Implementation Method]

[0079] In the past, material development was based on the experience and intuition of technicians. In recent years, technologies such as machine learning have been used to improve the efficiency of material development. On the other hand, when selecting materials or proportions from the prediction results obtained by machine learning, the final judgment needs to be made by professional technicians. Therefore, for the prediction results of machine learning, information indicating the reason for selecting the material or proportion is also required.

[0080] In order to explain the prediction results obtained by the machine learning model, explainable artificial intelligence (XAI) has been proposed. For example, as XAI technology, SHAP (Shapley Additive Explanations) and LIME (Local Interpretable Model-agnostic Explainations) can be listed. However, in the prior art, it takes time for users to understand the information output by XAI technology. In addition, it is difficult to use this information when it contains inoperable explanatory variables.

[0081] One embodiment of the present disclosure is a design support system that supports the design of a substance manufactured using a plurality of substances. Hereinafter, the substance to be designed is referred to as a "target substance", and the substance used to manufacture the target substance is referred to as a "material substance".

[0082] An example of the target substance in the present embodiment is a resin or an alloy, etc. An example of the material substance in the present embodiment is a monomer, a polymer or an additive, etc. The target substance and the material substance are not limited to these, and any substance can be applied as long as it is a substance made using multiple substances.

[0083] The design support system in this embodiment generates a plurality of design conditions (hereinafter also referred to as "peripheral design conditions") that change one or more design values ​​in the design conditions of the target substance (hereinafter also referred to as "baseline design conditions"), and displays the relationship between the changed design values ​​and one or more characteristic values ​​predicted in each peripheral design condition. The design conditions of the target substance include the proportion of each material, molecular weight and physical property values ​​(for example, polymer degree of polymerization and glass transition point, etc.), and manufacturing conditions (for example, processing temperature and processing time, etc.).

[0084] In particular, the design support system in this embodiment can display the relationship between the design value and the characteristic value even when multiple design values ​​are changed or multiple characteristic values ​​are predicted. Therefore, according to the design support system in this embodiment, it is possible to easily grasp the trend of characteristic value changes caused by changing the design value.

[0085] For example, the design support system in this embodiment predicts one characteristic value for the peripheral design condition in which two design values ​​in the reference design condition are changed, and draws a symbol represented by a color expression corresponding to the characteristic value on a plane with the changed design value as an axis. Therefore, according to the design support system in this embodiment, even when a plurality of design values ​​are changed, the changing trend of the characteristic value can be easily grasped.

[0086] In addition, for example, in the design support system of the present embodiment, when two characteristic values ​​are predicted for the peripheral design conditions that change one design value in the reference design condition, a symbol represented by a color expression corresponding to the second characteristic value is drawn on a plane with the changed design value and the first characteristic value as axes. Therefore, according to the design support system of the present embodiment, even when a plurality of characteristic values ​​are predicted, the changing trend of the characteristic values ​​can be easily grasped.

[0087] Furthermore, for example, in the design support system of the present embodiment, when two characteristic values ​​are predicted for the peripheral design conditions in which two design values ​​in the reference design conditions are changed, a symbol is drawn in which regions represented by color expressions corresponding to the respective characteristic values ​​are combined on a plane with the two changed design values ​​as axes. Therefore, according to the design support system of the present embodiment, even when a plurality of characteristic values ​​are predicted for the design conditions in which a plurality of design values ​​are changed, the changing trend of the characteristic values ​​can be easily grasped.

[0088] The design support system in this embodiment determines new reference design conditions based on the displayed prediction results, and can predict more than one characteristic value for peripheral design conditions that change more than one design value relative to the new reference design conditions. Therefore, according to the design support system in this embodiment, the design conditions that obtain good prediction results can be used as reference design conditions, and the peripheral design conditions that can obtain better prediction results can be repeatedly searched. Therefore, according to the design support system in this embodiment, the design conditions of the target substance can be effectively determined.

[0089] <Overall structure of the design support system>

[0090] Reference Figure 1 The overall configuration of the design support system in this embodiment will be described. Figure 1This is a block diagram showing an example of the overall configuration of the design support system in this embodiment.

[0091] like Figure 1 As shown, the design support system 1 in this embodiment includes a characteristic prediction device 10 and a user terminal 20. The characteristic prediction device 10 and the user terminal 20 are connected to each other via a communication network N1 such as a LAN (Local Area Network) or the Internet so as to be data communicable.

[0092] The characteristic prediction device 10 is an information processing device such as a personal computer, a workstation, or a server that visualizes the relationship between the design value and the characteristic value of the target substance according to a request from the user terminal 20. The characteristic prediction device 10 receives basic design conditions related to the target substance as a design object from the user terminal 20. The characteristic prediction device 10 predicts the characteristic value in the peripheral design conditions generated based on the basic design conditions, and sends information for visualizing the relationship between the design value and the characteristic value of the target substance (hereinafter also referred to as "visualization information") to the user terminal 20.

[0093] The user terminal 20 is an information processing terminal such as a personal computer, a tablet terminal, or a smartphone operated by a user. The user terminal 20 transmits the design conditions of the target substance to be designed to the characteristic prediction device 10 according to the operation of the user. The user terminal 20 displays the relationship between the design value and the characteristic value of the target substance to the user based on the visualization information received from the characteristic prediction device 10.

[0094] in addition, Figure 1 The overall structure of the design support system 1 shown is an example, and there may be various system structure examples depending on the use and purpose. For example, the characteristic prediction device 10 may be implemented by multiple computers or as a cloud computing service. In addition, for example, the design support system 1 may also be implemented by an independent information processing device that has the functions that the characteristic prediction device 10 and the user terminal 20 should have.

[0095] <Hardware structure of the design support system>

[0096] Reference Figure 2 The hardware configuration of the design support system 1 in this embodiment will be described.

[0097] Computer Hardware Structure

[0098] The characteristic prediction device 10 and the user terminal 20 in the present embodiment are realized by, for example, a computer. Figure 2 This is a block diagram showing an example of the hardware configuration of the computer 500 in this embodiment.

[0099] like Figure 2As shown, the computer 500 includes a CPU (Central Processing Unit) 501, a ROM (Read Only Memory) 502, a RAM (Random Access Memory) 503, a HDD (Hard Disk Drive) 504, an input device 505, a display device 506, a communication I / F (Interface) 507, and an external I / F 508. The CPU 501, the ROM 502, and the RAM 503 form a so-called computer. Each hardware of the computer 500 is connected to each other via a bus 509. In addition, the input device 505 and the display device 506 may also be connected to the external I / F 508 for use.

[0100] The CPU 501 is a computing device that realizes the overall control and functions of the computer 500 by reading programs or data from a storage device such as the ROM 502 or the HDD 504 to the RAM 503 and executing processing.

[0101] ROM502 is an example of a nonvolatile semiconductor memory (storage device) that can retain programs or data even when the power is turned off. ROM502 functions as a main storage device that stores various programs and data required for CPU501 to execute various programs installed in HDD504. Specifically, ROM502 stores data such as boot programs such as BIOS (Basic Input / Output System) and EFI (Extensible Firmware Interface) executed when the computer 500 is started, OS (Operating System) settings, network settings, etc.

[0102] RAM503 is an example of a volatile semiconductor memory (storage device) in which programs and data are deleted if power is turned off. RAM503 is, for example, DRAM (Dynamic Random Access Memory), SRAM (Static Random Access Memory), etc. RAM503 provides a work area that is expanded when CPU501 executes various programs installed in HDD504.

[0103] HDD504 is an example of a nonvolatile storage device that stores programs or data. Among the programs or data stored in HDD504, there are OS, which is the basic software for controlling the entire computer 500, and application programs that provide various functions on OS. In addition, computer 500 may use a storage device (e.g., SSD: Solid State Drive, solid state drive, etc.) that uses flash memory as a storage medium instead of HDD504.

[0104] The input device 505 is a touch panel, operation keys or buttons, a keyboard or a mouse for a user to input various signals, a microphone for inputting sound data such as sound, or the like.

[0105] The display device 506 is composed of a display such as a liquid crystal or an organic EL (Electro-Luminescence) that displays a screen, a speaker that outputs audio data such as sound, and the like.

[0106] The communication I / F 507 is an interface for connecting to a communication network and performing data communication by the computer 500 .

[0107] The external I / F 508 is an interface with an external device. The external device includes a drive device 510 and the like.

[0108] The drive device 510 is a device for setting a recording medium 511. The recording medium 511 mentioned here includes a medium that records information optically, electrically or magnetically, such as a CD-ROM, a floppy disk, an optical magnetic disk, etc. In addition, the recording medium 511 may also include a semiconductor memory that records information electrically, such as a ROM, a flash memory, etc. Thus, the computer 500 can read and / or write the recording medium 511 via the external I / F 508.

[0109] In addition, for example, various programs to be installed on HDD 504 are installed by setting distributed recording medium 511 on drive device 510 connected to external I / F 508, and reading various programs recorded on recording medium 511 by drive device 510. Alternatively, various programs installed on HDD 504 may be installed by downloading from another network different from the communication network via communication I / F 507.

[0110] <Functional structure of design support system>

[0111] Reference Figure 3 The functional configuration of the design support system in this embodiment will be described. Figure 3 It is a block diagram showing an example of the functional configuration of the design support system 1 in the present embodiment.

[0112] 《Functional structure of characteristic prediction device》

[0113] like Figure 3 As shown, the characteristic prediction device 10 in the present embodiment includes a model storage unit 100 , a reference design determination unit 101 , a peripheral design generation unit 102 , a characteristic prediction unit 103 , and a visualization unit 104 .

[0114] The reference design determination unit 101, the peripheral design generation unit 102, the characteristic prediction unit 103, and the visualization unit 104 are executed by the CPU 501. Figure 2 The model storage unit 100 is implemented by the program processing that is expanded from the HDD 504 to the RAM 503 as shown. Figure 2 The HDD 504 shown is implemented.

[0115] The learned characteristic prediction model is stored in the model storage unit 100. The characteristic prediction model is a machine learning model that uses each design value in the design conditions of the target substance as an explanatory variable and the characteristic value of the target substance as a target variable. An example of a machine learning model is a regression model. The structure of the machine learning model can be, for example, a random forest or a deep learning neural network.

[0116] The reference design determination unit 101 receives the design condition selected by the user's operation from the user terminal 20. The reference design determination unit 101 determines the received design condition as the reference design condition.

[0117] The peripheral design generation unit 102 generates a plurality of peripheral design conditions in which one or more design values ​​are changed relative to the reference design condition. The peripheral design generation unit 102 generates the peripheral design conditions by changing the design value of the reference design condition according to the change condition set in advance for each design value. The change condition can also be arbitrarily set according to the user's operation.

[0118] The characteristic prediction unit 103 predicts one or more characteristic values ​​for each peripheral design condition. The characteristic prediction unit 103 calculates one or more characteristic values ​​by inputting the design values ​​of the peripheral design conditions to the learned characteristic prediction model stored in the model storage unit 100 .

[0119] The visualization unit 104 generates visualization information for visualizing the relationship between the design value and the characteristic value of the target substance. The visualization unit 104 sends the generated visualization information to the user terminal 20. The visualization unit 104 visualizes the relationship between the design value and the characteristic value by arranging symbols represented by color expressions corresponding to the characteristic values ​​on a plane with the design value or the characteristic value as an axis. In this embodiment, the design value or the characteristic value represented by the axis of the plane and the characteristic value represented by the color expression can be arbitrarily selected.

[0120] <<Functional structure of user terminal 20>>

[0121] like Figure 3As shown, the user terminal 20 according to the present embodiment includes a reference design input unit 201 , a benchmark design input unit 202 , and a result display unit 203 .

[0122] The reference design input unit 201, the reference design input unit 202, and the result display unit 203 are executed by the CPU 501. Figure 2 This is achieved by the processing of the program expanded from HDD 504 to RAM 503 as shown.

[0123] The reference design input unit 201 receives input of reference design data according to the user's operation. The reference design data is data indicating reference design conditions that are candidates for the base design conditions. The reference design conditions included in the reference design data may be one or more. The reference design input unit 201 displays the received reference design data on the display device 506.

[0124] The reference design input unit 202 receives selection of reference design conditions in response to user operations. When the reference design input unit 202 selects a reference design condition for the first time, it selects the reference design condition from the reference design data received by the reference design input unit 201. When the reference design input unit 202 selects a reference design condition for the second or subsequent time, it selects the reference design condition from the peripheral design conditions included in the prediction results displayed by the result display unit 203.

[0125] The result display unit 203 receives the visualization information from the characteristic prediction device 10. The result display unit 203 displays the relationship between the design value and the characteristic value under the peripheral design conditions on the display device 506 based on the received visualization information.

[0126] <Processing steps of the design support system>

[0127] Reference Figures 4 to 17 The processing steps of the characteristic prediction method executed by the design support system 1 in this embodiment will be described. Figure 4 This is a flowchart showing an example of the processing procedure of the characteristic prediction method in this embodiment.

[0128] In step S1, the reference design input unit 201 included in the user terminal 20 receives input of reference design data in response to a user's operation. Next, the reference design input unit 201 displays the received reference design data on the display device 506.

[0129] Figure 5 FIG. 1 is a diagram showing an example of reference design data. Figure 5 As shown in FIG. 1 , the reference design data includes a plurality of reference design conditions representing design conditions of the target substance. Figure 5 In the example, the proportion of the material substances used to produce the target substance is used as the design condition.

[0130] The reference design data is input and displayed on a reference design input screen displayed on the display device 506. The reference design input screen is a screen for accepting input of reference design data and displaying reference design conditions included in the reference design data.

[0131] 《Reference design input screen》

[0132] Reference Figure 6 The reference design input screen in this embodiment is described. Figure 6 It is a diagram showing an example of a reference design input screen in this embodiment.

[0133] like Figure 6 As shown, the reference design input screen 1000 in this embodiment includes a file input field 1010 , a file reference button 1011 , an OK button 1012 , a clear button 1013 , and a reference design condition display field 1020 .

[0134] The user can input the reference design data by dragging and dropping the reference design data into the file input field 1010. The user can also input the reference design data by pressing the file reference button 1011 and selecting the reference design data on the displayed file selection screen.

[0135] When the user inputs reference design data, the contents of the reference design data included in the reference design data may be selectively displayed in the reference design condition display column 1020. Figure 6 In the reference design input screen 1000 shown, reference design conditions are displayed in a table format in a reference design condition display field 1020 , and a selection field 1021 is displayed for each reference design condition.

[0136] When the user selects a selection field 1021 corresponding to any reference design condition and presses the confirmation button 1012, the reference design condition selected in the selection field 1021 is selected as the base design condition. When the user presses the clear button 1013, the reference design condition displayed in the reference design condition display field 1020 is deleted.

[0137] return Figure 4 In step S2 , the reference design input unit 202 included in the user terminal 20 receives selection of a reference design condition in response to a user operation. The reference design input unit 202 transmits the selected reference design condition to the characteristic prediction device 10 .

[0138] In the characteristic prediction device 10 , the reference design determination unit 101 receives the reference design condition from the user terminal 20 . The reference design determination unit 101 transmits the received reference design condition to the peripheral design generation unit 102 .

[0139] In step S3, the peripheral design generation unit 102 of the characteristic prediction device 10 receives the reference design condition from the reference design determination unit 101. Next, the peripheral design generation unit 102 receives the selection of one or more design values ​​as the change object and one or more characteristic values ​​as the prediction object according to the user's operation. The selection of the design value and the characteristic value is performed on the characteristic prediction screen displayed on the display device 506.

[0140] In step S4 , the peripheral design generating unit 102 of the characteristic prediction device 10 generates a plurality of peripheral design conditions in which one or more design values ​​to be changed are changed relative to the reference design condition. The peripheral design generating unit 102 sends the generated plurality of peripheral design conditions to the characteristic prediction unit 103 .

[0141] 《Peripheral Design Generation Processing》

[0142] Reference Figure 7 The peripheral design generation process in this embodiment is described in detail. Figure 4 Step S4). Figure 7 This is a flowchart showing an example of the procedure of the peripheral design generation process in this embodiment.

[0143] In step S4-1, the peripheral design generation unit 102 determines the change conditions for each design value of the design condition. The change conditions can be set in advance for each design value, or can be set arbitrarily according to the user's operation. The change conditions in this embodiment are an upper limit value, a lower limit value, and a change range. That is, the change condition indicates that the design value to be changed is changed by a specified change range unit within a range above a specified lower limit value and below a specified upper limit value.

[0144] In step S4-2, the peripheral design generation unit 102 selects a design value to be changed from one or more design values ​​selected as the change object (hereinafter also referred to as a "changed design value"). The basis for selecting the changed design value from the design values ​​as the change object is arbitrary. For example, the name of the design value, the step selected as the change object, etc. may be arbitrary.

[0145] In step S4-3, the peripheral design generating unit 102 generates peripheral design conditions by changing the variable design value according to the variable condition relative to the base design condition. At this time, the design values ​​other than the variable design value are fixed. That is, a plurality of peripheral design conditions are generated in which only the variable design value is different from the base design condition.

[0146] If there are already generated peripheral design conditions, the peripheral design generating unit 102 generates peripheral design conditions by changing the variable design value according to the variable condition for each of the generated peripheral design conditions.

[0147] Figure 8 This is a diagram showing an example of peripheral design conditions in which one design value is changed relative to the reference design condition. Figure 8 An example is to select Figure 5 The reference design condition (No. 1) shown in FIG. 1 is used as the base design condition, and the "ratio of additive A" as an example of a design value is selected as an example of a variable design value. Figure 8 As shown, the peripheral design conditions (No. 1-1 to 1-5) are such that only the compounding amount of the additive A is changed relative to the reference design condition (No. 1), and other design values ​​are fixed.

[0148] return Figure 7 In step S4-4, the peripheral design generation unit 102 determines whether there is a design value that has not been selected as a variable design value (in other words, the peripheral design condition has not been generated) among the design values ​​selected as the variable object. When there is a design value that has not been selected as a variable design value ("Yes"), the peripheral design generation unit 102 returns the process to step S4-2. When all the design values ​​have been selected as variable design values ​​("No"), the peripheral design generation unit 102 advances the process to step S4-5.

[0149] In step S4-2 executed again, the peripheral design generating unit 102 selects a new variable design value from the design values ​​selected as the variable design values ​​and the design values ​​not selected as the variable design values, and then executes steps S4-3 and S4-4 again for the selected new variable design value.

[0150] Fig. 9 This is a diagram showing an example of peripheral design conditions in which two design values ​​are changed with respect to the reference design conditions. Fig. 9 An example is to select Figure 5 The reference design condition (No. 1) shown in the figure is used as the base design condition, and "the proportion of additive A" and "the proportion of additive B" as an example of the design value are selected as examples of the variable design value. Fig. 9 As shown, multiple peripheral design conditions (No.1-1-1 to 1-3-5) include peripheral design conditions in which only the ratio of additive A changes (No.1-1-1 to 1-1-5), peripheral design conditions in which only the ratio of additive B changes (No.1-2-1 to 1-2-5), or peripheral design conditions in which the ratios of additive A and additive B change (No.1-3-1 to 1-3-5), relative to the baseline design condition (No.1).

[0151] return Figure 7In step S4-5, the peripheral design generating unit 102 sends the peripheral design conditions generated in step S4-3 to the characteristic predicting unit 103. When step S4-3 is executed multiple times, all the peripheral design conditions generated each time are sent to the characteristic predicting unit 103.

[0152] return Figure 4 In step S5 , the characteristic prediction unit 103 included in the characteristic prediction device 10 receives a plurality of peripheral design conditions from the peripheral design generation unit 102 . Next, the characteristic prediction unit 103 reads out the learned characteristic prediction model from the model storage unit 100 .

[0153] Next, the characteristic prediction unit 103 predicts one or more characteristic values ​​as prediction targets for each peripheral design condition. The characteristic prediction unit 103 calculates one or more characteristic values ​​by inputting the design values ​​of each peripheral design condition into the learned characteristic prediction model. Then, the characteristic prediction unit 103 sends the prediction result to the visualization unit 104. The prediction result includes the peripheral design condition and the characteristic value predicted under the peripheral design condition.

[0154] Fig.10 Yes Figure 8 An example of the prediction results obtained by predicting the characteristic values ​​of 2 in each peripheral design condition shown in FIG. Fig.10 As shown, for each peripheral design condition (No. 1-1 to 1-5), "Material Property 1" and "Material Property 2" as examples of characteristic values ​​are predicted.

[0155] Fig.11 Yes Fig. 9 An example of the prediction results obtained by predicting one characteristic value in each peripheral design condition shown in FIG. Fig.11 As shown, for each peripheral design condition (No. 1-1-1 to 1-3-5), "Material Property 1" as an example of a characteristic value is predicted.

[0156] Fig.12 Yes Fig. 9 An example of the prediction results obtained by predicting two characteristic values ​​in each peripheral design condition shown in FIG. Fig.12 As shown, for each peripheral design condition (No. 1-1-1 to 1-3-5), "Material Property 1" and "Material Property 2" as examples of characteristic values ​​are predicted.

[0157] return Figure 4In step S6, the visualization unit 104 of the characteristic prediction device 10 receives the prediction result from the characteristic prediction unit 103. Next, the visualization unit 104 generates visualization information for visualizing the relationship between the design value and the characteristic value of the target substance. Subsequently, the visualization unit 104 sends the visualization information to the user terminal 20.

[0158] In the user terminal 20, the result display unit 203 receives the visualization information from the characteristic prediction device 10. The result display unit 203 visualizes the relationship between the design value and the characteristic value of the target substance based on the visualization information. Specifically, the result display unit 203 displays the relationship between the design value and the characteristic value of the target substance on the characteristic prediction screen displayed by the display device 506.

[0159] 《Example 1 of the characteristics prediction screen》

[0160] Fig.13 1 is a diagram showing a first example of a characteristic prediction screen in the present embodiment. The first example of the characteristic prediction screen is a characteristic prediction screen for predicting two characteristic values ​​in peripheral design conditions in which one design value is varied relative to a reference design condition.

[0161] like Fig.13 As shown, the characteristic prediction screen 1100 in the first example includes a design value selection field 1110 , an axis setting field 1120 , and a graph display field 1130 .

[0162] The design value selection column 1110 includes a first selection column 1111 and a second selection column 1112. In the first selection column 1111 and the second selection column 1112, variable design values ​​in the design conditions are displayed selectively. In the characteristic prediction screen of the present embodiment, in the first selection column 1111 and the second selection column 1112, a list of each design value is displayed so that it can be exclusively selected by a radio button. The first selection column 1111 and the second selection column 1112 may be displayed, for example, as a drop-down list or a list box that has variable design values ​​as selection items.

[0163] The axis setting column 1120 includes an X-axis setting column 1121, a Y-axis setting column 1122, a Z-axis setting column 1123, and an axis addition button 1124. In the X-axis setting column 1121, the design value selected in the design value selection column 1110 is selectively displayed. In the Y-axis setting column 1122, the design value selected in the design value selection column 1110 or the predictable characteristic value is selectively displayed. In the Z-axis setting column 1123, the predictable characteristic value is selectively displayed. When the user presses the axis addition button 1124, the Z-axis setting column 1123 is added.

[0164] In the characteristic prediction screen of the present embodiment, the selected design value or the predictable characteristic value is displayed in the X-axis setting column 1121, the Y-axis setting column 1122, and the Z-axis setting column 1123 in a manner that allows selection from a drop-down list. The X-axis setting column 1121, the Y-axis setting column 1122, and the Z-axis setting column 1123 may be displayed, for example, by using a list box or the like with the selected design value or the predictable characteristic value as a selection item.

[0165] like Fig.13 As shown in FIG. 1 , the first example of the characteristic prediction screen is a display example in which two characteristic values ​​are predicted under the peripheral design conditions in which one design value is varied. Fig.13 In the example, the X-axis is "the proportion of additive A" as an example of a design value, the Y-axis is "material property 1" as an example of a first characteristic value, and the Z-axis is "material property 2" as an example of a second characteristic value.

[0166] The graph display column 1130 has a plane having an X-axis 1131 and a Y-axis 1132, and a Z-axis 1133 represented by a color bar. In the graph display column 1130, symbols corresponding to the prediction results related to each peripheral design condition are arranged on coordinates corresponding to the X-axis 1131 and the Y-axis 1132, and are expressed in colors corresponding to the values ​​of the Z-axis 1133.

[0167] The color representation of the value of the Z axis can be color, grayscale, or a pattern represented by binary values. When the symbol is represented by color, at least one of the hue, brightness, and saturation can be changed according to the characteristic value. For example, two colors that differ by 180 degrees on the color wheel can correspond to the minimum and maximum values ​​of the characteristic value, and each symbol can be represented by a hue of an angle corresponding to the predicted characteristic value. When the symbol is represented by grayscale, the brightness can be changed according to the characteristic value. For example, the minimum value (0) and maximum value (255) of the brightness can correspond to the minimum and maximum values ​​of the characteristic value, and each symbol can be represented by a brightness corresponding to the predicted characteristic value.

[0168] In the graph display column 1130, the reference design conditions and the peripheral design conditions can be displayed in a distinguishable manner. In the characteristic prediction screen of this embodiment, the symbol corresponding to the reference design condition among the symbols arranged in the XY plane is displayed with a bold line on the periphery. The method of distinguishing the reference design condition and the peripheral design condition is not limited to this, and any method may be used as long as the user can visually recognize the symbol by different colors, shapes, sizes, etc.

[0169] In the first example of the characteristic prediction screen, since there is only one design value to be changed, the symbols corresponding to the prediction results under each peripheral design condition are arranged in a way that a line graph is drawn on the XY plane. The user can understand the change trend of the two characteristic values ​​caused by changing the design value by looking down at the coordinates and color expression of each symbol.

[0170] 《Example 2 of the characteristics prediction screen》

[0171] Fig.14 2 is a diagram showing a second example of the characteristic prediction screen in the present embodiment. The second example of the characteristic prediction screen is a characteristic prediction screen for predicting one characteristic value in peripheral design conditions in which two design values ​​are varied relative to the reference design conditions.

[0172] like Fig.14 As shown in FIG. 1 , the second example of the characteristic prediction screen is a display example in which one characteristic value is predicted under the peripheral design conditions in which two design values ​​are varied. Fig.14 In the example, the X-axis 1131 is the "proportion of additive A" as an example of the first design value, the Y-axis 1132 is the "proportion of additive B" as an example of the second design value, and the Z-axis 1133 is the "material property 1" as an example of the characteristic value.

[0173] In the second example of the characteristic prediction screen, since there are two changing design values, symbols corresponding to the prediction results in each peripheral design condition are arranged on the entire XY plane. The user can understand the changing trend of one characteristic value relative to the combination of two design values ​​by looking down at the coordinates and color expression of each symbol.

[0174] 《Example 3 of the characteristics prediction screen》

[0175] Fig.15 3 is a diagram showing a third example of the characteristic prediction screen in the present embodiment. The third example of the characteristic prediction screen is a characteristic prediction screen for predicting two characteristic values ​​in peripheral design conditions in which two design values ​​are varied relative to the reference design conditions.

[0176] like Fig.15 As shown in FIG. 1 , the third example of the characteristic prediction screen is a display example in which two characteristic values ​​are predicted under the peripheral design conditions in which the two design values ​​are varied. Fig.15 In the example, the X-axis 1131 is the "proportion of additive A" as an example of the first design value, the Y-axis 1132 is the "proportion of additive B" as an example of the second design value, the Z1 axis 1133-1 is the "material property 1" as an example of the first characteristic value, and the Z2 axis 1133-2 is the "material property 2" as an example of the second characteristic value.

[0177] In addition, the Z2 axis is displayed by pressing the axis addition button 1124. In this case, the Z axis in the first and second examples is the Z1 axis. When predicting three or more characteristic values, the Z3 axis, the Z4 axis, etc. can be added by further pressing the axis addition button 1124.

[0178] In the third example of the characteristic prediction screen, since there are two changing design values, the symbols corresponding to the prediction results under each peripheral design condition are arranged on the entire XY plane. Each symbol is composed of a graph obtained by combining an area represented by a color expression corresponding to the Z1 axis and an area represented by a color expression corresponding to the Z2 axis. When predicting more than three characteristic values, it is sufficient to arrange the symbols formed by combining the areas represented by the color expressions corresponding to each characteristic value on the XY plane. The user can grasp the changing trend of the combination of two characteristic values ​​relative to the two design values ​​by looking down at the coordinates of each symbol and the color expression of each area constituting the symbol.

[0179] Fig.16 : is a diagram showing an example of symbols in Example 3. The symbols in Example 3 are symbols that combine regions represented by color expressions corresponding to respective characteristic values. The symbols in Example 3 are preferably formed in the same shape for each region and arranged in a point-symmetrical manner.

[0180] Fig.16 (A) to (B) are examples of symbols showing two predicted values. Fig.16 In the symbol shown in (A), each area is formed by a triangle and arranged opposite to each other on the left and right. Fig.16 In the symbol shown in (B), each area is formed by a quadrilateral (trapezoid) and is arranged opposite to each other on the left and right.

[0181] Fig.16 (C) is an example of showing the symbols of the three predicted values. Fig.16 In the symbol shown in (C), each area is formed by a triangle and is arranged so that the angle formed by the center line of the adjacent triangle is 120°. Fig.16 (D) is an example of showing the symbols of the four predicted values. Fig.16 In the symbol shown in (D), each area is formed by a triangle and is combined and arranged in the top, bottom, left and right.

[0182] The symbols in Example 3 are not limited to Fig.16 The symbols in the third example can be expressed in a manner that allows easy understanding of the balance between the color and shape of the region corresponding to each characteristic value.

[0183] return Figure 4In step S7, the visualization unit 104 of the characteristic prediction device 10 determines whether the search performed by the user is completed. If the search is completed ("Yes"), the visualization unit 104 ends the characteristic prediction method. If the search is not completed ("No"), the visualization unit 104 returns the process to step S2.

[0184] Whether the search is finished is determined based on whether the user has operated the prediction results displayed by the visualization unit 104. If the user has operated to select a new reference design condition for the prediction results, the visualization unit 104 determines that the search has not been finished. If the user has not operated to select a new reference design condition for the prediction results, the visualization unit 104 determines that the search is finished.

[0185] In step S2 which is executed again, the reference design determination unit 101 determines new reference design conditions according to the user's operation, and then steps S3 to S7 are executed again for the determined new reference design conditions.

[0186] Fig.17 This is a diagram showing an example of prediction results in which new reference design conditions are selected from the prediction results and two characteristic values ​​are predicted based on the new reference design conditions. Fig.17 The example is to select the surrounding design conditions (No. 1-23) from the prediction results as the new reference design conditions, and select "the proportion of additive C" as an example of a design value as an example of a variable design value. Fig.17 As shown, for multiple peripheral design conditions (No.1-23-1 to 1-23-5), only the proportion of additive C is changed relative to the reference design condition (No.1-23), and other design values ​​are fixed. In addition, for each peripheral design condition (No.1-23-1 to 1-23-5), "Material Property 1" and "Material Property 2" as examples of characteristic values ​​are predicted.

[0187] [Modifications]

[0188] In the design support system 1 of the embodiment, a configuration is described for selecting new reference design conditions from the peripheral design conditions included in the prediction results. In the design support system 1 of the modification, a configuration is described for selecting new reference design conditions by editing each design value of the peripheral design conditions included in the prediction results.

[0189] In the design support system 1 in this modification, a design condition editing screen for editing each design value of the peripheral design condition is displayed on the display device 506 of the user terminal 20. The design condition editing screen determines and displays the peripheral design condition to be edited. For example, by selecting each symbol displayed in the graph display column 1130 of the characteristic prediction screen 1100, a design condition editing screen for editing each design value of the peripheral design condition corresponding to the symbol is displayed.

[0190] 《Design condition editing screen》

[0191] Reference Fig.18 The design condition editing screen in this modification example will be described. Fig.18 This is a diagram showing an example of a design condition editing screen in this modification.

[0192] like Fig.18 As shown in FIG. 1 , the design condition editing screen 1200 in this modification example includes a design condition editing bar 1210, an OK button 1211, and a cancel button 1212. In the design condition editing bar 1210, each design value of the peripheral design condition to be edited is displayed in an editable manner. In the design condition editing screen in this modification example, each design value of the peripheral design condition is displayed in a table format, and each design value is displayed in a directly editable manner.

[0193] When the user edits any design value in the design condition editing column 1210, the user can press the OK button 1211. When the user presses the OK button 1211, the reference design input unit 202 accepts the peripheral design condition edited in the design condition editing column 1210 as a new basic design condition. When the user presses the Cancel button 1212, the result display unit 203 closes the design condition editing screen 1200.

[0194] The process after selecting the new reference design condition is the same as that of the design support system 1 in the above-mentioned embodiment. That is, peripheral design conditions are generated in which one or more design values ​​are changed relative to the new reference design condition, and the prediction results including each peripheral design condition and one or more characteristic values ​​predicted for each peripheral design condition are displayed on the characteristic prediction screen 1100.

[0195] <Effects of implementation>

[0196] The design support system in this embodiment, when predicting one characteristic value for a plurality of peripheral design conditions in which two design values ​​in the reference design condition are changed, draws a symbol represented by a color expression corresponding to the predicted value on a plane with the changed design value as an axis. Therefore, according to the design support system in this embodiment, even when a plurality of design values ​​are changed, the changing trend of the characteristic value can be easily grasped.

[0197] The design support system in this embodiment, when two characteristic values ​​are predicted for a plurality of peripheral design conditions that change one design value in the reference design condition, draws a symbol represented by a color expression corresponding to the other characteristic value on a plane with the changed design value and one characteristic value as an axis. Therefore, according to the design support system in this embodiment, even when a plurality of characteristic values ​​are predicted, the changing trend of the characteristic value can be easily grasped.

[0198] The design support system in this embodiment, when two characteristic values ​​are predicted for a plurality of peripheral design conditions that change two design values ​​in the reference design condition, draws a symbol formed by combining regions represented by color expressions corresponding to the respective characteristic values ​​on a plane with the two changed design values ​​as axes. Therefore, according to the design support system in this embodiment, even when a plurality of characteristic values ​​are predicted for a design condition that changes a plurality of design values, the changing trend of the characteristic values ​​can be easily grasped.

[0199] The design support system in this embodiment can select a desired design condition from the displayed prediction results, use the selected design condition as a new reference design condition, and re-display the prediction result of the desired characteristic value for the peripheral design condition that changes the desired design value. Therefore, according to the design support system in this embodiment, it is possible to repeatedly search for the peripheral design conditions of the design condition that obtains a good prediction result, and it is possible to effectively determine the design conditions of the target substance.

[0200] [Replenish]

[0201] Each function of the above-mentioned embodiment can be realized by one or more processing circuits. Here, the "processing circuit" in this specification includes a processor that performs each function by software programming, such as a processor installed by an electronic circuit, or an ASIC (Application Specific Integrated Circuit), DSP (Digital Signal Processor), FPGA (Field Programmable Gate Array), or existing circuit modules designed to perform each function described above.

[0202] As mentioned above, although embodiment of this invention is described in detail, this invention is not limited to these embodiments, Various deformation|transformation and change are possible within the range of the summary of this invention described in a claim.

[0203] This application claims the priority of Japanese Patent Application No. 2022-119644 filed with the Japan Patent Office on July 27, 2022, the entire contents of which are incorporated herein by reference.

[0204]

Explanation of symbols

[0205] 1 Design support system

[0206] 10. Characteristic prediction device

[0207] 101 Benchmark Design Decision Department

[0208] 102 Peripheral Design and Generation Department

[0209] 103 Characteristics Prediction Department

[0210] 104 Visualization Department

[0211] 20 User Terminal

[0212] 201 Reference design input section

[0213] 202 Benchmark Design Input Section

[0214] 203 result display unit.

Claims

1. A characteristic prediction device, comprising: A reference design determination unit configured to determine reference design conditions representing design conditions in a target substance manufactured using two or more material substances; a peripheral design generating unit configured to generate a plurality of peripheral design conditions by changing one design value with respect to the reference design condition; a characteristic prediction unit configured to predict two or more characteristic values ​​in each of the peripheral design conditions; and The visualization unit is configured to arrange symbols corresponding to the peripheral design conditions in a color corresponding to the second characteristic value on a plane having the design value and the first characteristic value as axes.

2. A characteristic prediction device, comprising: A reference design determination unit configured to determine reference design conditions representing design conditions in a target substance manufactured using two or more material substances; a peripheral design generating unit configured to generate a plurality of peripheral design conditions by changing two design values ​​with respect to the reference design condition; a characteristic prediction unit configured to predict one or more characteristic values ​​in each of the peripheral design conditions; and The visualization unit is configured to arrange symbols corresponding to the peripheral design conditions in colors corresponding to the characteristic values ​​on a plane having the first design value and the second design value as axes.

3. The characteristic prediction device according to claim 2, wherein: The characteristic prediction unit is configured to predict two or more characteristic values. The visualization unit is configured to arrange symbols formed by combining regions represented by color expressions corresponding to the respective characteristic values ​​on coordinates corresponding to the peripheral design conditions.

4. The characteristic prediction device according to claim 3, wherein: The respective regions of the symbol are formed in the same shape and are arranged in a point-symmetrical manner.

5. The characteristic prediction device according to any one of claims 1 to 4, wherein: The reference design determination unit is configured to determine, as the reference design condition, the peripheral design condition corresponding to the symbol selected from the symbols displayed by the visualization unit. The visualization unit is configured to re-arrange, on the plane, the symbols corresponding to the peripheral design conditions generated based on the reference design conditions when the reference design conditions are re-selected.

6. The characteristic prediction device according to any one of claims 1 to 4, wherein: The reference design determination unit is configured to determine the reference design condition by editing the peripheral design condition corresponding to the symbol displayed by the visualization unit. The visualization unit is configured to re-arrange, on the plane, the symbols corresponding to the peripheral design conditions generated based on the reference design conditions when the reference design conditions are re-selected.

7. The characteristic prediction device according to any one of claims 1 to 6, wherein: The peripheral design generating unit is configured to generate the peripheral design condition by changing the design value within a range of a predetermined upper limit value and a predetermined lower limit value and within a predetermined variation range.

8. The characteristic prediction device according to any one of claims 1 to 7, wherein: The characteristic prediction unit is configured to predict the characteristic value using a machine learning model that uses the design value as an explanatory variable and the characteristic value as a target variable.

9. A method for predicting characteristics, comprising: performing the following steps by a computer: Determining reference design conditions representing design conditions in a target substance manufactured using two or more material substances; generating a plurality of peripheral design conditions by changing one design value with respect to the reference design condition; predicting two or more characteristic values ​​in each of the peripheral design conditions; as well as On a plane having the design value and the first characteristic value as axes, symbols corresponding to the peripheral design conditions are arranged in a color corresponding to the second characteristic value.

10. A method for predicting characteristics, comprising: performing the following steps by a computer: Determining reference design conditions representing design conditions in a target substance manufactured using two or more material substances; generating a plurality of peripheral design conditions by varying two design values ​​with respect to the reference design condition; predicting one or more characteristic values ​​in each of the peripheral design conditions; as well as On a plane with the first design value and the second design value as axes, symbols corresponding to the peripheral design conditions are arranged in colors corresponding to the characteristic values.

11. A program that causes a computer to execute the following steps: Determining reference design conditions representing design conditions in a target substance manufactured using two or more material substances; generating a plurality of peripheral design conditions by changing one design value with respect to the reference design condition; predicting two or more characteristic values ​​in each of the peripheral design conditions; as well as On a plane having the design value and the first characteristic value as axes, symbols corresponding to the peripheral design conditions are arranged in a color corresponding to the second characteristic value.

12. A program that causes a computer to execute the following steps: Determining reference design conditions representing design conditions in a target substance manufactured using two or more material substances; generating a plurality of peripheral design conditions by varying two design values ​​with respect to the reference design condition; predicting one or more characteristic values ​​in each of the peripheral design conditions; as well as On a plane with the first design value and the second design value as axes, symbols corresponding to the peripheral design conditions are arranged in colors corresponding to the characteristic values.

Images