# US 2020/0395102 A1 · Polymer physical property prediction device, recording medium, and polymer physical property prediction method

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2020/0395102 A1
- Document ID: US20200395102A1
- Status: Published application
- Kind: A1
- Application number: 17/008,999
- Filed: 2020-09-01
- Published: 2020-12-17
- Earliest claimed: 2018-03-09
- Assignees: Showa Denko K.K.
- Applicants: Showa Denko K.K.
- Inventors: Takuya MINAMI; Yoshishige OKUNO
- Web page: https://patentgrep.com/patent/US20200395102A1

## CPC

- G16C 20/30: Prediction of properties of chemical compounds, compositions or mixtures
- C08F 110/02: Ethene
- G01N 33/442: {Resins; Plastics}
- G06N 20/10: using kernel methods, e.g. support vector machines [SVM]
- G16C 20/70: Machine learning, data mining or chemometrics
- G16C: COMPUTATIONAL CHEMISTRY; CHEMOINFORMATICS; COMPUTATIONAL MATERIALS SCIENCE
- C08F: MACROMOLECULAR COMPOUNDS OBTAINED BY REACTIONS ONLY INVOLVING CARBON-TO-CARBON UNSATURATED BONDS
- G01N: INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES (measuring or testing apparatus or processes other than immunoassay, involving enzymes or microorganisms C12M, C12Q)
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS

## Abstract

A polymer physical property prediction device includes a processor, and a memory storing program instructions that cause the processor to read a structural unit from a storage unit and use the structural unit to calculate numbers each indicating how many substructures are in a polymer, the polymer being formed of repetition of the structural unit, calculate a number indicating how many atoms are in the structural unit, calculate number densities of the substructures from the numbers of substructures and the number of atoms in the structural unit, construct a regression model that predicts a physical property value by using an experimental value of a physical property of the polymer and the number densities of the substructures, input a polymer structure of which the physical property value is to be predicted, and predict the physical property value corresponding to the input polymer structure by using the regression model.

## Claims

### Claim 1

A polymer physical property prediction device comprising:
a processor; and
a memory storing program instructions that cause the processor to
read a structural unit from a storage unit and use the structural unit to calculate numbers each indicating how many substructures are in a polymer, the polymer being formed of repetition of the structural unit;
calculate a number indicating how many atoms are in the structural unit;
calculate number densities of the substructures from the numbers of substructures and the number of atoms in the structural unit;
construct a regression model that predicts a physical property value by using an experimental value of a physical property of the polymer and the number densities of the substructures, the experimental value being obtained from the storage unit;
input a polymer structure of which the physical property value is to be predicted; and
predict the physical property value corresponding to the input polymer structure by using the regression model.

### Claim 2 (depends on 1)

The polymer physical property prediction device as claimed in claim 1, wherein the program instructions cause the processor to further classify the substructures of the polymer using an algorithm, and the calculated numbers of substructures each indicate a number of each corresponding classified substructure present in the polymer.

### Claim 3 (depends on 1)

The polymer physical property prediction device as claimed in claim 1, wherein the program instructions cause the processor to further remove part of the calculated number densities of the substructures.

### Claim 4 (depends on 1)

The polymer physical property prediction device as claimed in claim 1, wherein the program instructions cause the processor to further generate a model structure by repeatedly using the structural unit obtained from the storage unit.

### Claim 5 (depends on 1)

The polymer physical property prediction device as claimed in claim 1,
wherein the program instructions cause the processor to further
generate a model structure by repeatedly using the structural unit obtained from the storage unit, and
remove part of a structure descriptor representing the calculated number densities of the substructures,
wherein the numbers of substructures and the number of atoms are calculated from the generated model structure.

### Claim 6 (depends on 1)

The polymer physical property prediction device as claimed in claim 1, wherein the program instructions cause the processor to calculate the numbers of substructures and the number of atoms in the polymer structure by using the input polymer structure, calculate the number densities of the substructures from the calculated numbers of substructures and the calculated number of atoms, and apply the calculated number densities to the constructed regression model to predict the physical property value of the polymer structure.

### Claim 7 (depends on 3)

The polymer physical property prediction device as claimed in claim 3, wherein the program instructions cause the processor to calculate the numbers of substructures and the number of atoms of the polymer structure by using the input polymer structure, calculate the number densities of the substructures from the calculated numbers of substructures and the calculated number of atoms, remove part of the calculated number densities of the substructures, and apply the number densities whose part is removed to the constructed regression model to predict the physical property value of the polymer structure.

### Claim 8 (depends on 4)

The polymer physical property prediction device as claimed in claim 4, wherein the program instructions cause the processor to generate the model structure by repeatedly using the input polymer structure, calculate the numbers of substructures and the number of atoms in the model structure, calculate the number densities of the substructures from the calculated numbers of substructures and the calculated number of atoms, and apply the calculated number densities to the constructed regression model to predict the physical property value of the polymer structure.

### Claim 9 (depends on 5)

The polymer physical property prediction device as claimed in claim 5, wherein the program instructions cause the processor to generate the model structure by repeatedly using the input polymer structure, calculate the numbers of substructures and the number of atoms in the model structure, calculate the number densities of the substructures from the calculated numbers of substructures and the calculated number of atoms, remove the part of the structure descriptor representing the calculated number densities of the substructures, and apply the number densities whose part is removed to the constructed regression model to predict the physical property value of the polymer structure.

### Claim 10

A non-transitory computer-readable recording medium having stored therein a polymer physical property prediction program for causing a computer to execute a process comprising:
reading a structural unit from a storage unit and using the structural unit to calculate numbers each indicating how many substructures are in a polymer, the polymer being formed of repetition of the structural unit;
calculating a number indicating how many atoms are in the structural unit;
calculating number densities of the substructures from the numbers of substructures and the number of atoms in the structural unit;
constructing a regression model that predicts a physical property value by using an experimental value of a physical property of the polymer and the number densities of the substructures, the experimental value being obtained from the storage unit;
inputting a polymer structure of which the physical property value is to be predicted; and
predicting the physical property value corresponding to the input polymer structure by using the regression model.

### Claim 11 (depends on 10)

The recording medium as claimed in claim 10, wherein the predicting the physical property value further includes calculating the numbers of substructures and the number of atoms in the polymer structure by using the input polymer structure, calculating the number densities of the substructures from the calculated numbers of substructures and the calculated number of atoms, and applying the calculated number densities to the regression model to predict the physical property value of the polymer structure.

### Claim 12

A polymer physical property prediction method executed by a computer, the method comprising:
reading a structural unit from a storage unit and using the structural unit to calculate numbers each indicating how many substructures are in a polymer, the polymer being formed of repetition of the structural unit;
calculating a number indicating how many atoms are in the structural unit;
calculating number densities of the substructures from the numbers of substructures and the number of atoms in the structural unit;
constructing a regression model that predicts a physical property value by using an experimental value of a physical property of the polymer and the number densities of the substructures, the experimental value being obtained from the storage unit;
inputting a polymer structure of which the physical property value is to be predicted; and
predicting the physical property value corresponding to the input polymer structure by using the regression model.

### Claim 13 (depends on 12)

The polymer physical property prediction method as claimed in claim 12, wherein the predicting the physical property value further includes calculating the numbers of substructures and the number of atoms in the polymer structure by using the input polymer structure, calculating the number densities of the substructures from the calculated numbers of substructures and the calculated number of atoms, and applying the calculated number densities to the regression model to predict the physical property value of the polymer structure.

### Claim 14 (depends on 12)

A polymer manufacturing method comprising:
obtaining a polymer structure having a physical property predicted by the polymer physical property prediction method as claimed in claim 12;
preparing materials for forming the polymer structure; and
obtaining a polymer represented by the polymer structure by reacting the materials.

### Claim 15 (depends on 1)

The polymer physical property prediction device as claimed in claim 1, wherein the physical property value is a value of a refractive index or a value of a glass transition point.

### Claim 16 (depends on 1)

The polymer physical property prediction device as claimed in claim 1, wherein the physical property value is a value of a band gap, a value of a relative permittivity, a value of a coefficient of linear expansion, or a value of a density.
