# US 2020/0287497 A1 · Abnormality determination system, motor control apparatus, and abnormality determination apparatus

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

- Number: US 2020/0287497 A1
- Document ID: US20200287497A1
- Status: Published application
- Kind: A1
- Application number: 16/879,760
- Filed: 2020-05-21
- Published: 2020-09-10
- Earliest claimed: 2017-11-28
- Assignees: Kabushiki Kaisha Yaskawa Denki
- Applicants: Kabushiki Kaisha Yaskawa Denki
- Inventors: Takeshi NAGATA; Takaaki SHOGAKI; Yasufumi YOSHIURA; Tadashi OKUBO; Toshinobu KIRA
- Web page: https://patentgrep.com/patent/US20200287497A1

## CPC

- H02P 29/024: Detecting a fault condition, e.g. short circuit, locked rotor, open circuit or loss of load
- G01R 31/343: {in operation}
- G05B 23/024: {Quantitative history assessment, e.g. mathematical relationships between available data; Functions therefor; Principal component analysis [PCA]; Partial least square [PLS]; Statistical classifiers, e.g. Bayesian networks, linear regression or correlation analysis; Neural networks}
- G06N 3/0464: Convolutional networks [CNN, ConvNet]
- G06N 3/08: Learning methods
- G06N 3/09: Supervised learning
- H02P 6/12: Monitoring commutation; Providing indication of commutation failure
- H02P 23/0018: {using neural networks}
- H02P: CONTROL OR REGULATION OF ELECTRIC MOTORS, ELECTRIC GENERATORS OR DYNAMO-ELECTRIC CONVERTERS; CONTROLLING TRANSFORMERS, REACTORS OR CHOKE COILS
- G01R: MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES (indicating correct tuning of resonant circuits H03J3/12)
- G05B: CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS (systems for controlling or regulating non-electric variables G05D; systems for regulating electric or magnetic variables G05F; control devices or systems insofar as characterised by mechanical features only G05G)
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS

## Abstract

An abnormality determination system includes a state quantity obtaining circuit and an abnormality determination circuit. The state quantity obtaining circuit is configured to obtain a state quantity associated with a mechanical system. The abnormality determination circuit is configured to, according to a learning content obtained in a machine learning process and based on the state quantity, determine as to at least one of an occurrence of an abnormality in the mechanical system, an occurrence position of the abnormality, and a cause of the abnormality.

## Claims

### Claim 1

An abnormality determination system comprising:
a state quantity obtaining circuit configured to obtain a state quantity associated with a mechanical system; and
an abnormality determination circuit configured to, according to a learning content obtained in a machine learning process and based on the state quantity, determine as to at least one of an occurrence of an abnormality in the mechanical system, an occurrence position of the abnormality, and a cause of the abnormality.

### Claim 2 (depends on 1)

The abnormality determination system according to claim 1,
wherein the mechanical system comprises
a motor,
a motor control apparatus configured to control the motor, and
a motor-driven mechanism configured to be driven by the motor, and
wherein the state quantity obtaining circuit comprises a time-series detection data obtaining circuit configured to obtain, as the state quantity, time-series detection data associated with input and output of the motor while the motor is driven.

### Claim 3 (depends on 2)

The abnormality determination system according to claim 2, wherein the abnormality determination circuit comprises an equipment characteristic estimation circuit configured to estimate an equipment characteristic estimated value of the mechanical system based on the time-series detection data.

### Claim 4 (depends on 2)

The abnormality determination system according to claim 2, wherein the abnormality determination circuit comprises
a data abnormality determination circuit configured to make a determination as to whether the time-series detection data has a data abnormality, and
an equipment abnormality determination circuit configured to, based on the determination made by the data abnormality determination circuit, make a determination as to whether at least one of the motor, the motor control apparatus, and the motor-driven mechanism has an equipment abnormality.

### Claim 5 (depends on 4)

The abnormality determination system according to claim 4, wherein the data abnormality determination circuit is configured to make the determination according to a Hotelling's T2 method.

### Claim 6 (depends on 5)

The abnormality determination system according to claim 5,
wherein the data abnormality determination circuit comprises
a false alarm rate determination circuit configured to determine a false alarm rate,
a chi-squared distribution calculation circuit configured to calculate a chi-squared distribution,
a threshold calculation circuit configured to calculate a data abnormality determination threshold based on the false alarm rate and the chi-squared distribution,
a sample mean calculation circuit configured to calculate a sample mean based on the time-series detection data,
a sample covariance matrix calculation circuit configured to calculate a sample covariance matrix based on the time-series detection data,
a Mahalanobis distance calculation circuit configured to calculate a Mahalanobis distance based on the sample mean, the sample covariance matrix, and the time-series detection data, and
a determination circuit configured to compare the data abnormality determination threshold and the Mahalanobis distance with each other to make a determination as to whether the time-series detection data has a data abnormality,
wherein the sample mean calculation circuit and the sample covariance matrix calculation circuit are configured to use time-series reference data obtained during a normal driving of the motor-driven mechanism,
wherein the Mahalanobis distance calculation circuit is configured to use time-series observation data obtained during an observation driving of the motor-driven mechanism, and
wherein the determination circuit is configured to make a determination as to whether the time-series observation data has the data abnormality.

### Claim 7 (depends on 4)

The abnormality determination system according to claim 4, wherein the equipment abnormality determination circuit comprises an equipment abnormality cause estimation circuit configured to estimate a cause of the equipment abnormality.

### Claim 8 (depends on 7)

The abnormality determination system according to claim 7, wherein the equipment abnormality cause estimation circuit is configured to, based on the learning content obtained in the machine learning process, estimate the cause of the equipment abnormality according to at least one of the equipment characteristic estimated value and pattern matching which is performed with respect to a time-series distribution pattern of the Mahalanobis distance.

### Claim 9 (depends on 4)

The abnormality determination system according to claim 4, wherein the equipment abnormality determination circuit comprises a motion abnormality determination circuit configured to make a determination as to whether the motor has a motion abnormality based on how the time-series detection data determined as having the data abnormality was obtained.

### Claim 10 (depends on 9)

The abnormality determination system according to claim 9, wherein the motion abnormality determination circuit comprises a motion abnormality occurrence determination circuit configured to make a determination as to whether the motion abnormality is occurring or not occurring.

### Claim 11 (depends on 10)

The abnormality determination system according to claim 10, wherein the motion abnormality occurrence determination circuit is configured to, based on a data abnormality frequency, make the determination as to whether the motion abnormality is occurring or not occurring.

### Claim 12 (depends on 9)

The abnormality determination system according to claim 9, wherein the motion abnormality determination circuit comprises a motion abnormality kind determination circuit configured to make a determination as to a kind of the motion abnormality.

### Claim 13 (depends on 12)

The abnormality determination system according to claim 12, wherein the motion abnormality kind determination circuit is configured to, based on the learning content obtained in the machine learning process, make the determination as to the kind of the motion abnormality using pattern matching performed with respect to a time-series distribution pattern of the Mahalanobis distance.

### Claim 14 (depends on 12)

The abnormality determination system according to claim 12, wherein the equipment abnormality cause estimation circuit is configured to estimate the cause of the equipment abnormality based on the determination made by the motion abnormality kind determination circuit.

### Claim 15 (depends on 4)

The abnormality determination system according to claim 4, wherein the equipment abnormality determination circuit comprises an equipment abnormality position estimation circuit configured to estimate an occurrence position of the equipment abnormality.

### Claim 16 (depends on 15)

The abnormality determination system according to claim 15, wherein the equipment abnormality position estimation circuit is configured to, based on the learning content obtained in the machine learning process, estimate the occurrence position of the equipment abnormality according to at least one of the equipment characteristic estimated value and the pattern matching which is performed with respect to the time-series distribution pattern of the Mahalanobis distance.

### Claim 17

A motor control apparatus comprising:
a circuit configured to supply electric power to a motor to drive a motor-driven mechanism;
a state quantity obtaining circuit configured to obtain a state quantity associated with at least one of the motor, the motor-driven mechanism, and the motor control apparatus; and
an abnormality determination circuit configured to, according to a learning content obtained in a machine learning process and based on the state quantity, determine whether at least one of the motor, the motor-driven mechanism, and the motor control apparatus has an abnormality.

### Claim 18

An abnormality determination apparatus comprising:
a state quantity obtaining circuit configured to obtain a state quantity associated with a mechanical system; and
an abnormality determination circuit configured to, according to a learning content obtained in a machine learning process and based on the state quantity, determine as to at least one of an occurrence of an abnormality in the mechanical system, an occurrence position of the abnormality, and a cause of the abnormality.

### Claim 19 (depends on 3)

The abnormality determination system according to claim 3, wherein the abnormality determination circuit comprises
a data abnormality determination circuit configured to make a determination as to whether the time-series detection data has a data abnormality, and
an equipment abnormality determination circuit configured to, based on the determination made by the data abnormality determination circuit, make a determination as to whether at least one of the motor, the motor control apparatus, and the motor-driven mechanism has an equipment abnormality.

### Claim 20 (depends on 19)

The abnormality determination system according to claim 19, wherein the data abnormality determination circuit is configured to make the determination according to a Hotelling's T2 method.
