HEALTH MANAGEMENT SYSTEM, FAULT DIAGNOSIS SYSTEM, HEALTH MANAGEMENT METHOD, AND FAULT DIAGNOSIS METHOD

• US 20150160098A1
• Filed: 10/30/2014
• Published: 06/11/2015
• Est. Priority Date: 11/01/2013
• Status: Active Grant

• Current Assignee: Hitachi Power Solutions Co., Ltd. (Hitachi, Ltd.), Tsuru Educational Foundation
• Original Assignee: Hitachi Power Solutions Co., Ltd. (Hitachi, Ltd.), Tsuru Educational Foundation
• Inventors: SHIBUYA, Hisae, NODA, Toujirou, SUZUKI, Tadashi, MIYAKOSHI, Naoki, KOBARI, Toshiaki, MIYABE, Shouzou, MAEDA, Shunji
• Granted Patent: US 9,933,338 B2
• Time in Patent Office: Days
• US Class Current: 1/1
• CPC Class Codes: G01M 99/00 Subject matter not provided for in other groups of this subclass
• CPC Class Codes: 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
• CPC Class Codes: G05B 23/0281 Quantitative, e.g. mathematical distance; Clustering; Neural networks; Statistical analysis
• CPC Class Codes: G05B 23/0283 Predictive maintenance, e.g. involving the monitoring of a system and, based on the monitoring results, taking decisions on the maintenance schedule of the monitored system; Estimating remaining useful life [RUL] preventive maintenance, i.e. planning maintenance according to the available resources without monitoring the system G06Q10/06

Abstract

An objective is to identify the health state of mechanical equipment and provide information usable for determining maintenance work timing or the like. A health management system includes a time-series data acquisition unit configured to acquire multi-dimensional sensor data and environmental data from mechanical equipment; a first discrimination unit configured to quantify the equipment state of the mechanical equipment by a statistical method using normal data as learning data; a second discrimination unit configured to quantify the health state indicating the performance or quality of the mechanical equipment by a statistical method using normal data; and an output unit configured to display and/or output to the outside the quantified equipment state and health state.

Claims

1. A fault diagnosis system of diagnosing a fault of mechanical equipment, the system comprising:

a time-series data acquisition unit configured to acquire, as time-series data, sensor data from a plurality of sensors installed in the mechanical equipment, or the sensor data and environmental data indicating an installation environment of the mechanical equipment;

a multi-dimensional vector calculation unit configured to calculate a multi-dimensional vector extending from a normal state to a fault state of the mechanical equipment by a statistical method using the time-series data as learning data;

a state estimation unit configured to estimate an equipment state or health state of the mechanical equipment using the multi-dimensional vector;

a starting point change detection unit configured to detect whether a starting point position of the multi-dimensional vector calculated by the multi-dimensional vector calculation unit changes by a predetermined value or more, the multi-dimensional vector calculated from time-series data acquired by the time-series data acquisition unit in a period before and after a maintenance work is performed on the mechanical equipment;

a sensor data extraction unit configured to extract, from the time-series data, sensor data that contributes to the change in the starting point position of the multi-dimensional vector; and

an output unit configured to display or output to an outside the extracted sensor data as an index indicating the health state of the mechanical equipment,

wherein the environmental data, if included in the time-series data, includes at least one type of data among event data indicating an operation state of the mechanical equipment, operation data indicating an operation time of the mechanical equipment, load data indicating a load state of the mechanical equipment, and maintenance history data indicating a history of the maintenance work of the mechanical equipment.

3. The fault diagnosis system according to claim 1, further comprising a feature transformation unit configured to transform a feature of the time-series data through independent component analysis of the time-series data,

wherein the multi-dimensional vector calculation unit calculates the multi-dimensional vector using the feature-transformed time-series data.

4. The fault diagnosis system according to claim 3, further comprising a trend component extraction unit configured to extract a trend component indicating a tendency of a change for the time-series data feature-transformed by the feature transformation unit,

wherein the multi-dimensional vector calculation unit calculates the multi-dimensional vector using the time-series data including the trend component.

5. The fault diagnosis system according to claim 3, wherein, in performing the independent component analysis on the time-series data, the feature transformation unit calculates a weight from a mixing coefficient matrix indicating a relationship between the time-series data and an independent component, and controls the size of a signal of the independent component using the weight.

6. The fault diagnosis system according to claim 1, wherein a statistical method of calculating the multi-dimensional vector is any one of a Gaussian process, any other regression procedure, a k-nearest neighbor (k-NN) method, a local subspace method, and any other recognition procedure,

in the system where claim 1 is referred to, the output unit displays the starting point of the multi-dimensional vector calculated by the multi-dimensional vector calculation unit so as to enable visual recognition of the change in the starting point on a time-series basis, whereas

the system where claim 1 is not referred to comprises an output unit configured to display the starting point of the multi-dimensional vector calculated by the multi-dimensional vector calculation unit so as to enable visual recognition of the change in the starting point on the time-series basis.

10. The fault diagnosis system according to claim 4, wherein, in performing the independent component analysis on the time-series data, the feature transformation unit calculates a weight from a mixing coefficient matrix indicating a relationship between the time-series data and an independent component, and controls the size of a signal of the independent component using the weight.

11. The fault diagnosis system according to claim 2, wherein a statistical method of calculating the multi-dimensional vector is any one of a Gaussian process, any other regression procedure, a k-nearest neighbor (k-NN) method, a local subspace method, and any other recognition procedure,

in the system where claim 1 is referred to, the output unit displays the starting point of the multi-dimensional vector calculated by the multi-dimensional vector calculation unit so as to enable visual recognition of the change in the starting point on a time-series basis, whereas

the system where claim 1 is not referred to comprises an output unit configured to display the starting point of the multi-dimensional vector calculated by the multi-dimensional vector calculation unit so as to enable visual recognition of the change in the starting point on the time-series basis.

12. The fault diagnosis system according to claim 3, wherein a statistical method of calculating the multi-dimensional vector is any one of a Gaussian process, any other regression procedure, a k-nearest neighbor (k-NN) method, a local subspace method, and any other recognition procedure,

in the system where claim 1 is referred to, the output unit displays the starting point of the multi-dimensional vector calculated by the multi-dimensional vector calculation unit so as to enable visual recognition of the change in the starting point on a time-series basis, whereas

the system where claim 1 is not referred to comprises an output unit configured to display the starting point of the multi-dimensional vector calculated by the multi-dimensional vector calculation unit so as to enable visual recognition of the change in the starting point on the time-series basis.

13. The fault diagnosis system according to claim 4, wherein a statistical method of calculating the multi-dimensional vector is any one of a Gaussian process, any other regression procedure, a k-nearest neighbor (k-NN) method, a local subspace method, and any other recognition procedure,

in the system where claim 1 is referred to, the output unit displays the starting point of the multi-dimensional vector calculated by the multi-dimensional vector calculation unit so as to enable visual recognition of the change in the starting point on a time-series basis, whereas

the system where claim 1 is not referred to comprises an output unit configured to display the starting point of the multi-dimensional vector calculated by the multi-dimensional vector calculation unit so as to enable visual recognition of the change in the starting point on the time-series basis.

14. The fault diagnosis system according to claim 5, wherein a statistical method of calculating the multi-dimensional vector is any one of a Gaussian process, any other regression procedure, a k-nearest neighbor (k-NN) method, a local subspace method, and any other recognition procedure,

in the system where claim 1 is referred to, the output unit displays the starting point of the multi-dimensional vector calculated by the multi-dimensional vector calculation unit so as to enable visual recognition of the change in the starting point on a time-series basis, whereas

the system where claim 1 is not referred to comprises an output unit configured to display the starting point of the multi-dimensional vector calculated by the multi-dimensional vector calculation unit so as to enable visual recognition of the change in the starting point on the time-series basis.

2. A fault diagnosis system of diagnosing a fault of mechanical equipment, the system comprising:

a time-series data acquisition unit configured to acquire, as time-series data, sensor data from a plurality of sensors installed in the mechanical equipment, or the sensor data and environmental data indicating an installation environment of the mechanical equipment;

a multi-dimensional vector calculation unit configured to calculate a multi-dimensional vector extending from a normal state to a fault state of the mechanical equipment by a statistical method using the time-series data as learning data;

a state estimation unit configured to estimate an equipment state or health state of the mechanical equipment using the multi-dimensional vector;

a starting point change detection unit configured to detect whether the starting point position of the multi-dimensional vector calculated by the multi-dimensional vector calculation unit changes by a predetermined value or more, the multi-dimensional vector calculated from time-series data acquired by the time-series data acquisition unit in a period before and after a maintenance work is performed on the mechanical equipment; and

an effectiveness evaluation unit configured to evaluate effectiveness of a work item carried out as the maintenance work by correlating the work item with the change in the starting point position of the multi-dimensional vector detected by the starting point change detection unit,

wherein the environmental data, if included in the time-series data, includes at least one type of data among event data indicating an operation state of the mechanical equipment, operation data indicating an operation time of the mechanical equipment, load data indicating a load state of the mechanical equipment, and maintenance history data indicating a history of the maintenance work of the mechanical equipment.

9. The fault diagnosis system according to claim 2, further comprising a feature transformation unit configured to transform a feature of the time-series data through independent component analysis of the time-series data,

wherein the multi-dimensional vector calculation unit calculates the multi-dimensional vector using the feature-transformed time-series data.

7. A fault diagnosis method for a device of diagnosing a fault of mechanical equipment, the method comprising:

a time-series data acquisition step of acquiring, as time-series data, sensor data from a plurality of sensors installed in the mechanical equipment, or the sensor data and environmental data indicating an installation environment of the mechanical equipment;

a multi-dimensional vector calculation step of calculating a multi-dimensional vector extending from a normal state to a fault state of the mechanical equipment by a statistical method using the time-series data as learning data;

a state estimation step of estimating an equipment state or health state of the mechanical equipment using the multi-dimensional vector;

a starting point change detection step of detecting whether the starting point position of the multi-dimensional vector calculated in the multi-dimensional vector calculation step changes by a predetermined value or more, the multi-dimensional vector calculated from time-series data acquired in the time-series data acquisition step in a period before and after a maintenance work is performed on the mechanical equipment;

a sensor data extraction step of extracting, from the time-series data, sensor data that contributes to the change in the starting point position of the multi-dimensional vector; and

an output step of displaying or outputting to an outside the extracted sensor data as an index indicating the health state of the mechanical equipment,

wherein the environmental data includes at least one type of data among event data indicating an operation state of the mechanical equipment, operation data indicating an operation time of the mechanical equipment, load data indicating a load state of the mechanical equipment, and maintenance history data indicating a history of the maintenance work of the mechanical equipment.

8. A fault diagnosis method for a device of diagnosing a fault of mechanical equipment, the method comprising:

a time-series data acquisition step of acquiring, as time-series data, sensor data from a plurality of sensors installed in the mechanical equipment, or the sensor data and environmental data indicating an installation environment of the mechanical equipment;

a multi-dimensional vector calculation step of calculating a multi-dimensional vector extending from a normal state to a fault state of the mechanical equipment by a statistical method using the time-series data as learning data;

a state estimation step of estimating an equipment state or health state of the mechanical equipment using the multi-dimensional vector;

a starting point change detection step of detecting whether the starting point position of the multi-dimensional vector calculated in the multi-dimensional vector calculation step changes by a predetermined value or more, the multi-dimensional vector calculated from time-series data acquired in the time-series data acquisition step in a period before and after a maintenance work is performed on the mechanical equipment; and

an effectiveness evaluation step of evaluating effectiveness of a work item carried out as the maintenance work by correlating the work item with the change in the starting point position of the multi-dimensional vector detected in the starting point change detection step,

wherein the environmental data includes at least one type of data among event data indicating an operation state of the mechanical equipment, operation data indicating an operation time of the mechanical equipment, load data indicating a load state of the mechanical equipment, and maintenance history data indicating a history of the maintenance work of the mechanical equipment.