Anomaly diagnosis system and anomaly diagnosis method

Patent metadata

Field	Value
Publication number	US-9465387-B2
Application number	US-201514982138-A
Country	US
Kind code	B2
Filing date	Dec 29, 2015
Priority date	Jan 9, 2015
Publication date	Oct 11, 2016
Grant date	Oct 11, 2016

Abstract

The anomaly diagnosis system includes the state measure calculator acquiring sensor data from sensors in a machine facility as time series data; an approximation formula calculator calculating a state measure being an index indicating a state of the machine facility, such as anomaly and a performance by a statistical method in which the time series data is used as learned data; and a state measure estimating unit estimating the state measures until future time using the approximation formula. Whenever the latest time series data is acquired, the reference period in which the time series data corresponding to the state measure referred to calculate the approximation formula by the reference period setting unit, is successively extended by addition of time when the latest time series data is acquired. The approximation formula calculator calculates the approximation formula using the state measure of the time series data acquired in the reference period.

First claim

The invention claimed is:

1. An anomaly diagnosis system diagnosing a state of a machine facility, comprising:
a time series data receiver acquiring sensor data as time series data from a plurality of sensors installed in the machine facility;
a state measure calculator calculating an anomaly measure or a performance measure as a state measure being an index indicating a state of the machine facility, by a statistical method in which the time series data is used as learned data, the anomaly measure being an index indicating a magnitude of deviation from a normal state of the machine facility, and the performance measure being an index indicating a performance of the machine facility;
an approximation formula calculator calculating an approximation formula approximating, with a polynomial expression, variation in the state measure calculated based on the time series data acquired from a past to the present, the approximation formula indicating future approximation of variation of the state measure;
a state measure estimating unit estimating the state measure until a predetermined future time using the approximation formula;
a reference period setting unit setting a reference period being a period for which the time series data corresponding to the state measure is acquired, the approximation formula calculator referring the reference period to calculate the approximation formula; and
an output which outputs a diagnose result of the state of the machine facility using the estimated state measure,
wherein the reference period setting unit sets, as the reference period, a first period including time when latest time series data is acquired or a second period shorter than the first period and including the time when the latest time series data is acquired, and
wherein the approximation formula calculator calculates the approximation formula using the state measure regarding the time series data acquired for the reference period set by the reference period setting unit.

2. The anomaly diagnosis system as claimed in claim 1 , wherein, whenever the time series data receiver acquires the latest time series data, the first period is successively extended by a time length defined by addition of time when the latest time series data is acquired.

3. The anomaly diagnosis system as claimed in claim 2 , wherein the second period has a predetermined time length including the time when the latest time series data is acquired.

4. The anomaly diagnosis system as claimed in claim 2 , wherein a period regarding which a future state measure is estimated using a first approximation formula is longer than a period regarding which a future state measure is estimated using a second approximation formula, the first approximation formula is the approximation formula calculated using the state measure of the time series data acquired for the first period, and the second approximation formula is the approximation formula calculated using a future state measure of the time series data acquired for the second period.

5. The anomaly diagnosis system as claimed in claim 2 , further comprising a filtering processor applying to the state measure a filtering process that calculates a maximum value, a minimum value, or a moving average of the state measure in a predetermined time width,
wherein the approximation formula calculator calculates the approximation formula using the state measure to which the filtering process is applied.

6. The anomaly diagnosis system as claimed in claim 1 , wherein the first period has a predetermined time length including the time when the latest time series data is acquired.

7. The anomaly diagnosis system as claimed in claim 6 , wherein the second period has a predetermined time length including the time when the latest time series data is acquired.

8. The anomaly diagnosis system as claimed in claim 6 , wherein a period regarding which a future state measure is estimated using a first approximation formula is longer than a period regarding which a future state measure is estimated using a second approximation formula, the first approximation formula is the approximation formula calculated using the state measure of the time series data acquired for the first period, and the second approximation formula is the approximation formula calculated using a future state measure of the time series data acquired for the second period.

9. The anomaly diagnosis system as claimed in claim 6 , further comprising a filtering processor applying to the state measure a filtering process calculating a maximum value, a minimum value, or a moving average of the state measure in a predetermined time width,
wherein the approximation formula calculator calculates the approximation formula using the state measure to which the filtering process is applied.

10. The anomaly diagnosis system as claimed in claim 1 , wherein the second period has a predetermined time length including the time when the latest time series data is acquired.

11. The anomaly diagnosis system as claimed in claim 1 , wherein a period regarding which a future state measure is estimated using a first approximation formula is longer than a period regarding which a future state measure is estimated using a second approximation formula, the first approximation formula is the approximation formula calculated using the state measure of the time series data acquired for the first period, and the second approximation formula is the approximation formula calculated using a future state measure of the time series data acquired for the second period.

12. The anomaly diagnosis system as claimed in claim 1 , further comprising a filtering processor applying to the state measure a filtering process calculating a maximum value, a minimum value, or a moving average of the state measure in a predetermined time width,
wherein the approximation formula calculator calculates the approximation formula using the state measure to which the filtering process is applied.

13. A method of diagnosing a state of a machine facility, comprising:
a time series data acquiring step that acquires sensor data as time series data from a plurality of sensors installed in the machine facility;
a state measure calculating step that calculates an anomaly measure or a performance measure as a state measure being an index indicating a state of the machine facility by a statistical method in which the time series data is used as learned data, the anomaly measure being an index indicating a magnitude of deviation from a normal state of the machine facility, and the performance measure being an index indicating a performance of the machine facility;
an approximation formula calculating step that calculates variation of the state measure calculated based on the time series data acquired from a past to the present by an approximation formula of a polynomial expression, the approximation formula indicating future approximate of the variation of the state measure;
a state measure estimating step that estimates the state measure until a predetermined future time using the approximation formula;
a reference period setting step that sets a reference period being a period for which the time series data corresponding to the state measure is acquired, the period being referred to calculate the approximation formula in the approximation formula calculating step; and
an outputting step which outputs a diagnose result of the state of the machine facility using the estimated state measure,
wherein in the reference period setting process step, the reference period is set to either of a first period including time when latest time series data is acquired or a second pe

Assignees

• Hitachi Power Solutions Co Ltd

Inventors

• Noda Toujirou
• Chikuma Shigeyoshi
• Kusano Masami
• Miyakoshi Naoki
• Suzuki Tadashi
• Shibuya Hisae

Classifications

• G05B23/0283 Primary

  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) · CPC title

• G05B23/0232

  based on qualitative trend analysis, e.g. system evolution · CPC title