Method for learning latest data considering external influences in early warning system and system for same

Patent metadata

Field	Value
Publication number	US-11276300-B2
Application number	US-201716328645-A
Country	US
Kind code	B2
Filing date	Aug 29, 2017
Priority date	Aug 29, 2016
Publication date	Mar 15, 2022
Grant date	Mar 15, 2022

Abstract

The present invention provides a method for learning latest data considering external influences in an early warning system, and the early warning system for same. The method for learning latest data considering external influences comprises the steps of: an early warning processing device categorizing device monitored variables according to external influences; and the early warning processing device differently applying a pattern learning method for each of the categorized monitored variables.

First claim

The invention claimed is:

1. A method for learning latest data considering external influence in an early warning system, comprising:
categorizing, performed by an early warning processing device, machine monitoring variables according to external environment influence; and
applying, performed by the early warning processing device, pattern learning method differently for each of the categorized monitoring variables,
wherein the step of categorizing according to the external environment influence includes:
calculating a degree of correlation between each of the machine monitoring variables and each of external environmental factors; and
categorizing the machine monitoring variables according to the calculated degree of correlation value,
wherein the categorizing the machine monitoring variables according to the calculated degree of correlation value includes:
according to a first criterion and a second criterion that an operator designates in advance, wherein the first criterion has greater value than the second criterion, when the degree of correlation value is equal to or more than the first criterion, categorizing the machine monitoring variables as a first group,
when the degree of correlation value is from equal to or more than the second criterion to equal to or less than the first criterion, categorizing the machine monitoring variables as a second group, and
when the degree of correlation value is equal to or less than the second criterion, categorizing the machine monitoring variables as a third group.

2. The method for learning latest data of claim 1 , wherein the external environmental factors include at least one of air temperature, air pressure, humidity and sea water temperature.

3. The method for learning latest data of claim 1 , wherein the applying the pattern learning method differently applies at least one of an automatic relearning method, a manual relearning method and a relearning unnecessary method to each of the monitoring variables of the first group, the monitoring variables of the second group and the monitoring variables of the third group.

4. The method for learning latest data of claim 3 , wherein the monitoring variables of the first group are stored in an automatic relearning database, and wherein a latest pattern learning is automatically performed for the external environmental factor and the monitoring variables of the first group, when difference between a change of the external environmental factor and a learning pattern constructed in advance is a predetermined level or more.

5. The method for learning latest data of claim 3 , wherein a latest pattern learning is manually relearned by an operator for the monitoring variables of the second group, when the early warning is determined to be from the external environmental factor when the early warning is generated.

6. The method for learning latest data of claim 3 , wherein the relearning is not performed for the monitoring variables of the third group.

7. An early warning system using a method for learning latest data considering external influence, comprising:
an early warning processing device configured to perform:
categorizing machine monitoring variables according to external environment influence; and
applying pattern learning method differently for each of the categorized monitoring variables,
wherein the categorizing according to the external environment influence includes:
calculating a degree of correlation between each of the machine monitoring variables and each of external environmental factors; and
categorizing the machine monitoring variables according to the calculated degree of correlation value,
wherein the categorizing according to the calculated degree of correlation value includes:
according to a first criterion and a second criterion that an operator designates in advance, wherein the first criterion has greater value than the second criterion, when the degree of correlation value is equal to or more than the first criterion, categorizing the machine monitoring variables as a first group,
when the degree of correlation value is from equal to or more than the second criterion to equal to or less than the first criterion, categorizing the machine monitoring variables as a second group, and
when the degree of correlation value is equal to or less than the second criterion, categorizing the machine monitoring variables as a third group.

8. The early warning system of claim 7 , wherein the external environmental factors include at least one of air temperature, air pressure, humidity and sea water temperature.

9. The early warning system of claim 7 , wherein the applying the pattern learning method differently applies at least one of an automatic relearning method, a manual relearning method and a relearning unnecessary method to each of monitoring variables greatly influenced by the external environmental factor, the monitoring variables of which influence from the external environmental factor is unclear, and the monitoring variables not influenced by the external environmental factor.

10. The early warning system of claim 9 , wherein the monitoring variables of the first group are stored in an automatic relearning database, and wherein a latest pattern learning is automatically performed for the external environmental factor and the monitoring variables of the first group, when difference between a change of the external environmental factor and a learning pattern constructed in advance is a predetermined level or more.

11. The early warning system of claim 9 , wherein a latest pattern learning is manually relearned by an operator for the monitoring variables of the second group, when the early warning is determined to be from the external environmental factor when the early warning is generated.

12. The early warning system of claim 9 , wherein the relearning is not performed for the monitoring variables of the third group.

Assignees

• Korea Hydro & Nuclear Power Co

Inventors

• Kim Dae Woong
• Choi Nam Woo
• Kim Bum Nyun
• Kim Hyoung Kyun
• Kim Ji In
• Kang Ju Hyung
• Park Yong Hoon
• Shin Gun Woong
• Kim Yang Seok
• Park Young Sheop
• Park Chi Yong
• Lee Byoung Oh
• Kim Jong Seog

Classifications

• G08B21/187 Primary

  Machine fault alarms · CPC title

• G08B31/00 Primary

  Predictive alarm systems characterised by extrapolation or other computation using updated historic data · CPC title

• G08B21/18

  Status alarms (G08B21/02 takes precedence) · CPC title

• G06N20/00

  Machine learning · CPC title

• G08B25/14

  Central alarm receiver or annunciator arrangements · CPC title