Plant abnormality prediction system and method

Patent metadata

Field	Value
Publication number	US-11113360-B2
Application number	US-201816231687-A
Country	US
Kind code	B2
Filing date	Dec 24, 2018
Priority date	Feb 7, 2018
Publication date	Sep 7, 2021
Grant date	Sep 7, 2021

Abstract

A system and method predict whether or not a plant is abnormal and perform an accurate prediction even if a modeling is executed in a state where the understanding for a target to abnormality determination is low, or when a person unfamiliar with system designs a prediction model. The system includes a correlation coefficient calculation unit for calculating a correlation coefficient for each of two tags among a plurality of tags; a relevant tag determination unit for determining a relevant tag for each tag of the plurality of tags by comparing the correlation coefficient with a reference value; and an independent tag determination unit for determining one or more among the plurality of tags as an independent tag based on the relevant tag. The relevant tag determination unit includes primary and second tag extraction sections for extracting primary and second tags for each tag of the plurality of tags.

First claim

What is claimed is:

1. A system for predicting plant abnormality, the system comprising:
a computer processor configured to:
receive a plurality of tags from a plant under operation;
calculate a correlation coefficient for each of two tags among the plurality of tags, the two tags among the plurality of tags being randomly set for input to the computer processor as by an operator unfamiliar with the system;
determine a relevant tag for each tag of the plurality of tags by comparing the correlation coefficient with a reference value, the relevant tag determined by extracting a primary tag for each tag of the plurality of tags and extracting a secondary tag for each tag of the plurality of tags, the extracted primary tag being a tag wherein an absolute value of the correlation coefficient is less than a first upper limit and is not less than a first lower limit;
remove tags from the extracted primary tag for each tag, the removed tags including at least one tag among the secondary tags; and
determine one or more among the plurality of tags as an independent tag based on the relevant tag,
wherein the relevant tags determined by the relevant tag determining are a final set of tags in which some tags have been removed from the extracted primary tag for each tag, the final set of tags determined by
specifying a relevant tag for each tag of the plurality of tags, the specified relevant tags having had tags removed from the extracted primary tag,
specifying a count of the relevant tags, and
determining a specific tag as the independent tag when the count of the relevant tags of the specific tag is less than a specific percentage relative to an entire number of tags, and

wherein the computer processor is further configured to perform k-NN-based prediction by using as an input the relevant tags for the specific tag determined as the independent tag.

2. The system of claim 1 , wherein the extracted secondary tag is a tag wherein an absolute value of the correlation coefficient is less than a second upper limit and is not less than a second lower limit.

3. The system of claim 2 , wherein the first lower limit is less than the second lower limit.

4. The system of claim 2 , wherein the first upper limit is equal to the second upper limit.

5. The system of claim 1 , wherein, for plural secondary tags classified into one or more groups, the removed tags include at least one secondary tag for each group.

6. The system of claim 1 ,
wherein the computer processor is further configured to determine the specific tag as a dependent tag when the count of the relevant tags of the specific tag is not less than the specific percentage relative to the entire number tags, and
wherein the computer processor is further configured to perform one of MLRM-based prediction, ensemble-based prediction, and the k-NN-based prediction by using as an input the relevant tags for the specific tag determined as the dependent tag.

7. A method for predicting plant abnormality in a plant system, the method comprising:
receiving a plurality of tags from a plant under operation;
calculating a correlation coefficient for each of two tags among the plurality of tags;
calculating a correlation coefficient for each of two tags among the plurality of tags, the two tags among the plurality of tags being randomly set for input to the plant system as by an operator unfamiliar with the plant system;
determining a relevant tag for each tag of the plurality of tags by comparing the correlation coefficient with a reference value, the relevant tag determined by extracting a primary tag for each tag of the plurality of tags and extracting a secondary tag for each tag of the plurality of tags, the extracted primary tag being a tag wherein an absolute value of the correlation coefficient is less than a first upper limit and is not less than a first lower limit;
removing tags from the extracted primary tag for each tag, the removed tags including at least one tag among the secondary tags; and
determining one or more among the plurality of tags as an independent tag based on the relevant tag,
wherein the relevant tags determined by the relevant tag determining are a final set of tags in which some tags have been removed from the extracted primary tag for each tag, the final set of tags determined by
specifying a relevant tag for each tag of the plurality of tags, the specified relevant tags having had tags removed from the extracted primary tag,
specifying a count of the relevant tags, and
determining a specific tag as the independent tag when the count of the relevant tags of the specific tag is less than a specific percentage relative to an entire number of tags, and

wherein the method further comprises performing k-NN-based prediction by using as an input the relevant tags for the specific tag determined as the independent tag.

8. The method of claim 7 , wherein the extracted secondary tag is a tag wherein an absolute value of the correlation coefficient is less than a second upper limit and is not less than a second lower limit.

9. The method of claim 8 , wherein the first lower limit is less than the second lower limit.

10. The method of claim 8 , wherein the first upper limit is equal to the second upper limit.

11. The method of claim 7 , wherein, for plural secondary tags classified into one or more groups, the removed tags include at least one secondary tag for each group.

12. The method of claim 7 ,
wherein the independent tag determining determines the specific tag as a dependent tag when the count of the relevant tags of the specific tag is not less than the specific percentage relative to the entire number tags, and
wherein the method further comprises performing one of MLRM-based prediction, ensemble-based prediction, and the k-NN-based prediction by using as an input the relevant tags for the specific tag determined as the dependent tag.

Assignees

• Doosan Heavy Ind & Construction Co Ltd
• Doosan Heavy Ind & Construction C

Inventors

• Kim Hyun Sik
• Park Jee Hun

Classifications

• G06F18/28 Primary

  Determining representative reference patterns, e.g. by averaging or distorting; Generating dictionaries · CPC title

• G06F17/15 Primary

  Correlation function computation {including computation of convolution operations (arithmetic circuits for sum of products per se, e.g. multiply-accumulators G06F7/5443; digital filters, e.g. FIR, IIR, adaptive filters H03H17/00)} · CPC title

• G06F2218/12

  Classification; Matching · CPC title

• G05B23/0218

  characterised by the fault detection method dealing with either existing or incipient faults · CPC title

• G05B23/0235

  based on a comparison with predetermined threshold or range, e.g. "classical methods", carried out during normal operation; threshold adaptation or choice; when or how to compare with the threshold · CPC title