System and method for issue detection of industrial processes

Patent metadata

Field	Value
Publication number	US-2018165384-A1
Application number	US-201615376149-A
Country	US
Kind code	A1
Filing date	Dec 12, 2016
Priority date	Dec 12, 2016
Publication date	Jun 14, 2018
Grant date	—

Abstract

A system is provided comprising a memory configured to store instructions and a processor configured to execute the instructions. The processor is configured to execute the instructions to receive sensor data comprising sensed operations for a machinery, the sensed operations sensed via one or more sensors disposed in the machinery, and to derive a first model matrix based on the sensor data. The processor is further configured to derive a covariance regression model based on the first model matrix, wherein the covariance regression model is configured to be executed to derive a predictive event based on operational machinery data as input.

First claim

1 . A system comprising:
a memory configured to store instructions; and a processor configured to execute the instructions to:
receive sensor data comprising sensed operations for a machinery, the sensed operations sensed via one or more sensors disposed in the machinery;
derive a first model matrix based on the sensor data; and
derive a covariance regression model based on the first model matrix, wherein the covariance regression model is configured to be executed to derive a predictive event based on operational machinery data as input.

2 . The system of claim 1 , wherein the processor is configured to execute the covariance regression model to derive the predictive event for the machinery.

3 . The system of claim 1 , wherein the first model matrix comprises a matrix

X
=

[

X

1
,
1

X

1
,
2

…

X

1
,
p

X

2
,
1

X

2
,
2

…

X

2
,
p

⋮

⋮

⋱

⋮

X

n
m

,
1

X

n
m

,
2

…

X

n
m

,
p

]

wherein each row of the matrix X comprises the sensor data for a different sensor of the one or more sensors, wherein p comprises a number of sensor data points and n m comprises a total number of sensors selected into the matrix X.

4 . The system of claim 3 , wherein the processor is configured to derive a second model matrix from the first model matrix, wherein the second model matrix comprises a higher degree of correlation between sensor data than the first model matrix; and wherein the processor is configured to derive the covariance regression model based on the second model matrix instead of based on the first model matrix.

5 . The system of claim 4 , wherein the processor is configured to derive the second model matrix by applying a Pearson correlation to the matrix X.

6 . The system of claim 4 , wherein the processor is configured to:
derive a third model matrix from the first model matrix, wherein the third model matrix comprises a higher degree of correlation between sensor data than the first model matrix; and derive a second covariance regression model based on the third model matrix, wherein the second covariance regression model is configured to be executed to derive the predictive event based on operational machinery data as input.

7 . The system of claim 6 , wherein the third covariance regression model is configured to be executed to derive a second predictive event based on operational machinery data as input.

8 . The system of 1 , wherein the processor is configured to derive the covariance regression model by executing, via the processor, a Covariance Regression Anomaly detection (CReAD) based on the first model matrix.

9 . The system of claim 8 , wherein executing, via the processor, the CReAD comprises deriving, via the processor, an euclidian distance d i (X i ,x)=√{square root over ((ΔX i,1 ) 2 +(ΔX i,2 ) 2 + . . . +(ΔX i,p ) 2 )} wherein X i comprises the first model matrix and wherein x comprises an observation vector x=[x 1 x 2 . . . x p ] wherein x 1 x 2 . . . x p comprise a plurality of measurements observed via the one or more sensors.

10 . The system of claim 1 , wherein the machinery comprises a gas turbine system and wherein the processor is configured to control the gas turbine system based on the predictive event.

11 . A method, comprising:
receiving, via a processor, sensor data comprising sensed operations for a machinery, the sensed operations sensed via one or more sensors disposed in the machinery; deriving, via the processor, a first model matrix based on the sensor data; and deriving, via the processor, a covariance matrix based on the first model matrix; and deriving, via the processor, a covariance regression model based on the covariance matrix, wherein the covariance regression model is configured to be executed to derive a predictive event based on operational machinery data as input.

12 . The method of claim 11 , comprising executing, via the processor, the covariance regression model to derive the predictive event for machinery.

13 . The method of claim 11 , wherein the first model matrix comprises a matrix

X
=

[

X

Assignees

• Gen Electric

Inventors

• Gandhi Devang Jagdish
• Heda Sanjeev Shyam
• Jiang Xiaomo
• Wile Bradley
• Metz Kevin
• D'Amato James John

Classifications

• G05B23/0254 Primary

  based on a quantitative model, e.g. mathematical relationships between inputs and outputs; functions: observer, Kalman filter, residual calculation, Neural Networks · CPC title

• Y02P90/02

  Total factory control, e.g. smart factories, flexible manufacturing systems [FMS] or integrated manufacturing systems [IMS] · CPC title

• G06F17/5009 Primary

  Physics · mapped topic

• G06F17/16

  Matrix or vector computation {, e.g. matrix-matrix or matrix-vector multiplication, matrix factorization (matrix transposition G06F7/78)} · CPC title

• G06F2217/16

  Physics · mapped topic