Self-healing adaptive controller for industrial asset abnormality accommodation

Patent metadata

Field	Value
Publication number	US-12504725-B2
Application number	US-202318353486-A
Country	US
Kind code	B2
Filing date	Jul 17, 2023
Priority date	Jul 17, 2023
Publication date	Dec 23, 2025
Grant date	Dec 23, 2025

Abstract

A training data store may contain training data associated with monitoring node values during normal operation of an industrial asset and simulated abnormal data. An offline model tuning platform accesses the training data from normal operation of the industrial asset and the simulated abnormal data in the training data store. Based on the training data from normal operation of the industrial asset, the simulated abnormal data, an abnormal operating condition, and a constrained optimization solution, controller tuning parameters are created for at least one tuned data-driven adaptive controller such that an operating condition of the industrial asset will move from the abnormal operating condition to a normal operation condition through a stable trajectory. An online monitoring platform receives a stream of current monitoring node values and, when the abnormal operating condition is detected, utilizes the controller tuning parameters to implement the at least one tuned data-driven adaptive controller.

First claim

The invention claimed is:

1 . A system to protect an industrial asset, comprising:
a training data store containing training data associated with monitoring node values during normal operation of the industrial asset and simulated abnormal data; an offline model tuning platform, coupled to the training data store, to:
access the training data from normal operation of the industrial asset and the simulated abnormal data in the training data store, and
based on the training data from normal operation of the industrial asset, the simulated abnormal data, an abnormal operating condition, and a constrained optimization solution, create controller tuning parameters for at least one tuned data-driven adaptive controller such that an operating condition of the industrial asset will move from the abnormal operating condition to a normal operation condition through a stable trajectory;
a controller reference library to store the tuning parameters created by the offline model tuning platform; and an online monitoring platform, coupled to the controller reference library, to receive a stream of current monitoring node values and, when the abnormal operating condition is detected, utilize the controller tuning parameters to implement the at least one tuned data-driven adaptive controller.

2 . The system of claim 1 , wherein the tuned data-driven adaptive controller comprises a closed loop system that resembles a desired stable reference model.

3 . The system of claim 1 , wherein the simulated abnormal data is associated with at least one of a physics-based model and a data-driven model.

4 . The system of claim 1 , wherein the simulated abnormal data is associated with a Generative Adversarial Network (“GAN”).

5 . The system of claim 1 , wherein the operating condition is associated with a normal space and an abnormal space separated by a decision boundary.

6 . The system of claim 5 , wherein the abnormal space is divided into a plurality of abnormal zones each associated with a different stable closed loop model reference.

7 . The system of claim 5 , wherein the normal space is divided into a plurality of normal zones each associated with a different achievable normal operating condition.

8 . The system of claim 1 , wherein the constrained optimization solution is represented by an optimization problem given by:

ρ
s

=

argmin
ρ

⁢

J
⁡
(
ρ
)

,

s
.
t
.

δ
⁡
(
ρ
)

≤

δ
N

J
⁡
(
ρ
)

=

‖
⁡
(

1
-
M

)

⁢

(

M
-

K
⁡
(
ρ
)

⁢

(

1
-
M

)

⁢
G
⁢

‖
2

δ
⁡
(
ρ
)

=

‖
⁢

M
S

-

K
⁡
(
ρ
)

⁢

(

1
-

M
S

)

⁢
G
⁢

‖
2

where G represents an open loop plant, M represents a model reference, K represents stabilizing for a controller of the plant G, K(ρ) is the controller with control parameter ρ, and M S is the closed loop plant stabilized by controller K S .

9 . The system of claim 1 , wherein the online monitoring platform decides between a resilient estimation accommodation and the tuned data-driven adaptive controller accommodation.

10 . The system of claim 1 , wherein the received stream of current monitoring node values is associated with a Supervisory Control And Data Acquisition (“SCADA”) data stream.

11 . The system of claim 1 , wherein the monitoring nodes include at least one of: (i) a sensor node, (ii) a critical sensor node, (iii) an actuator node, (iv) a controller node, and (v) a key software node.

12 . The system of claim 1 , wherein the abnormal operating condition is associated with at least one of a cyber-attack and a fault.

13 . The system of claim 1 , wherein the industrial asset is associated with at least one of: (i) a turbine, (ii) a gas turbine, (iii) a wind turbine, (iv) an engine, (v) a jet engine, (vi) a locomotive engine, (vii) a refinery, (viii) a power grid, and (ix) an autonomous vehicle.

14 . A computerized method to protect an industrial asset, comprising:
accessing, by a computer processor of an offline model tuning platform, training data from normal operation of the industrial asset and simulated abnormal data from a training data store; based on the training data from normal operation of the industrial asset, the simulated abnormal data, an abnormal operating condition, and a constrained optimization solution, creating controller tuning parameters for at least one tuned data-driven adaptive controller such that an operating condition of the industrial asset will move from the abnormal operating condition to a normal operation condition

Assignees

• Gen Electric

Inventors

• Achanta Hema K
• Abbaszadeh Masoud

Classifications

• G05B13/0265 Primary

  the criterion being a learning criterion · CPC title