Anomaly detection and anomaly-based control

• US 10,642,262 B2
• Filed: 02/27/2018
• Issued: 05/05/2020
• Est. Priority Date: 02/27/2018
• Status: Active Grant

• Current Assignee: Woodward Incorporated
• Original Assignee: Woodward Incorporated
• Inventors: Shahroudi, Kamran Eftekhari
• Primary Examiner(s): Perez-Velez, Rocio Del Mar
• Assistant Examiner(s): Choi, Alicia M.
• Application Number:
  US15/906,924
• Publication Number:
  US 20190265688A1
• Time in Patent Office: 798 Days
• Field of Search: 700108
• CPC Class Codes: G05B 19/41875 characterised by quality surveillance of production
• CPC Class Codes: G05B 2219/32191 Real time statistical process monitoring
• CPC Class Codes: G05B 2219/32222 Fault, defect detection of origin of fault, defect of product
• CPC Class Codes: G05B 23/024 Quantitative history assessment, e.g. mathematical relationships between available data; Functions therefor; Principal component analysis [PCA]; Partial least square [PLS]; Statistical classifiers, e.g. Bayesian networks, linear regression or correlation analysis; Neural networks
• CPC Class Codes: G05B 23/0254 based on a quantitative model, e.g. mathematical relationships between inputs and outputs; functions: observer, Kalman filter, residual calculation, Neural Networks
• CPC Class Codes: G05B 23/0275 Fault isolation and identification, e.g. classify fault; estimate cause or root of failure
• CPC Class Codes: G05B 23/0286 Modifications to the monitored process, e.g. stopping operation or adapting control
• CPC Class Codes: Y02P 90/02 Total factory control, e.g. smart factories, flexible manufacturing systems [FMS] or integrated manufacturing systems [IMS]

Abstract

A plant control system includes a plant system and a control system controlling the plant system. Runtime conditions of an operating point of the plant control system are received. The runtime conditions include a runtime state of the plant system, a runtime output of the plant system, and a runtime control action applied to the plant system. Reference conditions of a reference point corresponding to the operating point are determined. Stability radius measures of a state difference, an output difference, and a control action difference are computed. One or more of an observability anomaly indicator, health observability indicator, tracking performance anomaly indicator, tracking performance health indicator, controllability anomaly indicator, and controllability health indicator are determined based on respective spectral correlations between two of the stability radius measure of the output difference, the stability radius measure of the state difference, and the stability radius measure of the control action difference.

Claims

1. A plant control system, comprising:

a plant system comprising one or more pieces of equipment; and

a control system controlling the plant system, the control system comprising one or more computers and one or more computer memory devices interoperably coupled with the one or more computers and having tangible, non-transitory, machine-readable media storing one or more instructions that, when executed by the one or more computers, perform operations comprising;

receiving runtime conditions of an operating point of the plant control system, the runtime conditions of the operating point comprising a runtime state of the plant system, a runtime output of the plant system, and a runtime control action applied to the plant system;

determining reference conditions of a reference point corresponding to the operating point, the reference conditions of the reference point comprising a reference state of the plant system, a reference output of the plant system, and a reference control action applied to the plant system;

computing differences between the reference point and the operating point based on the reference conditions and the runtime conditions, the differences comprising;

a state difference between the runtime state and the reference state of the plant system,

an output difference between the runtime output and the reference output of the plant system, and

a control action difference between the runtime control action and the reference control action applied to the plant system;

computing a stability radius measure of the state difference, a stability radius measure of the output difference, and a stability radius measure of the control action difference;

determining an observability anomaly indicator based on a first spectral correlation between the stability radius measure of the output difference and the stability radius measure of the state difference, wherein the first spectral correlation is computed over a first frequency range, wherein the observability anomaly indicator indicates a likelihood of an anomaly occurring in the plant control system with respect to observability, wherein the observability indicates an ability of the control system to estimate a state of the plant system based on a measured output of the plant system;

determining an observability health indicator based on a second spectral correlation between the stability radius measure of the output difference and the stability radius measure of the state difference, wherein the second spectral correlation is computed over a second frequency range, wherein the observability health indicator indicates a likelihood of a normal operation of the plant control system with respect to the observability, and wherein the second frequency range is different from the first frequency range;

determining a tracking performance anomaly indicator based on a third spectral correlation between the stability radius measure of the output difference and the stability radius measure of the control action difference, wherein the third spectral correlation is computed over a third frequency range, wherein the tracking performance anomaly indicator indicates a likelihood of an anomaly occurring in the plant control system with respect to tracking performance, and wherein the tracking performance measures an ability of the control system to track a set point;

determining a tracking performance health indicator based on a fourth spectral correlation between the stability radius measure of the output difference and the stability radius measure of the control action difference, wherein the fourth spectral correlation is computed over a fourth frequency range, wherein the tracking performance health indicator indicates a likelihood of a normal operation of the plant control system with respect to the tracking performance, and wherein the fourth frequency range is different from the third frequency range;

determining a controllability anomaly indicator based on a fifth spectral correlation between the stability radius measure of the control action difference and the stability radius measure of the state difference, wherein the fifth spectral correlation is computed over within a fifth frequency range, wherein the controllability anomaly indicator indicates a likelihood of an anomaly occurring in the plant control system with respect to controllability, and wherein the controllability indicates an ability of the control system to impact a state of the plant system; and

determining a controllability health indicator based on a sixth spectral correlation between the stability radius measure of the control action difference and the stability radius measure of the state difference, wherein the sixth spectral correlation is computed over a sixth frequency range, wherein the controllability health indicator indicates a likelihood of a normal operation of the plant control system with respect to the controllability, and wherein the fifth frequency range is different from the sixth frequency range.

2. The plant control system of claim 1, wherein the plant system further comprises one or more actuators operable to change operations of the one or more pieces of equipment, and the operations further comprising sending instructions to the one or more actuators to dynamically change operations of the one or more pieces of equipment based on at least one of the observability anomaly indicator, the observability health indicator, the tracking performance anomaly indicator, the tracking performance health indicator, the controllability anomaly indicator, or the controllability health indicator.

3. The plant control system of claim 2, wherein:

the plant system comprises a valve;

the one or more actuators comprise a valve actuator that actuates the valve; and

sending instructions to the one or more actuators to dynamically change operations of the one or more pieces of equipment comprises sending instructions to the valve actuator of the plant system to dynamically change a position of the valve of the plant system.

4. The plant control system of claim 1, wherein the stability radius measure of the state difference comprises one or both of a stability radius of the state difference and a dimensionless rate of change of the stability radius of the state difference, and wherein the stability radius of the state difference comprises one or more orders of the rate of change of the stability radius of the state difference.

5. The plant control system of claim 1, wherein the stability radius measure of the output difference comprises one or both of a stability radius of the output difference and a dimensionless rate of change of the stability radius of the output difference, wherein the stability radius of the output difference comprises one or more orders of the rate of change of the stability radius of the output difference.

6. The plant control system of claim 1, wherein the stability radius measure of the control action difference comprises one or both of a stability radius of the control action difference and a dimensionless rate of change of the stability radius of the control action difference, wherein the stability radius of the control action difference comprises one or more orders of the rate of change of the stability radius of the control action difference.

7. The plant control system of claim 1, wherein the spectral correlation comprises a coherence.

8. The plant control system of claim 1, wherein the plant control system is a sub-system of a global system with a distributed control architecture, and the global system comprises at least two subsystems connected in series or in parallel, and wherein the operations are performed by each of the at least two subsystems internally.

9. A control method comprising:

receiving, by a control system of a plant control system that comprises a plant system and the control system controlling the plant system, runtime conditions of an operating point of the plant control system, the runtime conditions of the operating point comprising a runtime state of the plant system, a runtime output of the plant system, and a runtime control action applied to the plant system;

determining, by the control system, reference conditions of a reference point corresponding to the operating point, the reference conditions of the reference point comprising a reference state of the plant system, a reference output of the plant system, and a reference control action applied to the plant system;

computing, by the control system, differences between the reference point and the operating point based on the reference conditions and the runtime conditions, the differences comprising;

a state difference between the runtime state and the reference state of the plant system, and

an output difference between the runtime output and the reference output of the plant system;

computing a stability radius measure of the state difference, and a stability radius measure of the output difference;

determining an observability anomaly indicator based on a first spectral correlation between the stability radius measure of the output difference and the stability radius measure of the state difference, wherein the first spectral correlation is computed over a first frequency range, wherein the observability anomaly indicator indicates a likelihood of an anomaly occurring in the plant control system with respect to observability, wherein the observability indicates an ability of the control system to estimate a state of the plant system based on a measured output of the plant system;

determining an observability health indicator based on a second spectral correlation between the stability radius measure of the output difference and the stability radius measure of the state difference, wherein the second spectral correlation is computed over a second frequency range, wherein the observability health indicator indicates a likelihood of a normal operation of the plant control system with respect to the observability, and wherein the second frequency range is different from the first frequency range; and

sending instructions to an actuator of the plant system to dynamically change operation of the plant system based on the observability anomaly indicator and the observability health indicator.

10. The control method of claim 9, wherein the observability anomaly indicator and the observability health indicator are independent of each other.

11. The control method of claim 9, wherein:

the plant system comprises a valve;

the actuator comprises a valve actuator that actuates the valve; and

sending instructions to an actuator of the plant system to dynamically change operation of the plant system comprises sending the instructions to the valve actuator of the plant system to dynamically change a position of the valve of the plant system.

12. The control method of claim 9, wherein the instructions cause the actuator of the plant system to dynamically change the operation of the plant system to reduce the observability anomaly indicator, increase the observability health indicator, or both.

13. A control method, comprising:

receiving, by a control system of a plant control system that comprises a plant system and the control system controlling the plant system, runtime conditions of an operating point of the plant control system, the runtime conditions of the operating point comprising a plant system and a control system controlling the plant system, the runtime conditions of the operating point comprising a runtime state of the plant system, a runtime output of the plant system, and a runtime control action applied to the plant system;

determining reference conditions of a reference point corresponding to the operating point, the reference conditions of the reference point comprising a reference state of the plant system, a reference output of the plant system, and a reference control action applied to the plant system;

computing differences between the reference point and the operating point based on the reference conditions and the runtime conditions, the differences comprising;

an output difference between the runtime output and the reference output of the plant system, and

a control action difference between the runtime control action and the reference control action applied to the plant system;

computing a stability radius measure of the output difference, and a stability radius measure of the control action difference;

determining a tracking performance anomaly indicator based on a third spectral correlation between the stability radius measure of the output difference and the stability radius measure of the control action difference, wherein the third spectral correlation is computed over a third frequency range, wherein the tracking performance anomaly indicator indicates a likelihood of an anomaly occurring in the plant control system with respect to tracking performance, and wherein the tracking performance measures an ability of the control system to track a set point;

determining a tracking performance health indicator based on a fourth spectral correlation between the stability radius measure of the output difference and the stability radius measure of the control action difference, wherein the fourth spectral correlation is computed over a fourth frequency range, wherein the tracking performance health indicator indicates a likelihood of a normal operation of the plant control system with respect to the tracking performance, and wherein the fourth frequency range is different from the third frequency range; and

sending instructions to an actuator of the plant system to dynamically change operation of the plant system based on the tracking performance anomaly indicator and the tracking performance health indicator.

14. The control method of claim 13, wherein the tracking performance anomaly indicator and the tracking performance health indicator are independent of each other.

15. The control method of claim 13, wherein:

the plant system comprises a valve;

the actuator comprises a valve actuator that actuates the valve; and

sending instructions to an actuator of the plant system to dynamically change operation of the plant system comprises sending the instructions to the valve actuator of the plant system to dynamically change a position of the valve of the plant system.

16. The control method of claim 13, wherein the instructions cause the actuator of the plant system to dynamically change operation of the plant system to decrease the tracking performance anomaly indicator, increase the tracking performance health indicator, or both.

17. A non-transitory, computer-readable medium storing one or more instructions executable by a computer system of a control system to perform operations comprising, the operations comprising:

receiving runtime conditions of an operating point of a plant control system comprising a plant system and a control system controlling the plant system, the runtime conditions of the operating point comprising a runtime state of the plant system, a runtime output of the plant system, and a runtime control action applied to the plant system;

determining reference conditions of a reference point corresponding to the operating point, the reference conditions of the reference point comprising a reference state of the plant system, a reference output of the plant system, and a reference control action applied to the plant system;

computing differences between the reference point and the operating point based on the reference conditions and the runtime conditions, the differences comprising;

a state difference between the runtime state and the reference state of the plant system, and

a control action difference between the runtime control action and the reference control action applied to the plant system;

computing a stability radius measure of the state difference, and a stability radius measure of the control action difference;

determining a controllability anomaly indicator based on a fifth spectral correlation between the stability radius measure of the control action difference and the stability radius measure of the state difference, wherein the fifth spectral correlation is computed over within a fifth frequency range, wherein the controllability anomaly indicator indicates a likelihood of an anomaly occurring in the plant control system with respect to controllability, and wherein the controllability indicates an ability of the control system to impact a state of the plant system;

determining a controllability health indicator based on a sixth spectral correlation between the stability radius measure of the control action difference and the stability radius measure of the state difference, wherein the sixth spectral correlation is computed over a sixth frequency range, wherein the controllability health indicator indicates a likelihood of a normal operation of the plant control system with respect to the controllability, and wherein the fifth frequency range is different from the sixth frequency range; and

sending instructions to an actuator of the plant system to dynamically change operation of the plant system based on the controllability anomaly indicator and the controllability health indicator.

18. The non-transitory, computer-readable medium of claim 17, wherein the controllability anomaly indicator and the controllability health indicator are independent of each other.

19. The non-transitory, computer-readable medium of claim 17, wherein:

the plant system comprises a valve;

the actuator comprises a valve actuator that actuates the valve; and

sending instructions to an actuator of the plant system to dynamically change operation of the plant system comprises sending the instructions to the valve actuator of the plant system to dynamically change a position of the valve of the plant system.

20. The non-transitory, computer-readable medium of claim 17, wherein the instructions cause the actuator of the plant system to dynamically change operation of the plant system to decrease the controllability anomaly indicator, increase the controllability health indicator, or both.