System, method and computer program for pattern based intelligent control, monitoring and automation

• US 9,423,793 B2
• Filed: 12/23/2009
• Issued: 08/23/2016
• Est. Priority Date: 12/23/2008
• Status: Active Grant

• Current Assignee: Andrew Wong, Chung Lam Li
• Original Assignee: Andrew Wong, Chung Lam Li
• Inventors: Wong, Andrew, Li, Chung Lam
• Primary Examiner(s): Chaki, Kakali
• Assistant Examiner(s): Seck, Ababacar
• Application Number:
  US13/141,944
• Publication Number:
  US 20110320388A1
• Time in Patent Office: 2,435 Days
• Field of Search: 706/12
• US Class Current: 1/1
• CPC Class Codes: G05B 23/0232 based on qualitative trend analysis, e.g. system evolution
• CPC Class Codes: G05B 23/0278 Qualitative, e.g. if-then rules; Fuzzy logic; Lookup tables; Symptomatic search; FMEA
• CPC Class Codes: G06F 18/214 Generating training patterns; Bootstrap methods, e.g. bagging or boosting
• CPC Class Codes: G06F 18/2411 based on the proximity to a decision surface, e.g. support vector machines
• CPC Class Codes: G06N 20/00 Machine learning
• CPC Class Codes: G06N 5/025 Extracting rules from data
• CPC Class Codes: G06N 7/01 Probabilistic graphical models, e.g. probabilistic networks

Abstract

The present invention relates to control, monitoring, and automation. The present invention more specifically relates to pattern-based intelligent control, monitoring and automation. The invention performs pattern-based monitoring. It collects signal data from one or more signals. The signal data define signal data streams. It then transforms each of the signal data streams into trends. It also discovers patterns based on the trends within each signal data stream and/or across the signal data streams. The patterns are optionally used for diagnostics and root cause analysis, online plant monitoring and operation control, plant optimization, and other environments where a causal link or correlation may exist between related inputs, states and/or outputs.

Claims

1. A system for monitoring an industrial process, the system comprising:

a data collection utility for receiving a plurality of signal data streams from one or more sensors, each signal data stream comprising a series of signal data points representing at least one of an input, an output, and a state associated with a component of the industrial process; and

a data mining utility for determining a series of trends for each of the plurality of the signal data streams, each trend representing a change between at least two signal data points in the respective signal data stream, and further determining at least one pattern in the plurality of the signal data streams based on the series of trends;

wherein the at least one pattern comprises a first sequence of trends representing a first event associated with the industrial process, a second sequence of trends representing a second event associated with the industrial process, and a time delay between the first event and the second event; and

wherein the data mining utility is configured to predict an occurrence of the second event using the determined at least one pattern.

2. The system of claim 1, wherein the data mining utility is configured to determine the at least one pattern based on a detected correlation between the first and second sequences of trends in the plurality of signal data streams.

3. The system of claim 2, wherein the correlation between the first and second sequences of trends is detected based on the time delay.

4. The system of claim 3, wherein the first sequence of trends and the second sequence of trends are detected by the data mining utility to be correlated if:

each of the first and the second sequences of trends occurs at least twice, and that each time the occurrence of the second sequence of trends is delayed from the respective occurrence of the first sequence of trends by the time delay.

5. The system of claim 2, wherein the data mining utility is configured to determine the at least one pattern based on a pre-determined value representing a threshold of statistical interdependence between the first and second sequences of trends in the plurality of signal data streams.

6. The system of claim 1, further comprising a user interface for receiving user inputs and for controlling the industrial process, using the series of trends of each of the plurality of signal data streams and the at least one pattern.

7. The system of claim 6, wherein the user interface is configured to generate one or more of:

alert, log and report.

8. The system of claim 1, wherein the data collection utility receives the series of signal data points from the one or more sensors at predetermined time intervals and each of the first and second sequences of trends is defined by a plurality of signal data points at consecutive time intervals.

9. The system of claim 1, wherein the trend representing a change between the at least two signal data points in the respective signal data stream comprises a slope defined by the at least two signal data points.

10. The system of claim 1, wherein the trend comprises one of:

up, down, and flat.

11. A method for monitoring an industrial process, the method comprising:

receiving a plurality of signal data streams from one or more sensors, each signal data stream comprising a series of signal data points representing at least one of an input, an output, and a state associated with a component of the industrial process;

determining a series of trends for each of the plurality of the signal data streams, each trend representing a change between at least two signal data points in the respective signal data stream; and

determining at least one pattern in the plurality of the signal data streams based on the series of trends, wherein the at least one pattern comprises a first sequence of trends representing a first event associated with the industrial process, a second sequence of trends representing a second event associated with the industrial process, and a time delay between the first event and the second event; and

predicting an occurrence of the second event using the determined at least one pattern.

12. The method of claim 11, wherein the at least one pattern is determined based on a detected correlation between the first and second sequences of trends in the plurality of signal data streams.

13. The method of claim 12, wherein the correlation between the first and second sequences of trends is detected based on the time delay.

14. The method of claim 13, wherein the first sequence of trends and the second sequence of trends are detected by the data mining utility to be correlated if:

each of the first and the second sequences of trends occurs at least twice, and that each time the occurrence of the second sequence of trends is delayed from the respective occurrence of the first sequence of trends by the time delay.

15. The method of claim 12, wherein the at least one pattern is determined based on a pre-determined value representing a threshold of statistical interdependence between the first and second sequences of trends in the plurality of signal data streams.

16. The method of claim 11, further comprising receiving user inputs and controlling the industrial process, using the series of trends of each of the plurality of signal data streams and the at least one pattern.

17. The method of claim 16, further comprising generating one or more of:

alert, log and report.

18. The method of claim 11, wherein the series of signal data points is received from the one or more sensors at predetermined time intervals and each of the first and second sequences of trends is defined by a plurality of signal data points at consecutive time intervals.

19. The method of claim 11, wherein the trend representing a change between the at least two signal data points in the respective signal data stream comprises a slope defined by the at least two signal data points.

20. The method of claim 11, wherein the trend comprises one of:

up, down, and flat.