Diagnostic systems and methods for predictive condition monitoring

• US 7,539,597 B2
• Filed: 10/09/2003
• Issued: 05/26/2009
• Est. Priority Date: 04/10/2001
• Status: Expired due to Term

• Current Assignee: SmartSignal Corporation (Emerson Electric Company)
• Original Assignee: SmartSignal Corporation (Emerson Electric Company)
• Inventors: Pipke, R. Matthew, Wegerich, Stephan W., Wolosewicz, Andre
• Primary Examiner(s): TSAI, CAROL S W
• Application Number:
  US10/681,888
• Publication Number:
  US 20040078171A1
• Time in Patent Office: 2,056 Days
• Field of Search: 702/35, 702/81, 702/115, 702/127, 702/182, 702/183, 702/184, 702/185, 702/188, 703/2, 700/108, 700/110, 706/20, 707/104.1
• US Class Current: 702/185
• 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 23/0281 Quantitative, e.g. mathematical distance; Clustering; Neural networks; Statistical analysis
• CPC Class Codes: G06F 18/2433 Single-class perspective, e.g. one-against-all classification; Novelty detection; Outlier detection
• CPC Class Codes: G06F 2218/12 Classification; Matching

Abstract

A system for empirically diagnosing a condition of a monitored system. Estimates of monitored parameters from a model of the system provide residual values that can be analyzed for failure mode signature recognition. Residual values can also be tested for alert (non-zero) conditions, and patterns of alerts thus generated are analyzed for failure mode signature patterns. The system employs a similarity operator for signature recognition and also for parameter estimation. Failure modes are empirically determined, and precursor data is automatically analyzed to determine differentiable signatures for failure modes.

Claims

1. A monitoring apparatus for diagnosing faults in a system, comprising:

a kernel-based non-parametric model responsive to monitored parameter data from said system for generating estimates of the monitored parameter data;

an alert module disposed to produce parameter alerts in response to a comparison of said estimates to said monitored data; and

a failure identification module for identifying an impending failure mode in said system by matching said parameter alerts with at least one reference alert pattern associated with said failure mode.

2. An apparatus according to claim 1, wherein said alert module performs a sequential probability ratio test on at least one parameter.

3. An apparatus according to claim 1, wherein said alert module produces an alert when the difference of an estimate and corresponding monitored data exceeds a selected threshold.

4. An apparatus according to claim 1, wherein said kernel-based non-parametric model employs a Nadaraya-Watson kernel regression.

5. An apparatus according to claim 1, wherein said kernel-based non-parametric model is a similarity-based model.

6. An apparatus according to claim 5, wherein said failure identification module identifies an impending failure mode in said system by matching a combination of parameter alerts occurring together in a given time stamp to a reference alert pattern associated with said failure mode.

7. An apparatus according to claim 5, wherein said failure identification module identifies an impending failure mode in said system by matching a sequence in which at least some of said monitored parameters exhibit alerts over successive time stamps to a reference alert pattern associated with said failure mode.

8. An apparatus according to claim 5, wherein said failure identification module identifies an impending failure mode in said system by matching a combination of parameters that have exhibited alerts cumulatively over a period of time to a reference alert pattern associated with said failure mode.

9. An apparatus according to claim 5, wherein said similarity-based model uses a similarity function in connection with a set of reference observations of normal system behavior, whereby an elemental similarity between a monitored parameter and a corresponding value from said set of reference observations is proportional to the difference of the monitored parameter and the corresponding value, divided by an expected range for the monitored parameter;

and wherein similarity between an observation of monitored parameter data and an observation in said reference set is an average of the elemental similarities of each of said monitored parameters with their respective corresponding values.

10. A method for diagnosing faults in a monitored system, comprising the steps of:

comparing for similarity monitored parameter data from said system to reference parameter data characteristic of known behavior of said system;

generating estimates of the monitored parameter data based on the similarity comparison;

generating alerts in response to a comparison of said estimates to said monitored data;

and

identifying an impending failure mode in said system by matching said parameter alerts with at least one reference alert pattern associated with said failure mode.

11. A method according to claim 10, wherein said alert generating step comprises performing a sequential probability ratio test on at least one parameter.

12. A method according to claim 10, wherein said alert generating step comprises producing an alert when the difference of an estimate and corresponding monitored data exceeds a selected threshold.

13. A method according to claim 10 wherein said identifying step comprises matching a combination of parameter alerts occurring together in a given time stamp to a reference alert pattern associated with said failure mode.

14. A method according to claim 10 wherein said identifying step comprises matching a sequence in which at least some of said monitored parameters exhibit alerts over successive time stamps to a reference alert pattern associated with said failure mode.

15. A method according to claim 10 wherein said identifying step comprises matching a combination of parameters that have exhibited alerts cumulatively over a period of time to a reference alert pattern associated with said failure mode.

16. A method for diagnosing faults in a monitored system, comprising the steps of:

comparing for similarity monitored parameter data from said system to reference parameter data characteristic of known behavior of said system;

generating estimates of the monitored parameter data based on the similarity comparison;

differencing the estimates and the monitored data to generate residual data; and

comparing for similarity the residual data to reference residual data associated with a failure mode, as a diagnostic indication of said failure mode.

17. A method according to claim 16, further comprising the step of communicating a remedial control command to a control program for said system.

18. A method according to claim 16, further comprising the steps of:

recognizing an impending failure mode based on the residual similarity comparison step; and

controlling remedially said system responsive to recognition of an impending failure of said system.

19. A method according to claim 16, wherein said differencing step includes quantizing said residual data, and where said reference residual data is likewise quantized.

20. A method according to claim 18, wherein said step of comparing residual data for similarity comprises forming a residual observation comprised of a residual for each monitored parameter at a given time stamp;

generating a similarity measure using a similarity operator or said residual observation with an observation from said reference residual data associated with a failure mode; and

using said residual similarity measure to indicate the presence or absence of said failure mode.

21. A computer program product for diagnosing faults in a monitored system, comprising computer readable media with computer readable program instructions thereon comprising:

a kernel-based non-parametric model program module for causing a processor to generate estimates of parameter signals from said system, and difference the parameter signals and the estimates to generate residual signals;

a residual testing program module for causing a processor to generate alerts in response to said residual signals characterizing behavior of said system; and

a fault pattern detection program module disposed to cause a processor to receive said alerts and indicate a fault diagnosis in said system upon matching the alerts to a reference pattern associated with said fault.

22. A program product according to claim 21, wherein the residual testing program module performs a sequential probability ratio test on said residuals.

23. A program product according to claim 21, wherein the residual testing program module compares a residual to a threshold to generate an alert when the residual exceeds the threshold.

24. A program product according to claim 21, wherein said kernel-based non-parametric model program module employs a similarity-based model to generate said estimates.

25. A program product according to claim 24 wherein said similarity-based model uses elemental similarities.

26. A program product according to claim 24 wherein said similarity-based model is a radial basis function network.

27. A program product according to claim 21 wherein said kernel-based non-parametric model program module employs a kernel regression model to generate said estimates.

28. A program product according to claim 21, wherein said residual testing program module causes a processor to generate alerts in the form of quantized residuals.

29. A program product according to claim 24 wherein said fault pattern detection program module causes a processor to match a combination of alerts occurring together in a given time stamp to a reference pattern associated with said failure mode.

30. A program product according to claim 24 wherein said fault pattern detection program module causes a processor to match a sequence in which at least some of said monitored parameters exhibit alerts over successive time stamps to a reference pattern alert sequence associated with said failure mode.

31. A program product according to claim 24 wherein said fault pattern detection program module causes a processor to match a combination of parameters that have exhibited alerts cumulatively over a period of time to a reference pattern of cumulative alerts associated with said failure mode.

32. A diagnostic monitoring apparatus for a sensor-instrumented system selected from a process and a machine, comprising:

an operational model module for modeling said system and generating estimates for said sensors in response to receiving actual values of said sensors;

a differencing module for generating residual signals from said estimates and said actual values;

a reference library for storing failure modes and associated residual data values; and

a failure mode recognition engine disposed to compare generated residual signals with said residual data values to select and output a recognized failure mode for said system.

33. A diagnostic monitoring apparatus according to claim 32, wherein said failure mode recognition engine comprises a similarity operation module for generating a similarity score for a comparison of said generated residual signals with said residual data values in said reference library.

34. A diagnostic monitoring apparatus according to claim 33, wherein said failure mode recognition engine further comprises a failure mode decision module responsive to the similarity scores generated by the similarity operation module to select at least one failure mode to output.

36. An apparatus according to claim 32, wherein said differencing module generates residual signals as quantized residuals, and said reference library stores associated residual data values in a likewise quantized fashion.

35. A diagnostic monitoring apparatus for a sensor-instrumented system selected from a process and a machine, comprising:

an operational model module for modeling said system and generating estimates for said sensors in response to receiving actual values of said sensors;

a differencing module for generating residual signals from said estimates and said actual values;

a testing module disposed to receive said residual signals and generate alerts in response thereto;

a reference library for storing failure modes and associated alert signatures; and

a failure mode recognition engine disposed to compare generated alerts with said alert signatures to select and output a recognized failure mode for said system.

37. An apparatus according to claim 35, wherein said reference library stores associated alert signatures as combinations of alerts occurring together in a given time stamp across said sensors.

38. An apparatus according to claim 35, wherein said reference library stores associated alert signatures as sequences of alerts occurring over successive time stamps across said sensors.

39. An apparatus according to claim 35, wherein said reference library stores associated alert signatures as a combination of sensors that have exhibited alerts cumulatively over a period of time.