Statistical trend generator for predictive instrument maintenance

• US 6,453,279 B1
• Filed: 11/05/1999
• Issued: 09/17/2002
• Est. Priority Date: 11/05/1999
• Status: Expired due to Term

• Current Assignee: Schneider Electric Systems USA, Inc. (Schneider Electric SE)
• Original Assignee: The Foxboro Company (Schneider Electric SE)
• Inventors: Prasad, Mohan, DeMoura, David N.
• Primary Examiner(s): Hilten, John S.
• Assistant Examiner(s): VO, HIEN XUAN
• Application Number:
  US09/434,980
• Time in Patent Office: 1,047 Days
• Field of Search: 702/189, 702/116, 702/179, 702/191, 702/60, 702/183, 702/77, 376/215, 376/216, 376/217, 73/240.1, 73/250.3, 714/736, 714/30, 714/745, 714/737, 714/799
• US Class Current: 702/189
• CPC Class Codes: G05B 23/0232 based on qualitative trend analysis, e.g. system evolution
• CPC Class Codes: G05B 23/0283 Predictive maintenance, e.g. involving the monitoring of a system and, based on the monitoring results, taking decisions on the maintenance schedule of the monitored system; Estimating remaining useful life [RUL] preventive maintenance, i.e. planning maintenance according to the available resources without monitoring the system G06Q10/06

Abstract

A method for adaptively scheduling maintenance of a sensor (14) infers the need for maintenance of that sensor on the basis of an error signal indicative of a difference between a measurement signal generated by that sensor and an independently obtained reference signal that is correlated with the measurement signal (16). Although the reference signal is not identical to the measurement signal, the correlation between these two signals assures that when the sensor is functioning, the two signal will track each other over time. The method includes the step of generating this error signal and then monitoring that error signal to detect a triggering event indicative of a mismatch between the reference signal and the measurement signal. This mismatch is indicative of a probability that there exists a malfunction in the sensor. The value of this probability depends on the choice of the triggering event.

Claims

1. A method for detecting deterioration in a measurement sensor comprising

obtaining a measurement signal from the measurement sensor,

obtaining a reference signal from a reference sensor, the reference signal being obtained independently of the measurement signal and being at least partially correlated with the measurement signal,

generating an error signal from the measurement signal and the reference signal, and

monitoring the error signal to detect deterioration in the measurement sensor.

2. The method of claim 1 wherein generating an error signal comprises generating a difference signal representative of a selected difference between the reference signal and the measurement signal.

3. The method of claim 2 wherein generating an error signal further comprises rectifying the difference signal, thereby generating a rectified difference signal.

4. The method of claim 3 wherein generating an error signal further comprises evaluating a time rate-of-change of the rectified difference signal.

5. The method of claim 4 further comprising evaluating a cumulative sum of values of the time rate-of-change of the rectified difference signal.

6. The method of claim 2 wherein generating an error signal further comprises

filtering the difference signal with a first first-order filter having a first time-constant, thereby generating a first filtered output,

filtering the difference signal with a second first-order filter having a second time-constant different from the first time-constant, thereby generating a second filtered output, and

evaluating a difference between the first filtered output and the second filtered output.

7. The method of claim 1 wherein monitoring the error signal comprises accumulating a time sequence of values of the error signal.

8. The method of claim 1 wherein monitoring the error signal comprises detecting a triggering event.

9. The method of claim 8 wherein detecting a triggering event comprises detecting when a rectified difference between the measurement signal and the reference signal exceeds a threshold.

10. The method of claim 8 wherein detecting a triggering event comprises detecting when a time rate-of-change of a rectified difference between the measurement signal and the reference signal exceeds a threshold.

11. The method of claim 8 wherein detecting a triggering event comprises detecting when an accumulated sum of a plurality of values of a time rate-of-change of a rectified difference between the measurement signal and the reference signal exceeds a threshold.

12. The method of claim 6 wherein monitoring the error signal comprises determining when the value of the difference between the first filtered output and the second filtered output exceeds a threshold.

13. The method of claim 8 further comprising generating an alarm signal in response to the detection of the triggering event.

14. The method of claim 8 further comprising

determining a frequency of occurrence of the triggering event, and

generating an alarm signal in response to the frequency of occurrence of the triggering event.

15. The method of claim 8 further comprising generating an alarm signal upon the occurrence of a specified number of triggering events.

16. A system for detecting deterioration in a measurement sensor that generates a measurement signal comprising

a reference sensor for generating a reference signal independent of the measurement signal and at least partially correlated with the measurement signal,

an error signal generator in communication with the reference sensor and the measurement sensor for receiving the reference signal and the measurement signal and generating an error signal therefrom, and

a trend-spotter in communication with the error signal generator for detecting the occurrence of a triggering event on the error signal, the triggering event being selected to be indicative of the probability that the measurement sensor requires maintenance.

17. The system of claim 16 wherein the error signal generator comprises

a differencer to generate a difference signal from the measurement signal and the reference signal, and

a rectifier in communication with the differencer for generating an absolute value of the difference signal.

18. The system of claim 17 further comprising a differentiator in communication with the rectifier for generating a time derivative of the absolute value of the difference signal.

19. The system of claim 18 further comprising a summer in communication with the differentiator for accumulating successive evaluations of a time derivative of the absolute value of the difference signal.

20. The system of claim 17 further comprising

a first low-pass filter having a first time constant, the first low-pass filter being in communication with the differencer for generating a first filtered difference signal,

a second low-pass filter having a second time constant that differs from the first time constant, the second low-pass filter being in communication with the differencer for generating a second filtered difference signal, and

a second differencer in communication with first and second low-pass filters for generating a difference between the first filtered difference signal and the second filtered difference signal.

21. The system of claim 16 further comprising means for accumulating a time of values of the error signal.

22. The system of claim 16 wherein the trend spotter comprises means for detecting a triggering event.

23. The system of claim 22 wherein the means for detecting a triggering event comprises means for detecting when a rectified difference between the measurement signal and the reference signal exceeds a threshold.

24. The system of claim 22 wherein the means for detecting a triggering event comprises means for detecting when a time rate-of-change of a rectified difference between the measurement signal and the reference signal exceeds a threshold.

25. The system of claim 22 wherein the means for detecting a triggering event comprises means for detecting when an accumulated sum of a plurality of values of a time rate-of-change of a rectified difference between the measurement signal and the reference signal exceeds a threshold.

26. The system of claim 20 wherein the trend-spotter comprises means for determining when the value of the difference between the first filtered difference signal and the second filtered difference signal exceeds a threshold.

27. The system of claim 16 wherein the trend-spotter comprises means for generating an alarm signal in response to the occurrence of a triggering event.

28. The system of claim 16 wherein the trend-spotter comprises a counter for determining a frequency of occurrence of triggering event and alarm signal generating means in communication with the counter and for generating an alarm signal in response to the occurrence of a specified frequency of occurrence.