System and method for predicting component failures in large systems

• US 20060111857A1
• Filed: 11/23/2004
• Published: 05/25/2006
• Est. Priority Date: 11/23/2004
• Status: Active Grant

• Current Assignee: General Electric Company
• Original Assignee: General Electric Company
• Inventors: Bonissone, Piero Patrone, Rajiv, Vrinda, Osborn, Mark David, Asati, Mahesh Kumar, Shah, Rasiklal Punjalal
• Granted Patent: US 7,103,509 B2
• Time in Patent Office: Days
• US Class Current: 702/85
• CPC Class Codes: G01D 1/00 Measuring arrangements giving results other than momentary value of variable, of general application G01D3/00 takes precedence; in tariff metering apparatus G01D4/00; transducers not specially adapted for a specific variable G01D5/00
• CPC Class Codes: G01D 3/08 with provision for safeguarding the apparatus, e.g. against abnormal operation, against breakdown

Abstract

A method for predicting a time to failure of a component in a system is presented. The method comprises obtaining a set of data measurements related to the component. The set of data measurements are representative of a plurality of parameters including a plurality of leading parameters. The method comprises generating a prediction model based upon the leading parameters considered in combination. The prediction model is then used to predict the time to failure of the component based on a set of real-time measurements, wherein the plurality of parameters are processed to predict the time to failure for the component. Finally, a confidence level for the predicted time to failure is determined based upon the plurality of parameters.

Claims

1. A method for predicting a time to failure of a component in a system, the method comprising:

obtaining a set of data measurements related to the component, wherein the set of data measurements are representative of a plurality of parameters, including a plurality of leading parameters;

generating a prediction model based upon the leading parameters considered in combination;

using the prediction model to predict the time to failure of the component based on a set of real-time data measurements, wherein the plurality of parameters are processed to predict the time to failure for the component; and

determining a confidence level for the predicted time to failure based upon the plurality of parameters.

2. The method of claim 1, wherein the system is a superconducting magnet system.

3. The method of claim 2, wherein the leading parameters represent at least one of a heater duty cycle, a vessel pressure and a shield temperature associated with the superconducting magnet system.

4. The method of claim 3, further comprising one or more additional parameters, wherein the additonal parameters represent at least one of a water flow parameter and a water temperature parameter associated with the superconducting magnet system.

5. The method of claim 1, wherein generating a prediction model comprises:

processing the set of data measurements based on a set of rules related to the plurality of parameters and one or more failure modes related to the component;

applying a neuro-fuzzy prognostic technique to the set of processed data measurements to generate the prediction model; and

evaluating the generated prediction model based upon a portion of data measurements;

6. The method of claim 5, wherein the set of rules specify an acceptable rate of change for the plurality of parameters.

7. The method of claim 5, wherein the failure mode is an indication of impending failure related to the component.

8. The method of claim 5, further comprising accessing and processing the set of real-time data measurements.

9. The method of claim 8, comprising applying the set of rules to the set of real-time data measurements to generate an early alert signal for an impending failure mode related to the component.

10. The method of claim 1, further comprising automatically generating a service notification for service of the component based upon the failure prediction.

11. A system for predicting a time to failure for a component, the system comprising:

a data acquisition system, configured to obtain a set of data measurements related to the component, wherein the set of data measurements are representative of a plurality of parameters, including a plurality of leading parameters;

a training subsystem, configured to generate a prediction model based upon the leading parameters considered in combination; and

a runtime system configured to use the prediction model generated by the training system to predict the time to failure of the component based on a set of real-time measurements, wherein the plurality of parameters are processed to predict the time to failure for the component and a confidence level for the predicted time to failure.

12. The system of claim 11, wherein the training subsystem further comprises:

a failure identification module configured to process the set of data measurements from the data acquisition system, based on a set of failure modes related to the component;

a rule definition module configured to process the set of data measurements from the data acquisition system, based on a set of rules for the plurality of parameters, wherein the set of rules specify an acceptable rate of change for the plurality of parameters;

a prediction model configured to receive the set of processed data measurements from the failure identification module and the rule definition module and apply a neuro-fuzzy prognostic technique to the set of data measurements to predict a time to failure for the component; and

a validation engine configured to evaluate the generated prediction model based upon a portion of data measurements.

13. The system of claim 12, wherein the runtime subsystem further comprises a data analysis module, a rule definition execution module and a prediction model execution module.

14. The system of claim 13, wherein the data analysis module is configured to acquire and process the set of real-time data measurements.

15. The system of claim 13, wherein the rule definition execution module is configured to apply the set of rules from the rule definition module to the set of real-time data measurements to generate an early alert signal for an impending failure mode related to the component.

16. The system of claim 11, further comprising automatically generating a service notification for service of the component based upon the failure prediction.

17. The system of claim 13, wherein the prediction model execution module is configured to execute the generated prediction model to predict the time to failure of the component based on the set of real-time data measurements.

18. The system of claim 17, wherein the rule definition execution module is further configured to generate the confidence level for the predicted time to failure.

19. The system of claim 11, wherein the component is part of a superconducting magnet system.

20. The system of claim 19, wherein the leading parameters represent at least one of a heater duty cycle, a vessel pressure and a shield temperature associated with the superconducting magnet system.

21. The system of claim 19, further comprising one or more additional parameters, wherein the additonal parameters represent at least one of a water flow parameter and a water temperature parameter associated with the superconducting magnet system.

22. A method for predicting a time to failure of a coldhead component in a superconducting magnet system, the method comprising:

obtaining a set of data measurements related to the coldhead component, wherein the set of data measurements are representative of a plurality of parameters, and wherein the plurality of parameters include at least one of a heater duty cycle, a vessel pressure and a shield temperature associated with the coldhead component;

generating a prediction model based upon the leading parameters considered in combination;

using the prediction model to predict the time to failure of the coldhead component based on a set of real-time data measurements, wherein the plurality of parameters are processed to predict the time to failure for the component; and

determining a confidence level for the predicted time to failure based upon the plurality of parameters.

23. At least one computer-readable medium storing computer instructions for instructing a computer system for predicting a time to failure of a component in a system, the computer instructions comprising:

obtaining a set of data measurements related to the component, wherein the set of data measurements are representative of a plurality of parameters, including a plurality of leading parameters;

generating a prediction model based upon the leading parameters considered in combination;

using the prediction model to predict the time to failure of the component based on a set of real-time data measurements, wherein the plurality of parameters are processed to predict the time to failure for the component; and

determining a confidence level for the predicted time to failure based upon the plurality of parameters.

24. At least one computer-readable medium storing computer instructions for instructing a computer system for predicting a time to failure of a coldhead component in a superconducting magnet system, the computer instructuons comprising:

obtaining a set of data measurements related to the coldhead component, wherein the set of data measurements are representative of a plurality of parameters, and wherein the plurality of parameters include at least one of a heater duty cycle, a vessel pressure and a shield temperature associated with the coldhead component;

generating a prediction model based upon the plurality of parameters considered in combination;

using the prediction model to predict the time to failure of the coldhead component based on a set of real-time data measurements, wherein the plurality of parameters are processed to predict the time to failure for the component; and

determining a confidence level for the predicted time to failure based upon the plurality of parameters.