METHOD AND SYSTEM FOR ANALYSIS OF TURBOMACHINERY

• US 20120226474A1
• Filed: 03/02/2011
• Published: 09/06/2012
• Est. Priority Date: 03/02/2011
• Status: Active Grant

• Current Assignee: GE Infrastructure Technology, LLC (GE Vernova, Inc.)
• Original Assignee: General Electric Company
• Inventors: Trejo Sanchez, Adriana Elizabeth, Adhami, Mohammad Waseem, Vega Paez, Jose Leon, Perez Gamboa, Monica Lizbeth, Chavez Valdovinos, Juan Paulo
• Granted Patent: US 8,996,334 B2
• Time in Patent Office: Days
• US Class Current: 702/179
• CPC Class Codes: G05B 23/0275 Fault isolation and identification, e.g. classify fault; estimate cause or root of failure

Abstract

A method and a system for analyzing turbomachinery is provided. In one embodiment, a system for analyzing turbomachinery is provided. The system includes an intelligent turbomachinery tracking filter (ITTF) system configured to determine one or more performance shifts for one or more components of the turbomachinery based on a plurality of turbomachinery parameters. The system further includes a root cause analyzer configured to determine a root cause of the turbomachinery performance based on the one or more performance shifts. The one or more performance shifts include trended data.

Claims

1. A system for analyzing turbomachinery comprising:

an intelligent turbomachinery tracking filter (ITTF) system configured to determine one or more performance shifts for one or more components of a turbomachinery based on a plurality of the turbomachinery parameters; and

a root cause analyzer configured to determine a root cause of a turbomachinery performance based on the one or more performance shifts, wherein the one or more performance shifts comprise trended data.

2. The system of claim 1, wherein the root cause analyzer comprises a fuzzy logic-based engine configured to process the one or more performance shifts using fuzzy logic.

3. The system of claim 1, wherein the root cause analyzer comprises a physics model configured to model a mechanical behavior of the turbomachinery, a thermodynamic behavior of the turbomachinery, or a combination thereof.

4. The system of claim 1, wherein the root cause analyzer comprises a statistical model configured to model historical trends of the turbomachinery.

5. The system of claim 1, wherein the root cause analyzer comprises a knowledge-based system (KBS) having a plurality of knowledge rules, wherein the plurality of knowledge rules are configured to capture knowledge of a subject matter expert on the turbomachinery.

6. The system of claim 1, wherein the root cause analyzer comprises a trend quantifier configured to quantify trends in the trended data.

7. The system of claim 5, wherein the KBS comprises a forward chaining system, a backward chaining system, or a combination thereof.

8. The system of claim 1, wherein the root cause analyzer is configured to derive a probability of accuracy of the root cause.

9. The system of claim 1, wherein the turbomachinery comprises at least one of a turbine system, a pump, or a compressor.

10. The system of claim 1, wherein the turbomachinery parameters comprise at least one of a temperature, a vibration, a speed, a flow, a pressure, a fuel measure, a pollution measure, a geometry position, a clearance or an actuator position.

11. A method for analyzing a turbomachinery comprising:

measuring a plurality of turbomachinery parameters;

deriving a profile trend based on the plurality of turbomachinery parameters;

quantifying the profile trend to produce a quantified profile trend;

deriving a plurality of fuzzy values based on the quantified profile trend; and

deriving a root cause analysis of the profile trend based on applying a rule and a model to the quantified profile trend.

12. The method of claim 11, wherein deriving the profile trend comprises detecting a shift in the measured parameters relative to an expected value.

13. The method of claim 12, wherein detecting the shift in the measured parameters comprises a Kalman filtering, a tracking filtering, a regression mapping, a neural mapping, an inverse modeling, or a combination thereof, of the measured parameters.

14. The method of claim 11, wherein the quantifying the profile trend to produce a quantified profile trend comprises deriving a slope in the measured parameters, deriving an amount of noise in the measured parameters, deriving a shift in the measured parameters, deriving a maximum value of the measured parameters, deriving a minimum value of the measured parameters, or a combination thereof.

15. The method of claim 11, wherein deriving the plurality of fuzzy values comprises mapping a precise value and assigning a degree of truth.

16. The method of claim 11, wherein applying the rule and the model comprises applying a knowledge based system rule to the quantified profile trend and applying at least one of a physic model or a statistical model to the quantified profile trend.

17. A non-transitory machine readable media, comprising:

instructions configured to process sensor data to identify a shift in turbomachinery performance;

instructions configured to derive a profile trend based on the sensor data having the shift in turbomachinery performance; and

instructions configured to derive a first root cause of the shift in turbomachinery performance by applying fuzzy logic to the profile trend.

18. The non-transitory machine readable media of claim 17, wherein the instructions configured to derive the first root cause of the shift in the turbomachinery performance by applying the fuzzy logic comprise instructions for mapping a precise value to a fuzzy value.

19. The non-transitory machine readable media of claim 17, wherein the instructions configured to derive the first root cause of the shift in the turbomachinery performance by applying the fuzzy logic comprise instructions for applying a knowledge based system rule, a physics model, a statistical model, or a combination thereof.

20. The non-transitory machine readable media of claim 17, comprising instructions configured to derive a second root cause of the shift in turbomachinery performance by applying fuzzy logic to the profile trend, and comprising instructions configured to rank the first and the second root causes by assigning a first probability to the first root cause and a second probability to the second root cause, and ranking the first and the second root causes by comparing the first and second probabilities.