System and method for hybrid risk modeling of turbomachinery

• US 8,712,739 B2
• Filed: 11/19/2010
• Issued: 04/29/2014
• Est. Priority Date: 11/19/2010
• Status: Active Grant

• Current Assignee: GE Infrastructure Technology, LLC (General Electric Company)
• Original Assignee: General Electric Company
• Inventors: Jiang, Xiaomo, Farral, Christopher John, Zou, Tong
• Primary Examiner(s): Silver, David
• Application Number:
  US12/950,891
• Publication Number:
  US 20120130688A1
• Time in Patent Office: 1,257 Days
• Field of Search: None
• US Class Current: 703/2
• CPC Class Codes: F01D 21/003 Arrangements for testing or measuring for measuring vibrations G01H
• CPC Class Codes: F05D 2260/80 Diagnostics

Abstract

Systems and methods are disclosed herein for enhancing turbomachine operations. Such systems and methods include a hybrid risk model. The hybrid risk model includes a physics-based sub model and a statistical sub model. The physics-based sub model is configured to model physical components of a turbomachine. The statistical sub model is configured to model historical information of the turbomachine. The hybrid risk model is configured to calculate a turbomachine parameter.

Claims

1. A system for analyzing turbomachinery comprising:

a processor programmed to;

execute a hybrid risk model comprising a physics-based sub model and a statistical sub model, wherein the physics-based sub model is configured to model physical components of a gas turbine system by using a life parameter (LP) function F (metal temperature T_metal, stress σ

at a location of interest of the gas turbine system, fired hours per start of the gas turbine system)=remaining time before unplanned event occurrence based on a data set, and the LP function F is used by the processor to derive a physics-based maintenance factor (MF) derivation MF=SSF*1/NLP where SSF is a stress scaling factor and NLP is a normalized LP function F, and the statistical sub model is configured to model historical information of a gas turbine unit by calculating an actual fired hours for the gas turbine unit, and wherein the processor is configured to calculate an equivalent fired hour parameter Equivalent FH=MF*FH where FH is the actual fired hours by combining the MF derivation with the actual fired hours and to transform the equivalent fired hour parameter into a probability of retirement of a component of the gas turbine unit by predicting a probability of occurrence of the unplanned event based on a current number of fired hours for the gas turbine unit.

2. The system of claim 1, wherein the processor is configured to determine a remaining operational life of the component by using the probability of retirement.

3. The system of claim 1, wherein the processor is configured to apply a data mining to sensor data acquired for a fleet of the gas turbine class to derive the LP function F by executing a regression analysis comprising a linear or non-linear fit of a plurality of data points included in the sensor data, by classifying the plurality of data points as members of a group having a desired probability of having the remaining time before unplanned event occurrence, or a combination thereof.

4. The system of claim 1, wherein the probability of retirement comprises a lockwire tab retirement probability, an air cooling slot retirement probability, a wheel retirement probability, a blade retirement probability, or a combination thereof.

5. The system of claim 1, wherein the statistical sub model comprises a turbine system component installation history, a turbine system component utilization history, a turbine system fleet utilization history, a plurality of monitoring and diagnosis sensor data, or a combination thereof.

6. The system of claim 5, wherein the statistical sub model comprises a Weibull risk model configured to derive a survival function between a first inspection event of the gas turbine unit and a second inspection event of the gas turbine unit by using an interval censoring approach.

7. The system of claim 1, comprising an asset management system, wherein the asset management system collects turbine system data and uses the hybrid risk model and the collected turbine system data to manage turbine system components.

8. The system of claim 1, comprising a controller having the processor, and wherein the processor is configured to control the gas turbine unit.

9. A non-transient machine readable computer media comprising executable instructions configured to:

retrieve a data correlative of operations of a gas turbine system;

transform the data into an equivalent fired hour Equivalent_FH=MF*FH where FH is an actual fired hour for a gas turbine unit by executing a hybrid risk model comprising a physics-based sub model and a statistical sub model, wherein the physics-based sub model is configured to model physical components of a gas turbine system by using a life parameter (LP) function F (metal temperature T_metal, stress σ

at a location of interest of the gas turbine engine, fired hours per start of the gas turbine engine)=remaining time before unplanned event occurrence based on a data set, and wherein the LP function F is configured to derive a physics-based maintenance factor MF=SSF*1/NLP where SSF is a stress scaling factor and NLP is a normalized LP function F, and the statistical sub model is configured to analyze historical gas turbine unit information by calculating the actual fired hours for the gas turbine unit; and

,

transform the equivalent fired hour parameter into a probability of retirement of a component of the gas turbine unit by predicting a probability of occurrence of the unplanned event based on a current number of fired hours for the gas turbine unit.

10. The computer media of claim 9, comprising executable instructions configured to derive a remaining operational life of the component by using the probability of retirement.

11. The computer media of claim 9, comprising executable instructions configured to analyze sensor data acquired for a fleet of the gas turbine class to derive the LP function.

12. The computer media of claim 9, wherein the probability of retirement comprises a rotor wheel retirement probability.

13. The computer media of claim 12, wherein the rotor wheel retirement probability comprises a first stage wheel retirement probability, a second stage wheel retirement probability, a third stage wheel retirement probability, or combination thereof.

14. The computer media of claim 9, comprising executable instructions configured to predict a cooling air slot cracking, a lockwire tab cracking, a blade cracking or a combination thereof, by executing the hybrid model.

15. The computer media of claim 9, wherein the statistical sub model comprises a turbine system component installation history, a turbine system component utilization history, a turbine system fleet utilization history, a plurality of turbine system sensor data, or a combination thereof.

16. The computer media of claim 9, comprising an asset management system, wherein the asset management system collects turbine system data and uses the hybrid risk model and the collected data to manage turbine system components.

17. A method of creating and using a hybrid risk model comprising:

retrieving a data set correlative of operations of a gas turbine system;

transforming the data set correlative of operations by analyzing physical components of the gas turbine system to obtain a physics-based analysis by deriving a maintenance factor MF=SSF*1/NLP where SSF is a stress scaling factor and NLP is a normalized LP function F, wherein the MF is based on a life parameter (LP) function F (metal temperature T_metal, stress σ

at a location of interest of the gas turbine engine, fired hours per start of the gas turbine engine)=remaining time before unplanned event occurrence based on a data set;

analyzing historical data of a gas turbine unit to obtain the actual fired hours for the gas turbine unit;

integrating the physics-based analysis and the actual fired hours into a hybrid risk model comprising an equivalent fired hour parameter Equivalent_FH=MF*FH where FH is the actual fired hours by combining the MF derivation with the actual fired hours;

and;

transforming the equivalent fired hour parameter into a probability of retirement of a component of the gas turbine unit by predicting a probability of occurrence of the unplanned event based on a current number of fired hours for the gas turbine unit, wherein transforming the data and transforming the equivalent fired hour parameter are performed by a computing device.

18. The method of claim 17, comprising deriving the LP function by applying a data mining to sensor data acquired for a fleet of the gas turbine class.

19. The method of claim 17, comprising identifying a subset of monitoring and diagnosis (M&

D) variables useful in deriving the probability of retirement of the component from a set of M&

D variables based on an M&

D data acquired from a fleet of the gas turbine class by using a quadratic discriminant analysis (QDA) classification and wherein transforming the equivalent fired hour parameter into the probability of retirement of the component comprises combining the equivalent fired hour parameter and the subset of M&

D variables and then transforming the equivalent fired hour parameter and the subset of M&

D variables into the probability of retirement.