System and method for monitoring health of airfoils

• US 8,718,953 B2
• Filed: 04/28/2011
• Issued: 05/06/2014
• Est. Priority Date: 04/28/2011
• Status: Active Grant

• Current Assignee: GE Infrastructure Technology, LLC (GE Vernova, Inc.)
• Original Assignee: General Electric Company
• Inventors: Rajagopalan, Venkatesh, Badami, Vivek Venugopal, Prabhu, Rahul Srinivas, Behera, Ajay Kumar, Bhattacharya, Aninda
• Primary Examiner(s): Breene, John
• Assistant Examiner(s): BLOSS, STEPHANIE E
• Application Number:
  US13/096,244
• Publication Number:
  US 20120278004A1
• Time in Patent Office: 1,104 Days
• Field of Search: 702/39, 702/34, 702/35, 702/85, 702/184, 731/120.1
• US Class Current: 702/39
• CPC Class Codes: F01D 17/02 Arrangement of sensing elements
• CPC Class Codes: F01D 5/005 Repairing methods or devices
• CPC Class Codes: F05D 2260/83 Testing, e.g. methods, components or tools therefor

Abstract

A method for monitoring the health of one or more blades is presented. The method includes the steps of generating a signal representative of delta times of arrival corresponding to the rotating blade, generating a reconstructed signal by decomposing the signal representative of the delta times of arrival utilizing a multi-resolution analysis technique, wherein the reconstructed signal is representative of at least one of static deflection and dynamic deflection in the rotating blade.

Claims

1. A method for monitoring the health of a rotating blade, comprising:

selecting an appropriate wavelet based upon a signal representative of delta times of arrival corresponding to a rotating blade;

selecting a final decomposition level;

generating approximation coefficients and detailed coefficients utilizing a multi-resolution analysis technique and said appropriate wavelet until the final decomposition level is achieved;

equating the detailed coefficients to zero; and

generating a reconstructed signal from the approximation coefficients;

wherein the reconstructed signal is representative of static deflection in the rotating blade, and

wherein the multi-resolution analysis technique comprises a plurality of decomposition levels and each of said decomposition levels comprises at least one low pass filter and at least one high pass filter to generate the approximation coefficients and the detailed coefficients, respectively.

2. The method of claim 1, wherein generating the signal representative of the delta times of arrival comprises:

determining actual times of arrival of the rotating blade;

determining expected time of arrival of the rotating blade; and

determining delta times of arrival by subtracting the actual times of arrival from the expected time of arrival.

3. The method of claim 2, wherein the expected time of arrival of the rotating blade is a mean of respective actual times of arrival of one or more rotating blades in a turbine.

4. The method of claim 1, further comprising determining dynamic deflection of the rotating blade by subtracting the reconstructed signal representative of the static deflection from the signal representative of the delta times of arrival.

5. A method for monitoring the health of a rotating blade, comprising:

generating a signal representative of filtered delta times of arrival corresponding to the rotating blade;

selecting an appropriate wavelet based upon the signal representative of the filtered delta times of arrival;

selecting a final decomposition level;

decomposing the signal representative of the filtered delta times of arrival utilizing a multi-resolution analysis technique and the appropriate wavelet until the final decomposition level is achieved to generate approximation coefficients and detailed coefficients; and

generating a reconstructed signal utilizing the approximation coefficients, wherein the reconstructed signal is representative of static deflection in the rotating blade,

wherein the multi-resolution analysis technique comprises a plurality of decomposition levels and each of said decomposition levels comprises at least one low pass filter and at least one high pass filter to generate the approximation coefficients and the detailed coefficients, respectively.

6. The method of claim 5, further comprising determining a dynamic deflection in the rotating blade by subtracting the reconstructed signal from the signal representative of the filtered delta times of arrival.

7. The method of claim 5, wherein generating the signal representative of the filtered delta times of arrival comprises:

determining actual times of arrival of the rotating blade;

determining expected time of arrival of the rotating blade;

generating a signal representative of delta times of arrival by subtracting each of the actual times of arrival corresponding to the rotating blade from the expected time of arrival; and

filtering the delta times of arrival to generate the signal representative of the filtered delta times of arrival.

8. The method of claim 5, wherein the appropriate wavelet is an orthogonal wavelet or a bi-orthogonal wavelet, and has a compact support.

9. The method of claim 5, wherein the decomposition level is selected based upon the filtered delta times of arrival, a signal to noise ratio of the signal representative of the filtered delta times of arrival, and the length of delta times of arrival data.

10. The method of claim 7, wherein the expected time of arrival of the rotating blade is a mean of respective actual times of arrival of one or more rotating blades in a turbine.

11. A system, comprising a processing subsystem that:

generates a signal representative of delta times of arrival corresponding to a rotating blade based upon actual times of arrival of the rotating blade;

selects an appropriate wavelet based upon the signal representative of the delta times of arrival;

selects a final decomposition level;

decomposes the signal representative of the delta times of arrival utilizing a multi-resolution analysis technique and the appropriate wavelet until the final decomposition level is achieved to generate approximation coefficients and detailed coefficients; and

generates a reconstructed signal utilizing the approximation coefficients, wherein the reconstructed signal is representative of static deflection in the rotating blade,

wherein the multi-resolution analysis technique comprises a plurality of decomposition levels and each of said decomposition levels comprises at least one low pass filter and at least one high pass filter to generate the approximation coefficients and the detailed coefficients, respectively.

12. The system of claim 11 further comprising one or more sensors to generate signals that are representative of the actual times of arrival of the rotating blade.

13. The system of claim 11, further comprising an operator that selects the appropriate wavelet and the decomposition level.

14. The system of claim 11, wherein the appropriate wavelet is an orthogonal wavelet or a bi-orthogonal wavelet, and has a compact support.

15. The system of claim 11, further comprising at least one data repository that stores static deflection, delta times of arrival, actual times of arrival, intermediate results, or combinations thereof.

16. A non-transitory computer readable medium for a blade health monitoring system encoded with a program to instruct a computer to:

generate a signal representative of delta times of arrival corresponding to a plurality of rotating blades based upon actual times of arrival of the plurality of rotating blades;

select an appropriate wavelet in the signal representative of the delta times of arrival;

select a final decomposition level;

decompose the signal representative of the delta times of arrival utilizing a multi-resolution analysis technique and the appropriate wavelet until the final decomposition level is achieved to generate approximation coefficients and detailed coefficients; and

generate a reconstructed signal utilizing the approximation coefficients, wherein the reconstructed signal is representative of at least one of static deflection and dynamic deflection in the plurality of rotating blades,

wherein the multi-resolution analysis technique comprises a plurality of decomposition levels and each of said decomposition levels comprises at least one low pass filter and at least one high pass filter to generate the approximation coefficients and the detailed coefficients, respectively.

17. A method for monitoring the health of a rotating blade, comprising:

selecting an appropriate wavelet based upon a signal representative of delta times of arrival corresponding to a rotating blade, and selecting a final decomposition level;

generating approximation coefficients and detailed coefficients utilizing a multi-resolution analysis technique and an appropriate wavelet until the final decomposition level is achieved;

equating the approximation coefficients to zero; and

generating a reconstructed signal representative of dynamic deflection from the detailed coefficients;

wherein the reconstructed signal is representative of dynamic deflection in the rotating blade, and

wherein the multi-resolution analysis technique comprises a plurality of decomposition levels and each of said decomposition levels comprises at least one low pass filter and at least one high pass filter to generate the approximation coefficients and the detailed coefficients, respectively.