Method and apparatus for detecting compressor stall precursors

• US 20050132712A1
• Filed: 12/23/2003
• Published: 06/23/2005
• Est. Priority Date: 12/23/2003
• Status: Active Grant

• Current Assignee: General Electric Company
• Original Assignee: General Electric Company
• Inventors: Kande, Mallikarjun Shivaraya, Venkateswaran, Narayanan, Krok, Michael Joseph
• Granted Patent: US 7,596,953 B2
• Time in Patent Office: Days
• US Class Current: 60/772
• CPC Class Codes: F04D 27/001 Testing thereof; Determination or simulation of flow characteristics; Stall or surge detection, e.g. condition monitoring
• CPC Class Codes: F05D 2270/101 Compressor surge or stall

Abstract

A method of detecting onset of a gas turbine condition, such as compressor stall, includes receiving data indicative of an operating parameter of a compressor of the gas turbine. The method also includes performing a wavelet transformation on the data to generate wavelet transformed data. The wavelet transformation is configured to affect a processing characteristic regarding a performance of the wavelet transformation. Features indicative of onset of the gas turbine condition in the wavelet transformed data are then identified to provide an indication for controlling the gas turbine to prevent compressor stall from occurring. A system for detecting onset of compressor stall in a gas turbine includes a sensor for providing data indicative of an operating parameter of the compressor and a processor for performing a wavelet transform on the data to identify features of the optimized wavelet transformed data indicative of onset of stall.

Claims

1. A method of detecting onset of a gas turbine condition, which, if left uncorrected, may result in a malfunction of a gas turbine, said method comprising:

receiving data indicative of an operating parameter of a compressor of the gas turbine;

performing a wavelet transformation on the data to generate wavelet transformed data, said wavelet transformation configured to affect a processing characteristic regarding a performance of said wavelet transformation;

generating said wavelet transformed data; and

identifying features in said wavelet transformed data indicative of onset of the gas turbine condition.

2. The method of claim 1, wherein said processing characteristic is selected from the group consisting of processing speed and computational complexity.

3. The method of claim 1, wherein said wavelet transformation comprises truncating at least some of a set of wavelet coefficients generated by said wavelet transformation.

4. The method of claim 3, wherein said truncating comprises eliminating coefficients having a relatively smaller absolute value compared to coefficients having a relatively larger absolute value.

5. The method of claim 1, wherein said wavelet transformed data is selected from the group consisting of wavelet decomposition data and wavelet reconstruction data.

6. The method of claim 5, wherein said wavelet transformation comprises performing at least one level of decomposition to create wavelet decomposition data;

said at least one level of decomposition being performed without reconstructing said wavelet decomposition data.

7. The method of claim 1, wherein said wavelet transformation comprises serially performing component tasks of said wavelet transformation to spread said wavelet transformation out over a time period longer than a time period required to perform said component tasks in parallel.

8. The method of claim 1, wherein said wavelet transformation comprises using only one of the group consisting of wavelet approximation coefficients and wavelet detailed coefficients at each level of said wavelet transformation.

9. The method of claim 1, wherein performing said wavelet transformation on the data further comprises:

partitioning the data into respective data segments; and

sequentially performing said wavelet transformation on said respective data segments.

10. The method of claim 1, further comprising:

receiving additional data indicative of the operating parameter of the compressor of the gas turbine;

mixing said wavelet transformed data with the additional data to create mixed data; and

performing said wavelet transformation on said mixed data to generate wavelet transformed data based on said mixed data.

11. A method of detecting onset of a gas turbine condition, which, if left uncorrected, may result in a malfunction of a gas turbine, said method comprising:

receiving data indicative of an operating parameter of a compressor of the gas turbine;

performing a wavelet decomposition on the data to generate wavelet decomposition data; and

using only said wavelet decomposition data to identify features in said wavelet decomposition data indicative of onset of the gas turbine condition.

12. A method of performing a wavelet transformation on a signal comprising:

receiving data indicative of the signal;

performing a wavelet transformation on the data to generate a set of wavelet transform coefficients;

truncating at least some of said set of wavelet coefficients generated by said wavelet transformation; and

using a remaining some of said set of wavelet coefficients to perform an analysis of the data.

13. A method of performing a wavelet transformation on a signal comprising:

receiving a block of data indicative of the signal; and

serially performing component tasks of said wavelet transformation to spread said wavelet transform out over a time period longer than a time period required to perform said component tasks in parallel.

14. A system for detecting onset of a gas turbine condition, which, if left uncorrected, may result in a malfunction of a gas turbine, said system comprising:

a sensor for providing data indicative of an operating parameter of a compressor of the gas turbine; and

a processor, coupled to the sensor, comprising a first processing module configured to perform a wavelet transformation on the data to generate wavelet transformed data, said wavelet transformation configured to affect a processing characteristic regarding a performance of said transformation, and a second processing module for identifying features of said wavelet transformed data indicative of onset of the gas turbine condition.

15. The system of claim 14, wherein said first processing module is configured to truncate at least some of a set of wavelet coefficients generated by said wavelet transformation.

16. The system of claim 15, wherein said first processing module is further configured to eliminate coefficients having a relatively smaller absolute value compared to coefficients having a relatively larger absolute value.

17. The system of claim 14, wherein said first processing module is configured to perform at least one level of wavelet decomposition to create a wavelet decomposed representation of the data;

said at least one level of wavelet decomposition being performed without reconstructing said wavelet decomposed representation of the data.

18. The system of claim 14, wherein said first processing module is configured to serially perform component tasks of said wavelet transformation to spread said wavelet transformation out over a time period longer than a time period required to perform said component tasks in parallel.

19. The system of claim 14, wherein said first processing module is configured to use only one of the group consisting of wavelet approximation coefficients and wavelet detailed coefficients at each level of said wavelet transformation.

20. The system of claim 14, wherein said first processing module is configured to:

partition the data into respective data segments; and

sequentially perform said wavelet transformation on said respective data segments.

21. The system of claim 14, wherein said first processing module is configured to:

receive additional data indicative of said operating parameter of the compressor of the gas turbine;

mix said wavelet transformed data with the additional data to create mixed data; and

perform said wavelet transformation on said mixed data to generate wavelet transformed data based on said mixed data.