Wind turbine fault detection using acoustic, vibration, and electrical signals

• US 10,495,693 B2
• Filed: 06/01/2017
• Issued: 12/03/2019
• Est. Priority Date: 06/01/2017
• Status: Active Grant

• Current Assignee: GE Infrastructure Technology, LLC (GE Vernova, Inc.)
• Original Assignee: General Electric Company
• Inventors: Unnikrishnan, Jayakrishnan, He, Lijun, Matthews, Brett Alexander, Hao, Liwei
• Primary Examiner(s): Le, John H
• Application Number:
  US15/610,992
• Publication Number:
  US 20180348303A1
• Time in Patent Office: 915 Days
• Field of Search: 702 56, 702183, 702185, 702 48, 702 54, 324 711, 324228, 324522, 32476101, 73579, 73645
• CPC Class Codes: F03D 17/00 Monitoring or testing of wind motors, e.g. diagnostics testing during commissioning of wind motors F03D13/30
• CPC Class Codes: F03D 7/0296 to prevent, counteract or reduce noise emissions
• CPC Class Codes: F05B 2270/333 Noise or sound levels
• CPC Class Codes: F05B 2270/334 Vibration measurements
• CPC Class Codes: G01H 1/003 of rotating machines G01H1/10 takes precedence
• CPC Class Codes: G01H 3/12 by electric means G01H3/14 takes precedence
• CPC Class Codes: G01R 31/34 Testing dynamo-electric machines
• CPC Class Codes: G01R 31/343 in operation
• CPC Class Codes: Y02E 10/72 Wind turbines with rotation axis in wind direction

Abstract

Systems and methods for detecting faults are provided. A method for determining a fault condition for a component of a drivetrain in a wind turbine can include receiving an acoustic signal from an acoustic signal measuring device. The method can further include receiving a vibration signal from a vibration signal measuring device. The method can further include analyzing the acoustic signal to determine an analyzed acoustic signal. The method can further include analyzing the vibration signal to determine an analyzed vibration signal. The method can further include determining a fault condition for the component based at least in part on the analyzed acoustic signal and analyzed vibration signal. The fault condition can further be determined based at least in part on an analyzed electrical signal.

Claims

1. A method for determining a fault condition for a component of a drivetrain in a wind turbine, comprising:

detecting an audible acoustic signal with at least one acoustic signal measuring device positioned proximate to a mechanical component of the wind turbine;

detecting a vibration resulting from a rotation of a wind turbine component with at least one vibration signal measuring device positioned proximate to at least one of a drivetrain, a generator, or other device of the wind turbine,

transmitting the acoustic signal and the vibration signal to a controller;

receiving the acoustic signal from the at least one acoustic signal measuring device;

receiving the vibration signal from the at least one vibration signal measuring device;

analyzing the acoustic signal to determine an analyzed acoustic signal by extracting an envelope from the acoustic signal and estimating a power spectral density for the acoustic signal;

analyzing the vibration signal to determine an analyzed vibration signal by extracting an envelope from the vibration signal and estimating a power spectral density for the vibration signal;

determining a fault condition for the component based at least in part on the analyzed acoustic signal and analyzed vibration signal; and

performing a corrective action for the wind turbine based at least in part on the fault condition.

2. The method of claim 1, wherein analyzing the acoustic signal to determine an analyzed acoustic signal further comprises:

sampling the acoustic signal at a determined sampling rate,

providing the sampled acoustic signal to a high-pass filter, and

providing the filtered acoustic signal to a rectifier; and

wherein analyzing the vibration signal to determine an analyzed vibration signal comprises;

sampling the vibration signal at a determined sampling rate,

providing the sampled vibration signal to a high-pass filter, and

providing the filtered vibration signal to a rectifier.

3. The method of claim 1, wherein determining a fault condition comprises:

comparing the analyzed acoustic signal to a first threshold at a signature fault frequency, and

comparing the analyzed vibration signal to a second threshold at the signature fault frequency.

4. The method of claim 3, wherein a fault condition in the component is determined to exist when the analyzed acoustic signal exceeds the first threshold at the signature fault frequency and the analyzed vibration signal exceeds the second threshold at the signature fault frequency.

5. The method of claim 1, further comprising:

receiving an electrical signal from an electrical signal measuring device; and

analyzing the electrical signal to determine an analyzed electrical signal;

wherein the fault condition for the component is based at least in part on the analyzed electrical signal.

6. The method of claim 5, wherein analyzing each of the acoustic signal, the vibration signal, and the electrical signal comprises:

estimating a power spectral density for each of the acoustic signal, the vibration signal, and the electrical signal.

7. The method of claim 5, wherein analyzing each of the acoustic signal, the vibration signal, and the electrical signal comprises:

extracting one or more fault signature amplitudes for each of the acoustic signal, the vibration signal, and the electrical signal.

8. The method of claim 7, wherein the one or more fault signature amplitudes for each respective signal are extracted at one or more signature fault frequencies based at least in part each respective signal and a respective baseline signal for the wind turbine.

9. The method of claim 5, wherein analyzing each of the acoustic signal, the vibration signal, and the electrical signal comprises:

determining a weighting parameter for each of the acoustic signal, the vibration signal, and the electrical signal; and

wherein the fault condition for the component is determined based at least in part on the weighting parameter for each of the acoustic signal, the vibration signal, and the electrical signal.

10. The method of claim 9, wherein the weighting parameter for each of the acoustic signal, the vibration signal, and the electrical signal is determined based at least in part on a mean value and a variance value of a difference between each respective signal and a respective baseline signal at one or more signature fault frequencies.

11. A control system, comprising:

at least one acoustic signal measurement device positioned proximate to a mechanical component of a wind turbine and configured to detect an audible acoustic signal;

an electrical signal measurement device;

at least one vibration signal measurement device positioned proximate to at least one of a drivetrain, a generator, or other device of the wind turbine and configured to detect a vibration resulting from a rotation of the wind turbine component; and

one or more processors and one or more memory devices configured to store instructions that when executed by the one or more processors cause the one or more processors to perform operations, the operations configured to;

receive first data indicative of an audible acoustic signal from the at least one acoustic signal measurement device;

receive second data indicative of a vibration signal from the at least one vibration signal measurement device;

receive third data indicative of an electrical signal from the electrical signal measurement device;

process the first data, second data, and third data by estimating a power spectral density for each respective data to determine a first processed data, a second processed data, and a third processed data; and

determine a fault condition based at least in part on the first processed data, the second processed data, and the third processed data.

12. The control system of claim 11, wherein the one or more processors are configured to process each of the first data, second data, and third data by:

sampling at least one of the acoustic signal, the vibration signal, or the electrical signal at a determined sampling rate,

providing the sampled signal to a high-pass filter, and

providing the filtered signal to a rectifier.

13. The control system of claim 11, wherein the one or more processors are configured to process each of the first data, second data, and third data by extracting one or more fault signature amplitudes for each respective data.

14. The control system of claim 13, wherein the one or more fault signature amplitudes for each respective data are extracted at one or more signature fault frequencies based at least in part each respective data and data indicative of a respective baseline signal.

15. The control system of claim 11, wherein the one or more processors are configured to process each of the first data, second data, and third data by determining a weighting parameter for each respective data;

and

wherein the fault condition is determined based at least in part on the weighting parameter for each respective data.

16. The control system of claim 15, wherein the weighting parameter for each of the first data, second data, and third data is determined based at least in part on a mean value and a variance value of a difference between each respective data and data indicative of a respective baseline signal at one or more signature frequencies.

17. A wind turbine system, comprising:

a wind turbine comprising a drivetrain, the drivetrain comprising a plurality of components, the plurality of components comprising a plurality of bearings, races, and gears;

at least one acoustic signal measurement device positioned proximate to a mechanical component of the wind turbine and configured to detect an audible acoustic signal;

an electrical signal measurement device;

at least one vibration signal measurement device positioned proximate to at least one of a drivetrain, a generator, or other device of the wind turbine and configured to detect a vibration resulting from a rotation of the wind turbine component; and

one or more processors and one or more memory devices configured to store instructions that when executed by the one or more processors cause the one or more processors to perform operations, the operations comprising;

receive first data indicative of an audible acoustic signal from the acoustic signal measurement device;

receive second data indicative of a vibration signal from the vibration signal measurement device;

receive third data indicative of an electrical signal from the electrical signal measurement device;

process the first data, second data, and third data by estimating a power spectral density and extracting one or more fault signature amplitudes for each respective data to determine a first processed data, a second processed data, and a third processed data; and

determine a fault condition for at least one component in the plurality of components based at least in part on the first processed data, the second processed data, and the third processed data.

18. The wind turbine system of claim 17, wherein the one or more processors are configured to process each of the first data, second data, and third data by:

sampling at least one of the acoustic signal, the vibration signal, or the electrical signal at a determined sampling rate,

providing the sampled signal to a high-pass filter, and

providing the filtered signal to a rectifier.

19. The wind turbine system of claim 18, wherein the one or more fault signature amplitudes for each respective data are extracted at one or more signature fault frequencies based at least in part each respective data and data indicative of a respective baseline signal.

20. The wind turbine system of claim 17, wherein the one or more processors are configured to process each of the first data, second data, and third data by determining a weighting parameter for each respective data;

wherein the fault condition for the component is determined based at least in part on the weighting parameter for each respective data; and

wherein the weighting parameter for each respective data is determined based at least in part on a mean value and a variance value of a difference between each respective data and data indicative of a baseline signal at one or more signature frequencies.