Apparatus and amendment of wind turbine blade impact detection and analysis

1. A first apparatus comprising:

a first circuitry to capture and store a series of images or video of an object struck by a blade of a wind turbine;

a memory to store the series of images or video from the first circuitry;

one or more sensors to continuously sense vibration of the blade or for acoustic recordings;

a second circuitry to analyze a data stream in real-time from the one or more sensors and/or the stored series of images or video, and to apply machine-learning or statistical analysis to the data stream from the one or more sensors, wherein the machine-learning or the statistical analysis identifies a time instance of when the blade struck the object despite the object not leaving a surface defect on the blade; and

a communication interface to transmit data from the second circuitry to a second apparatus, wherein the second apparatus is in a cloud, and wherein the second circuitry is to:

pre-process the data stream with continuous wavelet transform (CWT);

generate a time marginal integration (TMI) graph by calculating an energy distribution in the CWT with respect to time; and

extract features from the TMI graph.

2. The first apparatus of claim 1 , wherein the second circuitry is to determine when the blade likely struck the object.

3. The first apparatus of claim 1 , wherein the second circuitry is to save a portion of the data stream that indicates when the blade likely struck the object.

4. The first apparatus of claim 3 , wherein the second circuitry is to discard another portion of the data stream that does not indicate when the blade likely struck the object.

5. The first apparatus of claim 1 , wherein the first circuitry includes one or more of: a visual camera, or an infrared camera, and wherein the visual camera or the infrared camera are installed on the blade of the wind turbine.

6. The first apparatus of claim 1 , wherein the first circuitry is to reuse memory space to store the series of images or video of the object struck by the blade.

7. The first apparatus of claim 1 comprises a battery to power the first circuitry, the second circuitry, the memory, the one or more sensors, and the communication interface.

8. The first apparatus of claim 1 comprises a power supply connected with a power system in a hub of the wind turbine.

9. The first apparatus of claim 1 , wherein the one or more sensors include one or more of: a 3-axis accelerometer, a piezoelectric contact microphone, gyroscope, or one or more sensors for acoustic recordings.

10. The first apparatus of claim 1 , wherein the communication interface includes one of Bluetooth, Wi-Fi, 5G, or LTE.

11. The first apparatus of claim 1 , wherein the one or more sensors includes a vibration sensor, which is to trigger the first circuitry to capture an image of the series of images.

12. A non-transitory machine-readable storage medium having machine-readable storage instructions that, when executed, cause one or more machines to perform a method comprising:

capturing a series of images or video of an object struck by a blade of a wind turbine;

storing the series of images or video in a memory;

continuously sensing, by one or more sensors, vibration of the blade or for acoustic recordings;

analyzing a data stream from the one or more sensors and/or the stored series of images or video by applying machine-learning or statistical analysis to the data stream from the one or more sensors;

identifying from the analyzed data stream and the applied machine-learning or statistical analysis when the blade struck the object, wherein the machine-learning or the statistical analysis identifies a time instance of when the blade struck the object despite the object not leaving a surface defect on the blade; and

transmitting a portion of the data stream to a device, wherein the device is to record an image of the object, wherein analyzing the data stream from the one or more sensors by applying machine-learning to control sensitivity of the one or more sensors, comprises:

pre-processing the data stream with continuous wavelet transform (CWT);

generating a time marginal integration (TMI) graph by calculating an energy distribution in the CWT with respect to time; and

extracting features from the TMI graph.

13. The non-transitory machine-readable storage medium of claim 12 having further machine-readable storage instructions that, when executed, cause the one or more machines to perform the method comprising:

discarding another portion of the data stream that does not indicate when the blade likely struck the object; and

reusing memory space to store the series of images or video of the object.

14. The non-transitory machine-readable storage medium of claim 12 , wherein transmitting the portion of the data stream to the device is via one of Bluetooth, Wi-Fi, 5G, or LTE.

15. A method comprising:

capturing a series of images or video of an object struck by a blade of a wind turbine;

storing the series of images or video in a memory;

continuously sensing, by one or more sensors, vibration of the blade or for acoustic recordings;

analyzing a data stream from the one or more sensors and/or the stored images or video by applying machine-learning or statistical analysis to the data stream from the one or more sensors;

identifying from the analyzed data stream and the applied machine-learning or the statistical analysis when the blade struck the object, wherein the machine-learning or the statistical analysis identifies a time instance of when the blade struck the blade despite the object not leaving a surface defect on the blade; and

transmitting a portion of the data stream to a device, wherein analyzing the data stream from the one or more sensors by applying machine-learning to control sensitivity of the one or more sensors, comprises:

pre-processing the data stream with continuous wavelet transform (CWT);

generating a time marginal integration (TMI) graph by calculating an energy distribution in the CWT with respect to time; and

extracting features from the TMI graph.

16. The method of claim 15 comprising discarding another portion of the data stream that does not indicate when the blade likely struck the object.

17. The method of claim 15 , wherein the one or more sensors include one or more of: a 3-axis accelerometer, a piezoelectric contact microphone, gyroscope, one or more sensors for acoustic recordings, or a vibration sensor, which is to trigger capturing of an image of the series of images.