Prediction and prevention of safety stop of a wind turbine

Patent metadata

Field	Value
Publication number	US-12359652-B2
Application number	US-202318500381-A
Country	US
Kind code	B2
Filing date	Nov 2, 2023
Priority date	Nov 2, 2022
Publication date	Jul 15, 2025
Grant date	Jul 15, 2025

Abstract

Disclosed is a method, performed by an electronic device, for control of operation of a wind turbine. The method comprises obtaining wind turbine data indicative of one or more alerting events of the wind turbine, wherein the wind turbine data comprises warning data indicative of a warning of the wind turbine, and/or alarm data indicative of an alarm state of the wind turbine. The method comprises obtaining sensor data from a plurality of sensors of the wind turbine indicative of operating conditions associated with the wind turbine. The method comprises predicting, based on the wind turbine data and the sensor data, an upcoming safety stop by applying a machine learning model to the wind turbine data and the sensor data. The method may comprise providing, based on the predicted upcoming safety stop, control data indicative of a controlled stop or a derating to be performed by the wind turbine.

First claim

The invention claimed is:

1. A method, performed by an electronic device, for control of operation of a wind turbine, the method comprising:
obtaining wind turbine data indicative of one or more alerting events of the wind turbine, wherein the wind turbine data comprises warning data indicative of a warning state of the wind turbine, and/or alarm data indicative of an alarm state of the wind turbine;
obtaining sensor data from a plurality of sensors of the wind turbine indicative of operating conditions associated with the wind turbine;
predicting, based on the wind turbine data and the sensor data, an upcoming safety stop by applying a machine learning model to the wind turbine data and the sensor data, wherein the safety stop is used when there is structural damage, a risk of structural damage, or a component failure; and
providing, based on the predicted upcoming safety stop, control data indicative of a controlled stop or a derating to be performed by the wind turbine.

2. The method according to claim 1 , wherein the machine learning model is trained based on historical data, wherein the historical data comprises one or more of: historical wind turbine data associated with a plurality of wind turbines, and historical sensor data associated with respective wind turbines of the plurality, wherein the historical wind turbine data is indicative of alerting events of respective wind turbines that led to a safety stop and/or an alarm state.

3. The method according to claim 2 , wherein the machine learning model is trained by determining, based on the historical data, one or more patterns in the historical data that led to the safety stop and/or the alarm state.

4. The method according to claim 3 , wherein determining, based on the historical data, one or more patterns that led to the safety stop and/or the alarm state comprises:
determining, based on the historical data, one or more features that led to the safety stop and/or the alarm state; and
labelling one or more data points of the historical data with an alerting label indicative of an upcoming safety stop.

5. The method according to claim 4 , wherein the one or more features comprise a combination of one or more data points indicative of one or more alarms from one or more predetermined alarm categories, and one or more data points indicative of one or more warnings from one or more predetermined warning categories.

6. The method according to claim 4 , wherein the one or more features comprise a temporal sequence of the data points.

7. The method according to claim 4 , wherein applying the machine learning model to the wind turbine data and the sensor data comprises:
mapping the wind turbine data and the sensor data to the one or more features;
determining, based on the mapping, one or more data sets corresponding to the one or more features; and
classifying the one or more data sets as early indicators of the upcoming safety stop.

8. The method according to claim 1 , wherein the machine learning model comprises a supervised machine learning model.

9. The method according to claim 1 , wherein the machine learning model is a neural network model, and wherein the method comprises training the machine learning model by adjusting one or more weights or other parameters of a mapping function of the neural network model based on historical data.

10. The method according to claim 1 , wherein applying the machine learning model to the wind turbine data and the sensor data comprises providing a likelihood that a safety stop is upcoming, and/or a confidence score associated with the likelihood that a safety stop is upcoming; and wherein predicting, based on the wind turbine data and the sensor data, the upcoming safety stop comprises predicting, based on the likelihood, and/or the confidence score, the upcoming safety stop.

11. The method according to claim 1 , the method comprising:
obtaining a feedback parameter indicative of an actual safety stop; and
updating, based on the feedback parameter, the machine learning model.

12. The method according to claim 1 , wherein the sensor data comprises one or more thresholds.

13. An electronic device comprising a memory circuitry, a processor circuitry, and an interface, wherein the electronic device is configured to control an operation of a wind turbine, the operation comprising:
obtaining wind turbine data indicative of one or more alerting events of the wind turbine, wherein the wind turbine data comprises warning data indicative of a warning state of the wind turbine, and/or alarm data indicative of an alarm state of the wind turbine;
obtaining sensor data from a plurality of sensors of the wind turbine indicative of operating conditions associated with the wind turbine;
predicting, based on the wind turbine data and the sensor data, an upcoming safety stop by applying a machine learning model to the wind turbine data and the sensor data, wherein the safety stop is used when there is structural damage, a risk of structural damage, or a component failure; and
providing, based on the predicted upcoming safety stop, control data indicative of a controlled stop or a derating to be performed by the wind turbine.

14. A wind turbine, comprising:
a tower;
a nacelle disposed on the tower;
a rotor extending from the nacelle and carrying a plurality of blades at one end thereof; and
an electronic device comprising a memory circuitry, a processor circuitry, and an interface, wherein the electronic device is configured to perform an operation, comprising:
obtaining wind turbine data indicative of one or more alerting events of the wind turbine, wherein the wind turbine data comprises warning data indicative of a warning state of the wind turbine, and/or alarm data indicative of an alarm state of the wind turbine;
obtaining sensor data from a plurality of sensors of the wind turbine indicative of operating conditions associated with the wind turbine;
predicting, based on the wind turbine data and the sensor data, an upcoming safety stop by applying a machine learning model to the wind turbine data and the sensor data, wherein the safety stop is used when there is structural damage, a risk of structural damage, or a component failure; and
providing, based on the predicted upcoming safety stop, control data indicative of a controlled stop or a derating to be performed by the wind turbine.

15. A computer readable storage medium storing one or more programs, the one or more programs comprising instructions, which when executed on any combination of one or more processors perform an operation controlling a wind turbine, the operation comprising:
obtaining wind turbine data indicative of one or more alerting events of the wind turbine, wherein the wind turbine data comprises warning data indicative of a warning state of the wind turbine, and/or alarm data indicative of an alarm state of the wind turbine;
obtaining sensor data from a plurality of sensors of the wind turbine indicative of operating conditions associated with the wind turbine;
predicting, based on the wind turbine data and the sensor data, an upcoming safety stop by applying a machine learning model to the wind turbine data and the sensor data, wherein the safety stop is used when there is structural damage, a risk of structural damage, or a component failure; and
providing, based on the predicted upcoming safety stop, control data indicative of a controlled stop or a derating to be performed by the wind turbine.

Assignees

• Vestas Wind Sys As

Inventors

• Thorsen Morten Tim
• Lyhne Casper Hillerup
• Nielsen Johnny
• Skovdal Mikkel Kåe

Classifications

• F05B2270/709

  with neural networks · CPC title

• F05B2270/20

  to optimise the performance of a machine · CPC title

• F03D7/0264

  for stopping; controlling in emergency situations (orientating out of wind F03D7/0208) · CPC title

• Y02E10/72

  Wind turbines with rotation axis in wind direction · CPC title

• F03D17/006

  Estimation methods · CPC title