Wind Turbine Wind Direction Estimation via Deflection Sensor
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current wind direction measurement methods in wind turbines lack accuracy and reliability, and there is a need for a backup system to prevent energy production loss due to sensor failures.
Innovation Solution
A method using a machine learning model trained with deflection and wind direction data from sensors, such as GNSS and wind direction sensors, to estimate wind direction and control the yaw angle of the rotor-nacelle-assembly, with the option to switch to the model's estimates in case of sensor faults, ensuring continuous operation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a primary wind direction sensor is used to control the wind turbine, then the control system can operate with simple architecture, but the system lacks reliability when the sensor develops a fault
Solution Approach 1:
The patent transforms the control approach by changing from direct sensor reading to using a machine learning model that processes multiple sensor parameters (deflection sensor data, wind speed, tower loads) to estimate wind direction. This parameter transformation enables redundant estimation without requiring duplicate primary sensors, thereby improving reliability while managing system complexity.
Solution Approach 2:
The patent introduces a machine learning model as an intermediary between the physical sensors and the control system. This intermediary processes deflection sensor data and other measurements to produce reliable wind direction estimates, acting as a mediator that enhances system reliability without directly increasing hardware complexity.
2Measurement precision
If traditional wind direction sensing is used, then the system structure remains simple, but measurement precision deteriorates during low wind speeds
Solution Approach 1:
The patent makes the deflection sensor serve multiple functions: it not only monitors structural integrity but also provides data for wind direction estimation through the machine learning model. This multi-functionality improves measurement precision across all wind conditions without adding dedicated sensors, thereby managing device complexity while enhancing accuracy.
Solution Approach 2:
The patent replaces the traditional mechanical wind direction sensor with a computational approach using a machine learning model that processes deflection sensor data. This substitution eliminates the mechanical sensing limitations at low wind speeds while using the same structural sensor, improving precision without proportionally increasing complexity.
3Reliability
If a backup wind direction system is implemented, then reliability improves against sensor failures, but the device complexity and cost increase
Solution Approach 1:
The patent enables the existing deflection sensor to serve as its own backup by using its data in the machine learning model for wind direction estimation. The system uses available structural data for dual purposes: structural monitoring and wind direction estimation, creating a self-sufficient backup mechanism that improves reliability without requiring separate backup sensors or systems.
Solution Approach 2:
The patent merges the structural monitoring function and wind direction estimation function into a single integrated system using the deflection sensor and machine learning model. By combining these functions, the patent creates a redundant estimation capability without the complexity of separate backup systems, as the same sensor and computational framework serve multiple purposes.
Data Source
AI summary
Systems and methods for estimating a direction of wind incident on a wind turbine, the wind turbine comprising a tower; a rotor-nacelle-assembly (RNA) carried by the tower; a deflection sensor configured to sense a position of the RNA or a deflection of the tower; and a wind direction sensor. One approach includes: obtaining deflection training data from the deflection sensor; obtaining wind direction training data from the wind direction sensor; training a machine learning model on the basis of the deflection training data and the wind direction training data in order to obtain a trained machine learning model; obtaining further deflection data from the deflection sensor; inputting the further deflection data into the trained machine learning model; and operating the machine learning model to output a wind direction estimate on the basis of the further deflection data.