Wind Turbine Backup Control Using Learned Setpoint Deltas
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Wind turbines face challenges in developing backup controllers due to the risk of cyberattacks, as model-based controllers require extensive domain knowledge and cannot be easily accessed, making it difficult to create a reliable backup system.
Innovation Solution
A learning-based backup controller that observes and learns control actions from the supervisory controller under normal operation, allowing it to take over and adjust setpoints by predicting deltas to maintain operation until the supervisory controller is available again, using machine learning algorithms like neural networks to handle model/parameter mismatches.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a model-based controller is used for wind turbine control, then control precision is improved, but device complexity and difficulty of creating backup controller increase
Solution Approach 1:
The patent creates a backup controller that copies the control behavior of the primary model-based controller by observing its inputs and outputs during normal operation. Instead of requiring a separate complex dynamic model, the backup controller learns to replicate the primary controller's decision-making process through machine learning, thereby reducing the complexity burden while maintaining control precision.
Solution Approach 2:
The backup controller is trained in advance during normal operation by observing the primary controller's inputs and outputs. This preliminary learning phase allows the backup controller to be ready immediately when needed, eliminating the need for complex real-time model derivation during emergency situations.
2Reliability
If a model-based backup controller is developed, then reliability is improved, but ease of manufacture deteriorates due to inaccessible dynamic models
Solution Approach 1:
Instead of requiring access to the primary controller's dynamic model for manufacturing purposes, the patent uses machine learning to copy the controller's behavioral patterns from operational data. This approach makes the backup controller manufacturable without needing proprietary model information, while still achieving reliable backup control functionality.
Solution Approach 2:
The backup controller learns control strategies by observing and analyzing the primary controller's own operational data. This self-service approach allows the system to generate its own training data from normal operations, eliminating the need for external model provision while ensuring reliability.
3Productivity
If the wind turbine control system is connected to network, then productivity is improved, but risk of cyberattack increases
Solution Approach 1:
The patent implements a backup controller that is trained and ready in advance to cushion against potential cyberattacks. When the primary networked controller is compromised, the pre-trained backup controller can immediately take over, providing a safety buffer that protects the system while maintaining productivity.
Solution Approach 2:
The backup controller acts as an intermediary between the networked primary controller and the wind turbine control actuators. During normal operation, it learns from the primary controller; during cyberattacks, it mediates control commands to ensure safe operation, thereby protecting the system while maintaining network connectivity benefits.
Data Source
AI summary
A method for providing backup control for a supervisory controller of at least one wind turbine includes observing, via a learning-based backup controller of the at least one wind turbine, at least one operating parameter of the supervisory controller under normal operation. The method also includes learning, via the learning-based backup controller, one or more control actions of the at least one wind turbine based on the operating parameter(s). Further, the method includes receiving, via the learning-based backup controller, an indication that the supervisory controller is unavailable to continue the normal operation. Upon receipt of the indication, the method includes controlling, via the learning-based backup controller, the wind turbine(s) using the learned one or more control actions until the supervisory controller becomes available again. Moreover, the control action(s) defines a delta that one or more setpoints of the wind turbine(s) should be adjusted by to achieve a desired outcome.