Wind Turbine MPC With Staged Control Outputs
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Model predictive control (MPC) methods for wind turbines face high computational demands due to the need for frequent re-solving of optimization problems, especially when a high resolution of blade azimuthal angle is required, leading to short time stages and increased computational requirements.
Innovation Solution
The method involves using a penalty parameter to constrain control outputs across multiple time stages, allowing only the initial set of control outputs to be implemented, with subsequent outputs penalized for deviation, and varying the length of time stages based on available computational resources, reducing peak computational requirements by implementing control outputs for a subset of time stages rather than just the initial stage.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If model predictive control is used to maximize power output and minimize loading, then wind turbine performance is improved, but computational requirements increase significantly
Solution Approach 1:
The prediction horizon is divided into multiple time stages, allowing the control problem to be solved in segments. Only control outputs for a subset of time stages are implemented, reducing the frequency of full optimization problem resolution while maintaining control effectiveness across the entire horizon.
Solution Approach 2:
Control outputs are determined for multiple future time stages in advance during the optimization process. These preliminary control outputs for subsequent time stages are then implemented without re-solving the optimization, reducing computational load while maintaining predictive control benefits.
2Measurement precision
If the sampling time is reduced to improve control resolution, then control accuracy is improved, but the time available to solve the optimization problem decreases
Solution Approach 1:
The time horizon is segmented into multiple stages with different implementation strategies. By implementing control outputs for multiple stages rather than re-solving at every sampling interval, the system maintains high control resolution while extending the effective solution time.
Solution Approach 2:
The control strategy maintains continuous control action by implementing predetermined control outputs for multiple time stages. This eliminates gaps in control while reducing the frequency of computationally intensive optimization solves, ensuring continuous useful control action without excessive computational burden.
3Measurement precision
If individual pitch control of blades is implemented with high azimuthal angle resolution, then blade control precision is improved, but the first time stage becomes shorter and computational demands increase
Solution Approach 1:
The control horizon for individual blade pitch control is divided into multiple time stages. By determining control outputs for several stages and implementing them sequentially without re-solving, the system maintains high azimuthal angle resolution for precise blade control while reducing the frequency of optimization solves and associated computational demands.
Solution Approach 2:
Pitch control commands for multiple future time stages are calculated in advance during the optimization process. These preliminary control actions for subsequent stages are implemented without re-solving the optimization problem, maintaining high control precision for individual blades while reducing computational burden.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
The invention provides a method for controlling a wind turbine. The method predicts behaviour of the wind turbine components for the time stages over a prediction horizon using a wind turbine model describing dynamics of the wind turbine, where the time stages include a first set of time stages from an initial time stage and a second set of time stages subsequent to the first set. The method determines control outputs, e.g. individual blade pitch, for time stages based on the predicted behaviour. The method then transmits a control signal to implement only the control outputs for each of the second set of time stages so as to control the wind turbine. Advantageously, the invention reduces both average and peak computational loads relative to standard predictive control algorithms.