Case Study: W2E Wind to Energy reduces wind turbine structural loads with MathWorks machine learning-enabled model predictive control

Reducing Structural Loads on Wind Turbines with Machine Learning and Model Predictive Control

The customer, W2E Wind to Energy GmbH, needed to minimize long-term structural wear and damage on wind turbines to optimize operational efficiency, a challenge that was difficult to manage with classical control strategies. They partnered with MathWorks, using Model Predictive Control Toolbox, MATLAB, and Simulink, to explore the development of a more advanced controller.

MathWorks provided the tools for a model-based design solution, enabling the engineers to integrate a machine learning model into a model predictive controller that proactively adjusts turbine settings to reduce loads. The solution, tested on a full-scale 3 MW turbine, successfully reduced the dynamics of the thrust force at higher wind speeds, demonstrating a promising path to decreasing structural wear while maintaining power output.

W2E Wind to Energy GmbH

Andreas Klein

W2E Wind to Energy GmbH