Machine learning-based identification of control algorithm for Chalmers test wind turbine

Electricity generation from wind energy is one of the most promising renewable energy sources, with a rapidly growing share in the power grid. To ensure both optimal power output and safe operation, control systems play a critical role in wind turbines. Chalmers University of Technology conducts research using its own test turbine, located on the island of Björkö of the coast of Gothenburg, which is equipped with numerous sensors. Based on the analysis of measurement data from these sensors, this thesis develops a graphical method to identify the turbine’s control algorithm using a torque–rotor speed diagram. Methods for detecting changes in the control algorithm and estimating applied control parameters are also considered. The applicability of basic machine learning techniques to support these methods is evaluated.

Most welcome!