MS#02.1 Data-driven flow modeling and control of wind farms

M. ABKAR¹, G.V. IUNGO², X. YANG³
¹ Aarhus University|² University of Texas at Dallas|³ Chinese Academy of Sciences

Wind farms and wakes

With the rapid expansion of wind farms and the availability of large datasets on turbine performance, loads, atmospheric conditions, and wakes, innovative data-driven modeling techniques and machine learning offer new and enhanced capabilities to predict wind-farm flows and optimize wind-farm operations. However, the complexity of fluid flows in wind farms poses unique challenges for data-driven modeling techniques, which are often developed for simplified, canonical problems. These challenges include the high-dimensional, multiscale nature of turbulence, geophysical and atmospheric effects, interactions among neighboring turbines and wind farms, and generally undisclosed design and control features of wind turbines.

Ultimately, effective data-driven models for wind-farm flows should provide a certain degree of interpretability, explainability, and generalizability, while complying with known first principles that govern underlying physical processes. This mini-symposium aims to foster a platform for exchanging insights and experiences in this rapidly evolving field. The primary goal is to address the current challenges and promote the advancement of data-driven methods as robust and reliable tools for wind-farm flow modeling and control.