Posters

Abstract

Like in many other domains, machine learning methods have had a tremendous impact on wind energy research in recent years. In wind turbine power curve modeling, for example, deep ANNs have shown significant advantages over conventional methods which translates into improvements of downstream tasks, such as performance assessment and forecasting or condition monitoring. However, despite their impressive predictive abilities, ANNs are often criticized as opaque black-boxes with no physical understanding of the system they describe. At the same time, Explainable AI (XAI) has developed into a major research area within machine learning offering a wide range of methods that increase transparency and interpretability of nonlinear, data-driven models. We will apply Shapley Values to analyse the strategies learned by deep ANNs trained on SCADA data from an onshore wind farm. Subsequently, we will present the benefits of the gained insights for power-curve modeling and related applications. These include the validation of models regarding their compliance with physical intuition as well as the explanation of deviations from the nominal power curve.