LAUSR

ABSTRACT Understanding and quantifying the wake effect plays an important role in improving wind turbine designs and operations, as well as wind farm layout planning. The majority of the current wake effect models are physics based, but these models have a number of shortcomings. Sophisticated models based on computational fluid dynamics suffer from computational limitations and are impractical for modeling commercial-sized wind farms, whereas simplified physics-based models are generally inaccurate for wake effect quantification. Nowadays, data-driven wake effect models are gaining popularity as the data from commercially operating wind turbines become available, but this development is still in its early stages. This study contributes to the general category of data-driven wake effect modeling that makes use of actual wind turbine operational data. We propose a wake effect model based on splines with physical constraints incorporated, which sets out to estimate wake effect characteristics such as wake width and wake depth under single-wake situations. Our model is one of the first data-driven models that provides a detailed account of the wake effect. Prediction accuracy of the proposed spline model, when compared with other alternatives, also confirms the benefit of incorporating the physical constraints in the statistical estimation.