ABSTRACT
This chapter discusses the abundance of wind farm operational data motivates the development of data science methods for analyzing and estimating wake effect. Understanding and quantifying the wake effect plays an important role in improving wind turbine designs and operations as well as wind farm layout planning. Hwangbo's et al. model is intended to estimate wake effect characteristics, such as wake width and wake depth, under single-wake situations arising between two turbines of which modeling assumptions are easier to justify. The chapter presents a study that quantifies annual wake power loss in actual wind turbine operations. Gaussian Markov random field (GMRF) is designed to perform well with more turbines since it benefits from the spatial modeling of multiple turbines at different locations. The GMRF model could be used to estimate the wake power loss indirectly by taking the difference of the maximum fitted value among all turbines and the power output fitted to a specific turbine.
