Fusion for modeling wake effects on wind turbines

Wind turbine wakes cause power reduction and structural loading. A data-driven approach is proposed for wake modeling to provide the azimuth, angular spread, and intensity of wakes. We introduce three wind speed difference definitions for wake analysis. The wake identification is automated using morphological imaging operators. The wake pattern is complicated by multiple neighboring turbines. Four fusion schemes are proposed to draw a complete picture. A similarity based clustering enhances the final fusion result by treating clusters, each with specific features, equally.