Thesis (Ph. D.)--University of Washington, 1996. / Vita. Includes bibliographical references (leaves [147]-157).

Vortex identification techniques are used to analyze the flow structure in a 4 x 3 array of scale model wind turbines. Q-criterion, Δ-criterion, and λ2-criterion are applied to Particle Image Velocimetry data gathered fore and aft of the last row centerline turbine. Q-criterion and λ2-criterion provide a clear indication of regions where vortical activity exists while the Δ-criterion does not. Galilean decomposition, Reynolds decomposition, vorticity, and swirling strength are used to further understand the location and behavior of the vortices. The techniques identify and display the high magnitude vortices in high shear zones resulting from the blade tips. Using Galilean and Reynolds decomposition, swirling motions are shown enveloping vortex regions in agreement with the identification criteria. The Galilean decompositions are 20% and 50% of a convective velocity of 7 m/s. As the vortices convect downstream, these vortices weaken in magnitude to approximately 25% of those present in the near wake. A high level of vortex activity is visualized as a result of the top tip of the wind turbine blade; the location where the highest vertical entrainment commences.