Turbulence structure in open channel flow with partially covered artificial emergent vegetation

An innovative approach was recently proposed to increase water depth and improve navigation in rivers by increasing overall roughness through an adaptive placement of artificial vegetation. In this approach, a section of the river width is covered with artificial vegetation constructed from an array of dampers that can be added or removed while maintaining sufficient width for navigational purposes. The goal of the work here is to explore the mean flow and turbulent momentum transport properties of such flow in a controlled flume setting, where the flow rate is steady, the vegetation placement is uniform, and the Froude number is maintained at sub-critical conditions. The focus is on vegetation-induced roughness alterations to the bulk flow statistics as well as the structure of turbulence impacted by the presence of such artificially engineered vegetation. The canopy flow experimentally explored here differs from the well-studied planar uniform emergent canopy flow because the interface between the vegetation and non-vegetation areas occurs along a lateral direction instead of the vertical direction. The flume experiments were conducted by partially covering the channel width with emergent artificial vegetation that is flexible and can interact with the flow. The quantities studied are the equilibration distance of the flow statistics from the leading vegetation edge as such distance is essential to any engineering design. Moreover, the mean flow and turbulence regimes at several longitudinal sections were analyzed using spectral analysis and conditional sampling to identify the main vortical structures and their contribution to momentum transport. Spectral analysis demonstrated that the sizes of the dominant vortices can be classified into stem-scale (small-scale) and shear-scale (larger scale) vortices as conventional in canonical canopy flows. The shear layer formed in the planar direction and its associated coherent vortices are spawned near the edge of the vegetated area in a manner analogous to a plane mixing layer. Quadrant analysis showed two contributions to the overall shear stress and lateral momentum exchange between the vegetated and non-vegetated zone, namely, ejections and sweeps. The measured ejection-sweep contributions are short-lived but significant to the overall turbulent momentum transport consistent with other canopy flow studies. Approximately 80% of the turbulent stress contributions at the interface occur within 30% of the time. Because the flume experiments were conducted at sufficiently high Reynolds number and sufficiently low Froude number, the scalability of the experimental findings here to much larger settings encountered in engineering practice are briefly discussed.

Duke Scholars

Author Gabriel G. Katul Civil and Environmental Engineering