Active grid turbulence generators that utilize rotating or flapping winglets can achieve higher turbulence levels in a controlled manner. Most active grid studies are experimental, and active grids in those were designed based on empirical data and calculations based on order-of-magnitude analysis. We used large-eddy simulations (LES) to assess the effects of square and circular winglet shapes and motion protocol. Differences in generated turbulence at distances of up to 5 mesh sizes from the active grid in the flow direction were analyzed in terms of turbulence intensity, Taylor-scale Reynolds number, spectra, isotropy level, and flow homogeneity. For these analyses, a two-plane active grid design was used to prevent full closure during operation in a wind tunnel. The numerical simulation’s validity was assessed using a combination of numerical benchmarks, physical parameters, and experimental validation, thereby facilitating the application of Large Eddy Simulation to the design of active grids. Rather than employing either random or prescribed motion protocols, this study developed an optimization method to determine the transient angular positions of the winglets. The optimization focuses on two objective functions: achieving a specified area-closure intensity and maintaining a target mean Rossby number. Simulations for the square wiglets were conducted at a constant intensity of area closure at two mean Rossby numbers, and one simulation for a circular winglet at the same lowest Rossby number and the same intensity of area closure. Five mesh size downstream of the active-grid, circular winglets generated approximately twice the Taylor-scale Reynolds number (Re_ 375) compared to square winglets (Re_ 180) at the same mean Rossby number. However, this increase in turbulence level was accompanied by higher levels of anisotropy and inhomogeneity in the turbulent kinetic energy and mean velocity fields. The downstream set of rods was found to be responsible for the higher velocity fluctuations in the lateral direction perpendicular to their rotation axis, both for square and circular grids. The global isotropy level for circular winglets showed a deviation of 25–30% from unity, which is significantly greater than that observed in square winglet configurations. The results revealed a clear trade-off between increasing turbulence level and preserving flow homogeneity and isotropy. Additionally, the implementation of a two-plane active grid system, combined with different winglet geometries and an optimization-based motion protocol, demonstrated significant potential for effectively modulating these flow characteristics.
Akardere et al. (Wed,) studied this question.