Turbulent Flows Over Macro-Scale Roughness Elements - From Biofouling Barnacles to Urban Canopies
MetadataShow full item record
Turbulent boundary layers developing over rough surfaces are frequently encountered in a variety of engineering and geophysical applications. Accurate modeling of boundary layer flow over such rough surfaces remains a challenge; direct numerical simulation (DNS) and wall-resolved large-eddy simulation (LES) are both computationally prohibitive in most cases. In this research a simulation methodology is employed, which uses an integral wall model coupled with a sharp-interface immersed boundary method and a modified rescale/recycle method, to conduct LES of developing turbulent boundary layer flows over surfaces with macro-scale roughness elements. The goal of this study is to develop and enhance our understanding of the effect of the roughness element geometry and arrangement on the mean flow response. The modeling effort is focused on two applications: (1) analysis of the effect of macro-biofouling, specifically acorn barnacles, on ship hulls and (2) study of flow over arrays of cuboidal roughness elements of various aspect ratios and arrangements. Such geometries are often used to explore the effect of atmospheric boundary layers over urban canopies. In addition to performing and analyzing wall-modeled large-eddy simulations of these flows, the research also delves into the development of phenomenological models for predicting the drag over such rough surfaces. For many practical problems involving rough surface boundary layer flows, flow simulations might not be feasible. Furthermore, in many practical situations, the primary interest is in predicting gross quantities like drag and not the details of the flow. For these reasons, simpler models and correlations which can predict the effect of the rough surface on the flow are highly attractive. In the current thesis, physics based phenomenological “sheltering” models are extended to high aspect-ratio rectangular-prism roughness elements and an attempt is also made to extend it for more general rough surfaces. These models can predict the drag and aerodynamic features on surfaces with canonical surface roughness elements in the order of 100 to 102 seconds depending on the surface complexity, and their predictions compare fairly well against the data from existing experiments and simulations and also from the present LES computations.