On the Formation Mechanisms of Artificially Generated High Reynolds Number Turbulent Boundary Layers

We investigate the evolution of an artificially thick turbulent boundary layer generated by two families of small obstacles (divided into uniform and non-uniform wall normal distributions of blockage). One- and two-point velocity measurements using constant temperature anemometry show that the canonical behaviour of a boundary layer is recovered after an adaptation region downstream of the trips presenting $$150~\%$$ higher momentum thickness (or equivalently, Reynolds number) than the natural case for the same downstream distance ( $$x\approx 3\,$$ m). The effect of the degree of immersion of the trips for $$h/\delta \gtrsim 1$$ is shown to play a secondary role. The one-point diagnostic quantities used to assess the degree of recovery of the canonical properties are the friction coefficient (representative of the inner motions), the shape factor and wake parameter (representative of the wake regions); they provide a severe test to be applied to artificially generated boundary layers. Simultaneous two-point velocity measurements of both spanwise and wall-normal correlations and the modulation of inner velocity by the outer structures show that there are two different formation mechanisms for the boundary layer. The trips with high aspect ratio and uniform distributed blockage leave the inner motions of the boundary layer relatively undisturbed, which subsequently drive the mixing of the obstacles’ wake with the wall-bounded flow (wall-driven). In contrast, the low aspect-ratio trips with non-uniform blockage destroy the inner structures, which are then re-formed further downstream under the influence of the wake of the trips (wake-driven).