Mechanical Engineering - Theses
Permanent URI for this collection
Search Results
1 - 2 of 2
-
ItemTurbulent flow over spatially varying roughness( 2023-03)Large-scale heterogeneity in surface roughness, such as uneven biofouling on ships, can lead to phenomena such as internal boundary layers (IBLs) and mean secondary flows that increase drag. Contemporary studies in the field often lack near-wall measurements either due to limitations in experimental methods or prohibitive computational costs for direct numerical simulation (DNS). As a result, contemporary studies on heterogeneous roughness model the near-wall flow using the equilibrium log-law without fully accounting for its validity in the presence of spatial heterogeneity. This thesis considers heterogeneities idealised as alternating strips of roughness and uses DNS in conjunction with an immersed-boundary method to resolve the complex surface geometries. This thesis explores two main themes: the equilibrium log-law and flow phenomena associated with heterogeneous roughness. For the former, the equilibrium log-law is found to yield accurate predictions of the wall shear stress when applied to spanwise-heterogeneous roughness. Where velocity measurements are not available, predictions of the skin-friction coefficient can be made in the asymptotic limits of small and large wavelengths of spanwise heterogeneity. The latter theme is explored by perturbing the spanwise heterogeneity through introducing a moving roughness geometry and through introducing an oblique angle to the heterogeneity. The conventional intuition of mean secondary flows as counter-rotating vortex pairs is found to not generalise outside of spanwise-heterogeneous roughness. Finally, IBLs are identified in all obliquely heterogeneous cases where the roughness is not aligned with the flow.
-
ItemPredicting turbulent heat transfer over rough surfaces( 2023-04)When fluids such as air or water flow over solid surfaces, a transfer of heat between the fluid and solid interface occurs. This heat transfer underpins numerous industrial and natural systems, making routine predictions for heat transfer of paramount interest. In practice this is no easy feat. The fluid flow is often turbulent and the underlying surface is often rough, both of which must be accounted for in accurate predictions. Unfortunately, current state-of-the-art models for this purpose are insufficient, as these models are usually restricted to empirical fits, which admit a diverse set of behaviours in their predictions. To rectify this current ambiguity in rough-wall heat transfer, we have systematically conducted high-fidelity direct numerical simulations (DNSs), aiming to unveil the physical mechanisms governing rough-wall heat transfer. In these comprehensive datasets, we cover a wide array of roughness regimes (controlled by the roughness Reynolds number k+), whilst simultaneously varying the working fluid (dictated by the Prandtl number Pr ). A further dataset is also considered where we vary both k+ and the roughness topography, which we control through the frontal solidity, defined as the frontal projected area divided by the total plan area. These simulations are conducted using the minimal channel framework (MacDonald et al., 2017), which can accurately resolve the near-wall roughness sublayer flow at affordable cost, thus enabling the comprehensive parameter sweep. We examine the disagreements which currently linger regarding the prediction of fully rough (high-k+ ) heat transfer. We focus on the fully rough phenomenologies. Although we find the mean heat transfer favours the scaling of Brutsaert (1975), the Prandtl–Blasius boundary-layer ideas associated with the Reynolds–Chilton–Colburn analogy of Owen & Thomson (1963); Yaglom & Kader (1974) can remain an apt description of the flow locally in regions exposed to high shear. Sheltered regions, meanwhile, violate this behaviour and are instead dominated by reversed flow, where no clear correlation between heat and momentum transfer is evident. The overall picture of fully rough heat transfer is then not encapsulated by one singular mechanism or phenomenology, but rather an ensemble of different behaviours locally. Building on this intuition of distinct heat transfer behaviours locally, we develop a predictive model which idealises the total heat transfer as being comprised of two competing heat transfer mechanisms: exposed heat transfer, which follows a Reynolds–Analogy-like scaling and sheltered heat transfer, which is spatially uniform. The summation of these contributions, each weighted by their individual area fractions yields the total rough-wall heat transfer, and is ultimately dependent on the roughness topography controlled by the frontal solidity. We show that a simple ray-tracing model parameterised solely by a sheltering angle is capable of capturing this sheltered and exposed area partition. Finally, we provide a view to transitionally rough, low-k + heat transfer. Here, we have demonstrated that the virtual-origin framework of Luchini (1996) can provide a valid avenue for the prediction of transitionally rough heat transfer.