Special Issue on the “Numerical Simulation of Rough-Wall Flows”

Dear Readers, This Special Issue features four articles on the numerical simulation and modelling of turbulent flows over rough-wall boundary layers. Roughness is a topic of crucial importance in many fields, ranging from hydroelectric-power generation, to naval hydrodynamics, to meteorology. Historically, experiments have been the principal tool that has shaped our understanding of the modifications of turbulence caused by roughness, and experimental data has been widely used to develop and validate the turbulence models used to close the Reynolds-Averaged Navier-Stokes (RANS) equations. Numerical simulations of rough-wall boundary-ayers that resolved the roughness were hampered by resolution requirements, set out in the review by Jiménez [Ann. Rev. Fluid Mech, vol 36, pp. 173–196 (2004)]. The increase in available computational power and the development of more advanced algorithms, however, have allowed direct and large-eddy simulations to begin having an impact. The data that can be provided by these techniques can answer questions that are difficult to address through experiments, which rarely have access to the region below the crest. This data, hopefully, can also lead to the development of more advanced turbulence models, through improved understanding of the interaction between the roughness sublayer and the outer flow. The issue is opened by a position paper by Paul Durbin, that addresses three matters that affect primarily the development of turbulence models, but also, to some extent, eddyresolving calculations. The paper discusses the limitations of models based on the doubleaveraging operation, compares drag models with boundary-condition modifications, and raises the troubling issue of the change in the value of the von Kármán constant caused by roughness. As mentioned above, one advantage of eddy-resolving simulations of rough-wall flows is the fact that the region below the roughness crest is accessible. This has allowed several researchers to study the effect of the form-induced velocity (the deviation of the timeaveraged velocity from the time and space-averaged – or Double-Averaged (DA) – one) and stresses on the turbulence. Mangavelli and Yuan investigate the role of these quantities on the statistics and on the turbulence structure. They consider channels in which an increase of the flow rate is followed by a steady period during which the flow rate remains constant. Their results highlight the role of the form-induced velocity gradients in generating pressure fluctuations that, in turn, affect the Reynolds-stress budgets. Busse and Jelly perform Direct Numerical Simulations of rough surfaces with very high values of skewness and kurtosis of the geometry. They find that the roughness function saturates for very high values of skewness, and that it is more sensitive to negative than