Abstract

Rough-wall turbulence is the rule rather than the exception in engineering and nature. Progress in understanding rough-wall turbulence has been difficult owing to the infinite number of roughness geometries, compounded by the prohibitive experimental or computational costs associated with dynamically characterising each roughness geometry. Consequently, roughness research has been a large-scale, decades-long effort in cataloging and empirical modelling. Even so, the accuracy of predictions| is questioned whenever unfamiliar roughness geometries are encountered or exacting designs are required. In a paradigm shift, I will discuss an alternative approach to understanding rough-wall turbulence in general. The method, namely minimal simulations, is originally developed in the context of smooth walls by Jiménez and Moin (1991) for the buffer layer and by Flores and Jiménez (2010) for the log layer. Recently, the method is extended by Chung et al. (2015) to rough-wall turbulence. The method provides a framework in which to directly simulate only the scales associated with roughness modification and is therefore fast, cheap and accurate. Because the method is specific to the roughness in question and is relatively cheap to evaluate, it lessens the need for empiricism. After presenting an overview of the method, I will discuss recent insights from the method applied to previously difficult-to-reach parameter spaces, including dense sinusoids and deep bars.