A restricted nonlinear (RNL) framework to enable computationally efficient investigations of high Reynolds number drag reduced flow
Abstract
reduce or increase skin-friction drag depending on the topology. One type of roughness element that is of particular interest due to its ability to both increase or decrease drag depending on the surface parameters is riblets. Riblets are roughnesselements that were inspired by the drag-reducing properties of shark-skin and are similarly elongated in the streamwise direction but can have a number of different topologies that vary in the spanwise direction. Flow over such surfaces has been widely studied, but a full understanding of the underlying phenomena remains elusive preventing the ability to exploit the underlying mechanism to improve vehicle design.A major challenge to developing the necessary knowledge is the computational resources needed for fully resolved simulations at the Reynolds numbers of interest. The computational requirements exceed those of smooth walls by an order of magnitude due to the wider range of physical scales that need to be captured. This computational complexity has limited the Reynolds numbers that have been investigated in fully resolved numerical simulations. The small size and scale of the roughness elements of ilicate extensive experimental studies. This work aims to bridge the knowledge gap through the development of a new suite of reduced order models that will enable us to isolate, analyze and reproduce the phenomena of interest at acceptablecomputational costs. In particular, we will develop new large eddy simulation (LES) tools that take advantage of the intrinsic order reduction and computational efficiency of the recently developed restricted nonlinear (RNL) model, which provides a reduced order representation of wall-boundedturbulent flows that accurately predicts low-order statistics, spanwise energy spectra, and turbulent kinetic energy budgets at greatly reduced computational costs.The RNL framework has recently been extended to flow over riblets. Simulation results indicate that this model accurately reproduces the mean velocity profile changes and secondary flow structures associated with changes in skin-friction drag at vastly reduced computational costs. The proposed yearlong project will extend our preliminary low Reynolds number studies to physically relevant Reynolds numbers through the following complementary research activities.(1) Developing an RNL-LES model and simulation tools for flow over spanwiseheterogeneous roughness elements (e.g. riblets). (2) Developing an RNL based LES wall model for flow over this class of roughness elements to improve the near surface accuracy of LES for these flows. The resulting suite of new RNL based tools will be used to systematically investigate the mechanisms underlying drag reduction in skin-friction drag observed in turbulent flow over riblets.The outcome of this work will be a suite of computationally tractable tools that both capture hydrodynamic phenomena around ship hulls, and provide a relatively inexpensive platform to evaluate the effects of changing the hull surface properties. The RNL based-tools can therefore play an important role in guiding the focus of experimental or high-fidelity computational investigations and, most importantly, reduce the required number of these expensive studies.Isolation of underlying flow physics in the simplified setting of the RNL dynamics can also be exploited to develop mechanistic models. Mechanistic models are valuable tools for characterizing the underlying hydrodynamics and providing new insights into the flow physics and flow control/manipulation in a range of applications. A better understanding of the flow physics in configurations that are known to red
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Jun 17, 2020
- Source ID
- N000142012535
Entities
People
- Dennice F. Gayme
Organizations
- Johns Hopkins University
- Office of Naval Research
- United States Navy