Quantifying uncertainties in direct-numerical-simulation statistics due to wall-normal numerics and grids
Peng E. S. Chen, Xiaowei Zhu, Yipeng Shi, Xiang I. A. Yang
Abstract
Direct numerical simulation (DNS) data are considered a reliable source of information and are commonly used to develop and validate models. This paper explores the uncertainties in DNS data for channel flow. We examine how the way we discretize the flow in the direction normal to the walls affects the results. Our findings indicate that basic measurements like the mean flow and the turbulence level are very accurate. However, there are uncertainties when looking at more complex measurements like the energy dissipation, which can be influenced by factors such as the numerical methods used, the distribution of points in the simulation, and the grid resolution.
Topics & Concepts
Direct numerical simulationDiscretizationDissipationGridTurbulenceComputer simulationFlow (mathematics)Computer scienceStatistical physicsMechanicsMathematicsPhysicsSimulationGeometryReynolds numberMathematical analysisThermodynamicsFluid Dynamics and Turbulent FlowsPlant Water Relations and Carbon DynamicsMeteorological Phenomena and Simulations