Explore chapters and articles related to this topic
Application of computational fluid dynamics to wind loading
Published in John D. Holmes, Seifu A. Bekele, Wind Loading of Structures, 2020
John D. Holmes, Seifu A. Bekele
Polyhedral (POLY) meshes: Polyhedral mesh can be easily applied to complex geometry and adopted to automatic mesh generation. The method is also known for its lower numerical diffusion. Polyhedral mesh has many faces, resulting in better gradient approximation than that with tetrahedral mesh, and reduced numerical diffusion than the application of hexahedral meshes to the same cases (Sosnowski et al., 2019). The polyhedral meshes for a computational domain in Figure 16.4 are shown in Figure 16.6. The same domain meshed with tetrahedral mesh is shown in Figure 16.7.
Hydraulic Methods for Unsteady Flows
Published in James L. Martin, Steven C. McCutcheon, Robert W. Schottman, Hydrodynamics and Transport for Water Quality Modeling, 2018
James L. Martin, Steven C. McCutcheon, Robert W. Schottman
A numerical-solution technique can be convergent but still introduce inaccuracies due to numerical diffusion. Numerical inaccuracies result from the discretization of the solution domain and truncation errors. This effect can be illustrated by transforming the difference equations back into a continuous solution using the Taylor series. For the modified Lax scheme (Equation 15), this produces (Grijsen 1986) () ∂Y∂t+ck∂Y∂x=Δx22Δt(αm−σ2)∂2Y∂x2+L
Modulation of secondary flows in curved serpentine micromixers
Published in Chemical Engineering Communications, 2022
Arees Qamareen, Mubashshir A. Ansari, Shah S. Alam, Anas Alazzam
It is well accepted that although numerical diffusion cannot be completely eliminated but some approaches can help minimize it. One technique is by employing a second order differencing scheme for discretizing the advection terms. The scheme is an upwind one with second order correction (Ansys 2009). High order approaches have obviously smaller numerical diffusion than low order ones. A high order method calculates the derivatives by using a large number of points hence are more accurate leading to less numerical diffusion. Another method which is used here is positioning the grid cells such that their edges are aligned either with the direction of streamlines or perpendicular to it. Even if very fine mesh is constructed, the numerical diffusion error cannot be completely exterminated. Numerical diffusion may originate due to any misalignment in even a few numbers of cells.
RANS-Based CFD Calculation for Pressure Drop and Mass Flow Rate Distribution in an MTR Fuel Assembly
Published in Nuclear Science and Engineering, 2021
N. L. Scuro, G. Angelo, E. Angelo, P. E. Umbehaun, W. M. Torres, P. H. G. Santos, L. O. Freire, D. A. Andrade
Avoid a first-order upwind numerical scheme, except to demonstrate qualitative behavior. At least second-order accurate discretization algorithms should be used in order to avoid degrading results by excessive numerical diffusion.