Browsing by Author "Wu, Tian-Cherng"

Now showing 1 - 2 of 2

Flux-blending schemes for interface capture in two-fluid flows
(2009-11-01)
Tsui, Yeng-Yung
;
Lin, Shi-Wen
;
Cheng, Tong-Ting
;
Wu, Tian-Cherng
;
機械工程學系
This paper is aimed at developing a robust algorithm (FBICS) to capture the interface between two immiscible fluids without the need of interface reconstruction. The advection equation of the volume fraction is solved using the fully conservative finite volume method. Determination of the convective flux through each cell face is based on blending of high resolution schemes and compressive schemes to preserve the sharpness and boundedness of the interface. The flux-blending practice is fulfilled with the use of flux limiters. Test on simple advection flow problems indicates that the well-known CICSAM and HRIC schemes lose accuracy as the Courant number increases. In contrast, the present method maintains high-accuracy performance for Courant numbers up to one. The capability of the method to cope with the complicated dynamics of free surface flows is demonstrated via calculation of the collapsing flow of a water column with an obstacle. (c) 2009 Elsevier Ltd. All rights reserved.
Use of Characteristic-Based Flux Limiters in a Pressure-Based Unstructured-Grid Algorithm Incorporating High-Resolution Schemes
(2009)
Tsui, Yeng-Yung
;
Wu, Tian-Cherng
;
機械工程學系
;
Department of Mechanical Engineering
A pressure-based procedure to solve problems ranging from incompressible to highly compressible flows is described. The method adopts the fully conservative finite-volume approach, for which the meshes can be of any topology. To handle the sharp change of the gradient in the regions near the shock or the solid wall, the convective flux is limited using high-resolution schemes such as the total variation diminishing (TVD) or the normalized variable diagram (NVD) scheme. The flux limiters are determined from the characteristic variables instead of the commonly used primitive or conservative variables. To enhance solution accuracy, the gradient is calculated using a linear reconstruction approach. The method is assessed and validated via testing on a number of flow problems, including inviscid flows through a convergent-divergent nozzle, inviscid and viscous flows past an airfoil, and viscous flows through a double-throat nozzle.