Integrated radial basis function methods for Newtonian and non-Newtonian fluid flows

Ho-Minh, Dao (2011) Integrated radial basis function methods for Newtonian and non-Newtonian fluid flows. [Thesis (PhD/Research)]

Text (Introductory Pages)

Download (261Kb)
Text (Whole Thesis)

Download (2076Kb)
[img] Other (Movie: Annulus-Isotherms)

Download (2740Kb)
[img] Other (Movie: Annulus-Streamlines)

Download (2974Kb)
[img] Other (Movie: Cavity-Isotherms)

Download (1904Kb)
[img] Other (Movie: Cavity-Streamlines)

Download (2068Kb)


In this PhD thesis, one-dimensional integrated radial basis function networks (1D-IRBFNs) are further developed for the simulation of viscous and viscoelastic flows in two dimensions. The thesis consists of two main parts.

In the first part, 1D-IRBFNs are incorporated into the Galerkin formulation to simulate viscous flows. The governing equations are taken in the streamfunction-
vorticity formulation and in the streamfunction formulation. Boundary conditions are effectively imposed with the help of the integration constants. The
proposed 1D-IRBFN-based Galerkin methods are validated through the numerical simulation of several benchmark test problems including free convection in a square slot and in a concentric annulus.

In the second part, 1D-IRBFNs are incorporated into the Galerkin and collocation formulations to simulate viscoelastic flows. The momentum and continuity
equations are taken in the streamfunction-vorticity formulation and two types of fluid, namely Oldroyd-B and CEF models, are considered. Flows in a rectangular duct and in straight and corrugated tubes are simulated to validate the proposed 1D-IRBFN-based Galerkin/Collocation methods.

Main attractive features of the proposed methods include (i) easy implementation; (ii) avoidance of the reduction in convergence rate caused by differentiation; and (iii) effective treatment of derivative boundary conditions. Numerical results show that the proposed methods are stable, high-order accurate and converge well. This study further demonstrates the great potential of using
RBFs in CFD.

Statistics for USQ ePrint 19533
Statistics for this ePrint Item
Item Type: Thesis (PhD/Research)
Item Status: Live Archive
Additional Information: Docotr of Philosophy (PhD) thesis.
Faculty/School / Institute/Centre: Historic - Faculty of Engineering and Surveying - No Department
Supervisors: Mai-Duy, Nam; Tran-Cong, Thanh
Date Deposited: 24 Aug 2011 04:05
Last Modified: 01 Aug 2016 02:33
Uncontrolled Keywords: radial basis function networks; Newtonian fluid flows; non-Newtonian fluid flows
Fields of Research : 09 Engineering > 0913 Mechanical Engineering > 091307 Numerical Modelling and Mechanical Characterisation

Actions (login required)

View Item Archive Repository Staff Only