Identifier

etd-04212010-111944

Degree

Doctor of Philosophy (PhD)

Department

Engineering Science (Interdepartmental Program)

Document Type

Dissertation

Abstract

The aim of this dissertation project is to extend the standard Lattice Boltzmann method (LBM) for shallow water flows in order to deal with three dimensional flow fields. The shallow water and mass transport equations have wide applications in ocean, coastal, and hydraulic engineering, which can benefit from the advantages of the LBM. The LBM has recently become an attractive numerical method to solve various fluid dynamics phenomena; however, it has not been extensively applied to modeling shallow water flow and mass transport. Only a few works can be found on improving the LBM for mass transport in shallow water flows and even fewer on extending it to model three dimensional shallow water flow fields. The application of the LBM to modeling the shallow water and mass transport equations has been limited because it is not clearly understood how the LBM solves the shallow water and mass transport equations. The project first focuses on studying the importance of choosing enhanced collision operators such as the multiple-relaxation-time (MRT) and two-relaxation-time (TRT) over the standard single-relaxation-time (SRT) in LBM. A (MRT) collision operator is chosen for the shallow water equations, while a (TRT) method is used for the advection-dispersion equation. Furthermore, two speed-of-sound techniques are introduced to account for heterogeneous and anisotropic dispersion coefficients. By selecting appropriate equilibrium distribution functions, the standard LBM is extended to solve three-dimensional wind-driven and density-driven circulation by introducing a multi-layer LB model. A MRT-LBM model is used to solve for each layer coupled by the vertical viscosity forcing term. To increase solution stability, an implicit step is suggested to obtain stratified flow velocities. Numerical examples are presented to verify the multi-layer LB model against analytical solutions. The model’s capability of calculating lateral and vertical distributions of the horizontal velocities is demonstrated for wind- and density- driven circulation over non-uniform bathymetry. The parallel performance of the LBM on central processing unit (CPU) based and graphics processing unit (GPU) based high performance computing (HPC) architectures is investigated showing attractive performance in relation to speedup and scalability.

Date

2010

Document Availability at the Time of Submission

Release the entire work immediately for access worldwide.

Committee Chair

Tsai, Frank T.-C.

Share

COinS