Kinetic approaches for the simulation of non-linear free surface flow problems in civil and environmental engineering

Abstract

This thesis focuses on the numerical simulation of non-linear free surface flow problems. Different simulation kernels based on the Lattice Boltzmann method (LBM) have been developed or extended, implemented, and, after validation, applied to a number of applications in civil and environmental engineering. The LB model solves viscous and turbulent flows, essentially representing similar physics as Navier-Stokes or reduced shallow water models, but with specific solver advantages concerning data locality and parallel computing. The first part of this thesis deals with numerical simulations on high-performance GPU (graphics processing unit) hardware. Validations and applications of a reduced LB model for solving the shallow water equations are presented. The resulting GPU kernel has shown to be applicable to state-of-the-art benchmark problems, dealing with wave propagation and wave run-up. Subsequently, the GPU implementation of a 3D numerical wave tank for the simulation of various applications in civil engineering is presented. The second main target of this thesis is to develop and apply a novel model based on an enhanced representation and advection of the phase interface for the simulation of more complex and demanding free surface flow problems. A volume-of-fluid (VOF) approach in combination with a piecewise linear interface reconstruction (PLIC) has been coupled with the LBM. The resulting hybrid model has been successfully validated against various benchmark experiments. Even a breaking wave during shoaling on a slope, which is a demanding test case for VOF solvers, was successfully simulated. Apart from the model development and validation itself, a coupling to a rigid body engine for the simulation of FSI problems has been established. Finally, several techniques for the coupling to a potential flow solver are discussed and validated, in order to generate realistic wave profiles and for the efficient simulation of wave run-up and wave breaking.