SC²S Colloquium - October 13, 2011
|Date:||October 6, 2011|
|Time:||15:00 am, s.t.|
Denis Jarema: Comparison and Coupling of a Lattice-Boltzmann Fluid-Structure Interaction Code with the Faxen Correction Approach (MA)
The number of applications that involve multiple physical models or multiple simultaneous physical phenomena is constantly growing. The main reason is that many interesting natural phenomena are complex and the only way to understand them is to couple different models. In this thesis, we treat a multi-physical problem of particle transport phenomena in a channel with a complex wall geometry and laminar fluid flow. We start from two independent ways of describing the particle transport. While a Lattice Boltzmann fluid-structure interaction code provides precise computations of the particle motion, a Navier-Stokes' equations solver with the Faxen's postprocessing force estimation is computationally fast. The idea of this thesis is to combine the best properties of the two approaches. To accomplish this goal we developed an application that exploits the precision of the Lattice Boltzmann regime, when the particle is close to the walls or moves in the dynamically changing flow and gains performance, from the Navier-Stokes' regime, when the particle does not experience the impact of the walls and the flow is almost steady. In the thesis we provide a theoretical model of the fluid forces acting on the particle. For the postprocessing force computations, we derive the Faxen's theorems for the two-dimensional system. This enables us to test our application with two-dimensional models before experimenting with more computationally demanding three-dimensional systems. We compare the computed forces with the ones obtained from the fluid-structure interaction simulations. The drift ratchet model was used for testing purposes. We implemented the application in the Peano framework, a framework that was initially developed to support multi-physical simulations and that provides all necessary tools to accomplish our goals.