ZURUECK HOCH VOR INHALT SUCHEN

» Back to overview
Proposing Institution

Institut für methodische Grundlagen
Project Manager

Prof. Dr. Michael Manhart
Arcisstr. 21
80333 München
Abstract
The project consists of direct numerical simulations (DNS) of turbulent flow under consideration of the flow microstructure which can not be resolved by the DNS grid for the flow. Two cases are considered: (i) the simulation of turbulent channel flow of a dilute polymer solution and a dilute suspension of rigid particles, respectively, and (ii) the simulation of micromixing processes in a turbulent flow through a T-mixer device. The first case investigates the turbulent drag reduction by dilute solutions of long-chained polymers in Newtonian fluids which can also be achieved in dilute suspensions of long fibres. It uses a Lagrangian Monte-Carlo method for describing the conformation of the polymers and fibres, respectively, on a microscopic level, which determines the non-Newtonian stress tensor acting on the fluid on a macroscopic level. A number of up to more than one billion single macromolecules is required to describe a sufficiently smooth stress tensor. The second case investigates the mixing process of two aqueous solutions in a T-mixer device which is used for the precipitation of particles in the nm-range. As a consequence of the high Schmidt-number (Sc=1900), the length scales of the concentration fluctuations become too small to be resolved by a computational grid. Therefore, a "Filtered-Density-Function"-approach (FDF) is used to describe the concentration fluctuations on the subgrid-level. The algorithm for the FDF-method shares many features with the Lagrangian Monte-Carlo method used to describe the conformation of macromolecules. By use of a high-performance computer, the description of turbulent channel flow of dilute suspensions of fibres on a microscopic level will be possible -- a task which was not accessible to numerical simulation up to now. The simulation of space-time concentration distributions by a combined DNS-FDF method will provide data on the mixing process of high-Schmidt-number flows that will provide valuable new insight into the process of micro-mixing.

Impressum, Conny Wendler