Boundary integral methods for fast and accurate simulation of droplets in two-dimensional Stokes flow

Time: Wed 2019-12-18 10.00

Location: F3, Lindstedtsvägen 26, Stockholm (English)

Subject area: Applied and Computational Mathematics, Numerical Analysis

Doctoral student: Sara Pålsson , Numerisk analys, NA

Opponent: Associate Professor Shravan Veerapaneni, University of Michigan, Michigan, USA

Supervisor: Professor Anna-Karin Tornberg, Numerisk analys, NA

Abstract

Accurate simulation of viscous fluid flows with deforming droplets creates a number of challenges. This thesis identifies these principal challenges and develops a numerical methodology to overcome them. Two-dimensional viscosity-dominated fluid flows are exclusively considered in this work. Such flows find many applications, for example, within the large and growing field of microfluidics; accurate numerical simulation is of paramount importance for understanding and exploiting them.

A boundary integral method is presented which enables the simulation of droplets and solids with a very high fidelity. The novelty of this method is in its ability to accurately handle close interactions of drops, and of drops and solid boundaries, including boundaries with sharp corners. The boundary integral method is coupled with a spectral method to solve a PDE for the time-dependent concentration of surfactants on each of the droplet interfaces. Surfactants are molecules that change the surface tension and are therefore highly influential in the types of flow problems which are considered herein.

A method’s usefulness is not dictated by accuracy alone. It is also necessary that the proposed method is computationally efficient. To this end, the spectral Ewald method has been adapted and applied. This yields solutions with computational cost O(N log N ), instead of O(N^2), for N source and target points.

Together, these innovations form a highly accurate, computationally efficient means of dealing with complex flow problems. A theoretical validation procedure has been developed which confirms the accuracy of the method.

urn.kb.se/resolve?urn=urn:nbn:se:kth:diva-264369

Page responsible:Kerstin Gustafsson
Belongs to: School of Engineering Sciences (SCI)
Last changed: Dec 03, 2019