Exploring Brownian Phenomena using Hydrodynamic Memory Kernels

Poster (konferens), 2025

At low particle to fluid density ratios observed in many microfluidic systems, Brownian motion needs to account for Drag, Added Mass and History Effects. A novel multiscale method based on the Generalised Langevin Equation (GLE) is presented in the poster, capable of modelling Brownian motion while incorporating these effects. The model uses short Direct Numerical Simulations (DNS) to determine hydrodynamic memory kernels for a spherical particle in a confined domain where the particle motion is to be studied, creating a library of memory kernels. These memory kernels are then used to model the hydrodynamic and Brownian forces needed in the GLE to simulate Brownian particle dynamics in the domain of interest. The method is validated and shown to be able to accurately capture the hydrodynamic and Brownian motion of a particle. Finally, the settling of a Brownian particle towards a wall under the influence of a conservative attractive force is studied using the new method. The variation in the particle impact region on the wall due to changes in the attractive force as well as changes in the distance of the particle to the wall is highlighted.