[Zurück]

D. Toneian:
""Multi-Particle Collision Dynamics" simulation of viscoelastic fluids";
Betreuer/in(nen): G. Kahl; Institut für Theoretische Physik, 2015; Abschlussprüfung: 18.11.2015.

In this thesis, the simulation technique called Multi-Particle Collision Dynamics (MPC) is extended such that it allows for the modeling of viscoelastic fluids. This is achieved by representing the fluid not by independent MPC particles, as would be the case in the original formulation of MPC, but rather by polymer-like aggregates of MPC particles. This way, the MPC particles do not propagate freely anymore, but instead are now subjected to an intra-polymer interaction potential. Said potential introduces elastic degrees of freedom, and is chosen to be quadratic in the separation of nearest neighbors, which results in polymers that are described by the Rouse model. This choice allows one to find, starting from a generalized Navier-Stokes equation in the limit of low Reynolds numbers, an analytic solution for components of the velocity autocorrelation in Fourier space, as will be shown in this work. Furthermore, a new implementation of the extended MPC algorithm, capable of highly parallel execution on a Graphics Processing Unit (GPU), is presented. Simulation data obtained from this implementation are compared to the theoretical results, and a remarkable agreement is demonstrated on both the qualitative and quantitative level for polymers containing up to 10 MPC particles. The velocity autocorrelation in Fourier space is examined for asymptotically large correlation times, and it is found that the fluid's behavior corresponds to that of a Newtonian fluid on sufficiently long time scales. Finally, for the case of trimers, the relationship between the interaction strength (spring constant) and the root-mean-square bond length of the polymer is derived analytically for a shear flow with arbitrary shear rate.

multi-particle collision dynamics, stochastic rotation dynamics, viscoelasticity, polymers

http://publik.tuwien.ac.at/files/PubDat_246611.pdf

Projektleitung Gerhard Kahl:
DFS