Molecular Dynamic Simulations of a Simplified Nanofluid

Abstract

This study describes the methodology that was developed to run a Molecular Dynamics Simulation (MDS) code to simulate the behaviour of a single nanoparticle dispersing in a fluid with a temperature gradient. A soft disk model described by the Lennard-Jones potential is used to simulate the system. The nanoparticle is assembled via the use of four subdomains of interatomic interactions and hence presents in full resolution the transfer of energy from the fluid-to-solid-to-fluid subdomains. A cluster computing system (HTCondor) was used to perform a large scale deployment of the MDS code. The obtained showcase results were successfully evaluated using three widely documented tests from the associated literature (Randomness, Radial Distribution and Velocity Autocorrelation Distribution Functions). It was discovered that the nanoparticle travels a larger distance in the fluid than the distance travelled by a fluid molecule (recovery region). The findings were confirmed by calculating the Green-Kubo self-diffusivity coefficient halfway through the simulation at which an enhancement of 156% was discovered in favour of the Nanoparticle. This might be the physical mechanism responsible for the experimentally observed thermal performance enhancement in nanofluids.