Applications of a particle simulation approach

Kabakcı, İsmail
The thesis is intended to utilize a particle simulation approach, introduced for simple particles, for engineering problems in order to study and understand fluid behavior at molecular level. First, an improvement in force potential estimation is proposed for the original method, which offers notable accuracy increase in simulations in terms of determination of position and momentum trajectories. Afterwards, the improved method is applied to heat diffusion and unidirectional fluid flow simulations. Within the context of the approach, instantaneous velocities of particles are calculated using simple algebraic equations instead of solving differential equations. Equations are derived from Newton’s 2nd Law of Motion and Lennard-Jones Force Potential Theory. For interactions taking place between unlike particles, Lorentz-Berthelot Combination Rule is used. The method is checked in terms of probability density function of speed distribution, distribution of velocity vector components and pressures at equilibrium state. In the scope of diffusion dynamics, thermal characteristics of particles and volume are tracked in order to perform equilibrium analyses. Furthermore, thermal conductivity coefficient is calculated. Finally, the variation of density between particles is investigated under unidirectional flow condition. Simulation results give Maxwell-Boltzmann and Gaussian distribution functions in terms of speed and velocity components respectively. Results on pressure calculation compromise with the classical equation of state. Thermal conductivity coefficient agrees with the experimental data. According to the unidirectional fluid flow simulations, the results imply the tendency of particles to stay closer with increasing unidirectional flow velocity.