Alejandro L. Garcia

Statistical error in particle simulations of hydrodynamic phenomena

We present predictions for the statistical error due to finite sampling in the presence of thermal fluctuations in molecular simulation algorithms. Specifically, we establish how these errors depend on Mach number, Knudsen number, number of particles, etc. Expressions for the common hydrodynamic variables of interest such as flow velocity, temperature, density, pressure, shear stress, and heat flux are derived using equilibrium statistical mechanics. Both volume-averaged and surface-averaged quantities are considered. Comparisons between theory and computations using direct simulation Monte Carlo for dilute gases, and molecular dynamics for dense fluids, show that the use of equilibrium theory provides accurate results.

Stabilization of thermal lattice Boltzmann models

A three-dimensional thermal lattice-Boltzmann model with two relaxation times to separately control viscosity and thermal diffusion is developed. Numerical stability of the model is significantly improved using Lax-Wendroff advection to provide and adjustable time step. Good agreement with a conventional fiitedifference Navier-Stokes solver is obtained in modeling compressible Rayleigh-Bénard convestion when boundary conditions are treated similarly.

CELL SIZE DEPENDENCE OF TRANSPORT COEFFICIENTS IN STOCHASTIC PARTICLE ALGORITHMS

Using the Green–Kubo theory, the dependence of the viscosity and thermal conductivity on cell size is obtained explicitly for stochastic particle methods such as direct simulation Monte Carlo (DSMC) and its generalization, the consistent Boltzmann algorithm (CBA). These analytical results confirm empirical observations that significant errors occur when the cell dimensions are larger than a mean free path.

The direct simulation Monte Carlo method

Numerical simulation of the hydrodynamics of gas flow and fluid flow is described using the Direct Simulation Monte Carlo method. (AIP)

Direct Simulation Monte Carlo for Thin Film Bearings

The direct simulation Monte Carlo (DSMC) scheme is used to study the gas flow under a read/write head positioned nanometers above a moving disk drive platter (the slider bearing problem). In most cases, impressive agreement is found between the particle‐based simulation and numerical solutions of the continuum hydrodynamic Reynolds equation which has been corrected for slip. However, at very high platter speeds the gas is far from equilibrium, and the load capacity for the slider bearing cannot be accurately computed from the hydrodynamic pressure.

Time step truncation error in direct simulation Monte Carlo

The time step truncation error in direct simulation Monte Carlo calculations is found to be O(Δt2) for a variety of simple flows, both transient and steady state. The measured errors in the transport coefficients (viscosity, thermal conductivity, and self-diffusion) are in good agreement with predictions from Green-Kubo analysis [N. Hadjiconstantinou, Phys. Fluids 12, 2634 (2000)].

Three-dimensional direct simulation Monte Carlo method for slider air bearings

The direct simulation Monte Carlo (DSMC) method is used to solve the three-dimensional nano-scale gas film lubrication problem between a gas bearing slider and a rotating disk, and this solution is compared to the numerical solution of the compressible Reynolds equations with the slip flow correction based on the linearized Boltzmann equation as presented by Fukui and Kaneko [molecular gas film lubrication (MGL) method] [ASME J. Tribol. 110, 253 (1988)]. In the DSMC method, hundreds of thousands of simulated particles are used and their three velocity components and three spatial coordinates are calculated and recorded by using a hard-sphere collision model. Two-dimensional pressure profiles are obtained across the film thickness direction. The results obtained from the two methods agree well with each other for Knudsen numbers as large as 35 which corresponds to a minimum spacing of 2 nm. The result for contact slider is also obtained by the DSMC simulation and presented in this paper.

On the validity of hydrodynamics in plane Poiseuille flows

Microscopic simulations of plane Poiseuille flow for a dilute gas are presented. Although the flow is laminar (Reynolds number ≈10) and sub-sonic, the temperature and pressure profiles measured in the simulations differ qualitatively from the hydrodynamic predictions. The results are in agreement with a recent theoretical analysis based on the asymptotic solution of the BGK model of the Boltzmann equation.

A Monte Carlo Simulation of Coagulation

A Monte Carlo simulation technique is described for the study of the coagulation of suspended particles. The method is computationally efficient since the particle trajectories are not used to determine coagulations. Instead, pairs of particles are assigned probabilities to coagulate and the evolution is computed as a stochastic Markov game. We also describe a simple analytic method to obtain the stationary distribution of sizes for the various mechanisms of relative particle motion. It is demonstrated that the simulation yields the correct stationary size distribution independent of initial condition.

Comparison of Kinetic Theory and Hydrodynamics for Poiseuille Flow

Comparison of particle (DSMC) simulation with the numerical solution of the Navier–Stokes (NS) equations for pressure-driven plane Poiseuille flow is presented and contrasted with that of the acceleration-driven Poiseuille flow. Although for the acceleration-driven case DSMC measurements are qualitatively different from the NS solution at relatively low Knudsen number, the two are in somewhat better agreement for pressure-driven flow.