Dissipative-particle-dynamics model for two-phase flows

Anupam Tiwari, Purdue University

Abstract

Dissipative particle dynamics (DPD) is an emerging mesoscopic computational method in which the interparticle forces arise because of coarse-graining at the molecular level. In the asymptotic limit of large length scales, the governing equations of continuum are reproduced, as demonstrated in this work. Furthermore, it is shown that for single-phase flows results obtained with the method lie within 3% of the results obtained using more conventional methods. In this work, a two-phase DPD model based on mean-field theory is developed. Phase segregation is simulated by the choice of an equation of state with a van der Waals loop. Surface tension is modeled by a term that depends on higher-order density gradients and accounts for long-range attractive forces. To benchmark the model, we have simulated several test problems. These are of a liquid-layer, a liquid cylinder to reproduce the Laplace-Young relation, small- and large-amplitude liquid cylinder oscillations, and capillary waves. The values of surface tension obtained from the liquid-layer, liquid cylinder and small-amplitude oscillations were found to be within about 5% of each other. The time period and decay rate for capillary waves were found to differ from the analytical solutions by 10% and 14%, respectively. The two-phase model developed in this work is extended to two components. Using this model, we have performed simulations of Rayleigh-Taylor instability for a wide range of density ratios. The results have been found to be within the error bars of experimental data and in good agreement with Youngs' phenomenological model. The model is also used to investigate thermally induced breakup of liquid nanocylinders and nanojets. In agreement with molecular dynamics (MD) simulations, we get the double-cone structures prior to breakup and the breakup time is found to be agreement with linear instability theory. The time evolution of minimum radius follows theoretical predictions available in literature. Higher temperatures are found to accelerate the breakup. In contrast to macroscopic observations, no satellite drops are observed for breakup at the nanolevel.

Degree

Ph.D.

Advisors

Abraham, Purdue University.

Subject Area

Mechanical engineering

Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server
.

Share

COinS