Direct Numerical Simulation of A Forced Micro Couette Flow using DSMC

The direct simulation Monte Carlo (DSMC) method is used in a time-dependent manner to simulate threedimensional micro Couette flows. An artificial forcing that mimics the centrifugal force in the Taylor problem has been applied to the flow. The sampled behaviors of the resulting flow, including the averaged properties and disturbances, are studied. The computations have been performed using a parallel computer cluster. The results presented include those with various channel heights, plate speeds, and the forcing level. These changes also result in changes in the flow Reynolds number and Knudsen number. Spatially coherent flow patterns can be identified in the averaged flow and the disturbance flow fields. The results indicate that the discrete approach can capture unsteady, three-dimensional vortical flow structures. In cases with strong forcing, the disturbance energy spectra show significant content above statistical scatter. INTRODUCTION Microelectromechanical systems (MEMS) technologies have had significant impacts in different areas such as biosciences, computer sciences, and telecommunication. Many of these realized devices use fluid as a working media and their designs have often been guided by correlations derived from the fluid flow behavior observed at macroscales. With the increasing demand for higher system complexity and performance, it is important to develop a better understanding of the fluid flow characteristics at the microscale. Many recent experiments in fluidic microchannels have reported significant differences in the heat and momentum transfer coefficients compared with those at the macroscale. For example, while the friction factor varies inversely with the Reynolds number when the Reynolds number is small, the proportionality constant does not agree with the conventional correlation. The friction factors are also found to diverge from this inverse proportionality, which at macroscale indicates a change of flow characteristics from that of a laminar flow to a turbulent flow, at smaller Reynolds numbers than those commonly observed in the corresponding large channels. This early change of flow characteristics has been attributed to the effects of, for instances, gas rarefaction and other surface mechanisms, the large change of the flow Reynolds number along the microchannels, and the likely experimental uncertainties in microscale measurements. Advanced measurement techniques such as molecular tagging can produce more detailed quantitative data for microflows, which may eventually lead to a better understanding of the apparent microflow transition. In this paper, results obtained in recent micro Couette gas flow simulations will be presented. The DSMC method pioneered by Bird [1] is used. With an artificially applied, adjustable forcing, the flow model is used to study capturing the unsteady, three-dimensional flow disturbances in microflows. The DSMC method has been applied to many different fluid dynamics problems. It has been widely used in the simulations of rarefied gas flow. The method has also recently been applied to calculate the fluid flow and heat transfer behaviors of microchannels.[2,3] DSMC has been employed to study the low-density limits of a number of flow instabilities. The centrifugal instabilities in Taylor-Couette flows have received the most attention.[4,5] The formation of Taylor vortices were clearly demonstrated for a range of Knudsen number in these two-dimensional