Proper design of thermal management solutions for future nano-scale electronics or photonics will require knowledge of flow and transport through micron-scale ducts. As in the macro-scale conventional counterparts, such micron-scale flow systems would require robust simulation tools for early-stage design iterations. It can be envisioned that an ideal Nanoscale thermal management (NSTM) solution will involve two-phase flow, liquid flow and gas flow. This study focuses on numerical simulation gas flow in microchannels as a fundamental thermal management technique in any future NSTM solution. A wellknown particle-based method, Direct Simulation Monte Carlo (DSMC) is selected as the simulation tool. Unlike continuum based equations which would fail at large Kn numbers, the DSMC method is valid in all Knudsen regimes. Due to its conceptual simplicity and flexibility, DSMC has a lot of potential and has already given satisfactory answers to a broad range of macroscopic problems. It has also a lot of potential in handling complex MEMS flow problems with ease. However, the high-level statistical noise in DSMC must be eliminated and pressure boundary conditions must be effectively implemented in order to utilize the DSMC under subsonic flow conditions. The statistical noise of classical DSMC can be eliminated trough the use of IP method. The method saves computational time by several orders of magnitude compared to a similar DSMC simulation. As in the regular DSMC procedures, the molecular velocity is used to determine the molecular positions and compute collisions. Separating the macroscopic velocity from the molecular velocity through the use of the IP method, however, eliminates the high-level of statistical noise as typical in DSMC calculations of low-speed flows. The conventional boundary conditions of the classical DSMC method, such as constant velocity free-stream and vacuum conditions are incorrect in subsonic flow conditions. There should be a substantial amount of backpressure allowing new molecules to enter from the outlet as well as inlet boundaries. Additionally, the application of pressure boundaries will facilitate comparison of numerical and experimental results more readily. Therefore, the main aim of this study is to build the unidirectional, non-isothermal IP algorithm method with periodic boundary conditions on the two dimensional classical DSMC algorithm. The IP algorithm is further modified to implement pressure boundary conditions using the method of characteristics. The applicability of the final algorithm in solving a real flow situation is verified on parallel plate Poiseuille and backward facing step flows in microchannels which are established benchmark problems in computational fluid dynamics studies. The backward facing step geometry is also of practical importance in a variety of engineering applications including Integrated Circuit (IC) design. Such an investigation in microchannels with sufficient accuracy may provide insight into the more complex flow and transport processes in any future Nanoscale thermal management (NSTM) solution. The flow and heat transfer mechanisms at different Knudsen numbers are investigated.


If this is your thesis or dissertation, and want to learn how to access it or for more information about readership statistics, contact us at STARS@ucf.edu

Graduation Date





Kapat, Jayanta S.


Doctor of Philosophy (Ph.D.)


College of Engineering and Computer Science


Mechanical, Materials, and Aerospace Engineering









Release Date

October 2018

Length of Campus-only Access


Access Status

Doctoral Dissertation (Open Access)