Direct Simulation Monte Carlo On Micro-and Nanoscale Gas Flow And Heat Transfer

The investigation of micro- and nanoscale gas flow and heat transfer is an important part for the further development of the Microelectromechanical Systems (MEMS) and nano science and technology (NEMS). The Knudsen number is large enough due to the decreased characteristic length and the classical macro-simulation methods are no longer suitable. The flow must be described in the view of molecular dynamics. The direct simulation Monte Carlo method (DSMC) is the one of the most widely used and exact numerical methods for high-Knudsen-number flow.Based on theoretical analysis, in the present dissertation a code of DSMC with C++ advanced computer language is designed and used in the cases of micro-gas fluid and two kinds of gas mixing. The physical phenomenon of the flow characters and heat transfer is studied in the point view of molecular dynamics. The following is the main clue of the work.Firstly, in order to use the DSMC method in the micro- and nanoscale gas flow and heat transfer, a code is developed with C++ advanced computer language, and a GUI interface is also developed for the code with QT 4.0. The code is efficient and correct when compared with the Bird's standard code written in Fortran. Because the C++ is based on the Object-Oriented technology, the code can be extended conveniently. Also, it is more safety.Secondly, the similarity is theoretically analyzed between microscale gas flows and normal-scale rarefied gas flows. In order to make sure the reliability of the similarity, some factors are discussed. When the rarefied gas dynamics is used in studying the micro-gas flow, the dilute gas condition (δ>> d) must be guaranteed. And also, This similarity theory is required that two of the three parameters ( Ma, Re and Kn) are equal. Except that, some factors are also considered, such as the size of grid, the numbers of molecular per cell, time step and so on.Thirdly, the micro-channel flow and cavity flow were invested by the DSMC code, and the results were discussed and analyzed. After these discussions, two kinds of gas mixing flow are simulated. In the results of binary gas mixture, the speed which is given at the inlet of each kind gas is a very important role, when the number density of each kind of gas is the same before entering the small channel. The full-mixed length will be longer with the increased speed. But this relation is non-linear.Fourthly, in order to solve the low speed flow in the micro-scale, the discrete velocity model (DVM) is brought out. It is different with the traditional DVM. And comparing this model with theory analysis and numerical solution, in the low speed flow, it shows excellent agreement when Knudsen number from slip region to transfer region. And also it took shorter time to get the solution.Fifthly, a hybrid method that combining the EPSM and IP methods is considered in solving the near continuum flow region. As comparing with theory analysis, it shows good agreement. It is a beginning for constructing hybrid method for saluting the low speed flow in near continuum region.