| Summary: | 博士 === 國立成功大學 === 機械工程學系碩博士班 === 92 === The TIP4P and Lennard-Jones potentials are used to predict the velocity profiles in the 3-D liquid water heat and mass transfer problems. These problems are investigated by the leap-frog method in the field of molecular dynamics. In these works, the wall boundary condition is considered to be the situation that the water is absorbed on the metal wall and is then formed to be flat ice. And, the periodic boundary condition is used under the infinite condition.
First, the TIP4P potential is used to predict the velocity profiles in the 3-D （about 100,000 molecules）liquid water lid-driven cavity flow. The vortices in the cavity are generated with the upper side wall moving with a constant speed. Two kinds of problems are investigated in this paper to demonstrate the feature of the velocity profiles and traced the particle in the system, one is the cavity flow problem with square cavity and the other is with V-shape cavity. The realistic parameters of the water molecule are adopted in this research.
In a very short time, the velocity profiles are evident that the vortices are driven by the moving top plate in all cases. And, the blow-up phenomena is observed in the small triangular cavity when the calculating time is long enough. In addition, the vortex-like profiles in the triangular cavity is stronger and more obvious than the ones in the rectangular cavity. Therefore, the strength of vortex would be affected by the variation of the geometry. It emerges from that the dynamic transport properties like the thermal conductivity, diffusion coefficient and shear stress, et al. would be varied by the variation of the geometry.
Next, the flow characteristics of the 3-D plane fully developed Poiseuille flow in the nano-channel driven by a constant external force are studied by the Lennard-Jones and TIP4P potentials. Both global effect （effective channel width） and local effect （wall boundary types） are examined to demonstrate the features of the distributions of velocity and its gradient in the system. When the effective channel width is less than a critical value, the numerical results show that the Navier-Stokes theory would be fail to predict the velocity distribution. Furthermore, the velocity profile at a virtual slip plane presents the slip condition. And, we can reason that the surface effect exists and will affect the shear stress in the nano-channel.
Finally, to understanding and treating the heat transfer problem is the most important purpose of this study. And, it will be a molecular dynamic analysis for Bernad convection across rectangular enclosures. The temperature drop between the top and bottom wall is applied to this model, i.e., the temperature of the top plate is lower than the bottom plate. The numerical result shows that the distribution of the temperature consists of the boundary layer and the isothermal layer. This result agrees with the earlier experimental study. But, the value of the maximum stream function does not change obviously since increases. Namely, the natural convection could not be driven by the buoyancy force induced by the temperature difference in nano-system. The thermal-jump is the reason of the formation of the temperature gradient in the solid-liquid interface. To get the reliable results of the heat transfer problem of Benard convection is the purpose of this research in order to build the basis of the work of molecular dynamics.
|
|---|
