Cad-based Three-dimensional Surface Grid Generation And Application,

Recently, more and more research works focus on the pre-process of 3D (3-dimensional) grid generation due to the complexity of 3D components and their combinations. In this thesis, methods of 3D surface grid generation and an application programming interface (API), which deals with closed manifold surface for UGCAD geometry and translates information of the geometry from UGCAD to that used for CFD (Computational Fluid Dynamics)-ready mesh generation, have been investigated and implemented. API is an important part of the pre-process for 3D grid generation. The Advancing Front Method (AFM) is employed to directly triangulate 3D surfaces.In seeking a robust and automated procedure, UGCAD and C++ are used as the modeling and programming environment. The API collects information of models from DXF (Drawing Exchange Format) of CAD system. CAD data is generally by its nature imprecise and the data structure does not match requirements by grid generation program. Thus, API combines the information of surfaces and edges, and then outputs information which is "grid-ready". That is, the data outputted from API can be used directly by the surface grid generation program.The grids are generated directly on 3D surfaces based on the advancing front technique. Background parameters are introduced to characterize the mesh distribution. And an approach is proposed to avoid the influence of the discontinuous parameterizationon smoothness during 3D surface meshing.The proposed method and treatment have been successfully applied to generate 3D surface unstructured grids for typical and complicated 3D models. Together with the existing FLOW_VR environment, the software developed in the thesis is used to simulate flow fields associated with internal explosions within a group of buildings. The initial flow pattern within multiple channels is also calculated and discussed. Numerical results indicate that the proposed method and treatment are feasible and efficient for 3D complicated flow simulations, and extend functions of the FLOW_VR.