Research On Grid Meshing Method For Free-Form Grid Structures

In recent years, a lot of free-form architectures appeared in many cities, bringing in splendid visual effect. An urgent problem comes along with it, that’s how to mesh these changeable free-form structures to show its aesthetics implications. Traditional spatial structure changes less in the architecture surface, so there’re always some popular meshing schemes. But for the rising free-form structures, no popular meshing schemes is available, for its changeable architecture surface. So, meshing method for arbitrary free-form structures has becoming a problem to be solved imminently.Advancing front technique starts with a given boundary mesh, naming front, and then creates one internal elements, and update the front, as long as the front is not empty. According the basic processing of AFT, this paper brings up a method which divides a curve with an approximate length. With this method, the mesh is obtained. Taking demands of architecture grid into consideration, every candidate element is graded and the best is chosen to be the advancing elements. Free-form surface advancing front technique aiming at meshing architecture grid is established and its practicability and reliability are testified by practical engineering and artificial architecture surface.The soul of Architecture must be presented by architecture grids, According to which, the free-form surface meshing based on guide line is put forward. The guide line is a pre-drawn curve on the surface, guiding grid trend. It’s more helpful to choose rejected curves in surface parameter domain to present curves which are on the surface. Translation method, offset method and scale method are raised up to process the guide line so that the guide line will cover the surface, which founds grid by dividing and linking. A new method named two guide lines and two end vertices is put forward to meet the demand that gird size changes with the surface. Merits of different processing method are pointed out and accuracy and applicability are proved by practical engineering and artificial architecture surface.It’s hard to get satisfying gird only by meshing, and optimization is needed. According to research status of architecture gird optimization, local optimization is raised up in this paper. Grids that should be optimized are chosen to be optimized by optimal degree of every point. Thanks to characteristic of NURBS surface, the Laplacian optimization is improved without shrinkage by few steps. Grids after topology optimized are then Laplacian optimized. By optimize local grids, the accuracy and practicability are proved. Compared with entire gird optimization, convergence of local optimization is proved to be better.