Studies On Mesh Feature Editing Techniques For Reverse Innovative Design

Design reuse plays a very important part in the product design. At present, parametric design, free-form deformation and reverse engineering are major methods for the design reuse of three-dimensional models. All have their own advantages and disadvantages and the scope of application. Recently, some scholars have proposed a new design reuse method called reverse innovative design for the problems in reverse engineering.Reverse innovative design, which is based on reverse engineering, combines the ideas of parametric design and free-form deformation. It has obvious advantages in the CAD/CAE system integration and design innovation over the usual methods. But currently, there is no systematic theoretical framework for this method, not to mention the sophisticated technology solutions. Thus, the research on the reverse innovative design technology is conducive to the development of design theory and its application in practical engineering. Therefore, in this thesis, an in-depth research on the mesh feature editing and its related technologies for reverse innovative design is given. All creative achievements in this thesis are listed as follows:·A new region growing algorithm is proposed to fast reconstruct a two manifold mesh from an unorganized point cloud. Starting with a seed face, the algorithm selects the best active edge to grow by the heuristic information. The best active edge is an active one which could make region growing easy and avoid geometry error. In each growing, among the candidate points, the best point is selected based on some criteria to create a triangle with the active edge. Add the triangle into the mesh if it passes the heuristic information and the geometric integrity test. Sample points are processed until a two-dimensional manifold mesh which recovers object shape is constructed. The algorithm doesn't need point normal, surface boundary, but only point position. It can deal with not only those non-uniform point clouds, but also the scan line type ones.·A new method is presented to interactively extract local sharp features and local transition features on the mesh. For sharp feature, first, some faces containing mesh feature are selected by the user with paintbrush and feature parameters are calculated from the selected faces. Then starting from the seed feature, similar features nearby are found according to the feature parameters and combined. This progress is iterated until there is no further expansion for the feature. The basic idea of the transition feature extraction is very similar to the one of the sharp feature extraction except that the analysis and extraction procedures for the transition features are embedded in the paintbrush selection. Furthermore, if the feature parameters are directly specified by the user, then this method can also be used to extract the global features of the mesh. The methods proposed in this paper are very intuitive and convenient. the method for the transition features can extract blend feature, chamfer feature, and even the transition feature with noise.·A new algorithm is proposed to construct sharp feature from the transition feature on the mesh. Pre-smoothing is first applied to feature region after the transition feature is extracted by the user. Then incrementally calculate the new face normals of the feature region according to the ones of the non-feature region and rectify the unreasonable normals. Finally, the vertex positions are gradually adjusted based on those new normals, which leads the actual normals of the faces to agree with their new ones. This method can deal with not only the blend features or chamfer feature on the mesh, but also the transition feature with noise.·A new algorithm is proposed to construct blend feature from the sharp feature line on the mesh. Although from the modeling point of view this operation is opposite to the sharp feature construction, there are lots of similarities between the constructing algorithm. The region to be blended is first constructed according to the sharp feature line and the blend radius specified by the user. Then, the target normals of the faces in the constructed region are estimated. Finally, the vertex positions are gradually adjusted based on those target normals, which leads the actual normals of the faces to agree with their target ones. The sharp feature line to be blended with our algorithm can either be open or closed. ·An iterative mesh optimization algorithm is proposed for finite element analysis. Given mesh is first refined to get enough degree of freedom for changing topology and geometry. Then within a given tolerance the mesh quality is improved by simplification and regularization. This progress is iterated until the mesh quality meets the requirement or a given maximum number of iterations is reached. The proposed algorithm can control the mesh properties flexibly and improve the mesh quality effectively. Thus, it can be used to generate high quality surface mesh for finite element analysis as a post-operator of the mesh editing in reverse innovative design.