Research On B-rep Solid Model Simplification Methods

In order to improve efficiency and quality of model processing, simplifying complex three-dimensional CAD model is very necessary in collaborative design, analysis and simulation, graphics rendering and other activities in product life cycle. With the increasing complexity of modern product and its involved three-dimensional CAD model, how to simplify an original complex CAD model automatically, efficiently to meet different application requirements has become a very important and challenging problem. This dissertation researches the boundary representation (B-rep) solid model simplification methods involved in the fields of product design and CAD/CAE integration, and has the following main contributions:1,B-rep model simplification based on region suppressionTo be able to efficiently and quickly deal with a simplified model of complex regional boundaries, a simplification method for B-rep models using regional suppression is presented. The method is applied on a pure geometric model without built-in features, and the simplified surfaces are divided into connected regions to be suppressed by the hybrid identification method. Then, the boundary loops of the connected regions are decomposed into co-surface loops, which is aided to reconstructive the simplified model using the Euler operations. Compared with of the existing feature-based simplification method, the simplified method can suppresses every complex unwanted region of a B-rep model as a whole so as to efficiently and robustly generate a simplified version of the original model that preserves the rough shape. In this method, the new surfaces that are used to cover each suppressed region are constructed based on the boundary loops decomposition of the suppressed region, rather than through the time-consuming face extension and intersection operations widely adopted by previous works.2,Recognition and suppression of feature-intersected blendsIn order to effectively deal with the complex blend intersected with other features, an efficient algorithm for recognizing and suppressing feature-intersected blends is presented. By analyzing the existing status of sharp edges before and after blend removal, from which the model topology can be anticipated. Sharp edges on a blend are classified into deleted edges, reserved edges and split edges. Blends with reserved edges are recognized as the feature-intersected. Furthermore, the sharp edge classification also indicates the model's topology change, and the final simplification model can then be reconstructed by modifying the boundaries of the faces adjacent to the blends. This method not only can handle more types of feature-intersected blends but also achieves high efficiency and robustness.3,Suppression of blend feature based on volume reconstructionIn order to effectively and uniformly deal with the conventional blends and the complex blends intersected with other feature, an algorithm for suppressing blend feature is proposed. The algorithm is mainly achieved through reconstructing blend volume, and avoids directly processing the complex blend boundaries. To suppress edge blends, blend volumes are first constructed and then combined with the original model using Boolean operations. In addition, vertex blends are also removed by modifying/reconstructing its adjacent face boundaries for finally reconstructing the simplification model. By further local geometric analysis, the algorithm can not only suppress the conventional blends of small radius, but also can suppress feature-intersected blends of big radius, and thus provides a uniform approach to deal with the different types of blends.Based on the above researches, a prototype system named BRMSS is developed to generate a simplified model. Experimental results demonstrate feasibility of the three simplified methods presented in this paper.