A Face Clustering Method for a Hexahedral Meshing

Finite element analysis has become a key technology for a design process of manufacturing industry. A hexahedral mesh is focused, because using a hexahedral mesh increases the quality of analysis. However it is very difficult problem to generate high quality hexahedral meshes, and there are many challenging research topics. Our goal is to develop a method to generate hexahedral meshes automatically to general volumes. Our method uses an intermediate model to recognize the input volume. The intermediate model is defined in the integer 3-dimensional space, and faces of the intermediate model are vertical to coordinate axes. Hexahedral mesh is generated by dividing the intermediate model into integer grids, and blocks of grids are projected into original volume. In this paper, we describe the method to generate a topology of the intermediate model. We use face clustering technique to generate the topology of the intermediate model. The faces of the input volume are clustered into 6 types; according to 3 coordinate axes and its direction, and clustered faces will be the faces of the intermediate model.