被覆最小自乗近似法を用いた線形構造解析

抄録

The voxel analysis method and the meshless method have been proposed to relieve the formidable task of mesh generation of traditional FEM. The model generation of the voxel analysis is very simple, but how to get reasonable accuracy and avoid the sharp increase of degrees of freedom is a problem of the voxel analysis. On the other hand, the meshless methods can avoid the mesh generation completely. However, some difficulties limit the freedom of meshless methods very much. One of these difficulties is how to easily guarantee the linear independence of approximation functions.Recently a new meshless approximation method, Cover Least Square Approximation, has been proposed, by which the linear independence conditions can be simply satisfied. In this paper, it is implemented for linear structural analysis by distributing the covers using multi-resolution voxel data.