MULTI-LEVEL-SUBSTRUCTURING-BASED SUBSPACE ITERATION IN THE EIGENVALUE EXTRACTION FOR VERY LARGE-SCALE STRUCTURES

For very large scale structures, a multi-level-substructuring-based subspace iteration strategy for eigenvalue extraction is proposed. The computational results of proposed method is independent of the substructure partitioning and assemblage order which may include some arbitrary combinations of the rotation, shifting or reflection of various substructure modules. Theoretically, it is an exact method. It has been shown to be very efficient through some practical application to large-scale complex structures. Two examples are given to show its excellent performance.