An efficient method of MLFMA combined with phase extracted basis functions

In order to reduce the memory of MLFMA combined with the phase extracted basis functions, a new refine-element scheme is presented in this paper. After refining every mesh of the target to a suitable size, a patch oriented grouping is used to ensure the memory cost for the basis pattern of MLFMA has been decreased significantly. Thanks to the guarantee of the addition theorem, the method displays no loss of the numerical accuracy. Compared with the conventional Octree-grouping, the method proposed in this paper can cut down the memory cost of the basis pattern by about 50%. Numerical examples have demonstrated the validity and effectiveness of the proposed scheme.