Multiscale Compressed Strategy with Merged Characteristic Basis Function Method for Analyzing the Targets with Sharp

Abstract

In this paper, a multiscale compressed algorithm with merged characteristic basis function method (CBFM) is proposed to analyze electrically large scattering problems. Refer to the excitation-independent CBFM based on singular value decomposition (SVD), limiting the range of incidence angles can reduce the number of generated characteristic basis functions (CBFs) with a negative impact on accuracy. The use of secondary CBFs (SCBFs) can decrease the number of the incident plane waves, but results in problematic number of CBFs. To balance the accuracy and the computing time, the merged CBFM by considering the mutual interaction of surrounding blocks is studied in this paper. The merged CBFs are generated by merging the SCBFs of each block with the primary characteristic basis functions (PCBFs). The number of incident waves and the number of CBFs are both decreased without reducing any accuracy. Meanwhile, the impedance matrix and the reduced matrix are compressed by a multiscale compressed algorithm to enhance the efficiency of the merged CBFM. The validity of the proposed method is illustrated with the systematic analysis of canonical geometries, and numerical results are given to demonstrate its efficiency and accuracy.