郑宇腾, 赵延文, 贾苗苗, 蔡强明, 聂在平. 求解多尺度目标电磁散射的积分方程快速算法[J]. 电波科学学报, 2014, 29(4): 601-605. doi: 10.13443/j.cjors.2013070501
      引用本文: 郑宇腾, 赵延文, 贾苗苗, 蔡强明, 聂在平. 求解多尺度目标电磁散射的积分方程快速算法[J]. 电波科学学报, 2014, 29(4): 601-605. doi: 10.13443/j.cjors.2013070501
      ZHENG Yuteng, ZHAO Yanwen, JIA Miaomiao, CAI Qiangming, NIE Zaiping. Fast method for solving electromagnetic scattering from multiscale targets with integral equation[J]. CHINESE JOURNAL OF RADIO SCIENCE, 2014, 29(4): 601-605. doi: 10.13443/j.cjors.2013070501
      Citation: ZHENG Yuteng, ZHAO Yanwen, JIA Miaomiao, CAI Qiangming, NIE Zaiping. Fast method for solving electromagnetic scattering from multiscale targets with integral equation[J]. CHINESE JOURNAL OF RADIO SCIENCE, 2014, 29(4): 601-605. doi: 10.13443/j.cjors.2013070501

      求解多尺度目标电磁散射的积分方程快速算法

      Fast method for solving electromagnetic scattering from multiscale targets with integral equation

      • 摘要: 多层快速卡特森展开算法(Multilevel Accelerated Cartesian Expansion Algorithm, MLACEA)可用于加速电小尺寸结构积分方程矩量法,且矩阵与矢量乘积运算计算复杂度为ON)量级; MLACEA和多层快速多级子算法(Multilevel Fast Multipole Algorithm, MLFMA)均基于八叉树分组结构,便于实现它们的混合快速算法MLACEA-MLFMA. 该混合算法可大幅度降低模拟含精细结构的电大尺寸目标宽带电磁散射问题的计算复杂度. 还详细阐述了求解电场积分方程的MLACEA算法及其与MLFMA算法的混合快速算法MLACEA-MLFMA算法;并通过计算实例对比分析了MLFMA算法与MLACEA-MLFMA混合算法的计算效率.

         

        Abstract: Multilevel accelerated Cartesian expansion algorithm (MLACEA) is applied on acceleration of method of moments to solve electrically small electromagnetic problems. Its computational complexity of matrix and vector multiplication is O(N). It is easy to combine multilevel fast multipole algorithm (MLFMA) with MLACEA, for MLACEA and MLFMA are both based on hierarchical Orc-tree data structure. Electrically large problems which contain complex and fine structure can be solved efficiently in wide band using the combined algorithm. Besides, MLACEA and MLACEA-MLFMA combined algorithm is described in detail. Memory and time consumption of MLACEA-MLFMA are compared with MLFMA by numerical experiment.

         

      /

      返回文章
      返回