宋凤丽,吴语茂. 基于自适应网格技术的快速物理光学方法[J]. 电波科学学报,2023,38(2):187-194. DOI: 10.12265/j.cjors.2022027
      引用本文: 宋凤丽,吴语茂. 基于自适应网格技术的快速物理光学方法[J]. 电波科学学报,2023,38(2):187-194. DOI: 10.12265/j.cjors.2022027
      SONG F L, WU Y M. The fast physical optics method based on the adaptive mesh technique[J]. Chinese journal of radio science,2023,38(2):187-194. (in Chinese). DOI: 10.12265/j.cjors.2022027
      Citation: SONG F L, WU Y M. The fast physical optics method based on the adaptive mesh technique[J]. Chinese journal of radio science,2023,38(2):187-194. (in Chinese). DOI: 10.12265/j.cjors.2022027

      基于自适应网格技术的快速物理光学方法

      The fast physical optics method based on the adaptive mesh technique

      • 摘要: 电大尺寸目标高频散射场的仿真一般使用物理光学(physical optics, PO)方法,该方法要求对目标使用尺寸为八分之一倍波长的三角网格进行剖分。然而,随着入射波频率增加,网格数目呈指数增长。本文提出基于一倍波长的二次曲面网格的快速PO(fast PO,FPO)方法,该方法使用线性多项式拟合振幅函数,能够有效降低面片数目。与二次多项式拟合振幅函数的快速物理光学方法相比,本文方法避免了菲涅尔积分的求解,且计算速度快1.22倍。为了进一步降低网格数目,提出了自适应网格技术的快速物理光学(FPO based on the adaptive mesh technique, AFPO)方法,允许使用比标准奈奎斯特采样率更少的网格数目对目标进行离散。通过后验误差估计和数值算例对所提出算法的精度和速度进行了验证,结果表明与采用均匀网格剖分的方法相比,AFPO方法所需的面片数目降低89.58%。

         

        Abstract: The physical optics (PO) method is generally used to simulate the high-frequency EM scattered fields from the large-scale scatterers. This method requires the target to be dissected using a triangular mesh with the size of 1/8 wavelength. Nevertheless, the number of meshes increases exponentially with the operating frequency. In this work, the fast physical optics method with the amplitude approximated by the linear polynomial (L-FPO) based on one-wavelength quadric mesh is proposed. The new method can effectively reduce the number of meshes. Compared with the fast physical optics method with the amplitude approximated by the quadratic polynomial (Q-FPO), L-FPO method avoids Fresnel integral calculation, and the calculation speed is 1.22 times faster than Q-FPO method. In order to further reduce the number of mesh, the fast physical optics method based on the adaptive mesh technique (AFPO) is proposed, which allows less meshes than standard Nyquist sampling theorem. The accuracy and speed of the proposed algorithm are verified by posteriori error estimation and numerical examples. Compared with the method using uniform mesh generation, the number of meshes required by AFPO method is reduced by 89.58%.

         

      /

      返回文章
      返回