大尺度电磁散射与逆散射问题的深度学习方法

大尺度电磁散射与逆散射一直是科学研究和工程应用的热点和难点,亟待发展将电磁模型与数据挖掘有机融合的高性能求解方法,针对此,提出了一种针对大尺度电磁散射与逆散射问题的深度学习模型.该模型不仅继承了深度神经网络结构简单、运算速度快等优点,而且还能高精度地解决大尺度电磁散射与逆散射问题.实验结果表明:文中提出的深度学习方法可为解决现有大尺度电磁测算融合和电磁逆散射的计算成本昂贵的难题提供新思路、开辟新方向.     

THE ELECTROMAGNETIC SCATTERING OF A RADIALLY INHOMOGENEOUS DIELECTRIC CYLINDER

A new method is demonstrated for solving the electromagnetic scattering problemof a radially inhomogeneous lossless or lossy dielectric cylinder by Fourier series expansion.Com-pared with other methods,the method is simpler in numerical calculation and can be used forsolving scattering problems of a cylinder due to an arbitrary incident electromagnetic beam.     

Hierarchically spectral algebraic multigrid method for electromagnetic scattering

A novel spectral algebraic multigrid (SAMG) method is developed for solving electromagnetic scattering problems. Unlike traditional algebraic multigrid methods, the SAMG is an algebraic two-grid cycle built on a preconditioned iterative method that is used as the smoother, and the grid transfer operators are defined using the spectral information based on high-order hierarchical basis functions. Numerical results show that this method is very effective in electromagnetic scattering and can reduce overall simulation time significantly.     

用FD－MEI计算横各向异性媒质柱的电磁散射特性

本文推导了二维横各向异性媒质中电磁散射的差分方程，得到了９点的差分格式，并将不变性测试方程与该差分方程结合计算了具有任意横截面、非均匀、有耗和一般张量形式的二维横各向异性媒质柱的电磁散射。文中给出一些典型数值算例。数值结果表明，该方法在分析计算此类问题时是有效的且具有一定的优势。     

解决电磁散射问题的现代高频技术

电大尺寸散射体的高频散射高效算法问题是一个重要而具有挑战性的问题.本文首先对解决电磁散射问题的现代高频方法的研究进展进行了综述.其次，对高频散射场的高振荡特性，提出了数值最速下降路径法来计算高频散射场.最后，综合讨论了高频驻相点贡献凯勒锥以及爬行波场.     

Survey of numerical methods for solution of large systems of linear equations for electromagnetic field problems

Many of the popular methods for the solution of large matrix equations are surveyed with the hope of finding an efficient method suitable for both electromagnetic scattering and radiation problems and system identification problems.     

二维多导体柱电磁散射特性的特征基函数法分析

特征基函数法是近两年提出来的一种求解电磁散射问题的有效方法,该方法使用的特征基函数不受传统矩量法离散尺寸的限制,因而可以大大减小要求解的矩阵方程。应用特征基函数法分析了二维多导体柱的电磁散射特性,计算了多个无限长导电椭圆柱和方柱的雷达散射截面,结果表明特征基函数法的计算结果与传统矩量法的计算结果吻合良好,而计算量却大为减少。     

采用分部外推BCGS-FFT方法快速求解电磁散射问题

为了快速求解电磁散射问题中具有震荡性、奇异性、慢收敛性的索末菲积分,提出了一种利用分部外推算法加速索末菲尾部积分计算,并结合稳定双共轭快速傅里叶变换(stabilized biconjugate gradient fast Fourier transform,BCGS-FFT)算法求解电磁散射问题场分布情况的新方法. 首先给出电场积分方程(electric field integral equation, EFIE)的表达形式,且在求解过程的索末菲积分中应用一种便捷的椭圆积分路径来最小化索末菲积分的震荡性与奇异性,在索末菲尾部积分使用Levin分部外推法来提高积分收敛速度,以此来快速填充并矢格林函数矩阵. 然后对新方法进行了多种数值实验,验证算法的精确度,并对比了新方法与传统BCGS-FFT方法的计算效率,发现在保持相同计算精度的条件下,新方法可节省20%～37%的计算时间. 该方法能应用于复杂散射体嵌入多层空间的电磁散射计算,为快速求解目标区域的电磁散射场提供了一种新的方法.     

高频电磁场问题中的并矢Green函数方法

高频电磁辐射、散射问题中激励与响应之间的线性关系成立。这一关系可用一并矢变换函数给出。因此高频电磁场问题关键在于寻求并矢变换函数。并矢格林函数理论能够给出典型电磁场问题激励与响应之间的线性关系，借助高频电磁场的局域性原理、传播与绕射遵循的费马原理，极易求得高频电磁场问题的并矢变换函数。本文讨论了高频电磁场问题中的并矢格林函数方法，研究了并矢格林函数方法在寻求并矢变换函数中的应用，得到了部分工程问题的理论结果。     

The tolerance method for electromagnetic scattering

The original electromagnetic field will be disturbed when the parameter of a local region is changed. To find the distribution of the disturbed electromagnetic field a boundary value problem, which is different from the original one, must be solved. A method to find the disturbed electromagnetic field from the known original one with a minimum amount of computation would be very significant. In this paper, we present a method, tolerance method, which can save significant computer time in the computation of the disturbed electromagnetic field. This method can be used for many complicated boundary value problems as well as for optimum design in engineering. An example of the computation of an electromagnetic scattering problem is given in this paper     

Numerical second-kind-integral-equation solutions of electromagnetic scattering problems

A new, highly accurate numerical method based on second-kind integral equations has been developed to solve electromagnetic scattering problems for closed conducting bodies in two dimensions. The method is approximately fourth-order convergent, owing to the use of accurate new quadrature formulas.<>     

高性能计算在目标电磁散射特性分析中的应用

基于高性能计算的电磁数值模拟在目标电磁散射特性分析中发挥着越来越重要的作用.由于任一种数值方法都有一定的适用范围，不能高效处理所有问题，因此，有必要发展和集成多种数值方法，形成能够为不同类型问题的雷达散射截面（radar cross section，RCS）计算提供高效解决途径的软件系统.文中在并行自适应结构/非结构网格应用支撑软件框架之上，充分考虑数值方法的可扩展性以及物理个性的可分离性，通过基于机理、数据的混合可计算建模和接口设计，以及算法的模块化开发，发展了多种用于RCS计算的数值方法，并将其集成到高性能电磁数值模拟软件系统JEMS中.数值算例表明了JEMS具有高效分析多种目标电磁散射特性的能力，并在大规模并行计算方面具有显著优势.     

电磁散射问题中的频域有限元算法

有限元法作为一种十分有效的数值方法,不仅适用于求解不规则的几何形状和边界条件,而且能计算解决复杂介质问题。首先推导了有限元应用于散射问题的变分公式,然后介绍了电磁散射算法。为验证有限元散射程序的准确性和对任意散射体的适应性,还完成了MIE级数解析求解球体散射问题的算法用于与数值结果对比。通过针对三种不同散射体与现有典型算例的比较,证明了有限元散射算法的有效性。     

目标电磁散射特性研究的若干热点和难点问题

尽管国内外在目标电磁散射特性研究领域取得了很大进展,但是从技术本身来说该专业领域仍然存在一些难点问题.同时,复杂环境下低散射目标的探测与识别等应用,以及太赫兹等新体制雷达的研发对目标电磁散射特性研究提出了新的需求、引入了新的问题.围绕这些应用需求和技术难点,本文从理论建模、散射测量和特征提取三个方面,归纳总结了目标电磁散射特性研究存在的若干热点和难点问题,期望为从事该方面研究的科研人员提供方向性的引导.     

基于节点编码的区域分解算法及其在二维散射中的应用

研究了一种高效率的基于节点编码的区域分解算法.将原始的求解区域分割为若干个相对独立的子区域,使原问题转化为若干个相对独立的子问题,通过求解公共边界上的场值,可以快速获得整个求解区域上的场值,极大地减少了存储量和计算量.此外,这种区域分解算法不仅能够快速、高效、并行地计算电大尺寸柱体的电磁散射,还特别适合于求解具有几何重复性特征的结构,如天线阵列、有限周期频率选择表面、PBG/EBG等的电磁仿真问题.数值算例验证了该方法的准确性和有效性.     

Radar Cross Section of General Three-Dimensional Scatterers

Two disparate approaches?the finite-difference time-domain (FD-TD) method and the method-of-moments (MOM) surface -patch technique-which permit highly realistic modeling of electromagnetic scattering problems are compared. New results of induced surface currents and radar cross section are presented for an important three-dimensional canonical cube scatterer. It is shown that a high level of agreement for the two modeling approaches is obtained for this scattering example.     

On iterative approaches for electromagnetic rough-surfacescattering problems

Iterative techniques developed for solving general systems of linear equations have been applied to systems resulting from electromagnetic rough-surface scattering problems. Previously used iterative procedures that model the multiple scattering of the electromagnetic energy are shown to be mathematically equivalent to the application of stationary iterative procedures to the system of equations resulting from the standard moment method. Convergence difficulties that are sometimes observed with these approaches are due to the inherent limitations of the stationary techniques. The performances of the stationary approaches are compared with that of several conjugate-direction-based nonstationary iterative procedures through the application to a series of scattering surfaces that yield rapidly changing conditioning of the moment method interaction matrix. The stationary algorithms give the quickest convergence when applied to the systems with the best conditioning, but the nonstationary techniques prove much more robust in other more ill-conditioned situations