系统工程与电子技术
繫統工程與電子技術
계통공정여전자기술
SYSTEMS ENGINEERING AND ELECTRONICS
2010年
4期
745-748,769
,共5页
分层粗糙面%前后向迭代法%谱积分加速法
分層粗糙麵%前後嚮迭代法%譜積分加速法
분층조조면%전후향질대법%보적분가속법
layered rough surface%forward-backward method%spectral acceleration algorithm
为快速获取分层粗糙面的电磁散射特性,提出一种结合前后向迭代算法(forward-backward method, FBM)与谱积分加速法(spectral accelerate algorithm, SAA)的快速算法(FBM/SAA),该算法的计算量和内存均与粗糙面离散剖分产生的未知量同量级(O(N)).建立了分层粗糙面上关于未知电流分布的电场积分方程并采用矩量法(method of moment, MoM)将其离散为矩阵方程;在用FBM对矩阵方程进行迭代求解过程中,采用SAA技术加速计算矩阵和矢量乘积以快速求解;将FBM/SAA应用于三层媒质分层粗糙面的双站散射系数的计算,计算结果与传统MoM和FBM相一致,证明了算法的有效性;分析了粗糙面参数不同情况下算法的收敛性,比较了传统MoM和FBM/SAA所耗费的CPU时间.结果表明,在计算较长分层粗糙面的散射时,FBM/SAA具有明显优势.
為快速穫取分層粗糙麵的電磁散射特性,提齣一種結閤前後嚮迭代算法(forward-backward method, FBM)與譜積分加速法(spectral accelerate algorithm, SAA)的快速算法(FBM/SAA),該算法的計算量和內存均與粗糙麵離散剖分產生的未知量同量級(O(N)).建立瞭分層粗糙麵上關于未知電流分佈的電場積分方程併採用矩量法(method of moment, MoM)將其離散為矩陣方程;在用FBM對矩陣方程進行迭代求解過程中,採用SAA技術加速計算矩陣和矢量乘積以快速求解;將FBM/SAA應用于三層媒質分層粗糙麵的雙站散射繫數的計算,計算結果與傳統MoM和FBM相一緻,證明瞭算法的有效性;分析瞭粗糙麵參數不同情況下算法的收斂性,比較瞭傳統MoM和FBM/SAA所耗費的CPU時間.結果錶明,在計算較長分層粗糙麵的散射時,FBM/SAA具有明顯優勢.
위쾌속획취분층조조면적전자산사특성,제출일충결합전후향질대산법(forward-backward method, FBM)여보적분가속법(spectral accelerate algorithm, SAA)적쾌속산법(FBM/SAA),해산법적계산량화내존균여조조면리산부분산생적미지량동량급(O(N)).건립료분층조조면상관우미지전류분포적전장적분방정병채용구량법(method of moment, MoM)장기리산위구진방정;재용FBM대구진방정진행질대구해과정중,채용SAA기술가속계산구진화시량승적이쾌속구해;장FBM/SAA응용우삼층매질분층조조면적쌍참산사계수적계산,계산결과여전통MoM화FBM상일치,증명료산법적유효성;분석료조조면삼수불동정황하산법적수렴성,비교료전통MoM화FBM/SAA소모비적CPU시간.결과표명,재계산교장분층조조면적산사시,FBM/SAA구유명현우세.
To investigate the scattering of layered rough surfaces, a fast calculation of electromagnetic scattering form layered surface based on the forward backward method (FBM) with spectral acceleration is presented. First, the electric field integral equations of the induced current on the layered rough surface are derived. Secondly, a hybrid approach to the FBM with spectral acceleration algorithm (SAA) is developed. As an example, a layered rough surface with two Gauss rough surfaces is studied. The result is compared with the MoM and FBM, which partly validates the formulation. The accuracy, efficiency and convergence of the method are then studied for rough surfaces with various parameters. As well, the CPU time of FBM/SAA is less than that of FBM.