發(fā)布時間：2018-05-04 23:20

  本文選題：積分方程方法 + 區(qū)域分解法��； 參考：《電子科技大學》2015年碩士論文

【摘要】：計算電磁學在當今工程界和學術(shù)界的諸多領(lǐng)域都有廣泛應用,對目標的電磁特性分析效率也越來越值得關(guān)注。隨著微波技術(shù)水平的不斷提升,實際工程對電磁分析計算方法提出了更高的要求,不僅需要求解包含復雜結(jié)構(gòu)的電大尺寸目標,還要能夠?qū)Χ喑叨饶繕诉M行高效電磁建模分析。本文主要研究積分方程區(qū)域分解法結(jié)合粒子群優(yōu)化算法在天線優(yōu)化設計中的應用。系統(tǒng)闡述了積分方程區(qū)域分解法原理,特別是使用非共型網(wǎng)格對復雜目標電磁特性進行快速正演計算。區(qū)域分解法基于分而治之的思想將待求目標整體分割為若干閉合子區(qū)域,先對子區(qū)域進行求解,再通過迭代計入?yún)^(qū)域之間的耦合,最終獲得正確的數(shù)值結(jié)果。此外,將上述方法與粒子群優(yōu)化算法結(jié)合,完成了簡化載機結(jié)構(gòu)電磁隱身和天線布局設計。本文首先介紹了優(yōu)化算法相關(guān)基礎(chǔ),特別是智能優(yōu)化算法的優(yōu)越性概述。進一步詳細闡述了粒子群優(yōu)化算法的算法原理、結(jié)構(gòu)及性能提升技術(shù),為將其應用到復雜電磁場問題的優(yōu)化設計打下基礎(chǔ)。然后研究了積分方程區(qū)域分解法的原理及實現(xiàn)方法。在積分方程矩量法的基礎(chǔ)上推導區(qū)域分解求解公式,子區(qū)域獨立地使用一定尺寸的網(wǎng)格剖分,區(qū)域內(nèi)迭代快速收斂,靈活控制網(wǎng)格剖分尺度。使用傳輸條件連接不同子區(qū)域,保證分解后目標電磁特性與原問題一致。突破傳統(tǒng)方法對交界面共形的要求,使用非共形接觸面處理技術(shù),完成非共形的積分方程區(qū)域分解方法,再運用投影接觸面判斷法,這使得目標的幾何建模具有極高的靈活性。最后將粒子群優(yōu)化算法與積分方程區(qū)域分解法相結(jié)合,將其用于簡化載機結(jié)構(gòu)電磁隱身優(yōu)化設計和天線布局中,優(yōu)化過程中避免重復幾何剖分,在保證計算精度的情況下,大大提高了復雜多尺度目標的電磁優(yōu)化效率。
[Abstract]:Computational electromagnetics is widely used in many fields of engineering and academia, and the efficiency of analyzing electromagnetic characteristics of target is more and more worthy of attention. With the continuous improvement of microwave technology, practical engineering has put forward higher requirements for electromagnetic analysis and calculation methods. It is necessary not only to solve electrically large size targets with complex structures, but also to be able to carry out efficient electromagnetic modeling and analysis of multi-scale targets. In this paper, the application of domain decomposition of integral equations and particle swarm optimization in antenna optimization design is studied. The principle of domain decomposition method for integral equations is systematically described, especially the fast forward calculation of electromagnetic characteristics of complex targets using non-conformal meshes. Based on the idea of divide-and-conquer, the domain decomposition method divides the whole target into several closed subregions. The subregions are solved first, and then the correct numerical results are obtained by iterating into the coupling between the regions. In addition, the above method is combined with particle swarm optimization algorithm to complete the design of simplified electromagnetic stealth and antenna layout. This paper first introduces the basis of optimization algorithm, especially the advantages of intelligent optimization algorithm. The principle, structure and performance improvement of particle swarm optimization (PSO) algorithm are described in detail, which lays a foundation for the application of PSO to the optimization design of complex electromagnetic field problems. Then, the principle and implementation of the domain decomposition method for integral equations are studied. Based on the method of moment of integral equation, the solution formula of domain decomposition is derived. The subregion uses meshes of certain size independently, the iteration in the region converges rapidly, and the meshing scale is controlled flexibly. The transmission conditions are used to connect different subregions to ensure that the electromagnetic characteristics of the decomposed target are consistent with the original problem. In order to break through the requirements of traditional methods for conformal interface, the non-conformal contact surface processing technique is used to complete the domain decomposition method of non-conformal integral equations, and then the projection contact surface judgment method is used, which makes the geometric modeling of the target have a high flexibility. Finally, the particle swarm optimization algorithm is combined with the integral equation domain decomposition method, which is used to simplify the electromagnetic stealth optimization design and antenna layout of the carrier structure. In the process of optimization, repeated geometric subsections are avoided, and the calculation accuracy is guaranteed. The electromagnetic optimization efficiency of complex multi-scale targets is greatly improved.
【學位授予單位】：電子科技大學
【學位級別】：碩士
【學位授予年份】：2015
【分類號】：TN820;TP18

【參考文獻】

相關(guān)博士學位論文 前1條

1 張麗平;粒子群優(yōu)化算法的理論及實踐[D];浙江大學;2005年

，

本文編號：1845133