本文選題：FDTD + 雷達散射截面積　； 參考：《電子科技大學(xué)》2015年碩士論文

【摘要】：FDTD(Finite Difference Time Domain)方法是電磁計算領(lǐng)域的一種非常重要的方法,它基于蛙跳方法求解麥克斯韋方程組的微分項。作為一種時域方法,FDTD被廣泛應(yīng)用于研究非均勻色散媒質(zhì)對電磁波的透射、散射作用。隨著人們對電磁問題研究的不斷深入,計算模型的空間尺寸不斷增大,復(fù)雜程度不斷升高。由于courant穩(wěn)定性條件和數(shù)值色散的限制了時間步長和空間步長的選取,這些復(fù)雜電磁的計算時間和內(nèi)存占用也隨著網(wǎng)格數(shù)目而增加。進而,本文在非均勻色散媒質(zhì)的計算中引入了非均勻網(wǎng)格技術(shù)以及并行計算技術(shù)來縮短計算時間、減少內(nèi)存使用。本文首先闡述了FDTD領(lǐng)域的重要原理,對于信號源、吸收邊界做了介紹。并通過金屬、介質(zhì)、金屬-介質(zhì)復(fù)合目標(biāo)的散射驗證了遠場外推的準(zhǔn)確性。針對非磁化等離子體、磁化等離子體,給出了PLRC方法的詳細推導(dǎo),并計算了分層、非均勻的非磁化等離子體、磁化等離子體的透射、散射問題。接著,針對非均勻網(wǎng)格方法,本文介紹了亞網(wǎng)格技術(shù)和緩變網(wǎng)格技術(shù)。將緩變網(wǎng)格技術(shù)引入到非均勻色散介質(zhì)的計算當(dāng)中,并給出了詳細的推導(dǎo)。接著,通過求解雙層、四層非均勻等離子體球的散射特性,驗證了緩變網(wǎng)格方法的數(shù)值準(zhǔn)確性。之后給出了緩變網(wǎng)格方法和均勻網(wǎng)格方法在計算時間、消耗內(nèi)存上的對比,很好的體現(xiàn)出該方法的有效性。最后,本文介紹了并行技術(shù)中的MPI方法和Open MP方法。再簡單對比兩種方法后,詳細的討論了Open MP方法以及在非均勻色散介質(zhì)介質(zhì)上的應(yīng)用。通過求解三層、五層非均勻等離子體立方體的散射特性,驗證了Open MP技術(shù)使用到色散媒質(zhì)計算中的可行性。之后,通過并行深度、并行結(jié)構(gòu)、并行調(diào)度的三個方面分別計算,詳細討論了并行性能的改善。
[Abstract]:The FDTD(Finite Difference Time domain method is a very important method in the field of electromagnetic computation. It is based on the leapfrog method to solve the differential subdivision of Maxwell's equations. As a time-domain method, FDTD is widely used to study the transmission and scattering of electromagnetic waves in inhomogeneous dispersive media. With the further study of electromagnetic problems, the space size and complexity of the computational model are increasing. Because the courant stability condition and numerical dispersion limit the choice of time step and space step, the computational time and memory footprint of these complex electromagnetic systems increase with the number of meshes. Furthermore, the non-uniform grid technology and the parallel computing technique are introduced to reduce the computing time and memory usage in the computation of non-uniform dispersive media. In this paper, the important principle of FDTD is introduced, and the signal source and absorbing boundary are introduced. The accuracy of far field extrapolation is verified by the scattering of metal, medium and metal-medium composite targets. For the unmagnetized plasma and magnetized plasma, the PLRC method is derived in detail, and the transmission and scattering problems of the layered, inhomogeneous and magnetized plasma are calculated. Then, this paper introduces subgrid technology and slow grid technology for non-uniform grid method. The slow grid technique is introduced into the calculation of inhomogeneous dispersive media, and a detailed derivation is given. Then, the numerical accuracy of the slow-varying grid method is verified by solving the scattering characteristics of two-layer and four-layer inhomogeneous plasma spheres. After that, the comparison of slow grid method and uniform grid method in computing time and memory consumption is given, which shows the effectiveness of this method. Finally, this paper introduces the MPI method and Open MP method in parallel technology. After comparing the two methods, the Open MP method and its application in inhomogeneous dispersive media are discussed in detail. By solving the scattering characteristics of three-layer and five-layer inhomogeneous plasma cubes, the feasibility of using Open MP technique in dispersive medium calculation is verified. Then, the improvement of parallel performance is discussed in detail by computing three aspects of parallel depth, parallel structure and parallel scheduling.
【學(xué)位授予單位】：電子科技大學(xué)
【學(xué)位級別】：碩士
【學(xué)位授予年份】：2015
【分類號】：TN011

【參考文獻】

相關(guān)期刊論文 前1條

1 金君;喬楠;;SMP集群上的混合并行計算[J];計算機教育;2007年07期

，

本文編號：1795098