應用於大型電磁散射體的平行計算之研究

時域有限差分法(Finite-Difference Time Domain, FDTD)是應用於電磁計算軟體中所常用的演算法，而這個演算法所適合處理的問題可以是複雜的幾何形狀或是問題中包含著多種不同的材料，然而FDTD在面對較龐大或更複雜的計算時仍需耗費大量時間及電腦資源，有藉於此，本論文利用MPI (Message Passing Interface)所提供的函式庫將FDTD程式平行化，而MPI的技術提供許多實用的函式來建構平行計算，不同於傳統的計算問題方式，在時間上只能將問題讓一顆CPU以序列方式一一處理，平行運算可以將問題切割給成許多部分，讓多顆CPU同步執行，來達到節省時間的效果。因此藉由平行FDTD演算法，我們所建構的平行計算環境亦得到很好的平行效率。本篇論文中，我們考慮一個波長尺寸大的問題，除了FDTD的計算以及實際量測結果互相比較之外，文中也比較了其他不同演算法的計算結果，而當我們處理這樣一個高頻問題時(問題中包含一個半波長天線以及一個波長尺寸大的金屬盒)，FDTD程式執行得到遠場結果則必須花費相當多的計算時間，藉由MPI技術，我們同樣使用平行化的FDTD程式來執行，結果顯示平行至三顆CPU計算時仍舊可以得到很好的平行效率。

關鍵字

時域有限差分法；平行運算

並列摘要

Finite-Difference Time Domain (FDTD) is one of the most popular algorithms in computational electromagnetics, and is well suited for analyzing problems with complex geometrical features as well as those containing arbitrarily inhomogeneous materials. However, this method will consume huge computing time and require a large amount of memory in complex questions. To reduce computer run time, we implement the parallel FDTD method using MPI library. The MPI technique survey many useful functions to implement the parallel computing. Instead of managing the sequential manner by using one single-CPU, the parallel procedures is divided the whole problem into several parts and processing with many CPUs and eventually save the computational time. By paralleling the FDTD code, we have high parallel efficiency on our cluster system. In this thesis, we consider an electrically large scattering problem, and comparing the measured result with the computed, which the numerical results are computed by FDTD program and different numerical approaches. When applied on this electrical large problem which consider a half-wave dipole antenna located near a electrically large metal box, FDTD program spent much time to obtain the far-field result. By employing MPI library, a parallel FDTD algorithm is implemented to treat the same problem and the parallel result remains have a high parallel efficiency when employing three parallel CPUs.

並列關鍵字

MPI ； FDTD

參考文獻

[1] A. Taflove and S. C. Hagness, Computational Electromagnetics – the Finite-Difference Time-Domain Method, Artech House, 2nd Ed., 2000.

[2] N. Wang, “Self-consistent GTD Formulation for conducting Cylinders with Arbitrary Convex Cross Section,” IEEE Trans. Antennas Propagat., vol. AP-24, no. 4, pp. 463-468, July 1976.

[4] W. Yu, R. Mittra, T. Su, Y. Liu and X. Yang, Parallel Finite-Difference Time-Domain Method, Artech House, 2006.

[7] Omar M. Ramahi, “Near- and Far-Field Calculations in FDTD Simulations Using Kirchhoff Surface Integral Representation,” IEEE transactions on antennas and propagation, vol 45, no. 5, MAY 1997.

[11] Danai Torrungrueng and E. H. Newman, “The Multiple Sweep Method of Moments (MSMM) Analysis of Electrically Large Bodies,” IEEE transactions on antennas and propagation, vol. 45, no.8, august 1997.

國際替代計量

應用於大型電磁散射體的平行計算之研究

未授權

主題瀏覽