本文選題：加權(quán)Laguerre多項(xiàng)式 + 時(shí)域有限差分��； 參考：《電子科技大學(xué)》2014年博士論文

【摘要】：本文研究了計(jì)算電磁學(xué)領(lǐng)域中一種新的無(wú)條件穩(wěn)定的快速時(shí)域數(shù)值計(jì)算方法——基于加權(quán)Laguerre多項(xiàng)式(weighted Laguerre polynomials,WLPs)的時(shí)域有限差分(finite-difference time-domain,FDTD)方法——的基本原理及應(yīng)用。WLP-FDTD算法在空間域采用Yee氏網(wǎng)格劃分和中心差分技術(shù)離散,在時(shí)間域采用加權(quán)Laguerre多項(xiàng)式作為基函數(shù)、Galerkin過(guò)程作為權(quán)函數(shù)處理時(shí)間變量。這樣,WLP-FDTD算法的電磁場(chǎng)分量在空間域和時(shí)間域分別計(jì)算,按照Laguerre多項(xiàng)式的階數(shù)步進(jìn)求解,不受Courant-Friedrich-Levy(CFL)時(shí)間穩(wěn)定性條件的限制。WLP-FDTD算法特別適合分析計(jì)算寬頻帶、復(fù)雜結(jié)構(gòu)和多尺度結(jié)構(gòu)的電磁特性問(wèn)題,相比傳統(tǒng)的時(shí)域計(jì)算方法,在計(jì)算效率方面有較大的優(yōu)勢(shì)。本文在已有WLP-FDTD算法框架下,進(jìn)一步完善了其基本理論、提出了改進(jìn)技術(shù)和擴(kuò)展了其應(yīng)用范圍:一、通過(guò)Laguerre域麥克斯韋方程中引入電磁場(chǎng)的傅立葉形式展開(kāi)式,對(duì)二維WLP-FDTD算法的數(shù)值色散進(jìn)行了分析,并從理論上分析了與數(shù)值色散有關(guān)的關(guān)鍵參數(shù)選取方法,導(dǎo)出了時(shí)間尺度因子與工作頻率的關(guān)系。通過(guò)分析多項(xiàng)式最大零根的特性,可以計(jì)算出為保證算法計(jì)算的準(zhǔn)確性所需要的步進(jìn)階數(shù)。然后把二維WLP-FDTD方法的數(shù)值色散分析推廣到三維,豐富了WLP-FDTD方法的基本理論。把包含有增長(zhǎng)因子的電磁場(chǎng)的傅立葉形式展開(kāi)式引入到Laguerre域麥克斯韋方程,從理論上證明了WLP-FDTD算法按階數(shù)步進(jìn)是無(wú)條件穩(wěn)定的。二、把具有四階精度的中心差分公式引入到WLP-FDTD算法中,推導(dǎo)得到高階WLP-FDTD算法,對(duì)高階WLP-FDTD的數(shù)值色散關(guān)系和穩(wěn)定性進(jìn)行了分析,并對(duì)算法的關(guān)鍵參數(shù)的確定進(jìn)行了定量描述。通過(guò)與低階WLP-FDTD算法的比較,高階WLP-FDTD算法具有數(shù)值色散誤差小、計(jì)算精度高和計(jì)算效率高的特點(diǎn)。三、把表征色散媒質(zhì)特性的輔助差分方程(auxiliary differential equation,ADE)運(yùn)用到WLP-FDTD算法中,得到了適合分析廣義色散媒質(zhì)模型的ADE-WLP-FDTD方法。同時(shí),把近似完全匹配層(nearly perfectly matched layer,NPML)引入到ADE-WLP-FDTD算法中,得到比傳統(tǒng)PML更好的吸收效果。提出的ADE-WLP-FDTD方法及其NPML,有效地?cái)U(kuò)展了傳統(tǒng)WLP-FDTD算法的使用范圍,能模擬復(fù)雜色散媒質(zhì)的電磁特性。四、將兩種快速求解WLP-FDTD算法的技術(shù)——因式分解技術(shù)和區(qū)域分解技術(shù)——引入到ADE-WLP-FDTD算法中,有效地提高了ADE-WLP-FDTD算法的計(jì)算效率。五、把廣義曲線坐標(biāo)系引入WLP-FDTD算法中,得到了適合模擬任意復(fù)雜曲面的非正交WLP-FDTD算法的計(jì)算格式。采用這種計(jì)算格式模擬復(fù)雜結(jié)構(gòu)的電磁特性問(wèn)題時(shí),可以在不增加計(jì)算量的情況下提高計(jì)算精度。
[Abstract]:In this paper, we study the basic principle and application of a new unconditionally stable fast time-domain numerical method in computational electromagnetics, a finite-difference time-domain FDTDmethod based on weighted Laguerre polynomials and weighted Laguerre polynomialsWLPs-and its application. WLP-FDTD algorithm In spatial domain, Yee's mesh division and central difference technique are used to discretize. In time domain, the weighted Laguerre polynomial is used as the basis function and the Galerkin process is used as the weight function to deal with the time variable. In this way, the electromagnetic field components of WLP-FDTD algorithm are calculated in the space domain and the time domain, respectively. According to the order step solution of the Laguerre polynomial, the time-stability condition of Courant-Friedrich-Levyn CFL is not restricted. WLP-FDTD algorithm is especially suitable for the analysis and calculation of wide frequency band. Compared with the traditional time-domain calculation method, the electromagnetic characteristics of complex structures and multi-scale structures have more advantages in computational efficiency. In this paper, the basic theory of WLP-FDTD algorithm is further improved, and the improved technique and its application are proposed. Firstly, the Fourier expansion of electromagnetic field is introduced into the Maxwell equation in Laguerre domain. The numerical dispersion of the two-dimensional WLP-FDTD algorithm is analyzed. The method of selecting the key parameters related to the numerical dispersion is theoretically analyzed and the relationship between the time scale factor and the working frequency is derived. By analyzing the property of the maximum zero root of the polynomial, the step number needed to ensure the accuracy of the algorithm can be calculated. Then, the numerical dispersion analysis of two-dimensional WLP-FDTD method is extended to 3D, which enriches the basic theory of WLP-FDTD method. The Fourier form expansion of electromagnetic field with growth factor is introduced into Maxwell equation in Laguerre domain. It is proved theoretically that the WLP-FDTD algorithm is unconditionally stable in order. Secondly, the central difference formula with fourth-order precision is introduced into the WLP-FDTD algorithm, and the high-order WLP-FDTD algorithm is derived. The numerical dispersion relation and stability of the high-order WLP-FDTD are analyzed, and the key parameters of the algorithm are quantitatively described. Compared with the low-order WLP-FDTD algorithm, the higher-order WLP-FDTD algorithm has the advantages of small numerical dispersion error, high accuracy and high efficiency. Thirdly, the auxiliary difference equation which characterizes the properties of dispersive media is applied to WLP-FDTD algorithm, and the ADE-WLP-FDTD method suitable for the analysis of generalized dispersive media model is obtained. At the same time, the approximate perfectly matched layer (NPML) is introduced into ADE-WLP-FDTD algorithm, and the absorption effect is better than that of traditional PMLs. The proposed ADE-WLP-FDTD method and its NPMLs effectively extend the application range of the traditional WLP-FDTD algorithm and can simulate the electromagnetic properties of complex dispersive media. Fourthly, two techniques of fast solving WLP-FDTD, factorization technique and domain decomposition technique, are introduced into ADE-WLP-FDTD algorithm, which can effectively improve the computational efficiency of ADE-WLP-FDTD algorithm. Fifthly, the generalized curvilinear coordinate system is introduced into the WLP-FDTD algorithm, and the non-orthogonal WLP-FDTD algorithm suitable for simulating arbitrary complex surfaces is obtained. When using this scheme to simulate the electromagnetic characteristics of complex structures, the calculation accuracy can be improved without increasing the computational complexity.
【學(xué)位授予單位】：電子科技大學(xué)
【學(xué)位級(jí)別】：博士
【學(xué)位授予年份】：2014
【分類號(hào)】：TM15

【相似文獻(xiàn)】

相關(guān)期刊論文 前4條

1 于開(kāi)平,鄒經(jīng)湘;求解結(jié)構(gòu)動(dòng)力響應(yīng)的一個(gè)新的高階算法[J];振動(dòng)與沖擊;1998年02期

2 崔旭明;秦玉文;;基于3次B樣條無(wú)條件穩(wěn)定的位移元子區(qū)間法[J];天津大學(xué)學(xué)報(bào);2007年10期

3 吳霞;周樂(lè)柱;;二次B-Spline時(shí)域基函數(shù)的TDFEM的應(yīng)用[J];哈爾濱工業(yè)大學(xué)學(xué)報(bào);2010年05期

4 ;[J];;年期

相關(guān)會(huì)議論文 前1條

1 孔永丹;褚慶昕;;一種具有較小相對(duì)誤差的無(wú)條件穩(wěn)定FDTD算法[A];2009年全國(guó)微波毫米波會(huì)議論文集（下冊(cè)）[C];2009年

相關(guān)博士學(xué)位論文 前4條

1 陳偉軍;無(wú)條件穩(wěn)定的快速時(shí)域算法及應(yīng)用研究[D];電子科技大學(xué);2014年

2 孔永丹;基于分裂步長(zhǎng)的無(wú)條件穩(wěn)定FDTD算法研究[D];華南理工大學(xué);2011年

3 杜磊;時(shí)域有限元電磁計(jì)算方法的研究[D];南京理工大學(xué);2010年

4 梁鋒;時(shí)域有限差分法及其在碳基互連線仿真中的應(yīng)用[D];武漢大學(xué);2011年

相關(guān)碩士學(xué)位論文 前1條

1 李琳;耗散介質(zhì)中麥克斯韋方程的無(wú)條件穩(wěn)定有限時(shí)域差分方法[D];山東師范大學(xué);2013年

，

本文編號(hào)：2002745