創(chuàng)建一個(gè)高效的仿真模型外文翻譯

1、畢業(yè)論文外文翻譯原文題目：creating an efficient simulation model譯文題目：創(chuàng)建一個(gè)高效的仿真模型在本章中，我們描述的數(shù)值模擬軟件的總體結(jié)構(gòu)實(shí)現(xiàn)一個(gè)高效的仿真模型，在不同的步驟。5.1數(shù)值模擬軟件的總體結(jié)構(gòu)數(shù)值模擬軟件由三個(gè)主要部件預(yù)處理，求解和后處理，如圖所示。 5.1。圖5.1。數(shù)值模擬軟件的主要組成部分前和后處理被納入了一個(gè)交互式圖形用戶界面（gui）。求解器啟動(dòng)，并從這個(gè)接口控制。雖然今天的用戶可能會(huì)覺得只有一個(gè)軟件，該司825創(chuàng)建一個(gè)高效的仿真模型軟件分為三個(gè)主要部分組成，從邏輯和說教感觀點(diǎn)。5.1.1預(yù)處理預(yù)處理器用于建立仿真模型，如圖。 5.2

2、。圖5.2?；镜姆抡婺Ｐ?，包括天線的幾何尺寸，材料性能，港口，邊界條件，為近場錄音盒我們在下面的列表給出一個(gè)簡短的概述，在處理的任務(wù)在模型生成的過程。每個(gè)項(xiàng)目的解釋在隨后的章節(jié)中更多的細(xì)節(jié)。一般定義：在這個(gè)階段，不同層次或組件定義組織的數(shù)據(jù)。此外，物理量的單位被選中，例如，長度以毫米為單位，在主頻。幾何對象的幾何形狀的定義。中有一般三種不同的方法，可以結(jié)合幾何進(jìn)入交互圖形用戶界面（gui），幾何導(dǎo)入數(shù)據(jù)從一個(gè)cad文件，或描述的對象是由一個(gè)宏語言。材料性能：介電性能被分配到幾何已在第一步為藍(lán)本的對象。材料選擇從數(shù)據(jù)庫中預(yù)定義的常用材料或新材料是指通過指定的介質(zhì)和磁學(xué)性質(zhì)。激勵(lì)：端口都定義為了

3、激發(fā)結(jié)構(gòu)和評估基于電路。5.1一般結(jié)構(gòu)的數(shù)值模擬軟件邊界條件：在不同的模擬量的外部邊界各種邊界條件適用于以代表自由空間，電場和磁場的墻壁或?qū)ΨQ平面。嚙合：離散成小分子結(jié)構(gòu)與均勻材料的性能。該模型的離散化是至關(guān)重要的一步，因?yàn)樗鼤?huì)影響所需的數(shù)值的努力和解決方案的準(zhǔn)確性。仿真參數(shù)：定義額外的參數(shù)控制解決方案的過程中，例如，要使用一種求解，定義一個(gè)endcriterion時(shí)間步進(jìn)算法，所需的精度，ar濾波，多激勵(lì)，端口模式。5.1.2求解求解計(jì)算電磁問題的近似解預(yù)處理過程中產(chǎn)生的數(shù)據(jù)的基礎(chǔ)上。該解決方案的過程是通常在一個(gè)所謂的日志文件記錄，列出了重要的細(xì)節(jié)，并模擬的統(tǒng)計(jì)信息，例如：解決方案需要時(shí)間，

4、數(shù)量遇到未知數(shù)，內(nèi)存的要求，以及警告和錯(cuò)誤在解決方案的過程中。5.1.3后處理后處理器是用來評估的求解結(jié)果。結(jié)果可表中顯示，一維，二維和三維圖。端口的相關(guān)參數(shù)：輸入阻抗等參數(shù)，散射從外地求解結(jié)果中提取參數(shù)，電壓和電流。這些參數(shù)需要指定結(jié)構(gòu)的行為當(dāng)連接到電路或其他rf元件?？梢暬翰煌膱龇植迹ǖ雀呔€和箭頭圖）近場數(shù)量顯示。場分布的可視化到結(jié)構(gòu)的工作方式，給用戶的見解。這種理解可以幫助改善結(jié)構(gòu)的性能，或幫助確定在該模型的錯(cuò)誤。天線參數(shù)：天線參數(shù)，如輻射方向圖，增益，方向性，半功率波束寬度，旁瓣抑制和輻射效率計(jì)算。其他參數(shù)：例如，評價(jià)特定吸收率（sar）和emc領(lǐng)域的優(yōu)勢，在一個(gè)指定的距離的相關(guān)數(shù)

5、據(jù)，如在比較emc限制。5創(chuàng)建一個(gè)高效的仿真模型。5.2幾何在建立仿真模型的第一步是一代三維結(jié)構(gòu)，如天線或一個(gè)過濾器的形狀。在最新的電磁建模軟件中，通常有三種方式存在實(shí)現(xiàn)這一目標(biāo)：通過圖形用戶界面（gui）的互動(dòng)建設(shè)，進(jìn)口的cad（計(jì)算機(jī)輔助設(shè)計(jì)）數(shù)據(jù)和通過宏語言對象的定義。5.2.1互動(dòng)建設(shè)，通過圖形用戶界面（gui）圖形用戶界面（gui），可幫助用戶創(chuàng)建交互式的他的模型幾何。 gui施工能力非常相似那些提供基本的電腦輔助設(shè)計(jì)（cad）軟件圖5.3。簡單三維物體一般建設(shè)開始與簡單對象的定義。更多結(jié)合和操作簡單的對象，可以產(chǎn)生復(fù)雜的對象。簡單的對象，可分為一，二，三維形狀。簡單的一維對象包括直

6、線，多邊形和樣條段。簡單的兩維對象包括三角形，長方形，圓形，橢圓形和封閉的多邊形。簡單的三維對象包括盒，圓筒，錐體和球體（見圖5.3）。根據(jù)軟件更復(fù)雜的基本如環(huán)形，螺旋，螺旋，橢圓形，與債券線的形狀?？梢孕薷倪@些簡單的對象，即，旋轉(zhuǎn)，移動(dòng)，鏡像，復(fù)制，縮放或拉伸。此外，布爾運(yùn)算，可應(yīng)用于重疊的元素。圖。 5.4影響減法，加法和路口顯示。如果我們繪制兩個(gè)重疊的像一個(gè)盒子和一個(gè)3d對象球體（圖5.4a），我們從包裝盒中減去球體形狀是一個(gè)球形，作為在圖中切出框。 54億。布爾操作另外結(jié)合兩圖所示的對象。 5.4c。如果我們相交兩個(gè)對象，我們得到一個(gè)對象框的共同部分組成和球體圖所示。 5.4d。圖5.

7、4。不同的布爾運(yùn)算修改的對象：（一）重疊的方塊和領(lǐng)域，（二）減法球體從框，（三）除了和（d）路口框和球旋轉(zhuǎn)二維實(shí)體。圖5.5顯示了基本的概念。圖。 5.5a12d臉是沿著一個(gè)矢量方向的擠壓。圖。 5.5（二）一個(gè)二維的臉被扭曲沿曲線圖。 5.5c二維面是圍繞一個(gè)軸旋轉(zhuǎn)。這些操作都非常在強(qiáng)大的創(chuàng)建復(fù)雜的幾何形狀。一般可以控制輸出其他選項(xiàng)，例如對象可以是錐形，即大小二維面有所不同，因?yàn)樗貍鞑サ臄D壓載體（圖5.5d）。有時(shí)gui包括先進(jìn)的功能，如對象的描述復(fù)雜的數(shù)學(xué)公式或邊緣平滑（四舍五入）。雖然已經(jīng)變得非常的幾何形狀的圖形強(qiáng)大的幾何應(yīng)保持盡可能的簡單。只有細(xì)節(jié)影響電磁行為應(yīng)列入。圓邊的結(jié)果865

8、創(chuàng)建一個(gè)高效的仿真模型。圖5.5。從二維輪廓3d對象的創(chuàng)建在一個(gè)有吸引力的模型，但導(dǎo)致的細(xì)離散結(jié)構(gòu)和從而增加了計(jì)算負(fù)擔(dān)。5.2.2通過宏語言的對象定義在初期的em模擬一個(gè)基于文本的模型的定義輸入文件是共同使用。今天，這種模型的創(chuàng)建仍然是有用的如果許多類似的幾何結(jié)構(gòu)進(jìn)行了分析。幾何特征可以定義變量和被輕易改變。因此，基于文本輸入文件是適合的參數(shù)研究和優(yōu)化程序。最em建模軟件包，提供基于文本的宏語言模型的定義。5.2.3導(dǎo)入cad（計(jì)算機(jī)輔助設(shè)計(jì)）數(shù)據(jù)在em建模，設(shè)計(jì)和微波分析，我們可以區(qū)分組件。在后一種情況下，它可能會(huì)使用cad數(shù)據(jù)的實(shí)際從生產(chǎn)過程中的微波元件。圖5.6顯示了作為一個(gè)5.2幾何8

9、7例如，一個(gè)天線模塊組成一個(gè)具有三維載體復(fù)雜形狀的天線板。機(jī)械cad軟件proengineer，solidworks或catia擁有許多更強(qiáng)大的功能，形狀復(fù)雜的機(jī)械設(shè)計(jì)對象比基本的cad納入em仿真工具軟件。為了從機(jī)械cad軟件導(dǎo)入模型可以使用標(biāo)準(zhǔn)化的數(shù)據(jù)交換格式，如step（產(chǎn)品模型數(shù)據(jù)交換標(biāo)準(zhǔn)）的iges（初始圖形交換規(guī)范）dxf（圖形交換格式）。圖5.6。cad模型的天線模塊雖然進(jìn)口的cad數(shù)據(jù)可能看起來造型非常優(yōu)雅的方式也有一些缺點(diǎn)。與cad數(shù)據(jù)的第一個(gè)問題是，許多包括機(jī)械重要的，但電無關(guān)的細(xì)節(jié)。這些cad細(xì)節(jié)導(dǎo)致更細(xì)的網(wǎng)格中，并因此增加模擬時(shí)間沒有顯著提高結(jié)果的準(zhǔn)確性。另一個(gè)問題與c

10、ad數(shù)據(jù)可能出現(xiàn)的幾何精度（精度）。如果cad系統(tǒng)和em建模軟件的圖形用戶界面，適用于不同程度幾何精度的意外可能會(huì)發(fā)生錯(cuò)誤。例如，小學(xué)連接對象顯示重疊區(qū)域是分開的。由此，我們得出這樣的結(jié)論：cad導(dǎo)入很少是一個(gè)“任務(wù)。 “隨后的模型簡化和數(shù)據(jù)愈合的繁瑣的任務(wù)。作為一個(gè)例子圖。 5.7a給出了一個(gè)天線模塊，包括小孔和差距。這些細(xì)節(jié)都包含在原有的機(jī)械cad文件原因（緊固），但電無關(guān)。圖5.7b所示刪除不必要的細(xì)節(jié)后的天線模塊。圖5.7c-f顯示如何的小細(xì)節(jié)，在精細(xì)網(wǎng)格的結(jié)果，導(dǎo)致較長的模擬倍。另一個(gè)例子是薄的金屬結(jié)構(gòu)。這些薄薄結(jié)構(gòu)，例如，圖微帶線帶導(dǎo)體。 5.8，可以看作一個(gè)二維（平面）對象。這種

11、建模方法模擬。圖5.7。從cad模型的詳細(xì)信息：（一）含有原始的cad數(shù)據(jù)的去除無關(guān)緊要的小細(xì)節(jié)，（二）簡化模型（細(xì)節(jié)），（cf）的影響正交和三角網(wǎng)格少未知數(shù)和增加細(xì)胞大小的機(jī)型相比，體積建模。5.2.4綜述當(dāng)設(shè)立一個(gè)仿真模型，意圖應(yīng)的幾何不從視覺角度來看，最準(zhǔn)確的結(jié)構(gòu)但最簡單的一個(gè)代表的電磁現(xiàn)象正在審議的結(jié)構(gòu)。此外，過多的細(xì)節(jié)可以放大計(jì)算的努力，一方面，導(dǎo)致容易出錯(cuò)，很難驗(yàn)證仿真模型。撤消修改圖5.8。 fdtd離散化微帶線的截面。簡化模型一個(gè)兩維板薄帶導(dǎo)體的代表性有時(shí)又何嘗不是如此明確，以決定是否有一定的幾何初步細(xì)節(jié)重大與否。在這種情況下，建議開始與簡單的模型，并驗(yàn)證這個(gè)細(xì)節(jié)的具體影響，在

12、一個(gè)單獨(dú)的仿真模型。5.3材料特性幾何結(jié)構(gòu)定義后，被分配到材料性能不同的對象。例如，在微帶結(jié)構(gòu)中，我們有一個(gè)金屬地平面和金屬的痕跡，基材是介質(zhì)和體積頂層上面是空氣。技術(shù)上重要的和廣泛使用的材料往往是預(yù)定在該軟件。他們可以直接從數(shù)據(jù)庫中選取。最重要的預(yù)定義的材料是真空和pec（理想導(dǎo)體）。如果列表中沒有提供一種特殊材料，它可以通過指定定義其相對介電常數(shù)介電常數(shù)，電導(dǎo)率，和相對滲透率r。代替的電導(dǎo)率的損耗角正切tan的可能定義。表5.1列出了一些通用材料的介電性能。不同的技術(shù)上有趣的介電性能的典型值從1 ghz至10 ghz頻率范圍內(nèi)的基板材料選項(xiàng)卡。 5.2。表5.1。不同材料的介電性能材料介電

13、常數(shù)/ s / mr真空1.00.01.0空氣1.0060.01.0特氟?。o損）2.20.0 1.0銅1.05.81071.0pec的1.01.0表5.2。不同的基板材料的介電性能的典型值在1-10千兆赫的頻率范圍材料介電常數(shù)的tanr氧化鋁9.5 - 10.50.0002 - 0.00031.0fr4的3.9 - 4.30.01 - 0.0251.0聚四氟乙烯2.0 - 2.20.001 - 0.0031.0撤消修改頻率相關(guān)的介電性能，可以進(jìn)入軟件在選定的頻率定義的屬性和使用分段線性在選定的頻率之間的頻率間隔的近似。有時(shí)頻率的依賴關(guān)系可以定義為封閉形式的功能或身體放松模型（debeye材料

14、）。各向異性審議確定各向異性張量。此外，非電參數(shù)是必要的具體計(jì)算在后加工的數(shù)量必須指定。例如，質(zhì)量密度需要以特定吸收率的計(jì)算（sar）的生物組織。高導(dǎo)電性的結(jié)構(gòu)，應(yīng)仿照完美電導(dǎo)體（pec），如果在材料的損失是微不足道的。它的優(yōu)點(diǎn)一個(gè)pec，努力減少計(jì)算：領(lǐng)域內(nèi)pec等于零，因此字段值也不需要計(jì)算基于卷的算法。雖然fdtd和fem是基于對三維量元素，持平或一維的pec的對象可以是創(chuàng)建。這是特別有用的薄金屬板或電線。5.4端口端口用于激發(fā)的被動(dòng)結(jié)構(gòu)和計(jì)算電路有關(guān)像散射參數(shù)，輸入阻抗，電壓和電流的數(shù)量。以下類型的端口，在大多數(shù)軟件包可供選擇：波導(dǎo)港口，平面波激勵(lì)和集中或集中的港口。結(jié)構(gòu)可以有多個(gè)端口

15、。在一個(gè)多模型軟件可以自動(dòng)計(jì)算出的散射矩陣的所有元素描述了不同的端口之間的相互作用。根據(jù)數(shù)值方法的應(yīng)用，這是由一系列的模擬只一個(gè)端口興奮的時(shí)間和所有其他模式相匹配的港口阻抗（例如，fdtd法）。其他方法（例如，mom）計(jì)算散射在一個(gè)單一的模擬矩陣。另一種方式來使用多個(gè)端口同時(shí)更激發(fā)多個(gè)端口。這可以用來激發(fā)沒有明確的天線陣列造型喂食網(wǎng)絡(luò)。在計(jì)算的同時(shí)激發(fā)的情況下散射參數(shù)是不可能執(zhí)行的方法一系列的模擬來確定散射參數(shù)。5.4.1波導(dǎo)端口撤消修改高頻元件通常是由波導(dǎo)饋，如微帶同軸或共面線。在仿真模型的進(jìn)料線延長計(jì)算空間的邊界，如圖。 5.9a-b。在邊界波導(dǎo)端口被定義，模擬半無限傳輸行向前行波送入。在

16、時(shí)域的情況下解決方案的過程（例如，時(shí)域有限差分法）來波進(jìn)入計(jì)算空間，互動(dòng)的結(jié)構(gòu)與入射功率的一小部分旅行通過端口備份。從傳入和傳出的信號散射參數(shù)可以計(jì)算。為了激發(fā)結(jié)構(gòu)，領(lǐng)域向前傳播模式上料線的波待定。由于總是比較一個(gè)模式，可以沿直線傳播，端口模式是由用戶指定。在大多數(shù)技術(shù)應(yīng)用的基本模式使用，例如，上同軸線tem波在矩形te10模式波導(dǎo)。在該行的橫截面的二維場模式計(jì)算之前，通過一個(gè)特殊的二維求解全三維分析。作為一個(gè)二維的例子準(zhǔn)tem微帶線的基本模式是港口領(lǐng)域的格局描繪圖。 5.9c。入射波進(jìn)入結(jié)構(gòu)是二維平面的大小一個(gè)關(guān)鍵點(diǎn)。如果只有基本模式感興趣的是，2d平面入射波進(jìn)入結(jié)構(gòu)沒有大到足以讓在港口區(qū)的

17、邊緣被扭曲。另一方面，如果2d平面上升，高階模式可以傳播。港口區(qū)的大小取決于線的幾何形狀，以及基板的介電常數(shù)作為頻率。對于許多實(shí)際問題，一個(gè)合理的起點(diǎn)港區(qū)是：五至十倍的基板的高度和五至十倍寬度跟蹤。在mom代碼，定義沒有計(jì)算量，微帶線可興奮的跟蹤邊緣。在不連續(xù)的高階模式出現(xiàn)。如果是高階模式非傳播模式，他們迅速腐爛，因?yàn)樗麄冄刂本€傳播。港口和連續(xù)性之間的距離應(yīng)該足以允許非蔓延（逝）模式衰減。如果高階模式設(shè)備的操作是必不可少的，必須考慮到這些模式在與這些模式相關(guān)的分析和s參數(shù)必須計(jì)算。5.4.2平面波激勵(lì)激發(fā)入射平面波的結(jié)構(gòu)圖所示。 5.10在各種應(yīng)用中非常有用。一個(gè)例子是計(jì)算雷達(dá)圖5.9。波導(dǎo)微

18、帶端口：（一）頂視圖，（二）透視圖（三）二維場模式（端口模式）入射波在側(cè)視圖截面（rcs）的散射對象。另一個(gè)例子是分析排入中的電磁兼容性（emc）領(lǐng)域。入射平面波的振幅通常描述電場矢量，極化矢量和傳播方向。內(nèi)部結(jié)構(gòu)，電流，電壓和當(dāng)?shù)氐碾娏τ^測點(diǎn)和磁場值可以被定義。此外，散射場進(jìn)行采樣，以計(jì)算對象的雷達(dá)截面。撤消修改圖5.10。平面波令人興奮的一個(gè)被動(dòng)的平面結(jié)構(gòu)。坡印廷矢量s指示方向的傳播和e和h的電場和磁場領(lǐng)域的組成部分，分別。5.4.3集總端口集中或集中的端口，可用于計(jì)算空間內(nèi)激勵(lì)結(jié)構(gòu)。他們通常都放在兩個(gè)導(dǎo)電部件如電壓或電流源模型和行為與定義的特征阻抗。圖5.11顯示了一個(gè)集中的港口，激發(fā)中

19、心饋電偶極。圖5.11。集總端口激動(dòng)人心的一個(gè)中心饋電偶極子天線集中可以計(jì)算空間內(nèi)使用的端口是多靈活的比通常是在邊界定義的端口波導(dǎo)問題域。此外，他們并不需要加強(qiáng)離散飼養(yǎng)結(jié)構(gòu)。因此，他們提供的通常是一個(gè)很好的起點(diǎn)為簡單的仿真模型，并可以取代波導(dǎo)端口，如果有更多的精確的模型或模式的分離是必需的。集中的港口限于相比是很小的喂養(yǎng)結(jié)構(gòu)波長。他們可以通過評估散射參數(shù)計(jì)算在港口的電壓和電流。撤消修改creating an efficient simulation modelin this chapter we describe the general structure of numerical model

20、ing softwareand the different steps towards an efficient simulation model.5.1 general structure of numerical modeling softwarenumerical simulation software consists of three main components preprocessor,solver, and postprocessor as shown in fig. 5.1.fig. 5.1. main components of numerical simulation

21、softwarethe pre- and postprocessor are incorporated into an interactive graphicaluser interface (gui). the solver is started and controlled from this interface.although today the user may feel only one software, the division of thesoftware into three main components makes sense from a logical and di

22、dacticpoint of view.5.1.1 preprocessorthe preprocessor is used to set up the simulation model as shown in fig. 5.2.fig. 5.2. basic simulation model including antenna geometry, material properties,ports, boundary conditions, box for nearfield recordingin the following list we give a short overview ov

23、er the tasks that are addressedwithin the process of model generation. each item is explained inmore detail in the subsequent sections.general definitions: at this stage different layers or components are definedto organize the data. furthermore, the units of physical quantitiesare selected, e.g., l

24、ength in millimeters, frequency in ghz.geometry: the geometrical shapes of the objects are defined. there are ingeneral three different approaches that can be combined: the geometry isentered interactively by a graphical user interface (gui), the geometricaldata is imported from a cad file, or the o

25、bjects are described by a macrolanguage.material properties: dielectric properties are assigned to the geometricalobjects that have been modeled in the first step. materials are selectedfrom a database of predefined commonly used materials or new materialsare defined by specifying dielectric and mag

26、netic properties.excitation: ports are defined in order to excite the structure and to evaluatecircuit-based quantities like scattering parameters, impedances, voltagesand currents.boundary conditions: at the outer boundary of the simulation volume differentkinds of boundary conditions are applied i

27、n order to represent freespace, electric and magnetic walls or planes of symmetry.meshing: the structure is discretized into small elements with homogeneousmaterial properties. the discretization of the model is a crucial step sinceit affects the required numerical effort and the accuracy of the sol

28、ution.simulation parameters: additional parameters are defined to control thesolution process, e.g., the kind of solver to be used, definition of an endcriterionfor time-stepping algorithms, desired accuracy, ar filtering, multiportexcitation, number of port modes.5.1.2 solverthe solver calculates t

29、he approximate solution of the electromagnetic problembased on the data generated during preprocessing. the solution process istypically documented in a so-called log-file that lists important details andstatistics of the simulation, for example: time need for the solution, number ofunknowns, memory

30、 requirements, as well as warnings and errors encounteredduring the solution process.5.1.3 postprocessorthe postprocessor is used to evaluate the results of the solver. the results canbe displayed in tables, 1d-, 2d- and 3d-plots.port-related parameters: parameters like input impedances, scatteringp

31、arameters, voltages and currents are extracted from the field solver results.these parameters are needed to specify the behavior of the structurewhen connected to a circuit or other rf components.visualization: field distributions (contour and arrow plots) of differentnearfield quantities are displa

32、yed. the visualization of field distributionsgives the user insights into the way the structure works. this understandingcan help to improve the performance of the structure or help to identifyerrors in the model.antenna parameters: antenna parameters like radiation pattern, gain, directivity,half-p

33、ower beam width, side lobe suppression and radiation efficiencyare calculated.additional parameters: for example, evaluation of specific absorption rate(sar) and emc related data like field strengths in a specified distancein comparison to emc limits.84 5 creating an efficient simulation model5.2 ge

34、ometrythe first step in setting up the simulation model is the generation of thethree-dimensional structure, e.g., the shape of an antenna or of a filter.in up-to-date electromagnetic modeling software usually three ways existto accomplish this goal: interactive construction via graphical user inter

35、face (gui), import of cad (computer aided design) data and object definition via macro language.5.2.1 interactive construction via graphical user interface (gui)a graphical user interface (gui) helps the user to create interactively thegeometry of his model. the gui construction capabilities are ver

36、y similar tothose that are available in basic computer aided design (cad) software.fig. 5.3. simple three-dimensional objectsusually the construction starts with the definition of simple objects. morecomplex objects can be generated by combining and manipulating simple objects.simple objects can be

37、classified into one-, two- and three-dimensionalshapes. simple one-dimensional objects include line, polygon and spline segment.simple two-dimensional objects include triangle, rectangle, circle, ellipseand closed polygon. simple three-dimensional objects include box, cylinder,cone and sphere (see f

38、ig. 5.3). depending on the software more complex basicshapes like toroid, spiral, helix, ellipsoid, and bond-wires are available.these simple objects can be modified, i.e., rotated, moved, mirrored,copied, scaled or stretched. furthermore, boolean operations can be appliedon overlapping elements. in

39、 fig. 5.4 the effects of subtraction, addition andintersection are shown. if we draw two overlapping 3d objects like a box and asphere (fig. 5.4a) and we subtract the sphere from the box the resulting shapeis a box with a spherical cut-out in it as shown in fig. 5.4b. the boolean operationaddition c

40、ombines the two objects as shown in fig. 5.4c. if we intersectthe two objects we get an object that consists of the common parts of the boxand sphere as shown in fig. 5.4d.fig. 5.4. objects modified by different boolean operations: (a) overlapping boxand sphere, (b) subtraction of sphere from box, (

41、c) addition and (d) intersectionof box and sphere、another way to create more complex 3d objects is to extrude, twist, orrotate 2d entities. figure 5.5 shows the basic concepts. in fig. 5.5a a 2d faceis extruded along a vector. in fig. 5.5b a 2d face is twisted along a curve andin fig. 5.5c a 2d face

42、 is rotated around an axis. these operations are verypowerful in creating complex geometries. usually the output can be controlledby additional options, for example the object can be tapered, i.e., the size ofthe 2d face varies as it propagates along the extrusion vector (fig. 5.5d).sometimes the gu

43、i includes advanced features like the description of objectsby complex mathematical formulas or the smoothening (rounding) of edges.although the graphical generation of geometrical shapes has become verypowerful, the geometry should be kept as simple as possible. only details thataffect the electrom

44、agnetic behavior should be included. rounded edges resultfig. 5.5. creation of 3d objects from 2d contoursin an appealing model but lead to a finer discretization of the structure andthus increase the computational burden.5.2.2 object definition via macro languagein the early days of em modeling the

45、 definition of a model by a text-basedinput file was common use. today this kind of model creation is still usefulif many structures with similar geometry are analyzed. geometrical featurescan be defined by variables and are easily changed. therefore, text-basedinput files are suited for parameter s

46、tudies and optimization routines. mostem modeling software packages provide a macro language for the text-baseddefinition of models.5.2.3 import of cad (computer aided design) datain em modeling we can distinguish between design and analysis of microwavecomponents. in the latter case it may be possi

47、ble to use cad data of actualmicrowave components from the production process. figure 5.6 shows as an5.2 geometry 87example an antenna module consisting of a three-dimensional carrier with acomplex-shaped antenna plate.mechanical cad software like proengineer, solidworks or catia possessmany more po

48、werful features for the design of complex-shaped mechanicalobjects than the basic cad tools that are incorporated in em simulationsoftware. in order to import models originating from mechanical cad softwareone can use standardized exchange data formats like step (standard for the exchange of product

49、 model data) iges (initial graphics exchange specification) dxf (drawing exchange format).fig. 5.6. cad model of an antenna modulealthough import of cad data may seem a very elegant way of modelingthere are some drawbacks. the first problem with cad data is that manymechanically important but electr

50、ically irrelevant details are included. thesecad details result in a finer mesh and therefore increase the simulation timewithout significantly improving the accuracy of the results. another problemwith cad data may arise from geometrical accuracy (precision). if the cadsystem and the gui of the em

51、modeling software apply different degreesof geometrical precision unexpected errors may occur. for example, primaryconnected objects show overlapping areas or are separated.from this we conclude that cad import is seldom a one-click task. thesubsequent model simplification and data healing are cumbe

52、rsome tasks.as an example fig. 5.7a shows an antenna module that includes smallholes and gaps. these details are included in the original cad file for mechanicalreasons (fastening) but are electrically irrelevant. figure 5.7b showsthe antenna module after the unnecessary details are removed. figures

53、 5.7cfshow how the small details result in finer meshes that lead to longer simulationtimes.another example concerns thin metallic structures. these thin structures,e.g., a strip conductor of a microstrip line shown in fig. 5.8, can be modeled asa two-dimensional (flat) object. this modeling approac

54、h leads to simulationfig. 5.7. removal of details from a cad model: (a) original cad data containingsmall irrelevant details, (b) simplified model (details removed), (c-f) effects onorthogonal and triangular meshingmodels with less unknowns and increased cell sizes compared to volumetricmodeling.5.2

55、.4 summarywhen setting up the geometry of a simulation model the intention shouldnot be to draw the most accurate structure from the visual point of viewbut the simplest one that represents the electromagnetic phenomena of thestructure under consideration. moreover, too many details can enlarge thec

56、omputational effort, on the one hand, and result in an error-prone, difficultto validate simulation model.fig. 5.8. fdtd discretization of cross-section of a microstrip line. simplified modelby representation of the thin strip conductor by a two-dimensional platesometimes is it not so clear to decid

57、e initially if a certain geometricaldetail is significant or not. in this case it is recommended to start with thesimple model and to validate the specific influence of this detail in a separatesimulation model.5.3 material propertiesafter a geometrical structure is defined, material properties are assigned todifferent objects. for example, in a microstrip structure we have a metallicground plane and metallic traces, the substrate is dielectric and the volumeabove the top layer is air.technically importan