怎么学习计算电磁学【QUORA】

链接

There are several resources. But it depends on what you actually want to learn...Let me explain:

First of all, There are many Numerical Modeling Techniques in Electromagnetic, and people are always confused about which one to choose. But if you are not able to decide, then best thing is to start with any of them (I am not talking about conventional theoretical approaches which cannot be done theoretically for bigger problem spaces). All numerical techniques can be used for all type of EM problems but major aspect is efficient solution, in terms of results and resources.

1. If you want to know about general idea about CEM, then its better you begin with Wikipedia itself, which has a great amount of details about numerical modeling in electromagnetics.

Computational electromagnetics

2. If you are not a beginner, then you might have already chosen, which numerical modeling you are interested in, like FDTD, FEM, MoM, FIT, etc.

If you have not chosen yet and want to decide which one to be considered for your numerical solutions of EM problems, then do not worry. Dr. Raymond Rumpf has created many lectures on CEM on youtube. So take a look at them.

If you want to go into full depth mathematical analysis then go through Dr. Cynthia Furse‘s videos on youtube.

I loved these two...I guess you may also like it at least. :)

3. If you have decided now, which numerical modeling you are interested in, then there are several books and videos available to learn each and every one of them.

There are many numerical modeling techniques, some of them you might have seen on Wikipedia itself.

But it depends on the EM problem, which decides choice the efficient way of numerical solution and numerical modeling method. So once you have decided it  according to your EM problem then only start working on specific numerical modeling.

I like mainly following Numerical Techniques:

I). Finite Difference Time Domain Method (FDTD)

There are youtube videos to understand it easily but I would rather prefer if you go through:

Prof. Kane Yee‘s fundamental paper (1966). This paper itself is more than sufficient to understand FDTD for EM problems.

http://ecee.colorado.edu/~mcleod...

And a paper by Prof. Taflove

http://www.ece.northwestern.edu/...

Trust me if you can understand these, it is more than sufficient to begin with FDTD. I started with these only. Then there are many papers by Prof. Allen Taflove, Prof. Karl S. Kunz and many others.

First good book on FDTD by Kunz  & Luebbers is like an epic for numerical scientists.

The Finite Difference Time Domain Method for Electromagnetics

Then there are many books by many professors, Prof. Taflove, Prof. Sadiku, Prof. Elsherbeni. You can just google them and search for free pdfs also. :) I will not put links here. :)

Even if you are interested in Videos only, then go through lecture series by Prof. C. Furse:

If you want to use commercially available Tools then:

i. FEKO -- https://www.feko.info/

ii. CST -- https://www.cst.com/

iii. FDTD Solutions -- Lumerical‘s Nanophotonic FDTD Simulation Software

Free FDTD codes and tools are

  • FDTD++: advanced, fully featured FDTD software, with included C++ source code, along with sophisticated material models and predefined fits as well as discussion/support forums and email support
  • openEMS (Fully 3D Cartesian & Cylindrical graded mesh EC-FDTD Solver, written in C++, using a Matlab/Octave-Interface)
  • pFDTD (3D C++ FDTD codes developed by Se-Heon Kim)
  • JFDTD (2D/3D C++ FDTD codes developed for nanophotonics by Jeffrey M. McMahon)
  • WOLFSIM (NCSU) (2-D)
  • Meep (MIT, 2D/3D/cylindrical parallel FDTD)
  • (Geo-) Radar FDTD
  • bigboy (unmaintained, no release files. must get source from cvs)
  • toyFDTD
  • Parallel (MPI&OpenMP) FDTD codes in C++ (developed by Zs. Szabó)
  • FDTD code in Fortran 90
  • FDTD code in C for 2D EM Wave simulation
  • Angora (3D parallel FDTD software package, maintained by Ilker R. Capoglu)
  • GSvit (3D FDTD solver with graphics card computing support, written in C, graphical user interface XSvit available)
  • gprMax (Open Source (GPLv3), 3D/2D FDTD modelling code in Python/Cython developed for GPR but can be used for general EM modelling.)

These are taken from Wikipedia...

II). Method of Moments (MoM)

This is also one of the oldest methods in electromagnetics for solving big problems.

There is one good book by Gibson: Search on google and first link... :)

https://www.google.de/url?sa=t&r...

I have read this book and its awesome...you need patience ... :)

There are many books with MATLAB coding of MoM. If you want to implement it with MATLAB, then Go Through:

Fundamentals of Electromagnetics with MATLAB, By Karl Erik Lonngren, Sava Vasilev Savov, Randy

Commercial Tools:

i. FEKO --https://www.feko.info/ --------Best available software for MoM

ii. NEC codes --Numerical Electromagnetics Code NEC2 unofficial home page

III). Finite Element Method (FEM)

For 2D solutions, to begin with, just go through:

http://deepblue.lib.umich.edu/bi...

There are many books on FEM for Electromagnetics, but I would rather prefer it for specific numerical problems only. Or maybe inside FEKO with hybridizing with MoM.

Commercial Tools:

i. FEKO --https://www.feko.info/ --------Best available software for MoM

ii. HFSS -- http://www.ansys.com/Products/Si...

IV). Finite Integration Technique (FIT)

It was proposed by Prof. Thomas Weiland. IT has basic papers written by Prof. Weiland. There is a commercially available tool company, which is founded by Prof. Weiland and his colleagues only, known as CST Studio Suite.

CST -- https://www.cst.com/

------------------------------------------------------------------------------

There are many enhancements in these numerical modeling techniques and hybridization with many other numerical techniques (MLFMM, PO, GO etc.) for specific applications. For more information use the Wikipedia page which I have posted first... :)

-----

I hope this is useful to someone... :)

时间: 2024-12-13 16:47:26

怎么学习计算电磁学【QUORA】的相关文章

深度学习计算模型中“门函数(Gating Function)”的作用

/* 版权声明:可以任意转载,转载时请标明文章原始出处和作者信息 .*/ author: 张俊林 看深度学习文献,门函数基本上已经是你必然会遇到的一个概念了,最典型的就是LSTM,首先上来你就得过得去"遗忘门""输入门""输出门"这三个门.门函数本身是个独立概念,不过LSTM使用多个门函数来组合出一个带有状态记忆的计算模型而已.随着LSTM大行其道,各种计算模型开始在计算过程中引入门函数的概念,相信这些论文你也没少看,其实这也是一种研究模式,比如

Python学习-计算BMI的小程序

示例: 小明身高1.75,体重80.5kg.请根据BMI公式(体重除以身高的平方)帮小明计算他的BMI指数,并根据BMI指数: 低于18.5:过轻 18.5-25:正常 25-28:过重 28-32:肥胖 高于32:严重肥胖 用if-elif判断并打印结果: # -*- coding: utf-8 -*- height = 1.75 weight = 80.5 bmi = 80.5/1.75*1.75 if bmi < 18.5:     print('过轻') elif bmi >= 18.

Caffe深度学习计算框架

Caffe | Deep Learning Framework是一个清晰而高效的深度学习框架,其作者是博士毕业于UC Berkeley的 Yangqing Jia,目前在Google工作.Caffe是纯粹的C++/CUDA架构,支持命令行.Python和MATLAB接口:可以在CPU和GPU直接无缝切换: 1 Caffe::set_mode(Caffe::GPU); Caffe的优势 上手快:模型与相应优化都是以文本形式而非代码形式给出. Caffe给出了模型的定义.最优化设置以及预训练的权重,

深度学习计算框架实现

参考与评述 参考书目<Deep Learning>Lan Goodfellow. 经典的深度学习框架是以计算图&梯度下降方法实现对前馈网络的有监督学习. 这里复现了前馈计算图的梯度计算实现. 一.前馈计算图实现 1. 前向与梯度计算 结果数组 (保存输入节点与计算节点的输出值,能够反映节点在计算方向的拓扑排序) 梯度数组 (保存输入节点与计算节点的梯度,能够反映节点在计算方向的拓扑排序) 连接图 (反映每个节点的父节点) 输出函数集合 (反映每个计算节点如何根据其输入得到输出) 梯度函

Vue.js 基础学习计算属性computed

我们要写一个成绩表如下 数学 90 物理 80 英语 70 <div id="app"> <table border="1"> <tr> <td>数学</td> <td>{{ math }}</td> </tr> <tr> <td>物理</td> <td>{{ physics }}</td> </tr>

MXNET:深度学习计算-自定义层

虽然 Gluon 提供了大量常用的层,但有时候我们依然希望自定义层.本节将介绍如何使用 NDArray 来自定义一个 Gluon 的层,从而以后可以被重复调用. 不含模型参数的自定义层 我们先介绍如何定义一个不含模型参数的自定义层.事实上,这和 "模型构造" 中介绍的使用 Block 构造模型类似. 通过继承 Block 自定义了一个将输入减掉均值的层:CenteredLayer 类,并将层的计算放在 forward 函数里. class CenteredLayer(nn.Block)

openstack学习-计算管理

实验通过openstack dashboard和openstack CLI两种方式管理Hypervisor.主机聚合.规格密钥对以及虚拟机组的测试,并测试虚拟机发放.生命周期管理以及快照和重建等 Openstack Dashboard操作 Hypervisor和主机聚合管理 主机聚合通过将主机组合到一起来把可用区域划分成逻辑单元.创建一个主机聚合,然后选择要放里面的主机. 使用admin用户登陆Openstack Dashboard界面.在导航栏选择"管理员-计算-虚拟机管理器",进入

计算电磁学方法

学习纠错码和相关的计算代数系统

过去的一年致力于IEEE 802.3bj的前向纠错编码模块(FEC),发现可以把我的对偶综合工作扩展到FEC上,包括FEC的形式化验证和对偶综合. 这就导致了我需要学习纠错码,而相关的多项式环和有限域操作需要在计算代数系统上进行操作,这就导致了我需要学习计算代数系统. 常见的计算代数系统由Singular(www.singular.uni-kl.de)和GAP(www.gap-system.org). Singular被几个相关的工作使用,比如发表于FMCAD12的Formal Verifica