Professional Documents
Culture Documents
Textbook Advanced Electromagnetic Computation Second Edition Dikshitulu K Kalluri Ebook All Chapter PDF
Textbook Advanced Electromagnetic Computation Second Edition Dikshitulu K Kalluri Ebook All Chapter PDF
https://textbookfull.com/product/principles-of-electromagnetic-
waves-and-materials-second-edition-dikshitulu-k-kalluri/
https://textbookfull.com/product/biota-grow-2c-gather-2c-cook-
loucas/
https://textbookfull.com/product/advanced-materials-for-
electromagnetic-shielding-fundamentals-properties-and-
applications-maciej-jaroszewski/
https://textbookfull.com/product/multivariable-advanced-calculus-
kuttler-k/
Time Series: Modeling, Computation, and Inference,
Second Edition Raquel Prado
https://textbookfull.com/product/time-series-modeling-
computation-and-inference-second-edition-raquel-prado/
https://textbookfull.com/product/advanced-r-second-edition-
hadley-wickham/
https://textbookfull.com/product/advanced-heat-transfer-second-
edition-f-naterer/
https://textbookfull.com/product/reinforcement-learning-an-
introduction-adaptive-computation-and-machine-learning-series-
second-edition-sutton/
https://textbookfull.com/product/introduction-to-probability-
second-edition-joseph-k-blitzstein/
Advanced Electromagnetic
Computation
Second Edition
Advanced Electromagnetic
Computation
Second Edition
Dikshitulu K. Kalluri
CRC Press
Taylor & Francis Group
6000 Broken Sound Parkway NW, Suite 300
Boca Raton, FL 33487-2742
This book contains information obtained from authentic and highly regarded sources. Reasonable efforts have been
made to publish reliable data and information, but the author and publisher cannot assume responsibility for the
validity of all materials or the consequences of their use. The authors and publishers have attempted to trace the copy-
right holders of all material reproduced in this publication and apologize to copyright holders if permission to publish
in this form has not been obtained. If any copyright material has not been acknowledged, please write and let us know
so we may rectify in any future reprint.
Except as permitted under U.S. Copyright Law, no part of this book may be reprinted, reproduced, transmitted, or
utilized in any form by any electronic, mechanical, or other means, now known or hereafter invented, including pho-
tocopying, microfilming, and recording, or in any information storage or retrieval system, without written permission
from the publishers.
For permission to photocopy or use material electronically from this work, please access www.copyright.com (http://
www.copyright.com/) or contact the Copyright Clearance Center, Inc. (CCC), 222 Rosewood Drive, Danvers, MA
01923, 978-750-8400. CCC is a not-for-profit organization that provides licenses and registration for a variety of users.
For organizations that have been granted a photocopy license by the CCC, a separate system of payment has been
arranged.
Trademark Notice: Product or corporate names may be trademarks or registered trademarks, and are used only for
identification and explanation without intent to infringe.
Preface...............................................................................................................................................ix
Acknowledgment............................................................................................................................xi
Author............................................................................................................................................ xiii
v
vi Contents
Part VI Appendices
Problems .......................................................................................................................................345
Index..............................................................................................................................................355
Preface
The second edition retains the flow and the flavor of the first edition, which was stated in
the preface to the first edition. Item 1 describes this flavor in a few words.
ix
x Preface
of the Laplace transform method as well as the location of the c urrent variable
at the center of the Yee’s cube to implement the constitutive equation of the
medium numerically by solving the associated auxiliary vector differential
equation. This algorithm and its illustration through the computation of the
fields in a time-varying magnetoplasma medium is one of the highlights of
Volume 2.
d. Section 20.5 is new to this edition. A one-dimensional FDTD solution of inter-
action of a pulsed wave with a switched plasma slab (given in Section 20.4 in
the first edition) is extended to the case of a switched magnetoplasma slab. It is
shown that the wiggler magnetic field in Section 20.4 is now converted to a
low-frequency whistler wave.
I thank Dr. Robert Kevin Lade for generating the results in Section 20.5, which are based on
his doctoral thesis.
xi
Author
xiii
Part V
Electromagnetic Computation
15
Introduction and One-Dimensional Problems
é ¶2 ¶2 ¶2 ù
ê 2 + 2
+ E + k 2E z = 0,
2ú z
(15.1)
ë ¶x ¶y ¶z û
k 2 = w2me. (15.2)
E z ( x ,y ,z ) = F ( x ,y ) e- jbz (15.3)
é ¶ 2F ¶ 2F ù 2
ê 2 + 2 ú + kc F = 0, (15.4)
ë ¶x ¶y û
where
kc2 = k 2 - b2 . (15.5)
3
4 Advanced Electromagnetic Computation
In Equation 15.6, Ñ 2t is the Laplacian operator in the transverse plane and λ is the eigen-
value. The waveguide problem is the eigenvalue problem involving a differential operator
and when a numerical technique is employed, the problem is converted into the familiar
form of
AX = lX , (15.7)
where A is a square matrix and λ’s are the eigenvalue and X’s are the corresponding eigen-
vectors. Equation 15.6 is a scalar second-order PDE and we have seen that for a rectangular
region 0 < x < a and 0 < y < b. One can solve it analytically by using separation of variable
technique and on imposing the PEC boundary conditions (Figure 15.1), one obtains
where
2 2
æ mp ö æ np ö
l mn = ( kc )mn = k x2 + k y2 = ç
2
÷ +ç ÷ . (15.9)
è a ø è b ø
Recognize in Equations 15.8 and 15.9 that the problem has double infinity number
of eigenvalues λmn and the corresponding eigenvectors are proportional to sin(mπx/a)
sin (nπy/b).
Many practical problems do not have such elegant analytical solution. For example,
Figure 15.2 shows a ridged waveguide used in practice to increase the bandwidth of
single-mode propagation of a rectangular waveguide. Numerical methods (finite differ-
ence, finite element and moment, and other methods) or approximate analytical methods
(perturbation or variational technique) are used to solve such problems [1–10]. We give
next integral formulation of an electromagnetic field problem. Suppose we wish to find the
capacitance of a square conducting plate located in free space with reference to sphere at
infinity (Figure 15.3).
~
PEC, Ez = 0, F = 0
y
n̂
ε, μ n̂
2 2
t F + kc F = 0
x
a
FIGURE 15.1
Rectangular waveguide: TM modes.
Introduction and One-Dimensional Problems 5
ε, μ
2 2
t F + kc F = 0
a x
FIGURE 15.2
Ridged waveguide.
P(x, y, z)
z
R
Rsp
–∞
S(x΄, y΄)
2a
PEC y
2a
x
FIGURE 15.3
Conducting plate in free space.
The key step in the solution is determining the surface charge density ρs(x, y) on the con-
ducting plate, assuming that the voltage of the plate is V0 volts. In terms of ρs(x′, y′), the
potential at a general point P(x, y, z) is given by
a a
rsdx¢dy¢
F ( x ,y ,z ) =
òò
- a- a
4peRsp
, (15.10)
where
( x - x¢ ) + ( y - y¢ )
2 2
Rsp = + z2 . (15.11)
6 Advanced Electromagnetic Computation
If the point P is on the plate, that is, z = 0, −a < x < a, and −a < y < a, the potential is a
constant given by V0:
a a
rs dx¢dy¢
VP = V0 =
òò 4pe
- a- a ( x - x¢ ) + ( y - y¢ )
2 2
( P on the plate ) . (15.12)
In Equation 15.12, the unknown ρs appears under the integral sign and hence
Equation 15.12 is called an integral equation.
15.3 One-Dimensional Problems
In the context of this discussion, some examples of one-dimensional problems of electro-
magnetics, are as follows.
a. Static problems:
d 2V r
dz 2
=- v
e
( one-dimensional Poisson’s equation ) . (15.13)
b. Time-harmonic problem:
d 2V
dz 2
+ k 2V = 0 ( transmission line equation ) . (15.14)
¶ 2V ¶ 2V
2
- me 2 = 0. (15.16)
¶z ¶t
Problems (a) and (b) are classified as equilibrium problems; the boundary conditions at the
endpoints, which bound the one-dimensional domain, are specified. Problem (c) is an
eigenvalue problem where the eigenvalue is l = kc2 of the differential operator with the
specified boundary conditions of the potential V on the endpoints of the domain. Problem
(d) is called propagation or marching problem. The boundary in the time domain is open
and the initial conditions and the boundary conditions at the endpoints are specified.
15.3.1 Finite Differences
Let y(x) be tabulated at equal intervals h and the table of values xn (independent variable)
and yn (dependent variable) are available, where
xn = x0 + nh, (15.17)
y n = y ( xn ) , n = 0, 1, 2, ¼. (15.18)
We can express the first derivative at a tabulated point in terms of the values of y at the
neighboring points. For example, if we approximate the tangent to the curve y(x) at P
(see Figure 15.4) by the Chord 1:
dy y - yn
» n +1 + O ( h). (15.19)
dx n h
yn+1
3 1
yn
2
P
yn–1
h h
x
xn–1 xn xn+1
FIGURE 15.4
Approximation of the first derivative.
8 Advanced Electromagnetic Computation
On the other hand, we can approximate the derivative by the Chord 2 and write
dy y - y n -1
» n + O ( h). (15.20)
dx n h
dy y - y n -1
dx n
» n +1
2h
+ O h2 . ( ) (15.21)
Note that the order of approximation in Equation 15.21 is stated to be O(h2). It would be
useful to prove the order of approximation in Equations 15.19 through 15.21 and find a
systematic approach to get higher-order approximation formulas; such an approach is
available through “finite difference calculus” [1].
We start by defining some operators and find the symbolic relationship between the
operators.
Displacement operator E:
y n+1 = Ey n ,
(15.22)
y n + 2 = E ( y n +1 ) = E 2 y n ,
\ y n + m = Em y n . (15.23)
Averaging operator μ:
1
my n = é y n +1/2 + y n -1/2 ùû . (15.24)
2ë
Forward-difference operator Δ:
Dy n = y n +1 - y n ,
(15.25)
Dy n = y n +1 - y n = Ey n - y n = ( E - 1) y n ,
\ D = E - 1. (15.26)
Backward-difference operator ∇:
Ñy n = y n - y n-1 ,
(15.27)
\Ñ = 1 - E-1 ,
E = (1 - Ñ ) .
-1
(15.28)
Central-difference operator δ:
dy n = y n +1/2 - y n -1/2 , (15.29)
D = ÑE = dE1/2 , (15.31)
1 2
m2 = 1 + d. (15.32)
4
Differential operator D:
dy
Dy = . (15.33)
dx
dy h 2 d 2 y
Ey ( x ) = y ( x + h ) = y ( x ) + h + +
dx 2 ! dx 2
æ h 2D 2 ö
= ç 1 + hD + + ÷ y ( x ) = e hD y ( x ) , (15.34)
è 2! ø
E=e ,hD
hD = ln E = ln ( 1 + D ) , (15.35)
= ln ( 1 - Ñ ) = - ln ( 1 - Ñ ) .
-1
(15.36)
é 1 1 ù
hy¢j = hDy j = - éëln ( 1 - Ñ ) ùû y j = êÑ + Ñ 2 + Ñ 3 + ú y j ,
ë 2 3 û
(15.37)
1é 1 2 1 3 ù
\ D = êÑ + Ñ + Ñ + ú .
hë 2 3 û
Let us terminate the series in Equation 15.37 with one term only:
1 1
Dy j » Ñy j = éë y j - y j -1 ùû . (15.38)
h h
Equation 15.38 is the same as Equation 15.20. We can improve the accuracy by taking two
terms in Equation 15.37:
1é 1 ù
Dy j » ê Ñ + Ñ2 ú y j ,
hë 2 û
Ñy j = y j - y j-1 ,
Ñ 2 y j = Ñ ( Ñy j ) = Ñ ( y j - y j-1 ) = ( y j - y j-1 ) - ( y j-1 - y j-2 ) , (15.39)
Ñ 2 y j = y j - 2 y j-1 + y j-2 ,
1é
\ Dy j »
h êë
( y j - y j-1 ) + 21 ( y j - 2y j-1 + y j-2 )ùú .
û
10 Advanced Electromagnetic Computation
TABLE 15.1
Relations between Symbolic Operators
E Δ δ ∇ hD
1 2 1
E E 1+Δ 1+ d + d 1 + d2 (1 − ∇)−1 ehD
2 4
1 2 1
Δ E−1 Δ d + d 1 + d2 ∇(1 − ∇)−1 ehD − 1
2 4
1
δ E1/2 − E−1/2 Δ(1 + Δ)−1/2 δ ∇(1 − ∇)−1/2 2 sinh hD
2
1 1
∇ 1 − E−1 Δ(1 + Δ)−1 - d2 + d 1 + d2 ∇ 1 − e−hD
2 4
1
hD log E log(1 + Δ) 2 sinh -1 d −log(1 − ∇) hD
2
æ 1 ö æ 1 ö
μ
2
(
1 1/2
E + E-1/2 ) ç 1 + 2 D ÷ (1 + D )
è ø
-1/2
1+
1 2
4
d ç 1 - 2 Ñ ÷ (1 - Ñ )
è ø
-1/2
cosh
1
2
hD
Source: Adapted from Thom, A. and Apelt, C.J., Field Computations in Engineering and Physics,
Van Nostrand Reinhold, London, U.K., 1961.
Simplifying, we obtain
1
Dy j »
2h ë
( )
é 3 y j - 4 y j -1 + y j - 2 ùû + O h 2 . (15.40)
Equation 15.40 is more accurate than Equation 15.38 but note that Equation 15.40 involves
more neighboring points than Equation 15.38. It can be shown that, using the forward-
difference formula,
1
Dy j =
2h ë
( )
é - y j + 2 + 4 y j +1 - y j ùû + O h 2 (15.41)
d2 y 1
2
dx j
= D2 y j = 2 éë y j +1 - 2 y j + y j -1 ùû + O h 2 .
h
( ) (15.42)
dx
+ x = 0, (15.43a)
dt
Introduction and One-Dimensional Problems 11
x (0) = 1 (15.43b)
t2 t3
x = e-t = 1 - t + - + . (15.44)
2! 3!
The exact solution is an infinite series and one can evaluate the series on a computer to
the accuracy desired. Let us suppose that we put the restriction that we cannot use more
than three terms and write an approximate solution
t2
xT ( t ) = 1 - t + . (15.45)
2
We will show that we can do better than Equation 15.45 if we write that x(t) is approxi-
mately equal to xa(t):
dx
R (t ) =
dt
(
+ x = 1 + c1 ( 1 + t ) + c2 2t + t 2 ) (15.47)
is minimized in some sense. Of course, R(t) will be zero for all t in the domain for an exact
solution of x. Let us examine several possible ways of defining the minimization of R(t).
æ1ö æ 1ö é 1 æ 1 ö2 ù
R ç ÷ = 1 + c1 ç 1 + ÷ + c2 ê 2 ´ + ç ÷ ú ,
è3ø è 3ø êë 3 è 3 ø úû
giving
4 7
c1 + c2 = -1. (15.48a)
3 9
5 16
c1 + c2 = -1. (15.48b)
3 9
12 Advanced Electromagnetic Computation
15.3.2.2 Subdomain Method
The domain is divided into as many subdomains as there are adjustable constants; in this
case two and use the criteria that the average value of the residue over each of the sub
domains is zero. Let the subdomains be 0 < t < 1/2 and 1/2 < t < 1. The equations for
determining c1 and c2 are obtained from
1/2 1
ò R (t )dt = 0, ò R (t ) dt = 0.
0 1/2
(15.51)
15.3.2.3 Galerkin’s Method
In Equation 15.46, we have used as our “basis functions” t0, t1, t2 which belong to a series
and we let the coefficient of t0 be 1 to satisfy the initial condition and we let the coefficients
c1 and c2 adjustable so as to minimize the residue R in some sense. The Russian engineer
Galerkin used the basis functions as the weights and defined the criteria as reducing the
weighted average of the residue over the entire domain to zero:
1 1
ò tR ( t )dt = 0;
òt R (t )dt = 0.
2
(15.53)
0 0
1 1
¶ ¶R
¶c1 ò
R2 ( t )dt = 2
0
¶c1 ò
R ( t )dt = 0.
0
Introduction and One-Dimensional Problems 13
1
¶R 4 9 38
ò ¶c
0
2
Rdt = + c1 +
3 4 15
c 2 = 0. (15.55b)
Figure 15.5 compares the approximate solutions by calculating the difference between
the exact solution x(t) = e−t and the various approximation solutions xa. The worst approxi-
mation is the truncated Talyor-series solution (Equation 15.45) for t > 0.4 and the others are
with in an error of 0.01 in the entire domain. The computational labor among the methods
1–4 is of increasing order. Collocation is the most direct. The four methods can be described
under the category of “method of weighted residuals” since the criterion used to obtain the
algebraic equations may be written as
1
òw (t ) R (t )dt = 0,
0
j j = 1, 2, (15.57)
where wj are the weight functions. The weight functions are sketched in Figure 15.6. The
weight function for the collocation method (also called point-matching technique) is
impulse functions. This may be readily seen from
1
òd (t - t ) R (t )dt = R (t ) ,
0
j j j = 1, 2. (15.58)
0.08
0.07
0.06
Assume c1 = –1, c2 = 0.5 for (15.45)
0.05
xa(t) – e–t
0.04
0.03 (15.45)
0.02
0 (15.55)
(15.52)
–0.01 t
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
FIGURE 15.5
Comparison of the approximation solution.
14 Advanced Electromagnetic Computation
W1 W2
Unit area
Unit area
t t
1 1 2 1
(a) 3 3
W1 W2
1 1
t t
1 1 1 1
(b) 2 2
W1 W2
1 1
t t
(c) 1 1
W1
W2
2 2
1 1
t t
(d) 1 1
FIGURE 15.6
Weight functions for various approximation techniques: (a) collocation, (b) subdomain method, (c) Galerkin’s
method, and (d) least squares.
The weight functions for the subdomain method are pulse functions as shown in
Figure 15.6. The basis functions in this example are called entire domain functions since
they are defined over the entire domain.
One can also use subdomain basis functions that are nonzero over the subdomain and
zero outside the subdomain. More will be said on this topic when we discuss the “moment
method.” In the previous section, we discussed finite differences and let us see whether
Equation 15.43 can also be solved using finite differences. Let us develop this algorithm for
the more general case of Equation 15.43 given in Equations 15.59a and 15.59b:
dx
= f ( t ,x ) , (15.59a)
dt
x ( 0 ) = x0 . (15.59b)
Another random document with
no related content on Scribd:
The Project Gutenberg eBook of Caillou et Tili
This ebook is for the use of anyone anywhere in the United States
and most other parts of the world at no cost and with almost no
restrictions whatsoever. You may copy it, give it away or re-use it
under the terms of the Project Gutenberg License included with this
ebook or online at www.gutenberg.org. If you are not located in the
United States, you will have to check the laws of the country where
you are located before using this eBook.
Language: French
CAILLOU ET TILI
PARIS
CALMANN-LÉVY, ÉDITEURS
3, RUE AUBER, 3
CALMANN-LÉVY, ÉDITEURS
DU MÊME AUTEUR
Format in-18.
J’ai alors regardé les petites filles et j’ai été obligé de constater
qu’il avait raison : quand elles se trouvent, en nombre, devant un
seul petit garçon, ce sont de petites rosses. Mais Caillou ne
s’aperçoit pas qu’elles ne font peut-être que se venger, car il est, lui
Caillou, complètement idiot quand il se trouve seul avec une seule
d’entre elles, ou deux tout au plus. Elles lui font la cour, et il ne s’en
aperçoit pas. Il est poli, mais il s’embête.
Ça doit tenir à deux choses. La première, c’est qu’elles sont
beaucoup plus intelligentes que lui pour leur âge, et moins actives.
Caillou est pour les jeux où l’on remue. Il a besoin d’épancher une
surabondance de force, et s’il parle en jouant c’est pour raconter des
choses absurdes et démesurées. N’oubliez pas que c’est lui qui
voulait m’écraser avec une charrette de vingt-neuf sous. Il est
instinctivement énorme, c’est-à-dire romantique, et la réalité l’ennuie.
Les petites filles ont au contraire le sens des charmes de cette
réalité, elles la voient d’une façon beaucoup plus aiguë et précise.
La seconde différence entre elles et Caillou, c’est qu’elles ont
l’instinct inné de la coquetterie et qu’il en est dépourvu. Caillou existe
pour les petites filles, tandis que les petites filles n’existent point pour
Caillou : ce point de dissidence est grave. Et plus elles sont petites,
plus il les méprise. Il n’aime que ce qui est grand.
… On vient de le conduire chez Jeanne, qui reçoit aussi Vivette.
Ils vont être trois, dans une nursery pour passer deux heures.
Caillou ne discute jamais la décision de ses parents ou de sa bonne,
quand on le mène dans un endroit qu’il ne connaît pas ; il n’a aucune
opinion préconçue. De plus on lui a dit : « Tu seras gentil, n’est-ce
pas ? » Il n’aime pas beaucoup ces avertissements, mais ils lui font
de l’effet. Toute parole agit sur lui, elle émeut sa volonté imaginative
et malléable. Vivette et Jeanne sont d’ailleurs très aimables avec lui.
Elles ne sont que deux. Ce n’est pas aujourd’hui « l’instinct ennemi »
du sexe contre un autre sexe qui parlera, c’est celui de la
coquetterie. Chacune voudrait être celle qui est remarquée, et
d’ailleurs on les a faites très belles. Seulement Vivette, qui est en
visite, a une capote blanche sur la tête, tandis que Jeanne, qui
reçoit, n’a qu’un ruban bleu. Et cela n’est pas sans l’inquiéter. Un
instinct primitif et sauvage porte en effet les enfants à mettre la
beauté, non pas dans les traits, mais dans ce qu’on y ajoute. Pour
une petite fille, une belle petite fille est celle qui a une belle robe.
Pour un Caillou, au contraire, le petit garçon enviable, n’eût-il pas de
jambes, sera celui qui a un aéroplane.
… Mais Caillou, une fois qu’il est dans la nursery, ne fait pas plus
attention à Jeanne qu’à Vivette. Il sent qu’elles n’ont pas de
mauvaises intentions à son égard, ce qui lui suffit ; il ne se soucie
pas du tout de savoir qu’elles veulent lui plaire. Il les traite donc de la
même manière. Ceci ne veut pas dire qu’il leur accorde
impartialement ses faveurs ; il reste lui-même, tout simplement. Il
s’amuse pour son compte et les deux petites filles le suivent, en
essayant de se faire remarquer. Parfois l’une met sa joue sur la joue
de Caillou, et Caillou l’embrasse. Alors l’autre fait de même, et
Caillou l’embrasse également, sans y trouver beaucoup de plaisir.
Mais il ne s’ennuie pas, il est à son aise.
Cependant on vient le chercher, pour dire bonjour à la maman de
Vivette. Il y va ingénument, sans grands regrets ni satisfaction
évidente. Je ne sais pas ce qu’on lui dit, je ne sais pas ce qu’il
répond, et ceci n’importe pas à l’histoire. Mais tout à coup on entend,
dans la nursery, des pleurs et des cris qui font retentir les murailles,
et les mères se précipitent.
Un sentiment obscur et puissant, quelque chose comme un
désespoir passionné, venait de s’emparer de Vivette et de Jeanne,
laissées à elles-mêmes. Ni l’une ni l’autre n’a réussi à vraiment
attirer l’attention de Caillou, et durant toute une heure leur amertume
en a grandi ; elles s’en rendent, sans même s’en douter,
réciproquement responsables. Voilà pourquoi, l’objet de leur rivalité
ayant disparu, la querelle a éclaté, sans qu’elles sachent pourquoi.