Ebook Numerical Solutions of Three Classes of Nonlinear Parabolic Integro Differential Equations PDF Full Chapter PDF

Numerical Solutions of Three Classes
of Nonlinear Parabolic
Integro-Differential Equations - eBook
PDF
Visit to download the full and correct content document:
https://ebooksecure.com/download/numerical-solutions-of-three-classes-of-nonlinear-
parabolic-integro-differential-equations-ebook-pdf/
Numerical Solutions of Three Classes of Nonlinear
Parabolic Integro-Differential Equations
Numerical Solutions of Three Classes
of Nonlinear Parabolic Integro-Differential
Equations
Temur Jangveladze
Ilia Vekua Institute of Applied Mathematics of
Ivane Javakhishvili Tbilisi State University
Tbilisi, Georgia
&
Georgian Technical University
Tbilisi, Georgia
Zurab Kiguradze
Ilia Vekua Institute of Applied Mathematics of
Ivane Javakhishvili Tbilisi State University
Tbilisi, Georgia
Beny Neta
Naval Postgraduate School
Department of Applied Mathematics
Monterey, CA, U.S.A.
Academic Press is an imprint of Elsevier
32 Jamestown Road, London NW1 7BY, UK
225 Wyman Street, Waltham, MA 02451, USA
525 B Street, Suite 1800, San Diego, CA 92101-4495, USA
The Boulevard, Langford Lane, Kidlington, Oxford OX5 1GB, UK
Copyright © 2016 Elsevier Inc. All rights reserved.
No part of this publication may be reproduced or transmitted in any form or by any

means, electronic or mechanical, including photocopying, recording, or any information
storage and retrieval system, without permission in writing from the publisher. Details
on how to seek permission, further information about the Publisher’s permissions
policies and our arrangements with organizations such as the Copyright Clearance
Center and the Copyright Licensing Agency, can be found at our website:
www.elsevier.com/permissions.
This book and the individual contributions contained in it are protected under copyright
by the Publisher (other than as may be noted herein).
Notices
Knowledge and best practice in this field are constantly changing. As new research and
experience broaden our understanding, changes in research methods, professional
practices, or medical treatment may become necessary.
Practitioners and researchers must always rely on their own experience and knowledge
in evaluating and using any information, methods, compounds, or experiments
described herein. In using such information or methods they should be mindful of their
own safety and the safety of others, including parties for whom they have a professional
responsibility.
To the fullest extent of the law, neither the Publisher nor the authors, contributors, or
editors, assume any liability for any injury and/or damage to persons or property as a
matter of products liability, negligence or otherwise, or from any use or operation of
any methods, products, instructions, or ideas contained in the material herein.
ISBN: 978-0-12-804628-9
British Library Cataloguing in Publication Data

A catalogue record for this book is available from the British Library
Library of Congress Cataloging-in-Publication Data

A catalog record for this book is available from the Library of Congress
For information on all Academic Press Publications

visit our website at http://store.elsevier.com/
PREFACE
This book is concerned with the numerical solutions of some

classes of nonlinear integro-di erential models. Some properties
of the solutions for investigated equations are also given. Three
types of nonlinear integro-di erential models are considered. Al-
gorithms of nding approximate solutions are constructed and
investigated. Results of numerical experiments with graphical
illustrations and their analysis are given. The book consists of
four Chapters.
In the rst Chapter three models (Model I, Model II and
Model III, that will be detailed further) to be discussed are
introduced and a brief history of integro-di erential equations
is given.
In the second Chapter, mathematical modeling of a process
of penetration of an electromagnetic eld into a substance by
integro-di erential models is described. A short description of
the integro-di erential equation that is a special model for one-
dimensional heat ow in materials with memory is also given
in the second Chapter. This model arises in the theory of one-
dimensional viscoelasticity as well. This Chapter closes with
some concluding remarks of the three investigated models.
The third Chapter is devoted to the numerical solution of
the initial-boundary value problems for models stated in the
previous Chapter. Semi-discrete schemes and nite-di erence
approximations, as well as nite elements are discussed. The
mathematical substantiation of all these questions for initial-
boundary value problems is given.
The questions of the realizations of algorithms investigated
in the third Chapter are discussed in the fourth Chapter. Re-
sults of the many numerical experiments with graphical illustra-
tions and their analysis are also given in this Chapter.
viii PREFACE
At the end of the book a list of the quoted literature and

indexes are given. The list of references is not intended to be
an exhaustive bibliography on the subject, but it is nevertheless
detailed enough to enable further independent work.
Each Chapter is concluded with a detailed section, entitled
"Comments and bibliographical notes," containing references to
the principal results treated, as well as information on important
topics related to, but sometimes not included in the body of the
text.
The authors believe that the book will be useful to scientists
working in the eld of nonlinear integro-di erential models. In
the opinion of the authors, the material presented in the book
is helpful for a wide range of readers engaged in mathematical
physics, in problems of applied and numerical mathematics, and
also MS and PhD students of the appropriate specializations.
Temur Jangveladze,
Zurab Kiguradze,
Beny Neta
ACKNOWLEDGMENTS
The rst author thanks Fulbright Visiting Scholar Program for

giving him the opportunity to visit U.S.A. and the Naval Post-
graduate School in Monterey, CA, U.S.A. for hosting him during
the nine months of his tenure in 2012-2013. The second author
thanks Shota Rustaveli National Scienti c Foundation of Re-
public of Georgia for giving him opportunity to visit U.S.A.
and the Naval Postgraduate School in Monterey, CA, U.S.A. for
hosting him during the four months of his tenure in 2013.
Abstract
This book is concerned with the numerical solutions of some

classes of nonlinear integro-di erential models. Some proper-
ties of the solutions of the corresponding initial-boundary value
problems studied in the monograph equations are given. Three
types of nonlinear integro-di erential models are considered. Al-
gorithms of nding approximate solutions are constructed and
investigated. Results of numerical experiments with tables and
graphical illustrations and their analysis are given. The book
consists of four chapters. At the end of the book a list of the
quoted literature and indexes are given. Each chapter is con-
cluded with a detailed section, entitled "Comments and biblio-
graphical notes," containing references to the principal results
treated, as well as information on important topics related to,
but sometimes not included in the body of the text.
Numerical Solutions of Three Classes of Nonlinear Parabolic Integro-Differential Equation.

http://dx.doi.org/10.1016/B978-0-12-804628-9.50008-9
© 2016 Elsevier Ltd. All rights reserved. 1
2 ABSTRACT
Key words: Electromagnetic eld penetration, Maxwell’s

equations, integro-di erential models, existence and uniqueness,
asymptotic behavior, semi-discrete and nite di erence schemes,
Galerkin’s method, nite element approximation, error estimate,
stability and convergence.
Chapter 1
Introduction
Abstract
The description of various kinds of integro-di erential equa-
tions and a brief history of their origin and applications are
given. The importance of investigations of integro-di erential
models is pointed out as well. Classi cation of integro-di eren-
tial equation is given. The main attention is paid on parabolic
type integro-di erential models. In particular, three types of
integro-di erential equations are considered. Two of them are
based on Maxwell’s equations describing electromagnetic eld
penetration into a substance. The third one is obtained by
simulation of heat ow. At the end of the chapter, as at the
end of each chapter, the comments and bibliographical notes is
given, which consists of description of references concerning to
the topic considered.
Key words: Electromagnetic eld penetration, Maxwell’s

system, heat ow equation, integro-di erential models.

http://dx.doi.org/10.1016/B978-0-12-804628-9.50001-6
© 2016 Elsevier Inc. All rights reserved. 3
4 CHAPTER 1. INTRODUCTION
In mathematical modeling of applied tasks di erential, inte-

gral, and integro-di erential (I-D, for short) equations appear
very often. There are numerous scienti c works devoted to the
investigation of di erential equations. There is a vast literature
in the eld of integral and integro-di erential models as well.
The di erential equations are connecting unknown functions,
their derivatives, and independent variables. On the other hand,
integral equations contain the unknown functions under an in-
tegral as well.
The term integro-di erential equation in the literature is
used in the case when the equation contains unknown function
together with its derivatives and when either unknown function,
or its derivatives, or both appear under an integral.
Let us recall the general classi cation of integro-di erential
equations. If the equation contains derivatives of unknown func-
tion of one variable then the integro-di erential equation is called
ordinary integro-di erential equation. The order of an equation
is the same as the highest-order derivative of the unknown func-
tion in the equation.
The integro-di erential equations often encountered in math-
ematics and physics contain derivatives of various variables;
therefore, these equations are called integro-di erential equa-
tions with partial derivatives or partial integro-di erential equa-
tions.
In the applications very often there are integro-di erential
equations with partial derivatives and multiple integrals as well,
for example, Boltzmann equation [66] and Kolmogorov-Feller
equation [288].
Volterra is one of the founders of the theory of integral and
integro-di erential equations. His works, especially in the inte-
gral and integro-di erential equations, are often cited till today.
The classical book by Volterra [469] is widely quoted in the liter-
ature. In 1884 Volterra [465] began his research in the theory of
CHAPTER 1. INTRODUCTION 5
integral equations devoted to distribution of an electrical charge

on a spherical patch. This work led to the equation, which in
the modern literature is called the integral equation of the rst
kind with symmetric kernel.
The work on the theory of elasticity became the beginning
research of Volterra leading to the theory of partial integro-
di erential equations. In 1909 Volterra [466] has studied a par-
ticular type of such equations and has shown that this integro-
di erential equation is equivalent to a system consisting of three
linear integral equations and a second order partial di erential
equations.
The rst examples of integro-di erential equations with par-
tial derivatives investigated in the beginning of the twentieth
century were in Schlesinger’s works [417], [418], where the fol-
lowing equation is investigated:
Z b
@U (x; y)
= f (x; y; s)U (x; s)ds:
@x a
Numerous works in the beginning of the twentieth century
were devoted to research of integro-di erential equations of vari-
ous kinds. The excellent bibliography in this case is given in the
classical book by Volterra [469]. In addition, Kerimov [271], the
editor of the Russian translation of this book, has updated (up
to 1970s) the list of references on integral and integro-di erential
equations.
Let us describe some classes of mathematical models of sec-
ond order promoting intensive research on partial integro-di er-
ential equations.
When we take into account hereditary phenomena, the ques-
tions of physics and mechanics lead to integro-di erential equa-
tions. A hereditary phenomenon occurs in a system when the
phenomenon does not depend only on the actual state of the
system but on all the preceding states through which the sys-
tem has passed; that is to say, it depends on the history of the

system and may therefore be called hereditary.
One of the important representatives of an integro-di er-
ential equations of elliptic type is the following equation con-
nected with the hereditary phenomenon [467]
Z tX3
@ 2 U (x; )
U (x; t) + fi (t; )d = 0;
0 i=1 @x2i
where x = (x1 ; x2 ; x3 ); is a classical three-dimensional Laplace

operator
@ 2 U (x; t) @ 2 U (x; t) @ 2 U (x; t)

U (x; t) = + + ;
@x21 @x22 @x23
and fi are known functions of their arguments. Let us note that
here and below everywhere instead of x; y; z we use x1 ; x2 ; x3 for
the designation of Cartesian coordinates.
The mathematical modeling of the vibrations of an elastic
chord in the case of linear hereditary process gives a hyperbolic
type integro-di erential equation [467]
t
@ 2 U (x; t) @ 2 U (x; t) @ 2 U (x; )
Z
= + (t; )d ; (1.1)
@t2 @x2 0 @x2
where is a known function of its arguments.

One of the most important representatives of integro-di er-
ential models is the following nonlinear equation describing string
vibration obtained by Kirchho [286] in 1876
" #
2
@ 2 U (x; t) @ 2 U (x; t)
Z
2 @U (x; t)
+ dx = 0; (1.2)
@t2 0 @x @x2
where is a known constant. Many authors investigated equa-

tion (1.2) and its natural generalizations:
"Z #
2
@ 2 U (x; t) @U (x; t) @ 2 U (x; t)
a dx =0 (1.3)
@t2 0 @x @x2
and
@ 2 U (x; t) 1
a A 2 U (t) AU (t) = 0; (1.4)
@t2
where a(S) a0 = const > 0 is a known function of its argu-
ment and A is a self-adjoint positive operator, i.e., A = A > 0.
The norm used in (1.4) is the one de ned on the range of the
operator A.
In investigating (1.3) and (1.4) type models it is su cient
to mention the following publications: [9], [24], [36], [37], [55],
[61], [65], [130], [309], [327], [328], [349], [355], [356], [357], [373],
[380], [383], [390], [394], [395] though this list is not complete.
Let us note that equations (1.3) and (1.4) are also called Kirch-
ho equations. They, along with some similar equations, de-
scribe important physical processes, among which are linear and
nonlinear dynamics of di erent dimensional bodies (see, for ex-
ample, [36], [355], [373] and [471]).
In other questions connected with hereditary phenomena,
one nds the integro-di erential equations of a parabolic type,
which were investigated by Evans [160]. These equations look
like
Z t 2
@U (x; t) @ 2 U (x; t) @ U (x; )
2
+ A(t; )d = 0; (1.5)
@t @x t0 @x2
where A is a known function of its arguments.

Integro-di erential equations of parabolic type arise in the

study of various problems in physics, chemistry, technology,
economics, etc. One very important problem of applied type
is generated by mathematical modeling of processes of electro-
magnetic eld penetration into a substance and is described by
the well-known Maxwell’s equations [300]. In the works [187]
and [188], complex system corresponding to nonlinear partial-
di erential equations was reduced to integro-di erential form. If
the coe cient of thermal heat capacity and electroconductivity
of the substance are highly dependent on temperature, then the
Maxwell’s system can be rewritten in the following form (see
[187] and [188]):
@W (x; t)
+r [a(S(x; t))r W (x; t)] = 0;
@t (1.6)
r W (x; t) = 0;
where Z t
S(x; t) = jr W (x; )j2 d : (1.7)
0
In system (1.6), r W and r W are the usual vector oper-
ators with respect to the variables x = (x1 ; x2 ; x3 ): Even one-
dimensional scalar version of this model is very complicated and
its investigation has been possible yet only for special cases. The
one-dimensional scalar case of the model (1.6), (1.7) has the fol-
lowing form
" Z ! #
t 2
@U (x; t) @ @U (x; ) @U (x; t)
= a d ; (1.8)
@t @x 0 @x @x
where a(S) a0 = const > 0 is again a known function of its

argument. Investigation of (1.6), (1.7), and (1.8) type models
began in the works [138], [187], and [188]. Since then many
scienti c publications were devoted to the investigation of exis-
tence and uniqueness of their solutions under various types of
initial and boundary conditions. In this respect, especially sig-
ni cant are the works [49], [50], [141], [146], [147], [219], [220],
[238], [254], [256], [302], [303], [304], [305], [306], [307], [322],
[333], [334], and reference therein. Authors of this book have
also made contribution in this direction, for example, see [137],
[138], [139], [140], [141], [145], [146], [147], [186], [187], [188],
[219], [220], [223], [238], [247], [248], [253], [254], [256], [261],
and reference therein.
Making certain physical assumptions in mathematical de-
scription of the above-mentioned process of penetration of elec-
tromagnetic eld into a substance, Laptev [306] has constructed
a new integro-di erential model, which represents a general-
ization of the system introduced in [187] and [188]. Founded
on Maxwell’s system the following parabolic integro-di erential
model is obtained
Z tZ
@W (x; t)
=a jr W (x; )j2 dxd W (x; t): (1.9)
@t 0
In the works [303], [305], and [306] for conditionally closed

operators, an operator scheme is constructed. This scheme is
used for (1.6), (1.7) type models to prove existence and unique-
ness of solution of initial-boundary value problems. In the above-
mentioned work [306] Laptev points out that for these so-called
averaged integro-di erential models (1.9), it is necessary to de-
velop a di erent, special approach.
In systems (1.6), (1.7), and (1.9), W = (W1 ; W2 ; W3 ) denotes
a vector, which is connected by a vector of a magnetic eld H =
(H1 ; H2 ; H3 ) and is the function of the variables (x1 ; x2 ; x3 ; t);
which we will shorten to (x; t).
The multidimensional scalar analogues of systems (1.6), (1.7),

and (1.9) have the following forms:
Z t
@U (x; t)
=r a jrU (x; )j2 d rU (x; t) (1.10)
@t 0
and
Z tZ
@U (x; t)
=a jrU (x; )j2 dxd U (x; t); (1.11)
@t 0
respectively.
In equations (1.10) and (1.11) we have x = (x1 ; : : : ; xn ) and
the vector operator rU is given by
@U (x; t) @U (x; t)
rU = gradU = ;:::; = (D1 U; : : : ; Dn U ) :
@x1 @xn
Some generalizations of the models (1.6)-(1.11) are given in
the works [137], [139], [145], [219], [305], and [306]. One kind of
these models has the forms:
n
@U (x; t) X
Di a(S(x; t)) jrU (x; t)jq 2
Di U (x; t)
@t i=1 (1.12)
= f (x; t)
and
n
@U (x; t) X
a(S(t)) Di jrU (x; t)jq 2
Di U (x; t)
@t i=1 (1.13)
= f (x; t);
where Z t
S(x; t) = jrU (x; )jq d
0
and Z tZ
S(t) = jrU (x; )jq d :
0
The models of type (1.6)-(1.13) are complex and have been
intensively studied by many authors. The existence and unique-
ness of global solutions of initial-boundary value problems for
equations and systems of type (1.6) - (1.13) were studied in
[49], [50], [137], [138], [139], [140], [141], [145], [186], [187],[188],
[219], [223], [238], [247], [248], [253], [261], [302], [303], [304],
[305], [306], [307], [322], [333], [334], and in a number of other
works as well.
The existence theorems that are proved in [137], [138], [139],
[140], [141], [145], [187], [188], and [219] are based on a-priori
estimates, modi ed Galerkin’s method and compactness argu-
ments as in [327], [328], [461], and [462] for nonlinear elliptic
and parabolic equations.
For equation (1.8) with nonhomogeneous right-hand side and
a(S) = 1 + S, or for equation (1.13) in the one-dimensional
case with q = 2, such theorem for rst initial-boundary value
problem is proved in section 3.6.
The asymptotic behavior as t ! 1 of the solutions of such
models has been the object of intensive research in recent years,
see [28], [29], [31], [32], [33], [34], [35], [139], [141], [145], [146],
[147], [217], [218], [219], [220], [223], [227], [228], [229], [231],
[232], [233], [234], [235], [236], [237], [238], [239], [240], [241],
[242], [243], [245], [246], [247], [248], [251], [252], [253], [254],
[256], [257], [261], [276], [277], [278], [280], [282], [283], and ref-
erence therein.
Another model considered in this book and studied by one
of the authors of this monograph [376] is
Z t
@U (x; t) @ @U (x; )
= a(t ) d + f (x; t): (1.14)
@t 0 @x @x
This equation serves as a very special model for one-dimens-

ional heat ow in material with memory [196], [345]. It also
arises in the theory of viscoelasticity [109], [344], [346]. It is an
example in the general theory of equations of the form
Z t
U_ (t) + a(t )g(U ( ))d = f (t) (1.15)
0
on a Hilbert space H with g a nonlinear bounded operator [51].

The dot above U means time derivative.
MacCamy [345] and later Sta ans [438] have discussed the
existence, uniqueness, boundedness and asymptotic behavior of
solutions of initial-boundary value problems for (1.14).
Many scienti c publications are dedicated to the investiga-
tion of characteristics of (1.14) and (1.15) type equations (see,
for example, [67], [70], [80], [110], [111], [162], [206], [433], and
reference therein).
Numerous scienti c works are devoted to construction of al-
gorithms for the numerical solution of initial-boundary value
problems for the above-stated models (see, for example, [20],
[30], [32], [35], [80], [96], [125], [139], [145], [148], [220], [223],
[227], [239], [241], [243], [244], [247], [248], [249], [250], [251],
[252], [253], [255], [256], [257], [258], [261], [274], [276], [277],
[279], [280], [281], [282], [283], [284], [318], [333], [376], [378],
[423], [424], and [501]).
The purpose of the present monograph is the continuation
of study and uniform description of results developed for the
numerical solution of integro-di erential models (1.6)-(1.14) and
their generalizations. It should be noted that main attention is
paid on theorems with proofs for corresponding one-dimensional

equations.
The book consists of four chapters. In the next chapter
mathematical modeling of a process of penetration of an elec-
tromagnetic eld into a substance by integro-di erential models
and the conclusion of investigated models such as (1.6)-(1.13)
is brie y given. We also give the mathematical model for heat
conduction in materials with memory (1.14). Some features of
this model are also given.
The third chapter is devoted to the numerical solution of
the initial-boundary value problems for the models stated in
the previous chapter. The semi-discrete schemes, nite di er-
ence approximations, Galerkin’s method schemes, and nite el-
ements are discussed. The mathematical substantiation of all
these questions for various types of initial-boundary value prob-
lems is given.
Results of the various numerical experiments with graphical
illustrations and their analyses are given in the fourth chapter.
At the end of the book a list of the quoted literature and
indexes are given.
The list of references is not intended to be an exhaustive bib-
liography on the subject, but it is nevertheless detailed enough
to enable further independent work.
Each chapter is concluded with a detailed section, entitled
"Comments and bibliographical notes," containing references to
the principal results treated, as well as information on important
topics related to, but sometimes not included in the body of the
text.
The authors believe that the book will be useful to the ex-
perts working in the eld of nonlinear integro-di erential mod-
els. In the opinion of the authors, the material presented in the
book is helpful for a wide range of readers engaged in mathemat-
ical physics, in problems of applied and numerical mathematics,
and also MS and PhD students of the appropriate specializa-

tions.
1.1 Comments and bibliographical

notes
Mathematical models of many natural phenomena and processes
can be described by the initial-boundary value problems posed
for nonstationary partial di erential and integro-di erential equa-
tions and systems of such equations. Investigation and numer-
ical solution of these problems are the actual sphere of mathe-
matical physics and numerical analysis. One such partial integro-
di erential model describes the process of electromagnetic eld
penetration into a substance. In the quasi-stationary approxi-
mation, the corresponding system of Maxwell’s partial di eren-
tial equations can be rewritten in integro-di erential form (1.6),
(1.7) (see [187] and [188]).
Mathematical models describing electromagnetic processes
and many relative phenomena are given in many scienti c pa-
pers and books, see, for example, [92], [136], [300], [359], [360],
[399], [419], [473], and references therein.
It is well known that electromagnetic eld di usion pro-
cesses and many other important practical processes are simu-
lated by Maxwell’s systems of partial di erential equations and
Maxwell’s-type systems as well (see, for example, [195], [197],
[290], [297], [298], [300], [360], [426], [485], [486], [487], [489],
[491], and references therein).
Many scienti c works are devoted to investigation of various
problems for Maxwell’s and Maxwell’s-type systems, see [4], [5],
[56], [60], [101], [106], [120], [121], [124], [131], [139], [144], [145],
[152], [158], [194], [195], [197], [208], [224], [225], [226], [247],
[253], [275], [276], [290], [362], [365], [367], [397], [409], [427],
1.1. COMMNETS AND BIBLIOGRAPHICAL NOTES 15
[452], [453], [454], [470], [475], [476], [477], [485], [486], [487],
[489], [491], and references therein.
As we have already pointed out, by using Maxwell’s system
[300] for mathematical simulation of the process of electromag-
netic eld penetrating into a substance, new class of integro-
di erential models (1.6), (1.7), (1.8), and (1.10) arises (see [187]
and [188]).
Integro-di erential equations arise in many other practical
processes as well, see, for example, [11], [21], [22], [36], [37],
[55], [59], [74], [84], [119], [127], [128], [164], [185], [187], [188],
[295], [298], [306], [310], [311], [312], [355], [356], [357], [363],
[373], [380], [384], [405], [425], [459], [463], [465], [466], [467],
[468], [469], [471], [484], and in a number of other works as well.
The motivation for studying integro-di erential problems
comes from the many physical models in such elds as elec-
tromagnetic wave propagation, heat transfer, nuclear reactor
dynamics, and thermoelasticity. Besides the integro-di erential
equations arise in many spheres of human activity as well. For
example, the second order fully nonlinear integro-di erential
equations are derived from the pricing problem of nancial deriva-
tives and optimal portfolio selection problem in a market [59].
In [84] nonlinear integro-di erential equations that arise from
stochastic control problems with purely jump Levy processes
are considered.
Many problems of modern science and engineering can be
described by partial integro-di erential equations. Since quite
a lot of these problems are time-dependent, most of them are
evolution equations and especially nonlinear evolution parabolic
equations, see [93], [177], [179], [180], [191], [216], [310], [311],
[312], [322], [332], [345], [346], [358], [451], and references therein).
Many scienti c works are devoted to investigation and nu-
merical solution of parabolic integro-di erential models, see [6],
[18], [45], [53], [62], [63], [69], [74], [83], [86], [87], [93], [98], [103],
[108], [115], [118], [128], [137], [139], [145], [155], [156], [163],
[175], [177], [181], [182], [191], [198], [247], [253], [263], [265],
[266], [267], [276], [279], [281], [306], [308], [310], [311], [312],
[313], [319], [320], [322], [323], [327], [353], [358], [370], [376],
[378], [389], [393], [427], [430], [432], [433], [447], [448], [449],
[450], [451], [456], [479], [483], [488], [492], [497], and references
therein.
Study of the models of type (1.6), (1.7), and (1.8) has begun
in the works [138], [187] and [188]. In these works, in particu-
lar, the theorems of existence of solution of the initial-boundary
value problem (with rst (Dirichlet) boundary conditions) for
scalar equation with one-dimensional space variable are proved.
Investigations of higher space dimensions for model (1.10) car-
ried out initially in [137] and [140].
In [306] some generalization of the system of type (1.6), (1.7)
was proposed. In particular, assuming that the temperature
of the considered body depends on the time variable, but in-
dependent of the space coordinates, then the same process of
penetration of a magnetic eld into the material is simulated
by the averaged (as the author of [306] has named it) integro-
di erential models (1.9) and (1.11).
Study of the models of type (1.9) and (1.11) has started in
the works [217] and [219].
One must note that some works were devoted to the study
of modeling of physical process of electromagnetic eld pene-
tration in the case of cylindrical conductors. In this case, the
above-mentioned integro-di erential model (1.6), (1.7), written
in cylindrical coordinates, was given in [148]. The work [333]
is devoted to the investigation of periodic problem for one-
dimensional (1.8) type model in cylindrical coordinates.
Interest in the above-mentioned integro-di erential (1.6) -
(1.11) models is increasing. Some generalizations of (1.10) and
(1.11) models, which have the forms (1.12) and (1.13) corre-
1.1. COMMNETS AND BIBLIOGRAPHICAL NOTES 17
spondingly, are given and studied in [219], [305], and in a num-

ber of other works as well. Equation (1.12) is investigated, for
example, in [145], [219], [305], and [306]. Equation (1.13) was
investigated in [145] and [219]. In the scienti c literature some
more general models have also appeared, see [137], [139], [145],
[219], [305], [306], and [322].
Many di erent kinds of initial-boundary value problems with
a variety of boundary and initial conditions are considered for
the above-mentioned integro-di erential equations. In the works
[246], [280], [281] investigation and numerical approximation of
problems with mixed boundary conditions, for (1.10) and (1.11)
type one-dimensional scalar models, are studied.
Let us also note that rst kind initial-boundary value prob-
lems with nonhomogeneous boundary condition on one side of
lateral boundary are also considered and studied in many works.
This type of the problem statement is dictated by mathematical
simulation of the physical processes, see, for example, [148] and
problem (2.64), (2.65) given in "Comments and bibliographical
notes" section in Chapter 2.
The theorems proved for investigating the asymptotic behav-
ior as time tends to in nity in some cases show the di erence be-
tween stabilization character of solutions with homogeneous and
nonhomogeneous boundary conditions of the rst kind initial-
boundary value problems. More precisely, in homogeneous case
stabilization has an exponential character, whereas in nonho-
mogeneous case it has power-like form.
The works [49], [50] are also worth mentioning, where in-
vestigation of inverse problems for multidimensional models of
(1.10) type is carried out.
Another integro-di erential model studied in this monograph
is (1.14). This model describes heat ow in material with mem-
ory [196], [345]. It also arises in the theory of viscoelasticity
[109], [344], and [346].
As a rule we cannot nd exact solutions of the considered

nonlinear di erential and integro-di erential models. Therefore,
particular attention should be paid to construction of numerical
solutions and to their importance for integro-di erential models.
The rst steps in this direction, for the models studied in this
monograph, are made in the works [139], [220], [274], [376], and
[378]. Now the research in this direction has intensi ed.
Let us note that the models being considered in this mono-
graph have arisen from practical tasks. But they can be con-
sidered as models, generalizing known nonlinear parabolic equa-
tions, which are studied in many known scienti c papers, books,
and monographs, see [76], [133], [134], [169], [199], [298], [327],
[332], [381], [385], [461], and references therein.
Many scienti c researches are devoted to numerical solution
of partial di erential and nonlinear parabolic equations as well,
see [23], [43], [104], [167], [171], [183], [193], [210], [213], [225],
[226], [301], [327], [400], [404], [407], [411], [415], [435], [439],
[440], [447], and the references listed in these papers and books.
As we have already mentioned the main part of integro-
di erential structures considered here has arisen from Maxwell’s
systems of partial di erential equations. There are many scien-
ti c papers and books on the numerical solution of Maxwell’s
systems and Maxwell’s-type systems as well, see [4], [5], [15],
[26], [27], [54], [68], [139], [142], [145], [153], [212], [224], [247],
[253], [260], [276], [314], [315], [316], [317], [336], [341], [369],
[434], [442], [472], [480], [493], [498], and references therein.
The results of these researches very often are used in construc-
tion and investigation of the numerical algorithms for the cor-
responding integro-di erential models.
The detailed description of investigation and numerical so-
lution of the above-mentioned models is given in Chapters 2-4.
More complete references and comments are given in a section
entitled "Comments and bibliographical notes" in each chapter.
Chapter 2
Mathematical Modeling
Abstract
The chapter consists of six sections. The chapter concerns
mathematical modeling of the investigated in the monograph
equations. Some mathematical features of these models are
studies as well. In the rst section general statement of di usion
process is given. Sections two and three are dedicated to a re-
duction of Maxwell’s equations to the integro-di erential mod-
els. Consequently, two types of integro-di erential equations
are obtained which are called Model I and Model II accordingly.
Both models in di erent physical assumptions describe process
of penetration of an electromagnetic eld into a substance. In
the third section Model III is considered, which represents a
special model for one-dimensional heat ow in materials with
memory. This model arises in the theory of one-dimensional
viscoelasticity as well. Next two sections are devoted to some
mathematical features of all three models above. The existence
and uniqueness properties of the solutions as well as asymp-
totic behavior of solutions of the appropriate initial-boundary
value problems are presented. This chapter closes with some
concluding remarks and bibliographical overview of the three

http://dx.doi.org/10.1016/B978-0-12-804628-9.50002-8
© 2016 Elsevier Inc. All rights reserved. 19
20 CHAPTER 2. MATHEMATICAL MODELING
investigated models.
Key words: Mathematical modeling, electromagnetic eld

penetration, Maxwell’s equations, heat ow in materials with
memory, one-dimensional viscoelasticity, integro-di erential mod-
els, existence and uniqueness, asymptotic behavior.
2.1 Electromagnetic di usion process

2.1.1 General statement of di usion process
Let us consider the phenomena which occur in a substance
placed in an external variable electromagnetic eld.
Suppose the electromagnetic eld and the currents satisfy
quasi-stationary conditions, see, e.g., [300] and [419]. Maxwell’s
system of equations is given by:
1 @( H)
= r E; (2.1)
c @t
r ( H) = 0; (2.2)
4
E = r H; (2.3)
c
where E = (E1 ; E2 ; E3 ) and H = (H1 ; H2 ; H3 ) are vectors of
electrical and magnetic elds, respectively, is the magnetic
permeability, is the conductivity of substance, and c is the
speed of light in vacuum.
In (2.3), following an assumption of quasi-stationarity, the
term proportional to @E=@t is omitted and Ohm’s law is used,
relating the vector E to a vector of density of a current J by
the following relation
J = E: (2.4)
2.1. ELECTROMAGNETIC DIFFUSION PROCESS 21
If at some moment of time the magnetic eld is solenoidal,

i.e., r ( H) = 0, then (2.2) is always satis ed. From (2.1) we
have
@
r ( H) = 0:
@t
We shall assume that the substance is isotropic with = 1
and the coe cient of electroconductivity depends on tempera-
ture, i.e., = ( ). In applications, the form of is power-
1
like, for example, for metals ; for homogeneous plasma
3=2
, and so on.
For de nition of temperature it is necessary to use the equa-
tion of balance of heat. Let us make the assumption that the
characteristic time of resistive di usion is much less than that of
heat transfer. Then, neglecting the e ect of heat conductivity,
the change in temperature is de ned only by Joule’s heating and
taking into account (2.4), we have
@
cv = EJ = E 2 ; (2.5)
@t
where cv is the speci c heat capacity of the environment. Thus
the factor of heat capacity can also depend on temperature (as
a rule, in a power-like form).
Equations (2.1), (2.3), (2.5) form the closed system for def-
inition of an electromagnetic eld and temperature under an
appropriate initial and boundary conditions.
The questions of existence and uniqueness of the solutions of
linear di erential problems ( = (x)), in general enough state-
ment, are considered in [136], [297], and in a number of other
works. Thus, in [297] the movement of environment is supposed
within the framework of magnetic-hydrodynamical approach. In
these works the transition from classical statement to the gener-
alized one is made. The requirement that the functions satisfy
equation (2.2) and boundary conditions is replaced with the re-
quirement of their belonging to special functional spaces. The

problem for equations (2.1), (2.3) is formulated in the variational
form, in particular, in terms of variational inequalities.
The study of system (2.1)-(2.3) can be made also on the
basis of the equations which have been written down only for
vectors E or H. As an example of similar research, see [504], in
which the nonlinearity caused by dependence of the magnetic
permeability on H is entered.
2.1.2 A reduction of system of nonlinear dif-

ferential equations to the integro-di er-
ential model (Model I)
We assume that a massive body is placed in a variable magnetic
eld. It is necessary to describe the distribution of the eld
inside the body.
Inserting the resistance = 1= in (2.4) we shall express
E = J and substitute in the penultimate equation (2.3), we
have
c
E= r H:
4
Substituting this expression in (2.1), we receive
@H c2
+ r ( r H) = 0: (2.6)
@t 4
A variable magnetic eld penetrating into a substance in-
duces a variable electrical eld, which causes occurrence of cur-
rents. The currents result in heating of the substance and in-
crease of its temperature, which in uences its resistance. From
reasonings given in [399], it follows that one order change of
temperature results in change of resistance on several orders,
so at large uctuations of temperature the dependence = ( )
must be taken into account. Last essential restriction, which
will be made, is connected to the assumption that the change in

body temperature under action of a current J submits to Joule’s
law
@
0 cv = J 2: (2.7)
@t
Here is density of the substance, 0 is the density at time t = 0,
and cv (as was already mentioned in 2.1.1) is its speci c heat
capacity. Generally they also depend on temperature. Equation
(2.7) does not take into account transfer of heat due to heat
conductivity and radiation. Number of other physical e ects
are not considered. However, from the mathematical point of
view system (2.6), (2.7) is complex enough.
Now we shall begin the reduction of system (2.6), (2.7) to a
system of nonlinear integro-di erential equations.
Let us rewrite equation (2.7) in the following form
cv @
0( ) = J 2:
( ) @t
Introducing the function

Z
cv
S( ) = 0( ) d ;
0
( )
we have
@S
= J 2:
@t
Let us assume that the process begins at the moment of time
t = 0 in which the temperature 0 of the substance is constant.
Integrating this equation on a segment [0; t]; we shall nd
Z t
S( (x; t)) S( 0 ) = J 2d :
0
The functions 0 ; , and cv are positive in physical sense;

therefore, the function S( ) is monotonically increasing. From
here follows that the inverse function = (S) is uniquely de-
termined, connected with S( ) by parity (S( )) = : Thus, it
is possible to write down
Z t
(x; t) = J 2d :
0
From (2.3), (2.4) we have
c
J= r H;
4
so, Z t
c 2
(x; t) = r H d :
0 4
Substituting this expression in (2.6) as argument of function
= ( ), we receive
t
c2
Z
@H c 2
+ r r H d r H = 0;
@t 4 0 4
r H = 0:
Let us introduce the notations
c2 c
a(S) = ( (S)); W = H
4 4
and rewrite the system in the following form:
Z t
@W
+r a jr W j2 d r W = 0;
@t 0 (2.8)
r W = 0:
If we consider a at eld of the form W = (0; 0; U ), where

the function U = U (x1 ; x2 ; t) depends on two spatial variables,
then
@U @U
r W = ; ;0
@x2 @x1
and (2.8) takes the form
Z t
@U
=r a jrU j2 d rU : (2.9)
@t 0
Let us note again that equations such as (2.8), (2.9) for the
rst time have arisen in [187] and [188]. In these works, as well
as in [137], [138], [139], [140], [141], [145], [186], [219], [223],
[238], [247], [253], and [261] together with other questions, the
uniqueness of the solutions of the initial-boundary value prob-
lems for equations (2.8) and (2.9) is given, under a rather general
assumptions on the function a = a(S).
Assume that 0 and cv are constants. Let us give examples
of function ( ), inducing a(S), see, e.g., [145], [304], and [306].
If ( ) = , > 1, then
a(S) = C1 (C0 S) 1;
where C0 and C1 are some positive constants. Thus, the power
growth of the resistance ( ) gives a coe cient determined only
on a nite interval. Let us notice that the physical substances
do not have such property.
If ( ) = , then
a(S) = C1 eC0 S ;
with positive constants C0 and C1 , i.e., the linear growth of the
function ( ) gives exponential function a(S). The linear growth
of resistance with temperature is characteristic of metals.

If ( ) = , 0 < < 1, then
a(S) = C1 (C0 + S) 1 ;
with positive constants C0 and C1 . Thus, under-linear growth
of the function ( ) gives sedate growth of the coe cient a(S).
If ( ) = , < 0, then
a(S) = C1 (C0 + S) 1 + > 0;

with positive constants C0 and C1 , i.e., a decreasing function
( ) gives decreasing functions a(S). This is a general rule in
case 0 and cv are constants. Suppose that function ( ) is
di erentiable for 0 . Then according to the de nition we
have
da c2 d d c2 d =d c2
= = = ( ) 0 ( ):
dS 4 d dS 4 dS=d 4 cv
From here it is obvious that the function a(S) decreases or

grows simultaneously with the function ( ). Let us note that
the decrease of resistance ( ) with growth of temperature is
characteristic of semiconductors in solid, gaseous, and plasma
phases, for which, as it was already remarked in 2.1.1, the the-
oretical formula ( ) = K 3=2 is valid.
2.2 On the averaged Model II

We shall begin now a derivation of the averaged equation [306]
describing again the process of penetration of an electromagnetic
eld into a substance.
2.2. ON THE AVERAGED MODEL II 27
Following [419] and as in section 2.1, we shall consider the

system of the Maxwell’s equations, describing the interaction of
an electromagnetic eld into a substance:
1 @H
= r E; (2.10)
c @t
r H = 0; (2.11)
4
J = r H; (2.12)
c
J = E: (2.13)
Joule’s law (2.7) is a localization of the law of allocation of
heat [419]
Z
dQ = JEdxdt: (2.14)
Here dQ is in ow of a thermal energy absorbed by a body in an

electromagnetic eld in time dt on a substance, occupying area
R3 . Joule’s law does not take into account the process of
transfer of heat inside a body, which is fair if one to consider
the temperature time-dependent but independent of the spatial
coordinates. Thus, in this case it is possible to write that =
(t) and dQ = mcv d , where m is the weight of the substance.
Therefore, (2.14) becomes
Z
d
mcv ( ) = JEdx:
dt
Using Ohm’s law E = ( )J and repeating the process of
exception, we shall receive the following analogue of system (2.8)
in the same designations:
Z tZ
@W 1
+r a jr W j2 dxd r W = 0;
@t j j 0
r W = 0:
Here j j is the volume of the substance. Due to averaging the
coe cient of this equation depends only on the variable t, and
consequently the equation can be rewritten in the form
Z tZ
@W
=a jr W j2 dxd W: (2.15)
@t 0
Here we used the known vector identity
r r W = W + r(r W ):
For a at eld W = (0; 0; U ), where U = U (x1 ; x2 ; t) is a
function of two spatial variables, system (2.15) becomes
Z tZ
@U
=a jrU j2 dxd U: (2.16)
@t 0
Laptev [306] remarked that the research of the so-called aver-

aged integro-di erential models (2.15) and (2.16) requires inde-
pendent theory than he applied for investigation of (2.8), (2.9),
and such type equations.
2.3 Mathematical Model III

The classical linear theory of heat conduction for homogeneous
and isotropic media is given by
@U @ 2U
=k 2;
@t @x
where U (x; t) is the absolute temperature and k > 0 is a con-
stant called thermal di usivity. This parabolic equation allows
a thermal disturbance at any point in the medium to be in-
stantly felt at other points, see [196]. Gurtin and Pipkin [196]
2.3. MATHEMATICAL MODEL III 29
have suggested a heat ow model based on a memory e ect in

the medium. The linear one-dimensional version of their the-
ory assumes that the internal energy, (x; t); and the heat ux,
q(x; t), satisfy the following relations:
Z 1
(x; t) = bU (x; t) + B( )U (x; t )d ; (2.17)
0
Z 1
@U (x; t )
q(x; t) = c( ) d ; (2.18)
0 @x
where b is the instantaneous speci c heat, B is an energy re-
laxation function, and c is the thermal conductivity (see [345]).
Thus is a functional of the history of the temperature and q
is a functional of the history of the temperature gradient. Mac-
Camy [345] treated a partially nonlinear version of the above
model by replacing (2.18) by
Z 1
@U (x; t )
q(x; t) = K( ) d : (2.19)
0 @x
Now assume that the material is at zero temperature and in-
ternal energy up to time t = 0. This is not restrictive since
we can incorporate nonzero initial values with the forcing term.
Incorporating (2.17), (2.19) with the balance of heat equation,
we have
Z t
@U (x; t) @U (x; )
b + B(t ) d =
@t 0 @
Z t (2.20)
@ @U (x; )
K(t ) d + r(x; t);
0 @x @x
where r(x; t) is the heat source. Let R(x; t) be the resolvant for
B, i.e., the function U given by
Z t
1
U (x; t) = (x; t) + R(t ) (x; )d
b 0
solves the equation

Z t
(x; t) = bU (x; t) + B(t )U (x; )d :
0
Now de ne a(t) and f (x; t) by

Z t
1
a(t) = K(t) + R(t )K( )d ;
b 0
Z t
1
f (x; t) = r(x; t) + R(t )r(x; )d ;
b 0
then (2.20) becomes

Z t
@U (x; t) @ @U (x; )
= a(t ) d + f (x; t); (2.21)
@t 0 @x @x
which is (1.14).
2.4 Some features of Models I and II

2.4.1 Existence and uniqueness of the
solutions
In the present section we give some features of those types of
equations considered in sections 2.1 and 2.2.
As models we shall consider the following two nonlinear par-
tial integro-di erential equations:
" Z ! #
t 2
@U @ @U @U
a d = f (x; t) (2.22)
@t @x 0 @x @x
and
2.4. SOME FEATURES OF MODELS I AND II 31
!
Z tZ 1 2
@U @U @ 2U
a dxd = f (x; t): (2.23)
@t 0 0 @x @x2
These equations are one-dimensional one-component case of

the equations arising at mathematical modeling of process of
penetration of an electromagnetic eld into a substance with
temperature-dependent coe cient of electroconductivity. As we
noted in section 2.1, equation (2.22) given here was o ered in
[187] and [188] at rst and then was generalized in numerous
works [49], [50], [137], [139], [141], [145], [148], [219], [303], [304],
[305], [306], [333], and [334]. Equation (2.23) describes the same
process as (2.22) and is rst given in [306].
Under certain conditions on the initial data for these equa-
tions the existence and uniqueness of the initial-boundary value
problems with rst kind boundary conditions are given.
For investigations of initial-boundary value problems dis-
cussed in the book, we use usual Sobolev spaces Wpk ( ), W kp ( ),
Lp ( ), C k ( ), Lq 0; T ; Wpk ( ) , Lq 0; T ; W kp ( ) and their
properties, see, for example, [7], [136], [175], [327], and refer-
ences therein. Some description of these spaces will be given in
section 3.6 as well.
In the domain QT = (0; T ) of the variables (x; t); where
= (0; 1) and T is a xed positive constant for equations (2.22)
and (2.23) we shall consider the following initial-boundary value
problem:
U (0; t) = U (1; t) = 0; t 2 [0; T ]; (2.24)

U (x; 0) = 0; x 2 : (2.25)
We shall search the solution of problem (2.22), (2.24), (2.25)
which satis es the identity
Z Z
@U
+ A(U ) V dxdt = f V dxdt; (2.26)
QT @t QT
where V is any function from the appropriate space (this space

for case a(S) = 1 + S will be speci ed in section 3.6), and
" Z ! #
t 2
@ @U @U
A(U ) a d :
@x 0 @x @x
The basic characteristics of equations such as (2.22) is that

higher derivatives of the nonlinear term depend on time inte-
gral. The coe cients a = a(S) in equation (2.22) contains inte-
gral determining nonlocal operator. In the theory of di erential
equations
Z one often uses the fact that the Volterra’s operator of
t
a kind U d is increasing [175]. However, in a combination
0
with derivative on spatial variable this property might be lost.
For example, the equation
Z t
@U @ @U
= a d
@t @x 0 @x
Z t
is parabolic but if we replace U d = V then the nonlinear
0
equation becomes hyperbolic
@ 2V @ @V
2
= a ;
@t @x @x
which generally does not possess global solutions.
The following existence and uniqueness statement takes place
for problem (2.22), (2.24), (2.25) (see [138]).
Theorem 2.1 If
a(S) = (1 + S)p ; 0 < p 1;
and f 2 W21 (QT ); then there exists a unique solution U of prob-

lem (2.22), (2.24), (2.25) satisfying the following properties:
U 2 L2p+2 (0; T ; W 12p+2 ( )); @U=@t 2 L2 (QT );
@ 2U
2 L2 (QT );
@x2
p @ 2U
T t 2 L2 (QT ):
@t@x
Theorem 2.1 and some of its generalizations for (2.22), (2.23)

type equations and for their multidimensional variants are proved
in [137], [138], [139], [141], [145], [187], [188], and [219], with ap-
plication of Galerkin’s and compactness methods [327], [328],
and [461].
It should be noted that Theorem 2.1 in case p = 1 will be
proved completely in section 3.6.
Below we discuss the asymptotic behavior of the solutions of
the initial-boundary value problems. The existence result can
also be proved based on the a-priori estimates given in the inves-
tigation of the asymptotic behavior of the problem for Models I
and II.
In [305] and [306] it was shown that equation (2.22) gives rise
to coercive monotone operators in the space W 1p (0; 1; L2 (0; T ))
but not in the traditional L2 (0; T ; W 1p (0; 1)). In mentioned works
by the same scheme the analogous results are obtained for (1.12)
type models.
2.4.2 Asymptotic behavior of the solutions as

time tends to in nity
This section is devoted to longtime behavior of the solutions
of initial-boundary value problems for the nonlinear integro-
di erential models (2.22) and (2.23). These equations arise at
mathematical modeling of process of penetration of an electro-
magnetic eld into a substance, with temperature dependent
coe cient of electroconductivity. These equations were derived
in sections 2.1 and 2.2.
The initial-boundary value problems with rst-type bound-
ary conditions are stated. Investigations, which are made in
[217], [218], [228], [231] are given. Let’s formulate the state-
ment for asymptotic behavior of the solutions of the problem
with homogeneous boundary conditions.
At rst for one nonlinear integro-di erential problem (2.22)
we shall give a-priori estimates of the solutions independent of
t. From these estimates the stabilization of the solution follows
as t ! 1.
In the cylinder Q = (0; 1); we consider problem (2.22),
(2.24), (2.25).
Let us assume that
a(S) a0 = const > 0:
If
2a0
;
C
where C here and below is a positive constant independent of t,
the following estimate takes place
t=2
jjU ( ; t)jj e jjU0 jj :
Here jj jj is a norm of the space L2 ( ), i.e.,

Z 1=2
jjU ( ; t)jj = U 2 (x; t)dx :
Sometimes, we will use subscript to indicate di erent kinds of

norms.
Thus, for a problem (2.22), (2.24), (2.25) as t ! 1 the sta-
bility of the solution has decaying exponential rate in the norm
of the space L2 ( ). The asymptotic behavior of the solution of
problem (2.22), (2.24), (2.25) as t ! 1 can also be proved in
the norm of the space W21 ( ).
Using a technique found in [270], in which the stabilization of
the solution of an initial-boundary value problem for the equa-
tions of a barotropic viscous liquid is established, the following
statement is proved in [217].
Let us assume that a = a(S) satis es a(S) a0 = const > 0
and the two additional restrictions: a0 (S) 0 and a00 (S) 0:
In this case the following a-priori estimates are true for all t
[218]:
Z tZ 2
@U
dxd C;
0 @x
Z t Z 2
d @U
dx d C:
0 d @x
Therefore,
Z 1 Z 2
@U
dxd C;
0 @x
Z 1 Z 2
d @U
dx d C:
0 d @x
Now, using well-known fact (see, for example, [25], [270], and
[296]) it is clear that
jjU ( ; t)jjW 1 ( ) ! 0; t ! 1:
2
Thus, stabilization of the solution of problem (2.22) is proved

in the norm of the space W21 ( ) under the following conditions
on coe cient a = a(S):
a(S) a0 = const > 0; a0 (S) 0; a00 (S) 0:

Note that such type of result is true for problem (2.23)-(2.25)
as well.
It is interesting to get the stabilization character in a stronger
norm as well. Let us also note that some results about exis-
tence, uniqueness, and asymptotic behavior of the solutions for
above-discussed integro-di erential models and that generaliza-
tions with several kinds of initial and boundary conditions are
studied in [28], [29], [31], [32], [33], [34], [35], [49], [50], [137],
[138], [139], [140], [141], [145], [146], [147], [186], [187], [188],
[217], [218], [219], [220], [222], [223], [227], [228], [229], [231],
[232], [233], [234], [235], [236], [237], [238], [239], [240], [241],
[242], [243], [245], [246], [247], [248], [249], [251], [252], [253],
[254], [256], [257], [261], [276], [277], [278], [280], [282], [283],
[302], [303], [304], [305], [306], [322], [333], and [334] among oth-
ers.
2.4.3 Rate of the asymptotic behavior of

solutions of Model I
The relations in subsection 2.4.2 give the stabilization of a non-
stationary solution with the norm of the space W21 ( ), but usu-
ally do not provide any information on the convergence rate.
Here we rectify this de ciency and give asymptotic behavior to

be also con rmed by the numerical experiments in Chapter 4.
Consider the following initial-boundary value problem:
@U @ @U
= a(S) ; (x; t) 2 Q; (2.27)
@t @x @x
U (0; t) = U (1; t) = 0; t 0; (2.28)

U (x; 0) = U0 (x); x 2 [0; 1]; (2.29)
where Z t 2
@U
S(x; t) = d : (2.30)
0 @x
Here a = a(S) and U0 = U0 (x) are given functions of their
arguments.
Let us estimate the rate of asymptotic behavior of the solu-
tion. In particular, let us show that the convergence occurring
in the relations in subsection 2.4.2 is also exponential.
Note that, in this section, we derive asymptotics with indi-
cation of order for k@U=@tk as well.
Let us note once again that below in this and in the next
sections C denote positive constants independent of t.
Theorem 2.2 If a(S) a0 = const > 0; a0 (S) 0; a00 (S)

0, and U0 2 W22 (0; 1) \ W k2 (0; 1); then the solution of problem
(2.27)-(2.30) satis es the relation
@U @U a0
+ C exp t :
@x @t 2
Proof. We multiply (2.27) by @U=@t and integrate the resulting

relation over the interval [0; 1]: By using the formula of integra-
tion by parts and taking into account the boundary conditions
(2.28), we obtain
2 Z 1 2
@U 1 @ @U
+ a(S) dx = 0: (2.31)
@t 2 0 @t @x
Let us di erentiate relation (2.27) with respect to t:
@ 2U @ @a(S) @U @ 2U
+ a(S) = 0: (2.32)
@t2 @x @t @x @t @x
We multiply (2.32) by U and integrate the resulting relation
over the interval [0; 1]:
Z 1 2 Z 1 2
@ U @a(S) @U
2
U dx + dx
0 @t 0 @t @x
(2.33)
Z 1 2
1 @ @U
+ a(S) dx = 0:
2 0 @t @x
It follows from (2.31) and (2.33) that
Z 1 2 2 Z 1 2
@ U @U @ @U
2
U dx + + a(S) dx
0 @t @t 0 @t @x
Z 1 2
@a(S) @U
+ dx = 0;
0 @t @x
which, in view of the relation
Z 1 2 2
@ U @U 1 d2
2
U dx + = kU k2 ;
0 @t @t 2 dt2
acquires the form
1 2
1 d2
Z
d @U
2
kU k2 + a(S) dx = 0: (2.34)
2 dt dt 0 @x
Let us multiply (2.32) by @U=@t and integrate the resulting

relation over the interval [0; 1]. By using the formula of integra-
tion by parts, the boundary conditions (2.28), and Poincare’s
inequality
@U @ 2U
;
@t @t @x
we obtain
2 2 Z 1 4
d @U @U 1 @ @U
+ 2a0 + a0 (S) dx 0: (2.35)
dt @t @t 2 0 @t @x
Multiplying (2.27) by U and integrating the resulting rela-
tion over the interval [0; 1]; we obtain
Z 1 2
1 d 2 @U
kU k + a(S) dx = 0; (2.36)
2 dt 0 @x
which, together with Poincare’s inequality and the condition
a(S) a0 ; implies that
1d
kU k2 + a0 kU k2 0: (2.37)
2 dt
We multiply relations (2.34), (2.36), and (2.37) by ; , and
, respectively, where ; , and are positive constants. By
summing the resulting relations and inequality (2.35), we obtain
Z 1 2
+ d 2 @U
kU k + a(S) dx + a0 kU k2
2 dt 0 @x
1 2 2
d2
Z
d @U d @U
+ 2
kU k2 + a(S) dx +
2 dt dt 0 @x dt @t
2 Z 1 4
@U 1 0 @ @U
+2a0 + a (S) dx 0;
@t 2 0 @t @x
Another random document with
no related content on Scribd:
So I found no difficulty in arranging with Mr. Holley to take a trip
with me, and visit some of my engines in operation, for the purpose
of forming a judgment as to its suitability for the use of his clients.
This he agreed to do as soon as he had finished the report of his trip,
on which he was then engaged. Our inspection took in the engines
running in New York and Brooklyn and vicinity and in New England,
finishing with the engine at the Arlington Mills in Lawrence. They
were all found to be on their best behavior, but Mr. Holley told me
that the engine at Lawrence, which was running there at its intended
speed of 150 revolutions per minute, impressed him more than all
the rest put together; not that it was doing any better, for they all ran
equally well, but solely because it was larger. It made him awake to
the great possibilities of the engine.
On his return Mr. Holley prepared a report on the performance of
the engine, and cordially endorsed it as sure of ultimate general
adoption. But he found capitalists to be absolutely dead. Not even
his great influence could awaken in them the least interest. The time
for the promoter had not yet come. And still my success in winning
Mr. Holley’s support proved to be vital to my subsequent progress.
As a last possible resort I finally thought of Mr. Phillips of Newark.
The firm of Hewes & Phillips had become dissolved by the death of
Mr. Hewes, and so, by purchase of Mr. Hewes’ interest from his
heirs, Mr. Phillips was the sole proprietor of the largest engineering
works in New Jersey. That concern had some time before the death
of Mr. Hewes given up the manufacture of steam-engines, a style
made by them having proved unsuccessful, and confined
themselves to making machine tools. In this line their business was
exceedingly dull, being disastrously affected by the depressed and
stagnant condition of the times.
I found Mr. Phillips ready to listen to me. He said that what he
knew about the engine was favorable, although he had not heard of
it for the last two or three years, but he was willing to consider a
proposition to take up its manufacture. I told him frankly that I had no
proposition of that kind to make. I wished to get the manufacture of
the engine revived, but to retain the business in my own hands, to
carry it on myself in my own name, with the view of gaining for the
engine a reputation that would enable me to command the capital
necessary to establish its manufacture in works that I had long
before planned for that purpose, and in which I could devote myself
to the development and building up of the business; that I hoped to
be able to reach this point in the course of two or three years, when
probably the anticipated financial revival would fill his works with
business in his own line of toolmaking.
He said that my proposal was entirely inadmissible, that he could
not permit any independent business to be carried on in his
establishment, and stated firmly the impossibility of any arrangement
of the kind I suggested, which would be something quite unheard of.
I stood firmly on my own position, but was obliged to leave him
without any sign of yielding on his part. The negotiation was,
however, renewed, exactly how I cannot now recall, but it ended in
my carrying my point. We finally concluded a bargain, in which I held
onto the business, but, of course, had to insure to him pretty much
all the profits. This I did not mind, my object was to obtain a position,
which it will be seen I fully accomplished, but did not know what to
do with it. I was conscious that I could never have made this
arrangement but for the extreme stagnation of the times; but was not
aware of an additional reason which impelled Mr. Phillips to agree to
my terms, when he found he could not do any better. What this
reason was will appear pretty soon.
The arrangement was to go into effect as soon as I got an order.
This was my next job. I learned that Mr. Peters, a manufacturer of
high-grade knit fabrics in Newark, all which, by the way, were sold by
him to importers in New York, was carrying on also a manufacture of
light oilcloths in Newark in temporary quarters, and was building a
large structure for this purpose in East Newark, the building now and
for many years past occupied by the Edison lamp manufactory, and
was in the market for an engine. I called on Mr. Peters, and got from
him the privilege of submitting an estimate for this engine. For this
purpose I went to his then present works, and measured the amount
of power he was using, and found that one of my 8×16 engines
would give him that power with the additional amount he wished to
provide for.
On calling with my estimate early one morning, I found Mr. Peters
ready to bow me out. He told me that he had been informed that the
high-speed engines had proved a failure, and the manufacture of
them had been abandoned three or four years ago. I said to him,
“Mr. Peters, I would like to make you a proposition.” He replied that
he would hear it.
I then said, “Your engineer, Mr. Green, I suppose never saw a
high-speed engine, but he strikes me as a fair-minded, cool-headed
man. I have three engines made by me in Harlem, and which have
been running from four to six years, two in New York and one at the
J. L. Mott Iron Works at Mott Haven. These can all be visited in one
trip. I propose that you send Mr. Green to see them in operation, and
talk with the engineers and owners and learn all about them, and
that you suspend your decision until you get his report.” “That is a
fair offer,” said he. “I will send him to-day.” I called again the next
day, and found Mr. Peters ready to throw the order into my hands.
Mr. Green told me afterwards what his impressions were. In the most
cool manner, entirely free from any excitement, he said: “My only
wonder is that everybody does not use this engine and that all
builders don’t make it. I got the same report everywhere. Would not
have anything else. Costs less money, occupies less space, burns
less coal, needs less attention, never cost a cent for repairs, never
anything the matter, never varies its speed.”
And so I began business in Mr. Phillips’ shop, where I continued
for four years, the most delightful period in my active life. I had Mr.
Goodfellow in his old place as my foreman, and three or four of my
best men back again at the work they loved. Everything went
smoothly and harmoniously, and the business grew steadily until the
orders thrust upon me became larger than I could have filled if I had
had the whole works to myself. In re-introducing the engine to the
public, I determined to change its name. I had been asked
occasionally what I had to do with the Allen engine. It struck me that
I had a good deal to do with it. Starting from Mr. Allen’s single
eccentric link motion, and four-opening equilibrium valve and my
own governor, I had, with the help which I have been happy to
acknowledge, created the high-speed engine, had solved every
problem, theoretical and practical, which it involved, and designed
every part of it. So I felt it to be proper that it should thereafter be
known as the Porter-Allen engine.
The following incident illustrates the ease with which everything
down to the smallest detail may unconsciously be prepared to insure
a disaster at some time.
Mr. Peters’ engine-room was a long, narrow room on one side of
the boiler-room, from which was the only entrance to it directly
opposite the guide-bars of the engine. The door opened inward, and
the latch was not very secure. They burned soft coal, which was
wheeled in on an elevated plank and dumped into a heap in front of
the furnace.
One day, about a year after the engine was put in, there was a
great wind blowing. A gust of unusual force blew the engine-room
door open at the instant when a barrowful of coal was being
dumped, and carried a cloud of its dust over the guide-bars. The
engine was soon brought to a standstill. All the faces of cross-head
and guide-bars were deeply scored. It was found, however, that
when these were cleaned up and scraped over to remove all
projections that they ran as well as ever, the grooves proving good
oil distributors, but they were not so pretty to look at.
One day, two or three weeks after we commenced work on this
engine, Mr. Phillips’ bookkeeper came to me and said: “Mr. Peters’
engine is contracted to be running on the first of May, is it not?”
“Yes.” “Do you think it will be ready?” I replied that the work was in a
good state of forwardness, and I thought most likely it would be
running before that time. I should say that was a size for which I had
made the revised drawings already, and the old cylinder pattern had
been readily altered to the new style. “Well,” said he, “Mr. Phillips is a
little short to-day, and he would be much obliged if you would give
him your note for a thousand dollars to come due, say, the fifteenth
of May.” So I gave him the note, the engine was ready on time,
accepted and paid for, and the note met at maturity.
This was the beginning of a uniform process, which continued for
four years. It was disclosed that Mr. Phillips’ financial position was
the same as my own, neither of us had a cent of money. The way we
managed was this. I always afterwards required payments in
instalments, one quarter with the order, one quarter when the engine
was ready for shipment, and the balance when running satisfactorily.
Thus with my notes we got along famously. My orders were always
from first-class parties, engines always ready on time, always gave
satisfaction, and promptly paid for. I had many thousands in notes
out all the time, and never had to renew a note. Mr. James Moore of
Philadelphia, the celebrated builder of rolling mill machinery, once
long after remarked to me, “I keep my bank account in the shop.” It
occurred to me that I had always done the same thing.
Directly after we got running I received a letter from William R.
Jones, superintendent of the Edgar Thompson Steel Company,
running a rail mill recently started at Braddocks by Carnegie
Brothers, saying that they were in need of an engine to drive a
circular saw at a very high rate of speed to cut off steel rails cold.
They had been recommended by Mr. Holley to get one of mine, and
if I could furnish a suitable engine immediately he would order it.
Fortunately I could. While I was building engines in Harlem, the city
of Washington, D. C., went into the system of wooden pavements,
and the contractor obtained an engine from me for sawing up the
blocks. About the very time I received Mr. Jones’ letter I had learned
that the wooden pavement system was being abandoned in
Washington for asphalt and the sawing-mill was closed. I at once
wrote to the contractor making him an offer for the engine. I received
by return mail a reply accepting my offer, and adding most
complimentary words concerning the engine. These I remember
closed by saying that his admiration of it was such that if he were
able he would put the engine in a glass case and keep it there as
long as he lived.
The engine proved just right for Mr. Jones’ use. I went myself to
Braddocks to see it started. All were much interested in the governor
action, I as much as any one, for I had never before seen this
particular application of it. In sawing through the head and web and
bottom flange of the rail, the width of section being cut varied
continually, and the gentle rising and falling of the counterpoise,
adjusting the power to the resistance, while the engine kept, so far
as the eye could detect, a uniform motion, had about it a continual
fascination. The success of this engine brought me several orders
for governors, the most important of which was one from Mr. Jones
himself for governors and throttle valves for his blooming mill and
rail-mill engines. I got up for him balanced piston valves which
operated perfectly. In iron valves and seats of this character it had
been found, where the steam contained primed water, that their
edges wore rounded, and their action in regulating the motion
became less and less satisfactory. I knew that these boilers primed
badly, and avoided this defect by setting brass rings in the edges.
The following illustrations show this regulating valve which I
designed and made in two sizes.
Mr. Porter’s Regulating Valve.
The brass liner for the lower seat was passed through the upper
seat by being made thinner than the upper liner. Those for the valve
were made ¹⁄₈ inch too long, and guttered in the lower edge. They
were then driven down by a set and sledge on an anvil. By going
around them three times the lower edges were spread out to fill the
chamfer, and the flanges brought down to their seats. Those for the
lower valve were put in in halves.
William R. Jones
CHAPTER XXIII
Experience as Member of the Board of Judges At the Philadelphia Centennial

Exhibition.
ne day in April I was surprised to receive by mail a

commission as a member of the Board of Judges in
Group Twenty of the Philadelphia International
Exhibition. I was at a loss to know how I got it, but
learned afterwards that I had been appointed on the
recommendation of Mr. Holley, who was consulted by
the commissioners about the judges in several groups. The
exhibition was opened on May 1, but the judges were not to
assemble until the 24th, and on that day we had quite a ceremony in
the judges’ hall. The American judges were seated at one side of the
hall and rose to receive the foreign judges who filed in from some
place where they had been corralled, while a fine band played the
national airs of all nations that had any airs. After a time spent in
welcoming and responsive addresses, we were marched to a large
café and given luncheon, after which the different groups were
organized. There I had the pleasure of first meeting Mr. James
Moore, also Professor Reuleaux of Berlin and Colonel Petroff of St.
Petersburg; and Emil Brugsch the interesting Egyptian
commissioner, also serving as a judge in our group. I observed that
these foreigners used the English language more accurately than I
did. We organized by the election as president of Horatio Allen,
formerly president of the Novelty Iron Works (then extinct), he being
the oldest and the biggest man among us. Under Mr. Allen’s
administration we had a fine illustration of how not to do anything—of
endless preparation and never getting to work. He had an
interminable series of subjects for discussion and was accustomed
to say: “These questions must be all settled before we can enter
upon the discharge of our duties, gentlemen.” This had the effect
upon our foreign judges that they absented themselves from our
meetings. I remember Mr. Moore saying to me: “Porter, if you and I
had had this work to do we would have had it half done by this time.”
Directly after that Mr. Moore resigned, ostensibly pleading want of
time to attend to it, but really disgusted at the waste of time. Our
work was in a state of chaos. The field was very extensive, as it
embraced all exhibits pertaining to steam and water except
locomotives. One morning I came to the meeting with a copy of the
catalogue on which I had divided the exhibits into three classes,
lettered A, B, and C: class A embraced steam-engines and their
accessories, class B boilers and their accessories and class C
pumps and their accessories; I had prefixed these letters to the
names of all our exhibits according to their class. At this meeting, at
which I had procured the attendance of the foreign judges, this
classification was unanimously adopted, and the judges formed
themselves into these classes accordingly. Our work was then
undertaken in earnest; it was found to be really too extensive to be
accomplished otherwise.
Mr. Charles E. Emery was appointed a judge to fill the vacant
place made by Mr. Moore’s resignation, and he proved most efficient.
As is well known, medals were not awarded, but brief written reports
were made on those exhibits which were deemed most deserving;
these reports were signed by all the judges.
Professor Francis Reuleaux
The firm of E. P. Allis & Co. of Milwaukee, exhibited a sawmill. This

exhibit consisted of two large circular saws, each driven by a
horizontal engine. The two engines were united by a common shaft
on the ends of which the cranks were set at right angles with each
other. The center lines of these engines were nearly 20 feet apart;
the shaft carried two belt drums 8 or 10 feet in diameter, one of them
near to the bed of each engine; at the middle of the shaft was a fly-
wheel about 16 feet in diameter. The rim of this fly-wheel was in
eight or ten segments, with an arm attached to the middle of each
segment; the segments were bolted together and the arms were
bolted to a hub on the shaft. The saws were set behind the cylinders,
and the belts were carried from the drums on the shaft past the
cylinders to smaller drums on the saw arbors. On starting these
engines the two bearings of the main shaft heated so badly that the
engines had to be stopped. The gentleman in charge of the exhibit
applied to me for advice. I told him that although his shaft was large
it was long, and the weight of the fly-wheel bent it so much that the
two journals ran on the inner edges of their bottom boxes, which
caused the heating. I told him he did not need the fly-wheel at all; the
cranks being quartering, the momentum of the belt-drums was amply
sufficient to maintain uniform motion, and I advised him to take off
the fly-wheel. This he did at once, leaving only the hub on the shaft;
the engines then ran with cold bearings and uniform motion
throughout the exhibition. They had made a cut-off gear for these
engines, but it was found not to suit the purpose and was taken off.
This firm then did a great stroke of business: they came to the
sensible conclusion that they could do a great deal better than to
attempt to work out a new system of engineering for themselves, so
they offered to Mr. Edwin Reynolds, the manager of Mr. Corliss’
works, and to his head draftsman, inducements sufficient for them to
leave Mr. Corliss’ employment and take the same positions in the
Allis works at Milwaukee for the manufacture of the Corliss engine
there. With the magnificent result of this action the engineering world
is familiar.
We had all sorts of queer experiences. One day I was demanded
by Mr. Jerome Wheelock to tell him why the engine exhibited by him
was not a perfect engine. I glanced over the long slender bed, a
copy of the Corliss bed without its rigidity, and declined to answer his
question. Mr. Emery was more compliant; on receiving the same
demand, he kindly pointed out to Mr. Wheelock one respect in which
his engine could hardly be considered perfect; the steam was
exhausted into a large chamber embracing the lower half of the
cylinder from end to end. This comparatively cold bath produced the
condensation of a large quantity of the entering steam. From the
middle of this chamber a pipe took away the exhaust from the
opposite ends of the cylinder alternately. Mr. Wheelock admitted the
defect, and said in future he would avoid it, so, as I learned, having
two exhaust pipes instead of one, he gave to each pipe one half the
area of the single one.
I had the pleasure of renewing my acquaintance with Professor
Sweet, who was superintending the exhibit of the mechanical work of
his boys at Cornell; this was very creditable and included quite a
show of surface plates.
The Corliss engine in this exhibition was far the most imposing,
and to the multitude the most attractive single exhibit ever shown
anywhere. It consisted of two distinct engines, each having a
cylinder 40 inches in diameter, with 10 feet stroke of piston, the
motion of which was transmitted through cast-iron walking beams to
cranks set at right angles with each other on the opposite ends of a
common shaft. This shaft made 36 revolutions per minute and
carried a gear-wheel 30 feet in diameter; this wheel engaged with a
pinion 10 feet in diameter on the line of shaft under the floor, giving
to this shaft a speed of 108 revolutions per minute.
One day I said to Professor Sweet: “Do you know, Professor, that
an engine with a single cylinder of the same bore as these and 5 feet
stroke directly connected with a line shaft and making 150
revolutions per minute, with a fly-wheel 10 or 12 feet in diameter,
would exert more power than is afforded by this monster and would
run with far greater economy, because the internal surfaces to be
heated by the condensation of the entering steam would be one
piston instead of two, two heads instead of four, and 5 feet length of
exposed cylinder instead of 20 feet?” He replied: “That is all very
true, but how would you get the steam in and out of the cylinder
properly with a piston travel of 1500 feet per minute?” I was not
prepared to answer that question on the instant, but I afterwards
found no difficulty about it.
The accompanying figures illustrate this engine and my high-
speed equivalent drawn to the same scale; it will be seen that the
small engine occupies about one tenth of the floor space needed for
the large one, and would cost less than ten per cent. of the money. It
would also have a more nearly uniform motion, the impulses
received by the crank being 300 per minute, against only 144 per
minute received by both cranks of the large engine, besides which in
the latter the full force of the steam is exerted at the commencement
of each stroke and falls to nothing at the end, while in the smaller
engine, by the inertia of the reciprocating parts, the forces exerted at
the opposite ends of the stroke would be practically equalized. The
reader will doubtless inquire, as Mr. Green did why, with these
enormous advantages, does not everybody use the high-speed
engines and every builder make them?
The Corliss Engine Exhibited at the Centennial Exhibition.
Porter-Allen Engine Equal in Power to the Exhibited Corliss Engine.
At this exhibition the Bell telephone was first shown to a select

company, among which were President Grant and Dom Pedro, the
last emperor of Brazil. This exhibition was given on Sunday, that
being the only day when silence could be had. Human speech, both
in talking and singing, was transmitted through the whole length of
the main building, about 1800 feet; it has since been transmitted
somewhat further.
The exhibitors of hand pumps all talked about the ease with which
their own pumps could be worked; one man touched bottom in this
respect. He had set his pump so that the spout was nearly on a level
with the surface of the pool from which it drew its water; he boldly
claimed that his pumps required no power at all. I was invited, as I
suppose multitudes were, to take hold of the handle and see for
myself that his claim was true. I never heard of but one man who I
think would be satisfied with this demonstration; that was the
engineering editor of the New York Tribune. Shortly before this he
had published an account of a wonderful pump invented by a Mr.
George, which he concluded by saying that the superiority of Mr.
George’s pump lay in the fact that at each stroke not the whole
column of water had to be lifted, but only that which was to be
discharged. We had a waterfall maintained by a centrifugal pump,
which received its water on one side only; the maker evidently
knowing nothing about the method of balancing these pumps by
admitting the water equally on the opposite sides.
The boiler-makers abounded. My old acquaintance, the Harrison
boiler, turned up. Mr. Allen urged a favorable award to Mr. Harrison
because of the motives of humanity by which he knew Mr. Harrison
was actuated in designing that boiler. A Mr. Pierce invited all the
judges to visit his boiler and hear him explain it. He informed us that
this boiler had been the subject of three scientific tests by Professor
Thurston, but he did not tell us the results of those tests.
As we were coming away Professor Reuleaux said to me: “That is
foolishness, isn’t it?”
An inventor named Smith came several times to our judges’ room
to urge upon us the merits of his boiler. He had two on exhibition,
one in use in the boiler-house and the other in Machinery Hall; these
were quite different from each other. One day not long after the close
of the exhibition I received a note from a stranger requesting me to
call upon him at the Astor House. I thought, “This man doubtless
wants an engine, but his time is too precious to come out to
Newark,” so at the hour appointed I was there. When I entered the
room the first object I saw was a sectional model of this Smith boiler,
and I found that the gentleman wanted to know our reasons for
overlooking that boiler. I replied to him that I had a question to which
I would like an answer at his earliest convenience; we observed that
the two boilers exhibited by Mr. Smith were quite different from each
other, and I saw that this model differed in essential details from both
of them, and I would like to know which one he wished us to approve
of and bade him good afternoon.
One day afterwards I happened to be in Mr. Holley’s office in New
York when a man came in with a drawing of a boiler which he wished
Mr. Holley to recommend. Mr. Holley turned him over to me, and he
explained to me that the great novel feature of his boiler was that the
feed-water was admitted by spraying it into the steam space, thus
avoiding the cooling of any part of the boiler by its admission at one
point; so I found one freak boiler that was not at the exhibition.
We had a fine exhibit of steam fire-engines. I think every maker in
this country was represented, and we had a trial of these engines
lasting three or four days. The committee desired to make a
thorough comparative test of their performance, but the man (a
lieutenant in the navy) appointed to keep the record put down so few
items that we found we had no record at all. We could only guess
how he came to do this.
An exhibitor from Canada brought an engine that presented a very
fine appearance; it was made up of a collection of what he believed
to be the best features of every steam-engine made in the United
States. The experts looked his machine over and saw where he had
got every one of them, but his different appropriations did not work
well together; his engine broke down every day and he worked all
night to be ready for the next day’s trial. It afforded a good
commentary on the narrow-minded laws of Canada, which forbade a
citizen of the United States from taking out a patent there.
The show of steam-engines was not large, and the indicator was
not applied to any engines, so I had no use for the indicators I had
imported from England. If I remember rightly, we had only two
engines from abroad, one of these sent by the Government of Brazil.
This was what was called a “table” engine, in which the cylinder
stands on a table in a vertical position and two connecting-rods
extend down from the cross-head and connect with the crank under
the table. It was copied from a Scotch elementary drawing-book from
which I learned mechanical drawing. One of these engines had been
made by Mr. Hoe to drive the press of the New York Daily Times
when that paper was started in 1851 or 1852. The other foreign
engine was made by a Brussels manufacturer with the assistance of
the Belgian Government. It had an American cut-off which was used
by Mr. Delamater on his engines, and it had the eccentric between
the main bearing and the crank, giving to the latter therefore three or
four inches of unnecessary overhang; it had my condenser, which I
learned was then coming into considerable use on the Continent.
Col. Alexis Petroff
The only American engines I now recall besides the Corliss were
the Buckeye and the Brown engines, and our awards to these
engines did not do them any harm; the Corliss engines were not
within our jurisdiction and we were not permitted to say anything
about them; Mr. Corliss was not a competitor but a patron of the
exhibition.
Mr. Frederick E. Sickels made an extensive exhibit of his various
inventions, the models of which had been loaned to him for that
purpose by the Patent Office. Only two of these inventions came
within our province: the first was what is known as the celebrated trip
cut-off, patented by him in the year 1842; the latter an arrangement
patented in 1848. The former invention was an improvement on the
Stevens cut-off, already in general use in steamboats on our Eastern
waters. The Stevens invention was applied to equilibrium valves,
rising and falling in a direction vertical to their seats. It enlarged the
opening movement of the valve in a degree increasing as the speed
of the piston increased, by means of the device known as the wiper
cam; but the closing motion of the valve, being the reverse of the
opening movement, grew slower and slower, until the valve was
gently brought to its seat. It was found that during the closing of the
port a great deal of steam blew into the cylinder through the
contracting openings, with very little addition to the useful effect. Mr.
Sickels conceived the idea of liberating the valve just before the
opening movement was completed and letting it fall instantly to its
seat, which would effect a sharp cut-off and a great economy in the
consumption of steam. This action involved the difficulty that the
valves would strike their seats with a violent blow, which would soon
destroy both. This difficulty Mr. Sickels met by the invention of the
dash-pot. This apparatus performed two functions: when its piston
was lifted above the water it left a vacuum under it, so the pressure
of the atmosphere on this piston was added to the weight of the
valve and the pressure of the steam on it to accelerate its fall. This
was arrested by the piston striking the surface of the water just in
time to prevent the valve from striking its seat, but not soon enough
to prevent the complete closure of the port. This nice point was
determined by the ear. The engineer first let water out of the dash-
pot gradually, until he heard the valve strike its seat faintly; then he

Ebook Numerical Solutions of Three Classes of Nonlinear Parabolic Integro Differential Equations PDF Full Chapter PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Ebook Numerical Solutions of Three Classes of Nonlinear Parabolic Integro Differential Equations PDF Full Chapter PDF

Uploaded by

Copyright:

Available Formats

Numerical Solutions of Three Classes

Copyright © 2016 Elsevier Inc. All rights reserved.

No part of this publication may be reproduced or transmitted in any form or by any

British Library Cataloguing in Publication Data

Library of Congress Cataloging-in-Publication Data

For information on all Academic Press Publications

This book is concerned with the numerical solutions of some

At the end of the book a list of the quoted literature and

The rst author thanks Fulbright Visiting Scholar Program for

This book is concerned with the numerical solutions of some

Numerical Solutions of Three Classes of Nonlinear Parabolic Integro-Differential Equation.

Key words: Electromagnetic eld penetration, Maxwell’s

Key words: Electromagnetic eld penetration, Maxwell’s

Numerical Solutions of Three Classes of Nonlinear Parabolic Integro-Differential Equation.

In mathematical modeling of applied tasks di erential, inte-

integral equations devoted to distribution of an electrical charge

tem has passed; that is to say, it depends on the history of the

where x = (x1 ; x2 ; x3 ); is a classical three-dimensional Laplace

@ 2 U (x; t) @ 2 U (x; t) @ 2 U (x; t)

where is a known function of its arguments.

where is a known constant. Many authors investigated equa-

where A is a known function of its arguments.

Integro-di erential equations of parabolic type arise in the

where a(S) a0 = const > 0 is again a known function of its

In the works [303], [305], and [306] for conditionally closed

The multidimensional scalar analogues of systems (1.6), (1.7),

This equation serves as a very special model for one-dimens-

on a Hilbert space H with g a nonlinear bounded operator [51].

paid on theorems with proofs for corresponding one-dimensional

and also MS and PhD students of the appropriate specializa-

1.1 Comments and bibliographical

spondingly, are given and studied in [219], [305], and in a num-

As a rule we cannot nd exact solutions of the considered

Numerical Solutions of Three Classes of Nonlinear Parabolic Integro-Differential Equation.

Key words: Mathematical modeling, electromagnetic eld

2.1 Electromagnetic di usion process

If at some moment of time the magnetic eld is solenoidal,

quirement of their belonging to special functional spaces. The

2.1.2 A reduction of system of nonlinear dif-

will be made, is connected to the assumption that the change in

Introducing the function

The functions 0 ; , and cv are positive in physical sense;

If we consider a at eld of the form W = (0; 0; U ), where

of resistance with temperature is characteristic of metals.

a(S) = C1 (C0 + S) 1 + > 0;

From here it is obvious that the function a(S) decreases or

2.2 On the averaged Model II

Following [419] and as in section 2.1, we shall consider the

Here dQ is in ow of a thermal energy absorbed by a body in an

Here we used the known vector identity

Laptev [306] remarked that the research of the so-called aver-

2.3 Mathematical Model III

have suggested a heat ow model based on a memory e ect in

solves the equation

Now de ne a(t) and f (x; t) by

then (2.20) becomes

2.4 Some features of Models I and II

These equations are one-dimensional one-component case of

U (0; t) = U (1; t) = 0; t 2 [0; T ]; (2.24)

where V is any function from the appropriate space (this space

The basic characteristics of equations such as (2.22) is that

a(S) = (1 + S)p ; 0 < p 1;

and f 2 W21 (QT ); then there exists a unique solution U of prob-