ques in numerical heat transfer the con- subsequent chapters can be sensibly general computer program.

TEACH-C,
tent should be of use and interest to a dealt with. used to stimulate heat conduction and
large sector of the research community. The fourth chapter is concerned with species diffusion. The book is intended
The packaging is good, with typeset text numerical implementation of the trans- to strengthen the traditional methods of
and professionally produced illust- formed equations on a regular grid. The teaching heat transfer and fluid mecha-
rations, giving the book a solid and res- numerical formulation considered is nics at undergraduate and postgraduate
pectable image, very important in what based upon finite difference methods but level by the more effective use of the
could become a useful reference work. foeusses upon the finite or control computer in an instructional mode.
volume approach to ensure conserva- Furthermore, it describes the physical
Ft. Voller tion. The only equation considered is and mathematical model on which the
that describing generalized convection- program is based.
diffusion and then by example to illus- The presentation format, layout,
Numerical grid generation-- trate the effects of the transformation. cover design, and the size of the book
foundations and applications Of course, the actual integration takes are pleasing for both students and
J. F. Thompton, S. Warsi and place over the transformed (i.e. regular) teachers. The program TEACH-C is
C. W. Mastin grid so the numerical approximations applicable to heat conduction (or species
North Holland, Amsterdam, 1985, are rather straightforward. The chapter diffusion) problems which are transient
483 pp. 150Dfl does, however, contain some helpful or in the steady state, plane or axisym-
details on coping with interfaces metric, with uniform or variable thermal
This book tackles a subject of growing between subregions of the finite volume conductivity and with distributed heat
importance for those involved in scienti- approach. The discussion on numerical sources.
fic and engineering computation. As its approximation is completed by a section In the field of fluid mechanics it can
title indicates, the book aims to provide on truncation error, which I did not find be used to calculate irrotationaI flows,
the mathematical foundations of the particularly enlightening. flows in porous media and lubricant
numerical generation of meshes and so The question of how to actually locate films. Although limited in applicability
it is not a trivial read. Having said this, the nodal points in the physical region for real world problems, it should prove
the book and the subject will be attrac- from a transformation of those regular useful to students who, after learning
tive to applied mathematicians because nodes in the rectangular integration how to adapt it to solve various pro-
it involves a number of classic topics, domain is divided into elliptic, parabolic blems, will be ready to use one of the
including vector and tensor analysis, and hyperbolic generation systems. In more complicated packages available for
complex analysis, elements of applied elliptic systems, the source term of the the simultaneous solution of both fluid
functional analysis, as well as numerical Poisson equation can be used to flow and heat transfer phenomena.
analysis. influence the variation of the density of Section 1 of the book provides an
The objective of the book is to des- grid cells in a region. A detailed discus- introduction, by outlining the capabili-
cribe the techniques used to generate sion follows of how to determine the grid ties of TEACH-C, discussing the layout
meshes to approximate complex regions density distribution for any particular of the user guide to the program and
within which the behaviour of a variable region; this includes the particular pro- by making suggestions on how to use it
is governed by an appropriate partial blems that three-dimensional regions in a computer-aided learning course.
differential equation. There are a bring. The subject is completed by a sur- Section 2 is devoted to the description
number of ways in which meshes can be vey of algebraic grid generation systems. of the differential equation solved by the
generated to cover a region. The meth- The book concludes with three program and to the derivation and solu-
ods with which this book is concerned chapters on special topics: orthogonal tion of the finite difference equations,
are those associated with the mapping systems, conformal mapping, and adap- to which the original equation is trans-
of a uniform mesh on a simple rectangu- tive grids. I found each interesting and formed. The presentation is clear and
lar shape onto the complex irregular helpful in its own way, but was particu- covers all relevant topics such as compu-
integration domain. Such an approach larly grateful for one of the appendices tational grid treatment of boundary con-
provides a means of achieving meshes which provided a program, together ditions, solution of difference equations,
which fits the integration domain very with a guide through it. convergence, stability, and accuracy. I
precisely. The net result of this is that I enjoyed reading the book and cer- found this section the most satisfying in
the governing equations then have to be tainly feel that justice has been done to the book.
expressed and solved in curvilinear coor- a ditt~cult subject. I would have been Section 3 describes the computer pro-
dinates. The reason for taking this route interested, however, to see how such gram, starting with its capabilities and
is obvious, of course, but whilst one grid distribution techniques actually limitations. The overall structure of the
gains accuracy of representation of the contribute to the accurate solution of an program is then described and the most
region, there is a heavy cost associated important partial differential equation important FORTRAN symbols defined.
with the solution procedure. Certainly (representing convection-diffusion, Attention is then turned to describing
the move from regular to irregular say). I suspect this book will become a the problem solved by the basic
meshes has nontrivial implications for useful reference text for some years to TEACH-C, e.g. a transient 2D conduc-
the finite domain method. come. tion problem. Using this example, the
The book develops the approach in authors describe the various subroutines
a fairly readable way, covering simply M. Cross and their operations. The section ends
and multiply-connected, as well as com- with suggestions for preparation and
posite, regions. Also mentioned are operation of the program. Closing
overlaid grids, but I have to say that, Computer-aided engineering--heat remarks, references (albeit limited) and
although the ability to generate such dis- transfer and fluid flow algebraic notation conclude the main
tributions may be helpful, their utility A. D. Gosman, B. E. Launder and G. part of the text. The appendices provide
must surely be rare and rather restricted. J. Reece four problems:
The mathematics becomes interesting in Ellis Horwood, Chichester, 1985, 179 • unsteady 1D conduction processes
chapter 3 when transformation relations pp., £18.50 • 2D conduction processes
are considered. Here, the vector and • potential flow
tensor analysis is far from trivial and This document, in the form of a ring- • fully-developed flow in noncircular
requires very careful study before the bound booklet, provides a guide to a ducts

462 Appl. Math. Modelling, 1986, Vol. 10, December