cambridge texts in applied mathematics j. m. ottino-the kinematics of mixing_ stretching, chaos, and...
TRANSCRIPT
-
7/25/2019 Cambridge Texts in Applied Mathematics J. M. Ottino-The Kinematics of Mixing_ Stretching, Chaos, And Transport -
1/387
The kinematics of mixing:
stretching, chaos, and transport
J . M .
O T T I N O
Dcprrrtrnc~nt01' C l ~ r r , ~ i r , u l11qirlrc~ri11g,
Unir .c~r.s i t j
1'
Mu.s.srrc~hu.sc~tt.s
CAMBRIDGE
UNIVERSITY PRESS
-
7/25/2019 Cambridge Texts in Applied Mathematics J. M. Ottino-The Kinematics of Mixing_ Stretching, Chaos, And Transport -
2/387
P U B L I S H E D B Y T H E P R E S S S Y N D I C A T E O F T H E U N I V E R S I T Y O F C A M B R I D G E
The Pitt Building, Trumpington Street, Cambridge CB 2 IRP, United Kingdom
C A M B R I D G E U N I V E R S I T Y P R E S S
The Edinburgh B uilding, Cam bridg e CB 2 2RU, United Kingdom
40 West 20th S treet, N ew York, NY 1001 1-421
I ,
USA
I 0 Stamford R oad. Oakleigh. Melbourne 3 166, Australia
O
Cambridge University Press I989
This book is in copy right. Sub ject to statutory exception
and to the provisions of relevant collective licensing agreements,
no reproduction of any part may take place without
the written permission of Cambridge University Press.
First published 1989
Reprinted 1997
Printed in the United Kingdom at the University Press. Cambridge
A catalogue record for this book i s ava ilable from the British Library
Library of Co ngress Cataloguing in Publication data
Ottino,
J.
M.
The kinematics of mixing : stretching, chao s, and transport 1
J.
M.
Ottino.
p. cm.
Bibliography : p.
Includes index.
ISBN 0 521 36335 7. ISBN 0 52 1 36878 2 (paperback)
I. Mixing. I . Title.
TP156.M5087 1989
660.2 '84292-dc 19
88-30253
ISBN 0 521 36335 7 hardback
ISBN 0 52 I 368 78 2 paperback
-
7/25/2019 Cambridge Texts in Applied Mathematics J. M. Ottino-The Kinematics of Mixing_ Stretching, Chaos, And Transport -
3/387
Contents
I
Introduction
I I Physical picture
1.2 Scope an d early works
1.3 Appl icat ions an d geometr ical s t ructure
1.4 Approach
Notes
2 Flow, trajectories, and deformation
7.1 Flow
7.7
Velocity. ;icceler;rtion. Lagrangian and Eulerian viewpoints
2.3 Extension to mul t icom ponent m edia
7.4
('l~issical me ans for visuali/ation of flows
2.4.1
Ptrr/ic,lr prrfll. or hii, or frcijec'for.~
2.4.2 S / r ~ ~ t r ~ l l l i ~ ~ ~ ~ . s
-7.4.3
Srrc~trlilirlc~.~
2 . 5 Steady and periodic flows
7 . 0
1)eformation gradient and velocity gradient
2.7 Kinem atics of deform ation-strain
7.X M ot io n a ro u nd a point
2.9
Kinemat ics o f defo rma t ion: rate of s train
7.10 Rates of cha nge of mater ial integrals
2.1 I'hysical me an ing of
Vv. (V v) ' .
a n d
D
Hibliography
Notes
3
('onserration equations, change of frame, and vorticity
' . I
Principle of conservation of mass
.
2
F'l.inciple o f con serv ation o f linear m o m en tu m
3.-;
7'r:iction
t (n .
x.
1 )
3.4 ('a uchy 's eclu:rtion o f motion
-
7/25/2019 Cambridge Texts in Applied Mathematics J. M. Ottino-The Kinematics of Mixing_ Stretching, Chaos, And Transport -
4/387
Contents
Principle of conservalion of angular momentum
Mechanical energy equa tion an d the energy equation
Ch ang e of frame
3.7.1 Objecticity
3.7.2 Velocity
3.7.3 Accelerarion
Vorticity distribution
Vorticity dynamics
Macroscopic balance of vorticity
Vortex line stretching in inviscid fluid
Streamfunction and potential function
Bibliography
Notes
Computation of stretching and efficiency
Efficiency of mixing
4.1.1 Properties of e, and e,
4.1.2
Typical behavior of the eflciencj
4.1 .3 Flow classification
Examples of stretching and efficiency
Flows with a special form of
V v
4.3.1 Flows with D( Vv )/D t=
0
4.3.2 Flows with D( Vv )/ Dt small
Flows with a special form of
F ;
motions with constant stretch history
Efficiency in linear three-dimensional flow
The importance of reorientation; efficiencies in sequences of flows
Possible ways to improve mixing
Bibliography
Notes
Chaos
in
dynamical systems
Introduct ion
Dynamicdl systems
Fixed points and periodic points
Local stability and linearized maps
5.4.1 Dejinilions
5.4.2
Stability of area preserriny two-dimensional maps
5.4.3 Fumi1ie.s
c
periodic points
Poincare sections
Invariant subspaces: stable an d unstable manifolds
Str uc tura l stability
Signatures of chaos: homoclinic and heteroclinic points. 1-iapunov
exponents an d horseshoe maps
-
7/25/2019 Cambridge Texts in Applied Mathematics J. M. Ottino-The Kinematics of Mixing_ Stretching, Chaos, And Transport -
5/387
C o n t e n t s
vii
5.8.1
Homoi,linic lend heteroclinic connections und points 1 1 1
5.8.2
Sensiticity t o ir~itiul onditions (end Li up un o~ : xpone nts
116
5 . 8 . 3 florseshoe mups
117
5.9 Su m m ary of definitions
of
chaos 124
5.10 Possibilities in higher dimensions
124
Bibliography 125
Notes 128
6
Ch aos in l lamilton ian systems 130
6.1 Intr od uc tion 130
6.2 Ham ilton's eq ua tion s 131
6.3 Integrability of Ilam ilton ian systems 132
6.4 Ge neral stru ctu re of integrable systems 133
6.5 ph ase spa ce of Harn iltonian systems 135
6.6 Pha se space in periodic Ha m ilton ian flows: Poinc are sections an d tori 136
6.7 1,iapunov expo nents 138
6.8 l lom oclinic an d heteroclinic points in H am iltonian systems 140
6.9
Pertu rbatio ns of Harnil tonian systems: M elnikov's m etho d 141
6.10 Behavior near elliptic points 143
6.10.1 PoincurP--H i r k h o th eo rem
144
6.10.2 Th e Kolmogoroa Arnold-Moser theorem (t he K A M theorem )
146
6.10.3 7'he tw ist theorr>m 148
6.1 1 General quali tat ive p icture of ncar integrable chao tic Ham iltonian
systcms 148
Bibliography 152
Notes 152
7
M ixing and chaos in two-dimensional time-periodic flows
7.1
Int roduct ion
7.2 T he tendril-whorl flow
7.2.1
1,occrl c~nu1y.si.s: octcction und slccbilitj o f perioil-1 and period-2
periodic points
7.2.2
(;loha1 unul~.s i.s ind interccctions be tw ee n munifi)kd.s
7.2.3 Formutior1 of horseshoe mccps in the T W mcip
7.3
The blinking vortex flow (BV)
7.3.1 Poinc,ur; seer ions
7.3.2
S luh i li tj o f period-I periodic poitlts ctnd conjugate) lines
7 . 3 . 3
Hor.se.shoe mrcps iri the BV flow
7.3.4
Liupunot. c~.uponenls, rerage cflic.iencbp, [end 'irrecers ihilit y'
7.3.5 Mclcro.sc~opic 1isper.sion o f tracer purtic~les
7.4 Mixing in a journal bearing flow
7.5
Mixing in cavity flows
Bibliography
Notes
-
7/25/2019 Cambridge Texts in Applied Mathematics J. M. Ottino-The Kinematics of Mixing_ Stretching, Chaos, And Transport -
6/387
. .
v l l l
Contents
Mixing and chaos in three-dimensional and open flows
In t roduct ion
Mixing in the part i t ioned-pipe mixer
8.2.1 Appro.uirntrtcl 1 ~ ~ l 0 c i t .vlirld
8.2.2
Poirlcrrrc; .sec,tiort.s trrlil three-dimertsiond struc'turr
8.2.3 E\-it tirile dist rib ut ions
8.2.4
L~ocrrl trercllir~g01' rnirr~rirrl ir1e.s
Mixing in the eccen tr ic hel ical an nu la r mixer
Mixing and dispers ion in a furrowed channel
Mixin g in the Kelvin ca t eyes flow
Flows near walls
Streamlines in an inviscid flow
Bibl iography
Notes
Epilogue: diffusion and reaction in lamellar structures and
microstructures in chaotic flows
Transport at s t r iat ion thickness scales
9.1 I
Pirrirrllrter.~rrrttl rtrrirrb1e.s chtrrtrc~rrrizirly rtrrlsporr ut .srnrrll
9.1.2 Re~jir l lc~.~
Comp l ica t ions and i l lus t ra t ions
9.2.1
Distortiorls of Iirrnrllirr structures trrttl rli.stributiort c1/1;.r.t.s
9.2.2 Illustrtrriorts
Passive an d act ive microst ructures
9.3.1 E.~perirnc~rttrrlt~rr1ir.s
9.3.2
Th ro rr ri c~ rl r~rt1ir.s
Active microst ructures a s proto type s
Bibl iography
Notes
Append ix: C ar tes ian oec tors and tensors
Lis t of' ,f iequently used sy mbo ls
Author index
Color sec t ion bc tn~c~enages 153 and 154
-
7/25/2019 Cambridge Texts in Applied Mathematics J. M. Ottino-The Kinematics of Mixing_ Stretching, Chaos, And Transport -
7/387
Preface
1.11~bjective of this book is to present a unified treatment of the mixing
o f f luids from a kinematical viewpoint . T he aim is to provide a conceptual ly
clear basis from which t o launch analysis an d t o facilitate the und erstan ding
c )f th c ume rous mixing problemsenc ountered in nature an d technology.
['rcsently, the study of fluid mixing has very little scientific basis;
processes and phenomena are analyzed on a case-by-case basis without
a n y
at tempt to discover general i ty . For example, the analysis of mixing
and 's t i rr ing ' of co nta m ina nts an d t racers in two-dimensional geophysical
flows such as in oceans; the mixing in shear flows and wakes relevant to
aeronaut ics a n d co m bu st io n; the mixing of fluids under the Stokes 's regime
generally e nc oun tere d in the 'blending' of viscous liquids such a s poly m ers;
and
the mixing of diffusing and reacting fluids encountered in various
lqpes of chemical reactors share l i t t le in common with each other, except
possibly the nearly universal recognit ion among researchers that they are
\cry difficult problems.'
Phcre are, however, real s imilar i t ies among the various problems and
the possible benefits from an overall at tack on the problem of mixing
uhing a general viewpoint are substant ial .
7'110poir~t
1'
c.ie\\. rrdoptetl Ilc>r.r s rl~trr fi.oru o kir~c~rntrticalir\z.point ,fluid
/ ~ r i . \ i r l ~ ~
s tho c:[jic,icv~rsrrc~rc~hingrrltl ,fbl tlir lg of ' r?~utrriu l irlr.s urld .surCfucc~.s.
S~lch problem corresponds to the solut ion of the dynamical system
tlx,'tlr = V ( X ,
t ) ,
l~11cre he right ha nd side is the Eulerinn velocity field ( a so luti on o f the
N al ier Stokes equa tions . for ex am ple) an d the initial condit ion corresp onds
the init ial configuration of the l ine or surface placed or fed into the
tlolv ( S represents the location of the init ial condit ion
x
=
X ) .
Seen in this
l ight. the problem can be formulated by merging the kinematical founda-
tions of fluid mechanics (Chapters 3 a n d
3 )
an d the theory of dynam ical
s \lems (C ha pt er s 5 ; ~ n d
).
The approach adopted here i s to analyze
siml'lc protypical f lows to enha nc e intuit ion a n d to extract conclusions
-
7/25/2019 Cambridge Texts in Applied Mathematics J. M. Ottino-The Kinematics of Mixing_ Stretching, Chaos, And Transport -
8/387
of general validity.
Mi.uincl i.s strc~tchinyund ,jOMiny and srrc~tchinycrnd
,j i) ldir~g s r l ~ r f ingc~rprint01' chlros. Relatively simple flows can act as
prototypes of real problems and provide a yardst ick of reasonable
expectations for the completeness of analyses of more complex flows.
Undoub t ed l y , I expect that such a program would facil i tate the analysis
of mixing problem s in chem ical, mec hanical , an d aero nau tical engineering,
physics, geophysics, oc ea no gra ph y, etc.
Th e plan of the book is the following: C ha pt er 1 is a visual sum m ary to
motivate the rest of the presentat ion. In Chapter 2 I have highlighted,
whenever possible, the relat ionship between dynamical systems and
kinematics as well as the usefulness of studying fluids dynamics start ing
with the concepts of mot ion a n d , f low.2 Mixing should be embedded in a
kinematical foundat ion. H owever,
I
have avo ided references to curvil inear
co-ordinates and different ial geometry in Chapters 2 and
4,
even though
it could have m ad e the pres entat ion of som e topics m ore satisfying but the
ent i re presentat ion sl ight ly uneven and considerably more lengthy. The
chapter on f lu id dynamics (Chapter 3 ) i s br ief and convent ional and
stresses conceptual points needed in the rest of the work. The dynamical
systems presentat ion (C ha pte rs 5 a n d
6 )
includes a list of topics which
I
have fou nd useful in mixing studies an d should not be reg arded as a
balanced introduct ion to the subject . In this regard, the reader should
note that most of the references to dissipative systems were avoided in
spite of the rather transparent connection with fluid flows.
A few words of caution are necessary. Mixing is intimately related to
flow visualization a n d the ma terial presented here indicates the price on e
has to pay to understand the inner workings of determinist ic unsteady
(albeit generally periodic in this w o rk ) two-dimen sional flows an d three-
dimensional flows in general. How ever, we should no te tha t the geometrical
theory used in the analysis will not carry over when v itself is chaotic.
Though mixing is st i l l dependent upon the kinematics, the basic theory
for analysis would be considerably different. Also, even th oug h m any of the
examples presented here pertain to what is sometimes called 'Lagrangian
turbulence', the reader might find a disconcerting absence of references
to convent ional (o r Euler ian ) turbulenc e. In this regard
I
have decided to
let the reader establish possible connections rather than present some
feeble ones.
I
give full citation to articles, books, and in a few cases,
i f
an idea is
unpubl ished, conferences. When only a last name and a date is given,
part icular ly in the case of problems or examples, and the name does not
ap pe ar in the bibliography, it serves to indicate the so urc e of the p roblem
-
7/25/2019 Cambridge Texts in Applied Mathematics J. M. Ottino-The Kinematics of Mixing_ Stretching, Chaos, And Transport -
9/387
or idea. I t is important to note also that some sections of the book can be
rcgnrded as work in progress and that complete accounts most likely will
(allow, expanding over the short descriptions given here; a few of the
problems, those at the level of small research papers are indicated with
an asterisk
( * ) .
In several passages
I
have pointed out problems that need
work. Ideally, new questions will occur to the reader.
Preface to the Second Printing
In preparing the second printing of this book, several typographica l and
formatting errors have been corrected. The objectives of the book
expressed in the original preface remain unchanged. Owing to space con-
straint limitations the amount of material covered remains approximately
the same. The reader interested in the connection of these ideas with
turbulence will find some leads in the article 'Mixing, chaotic advection,
and turbulence',
Annual Reviews
of
Fluid Mechanics,
22,
207-53 (1990);
a succinct summary of extensions of many of the ideas outlined in this
book is presented in 'Chaos, Symmetry, and Self-Similarity: Exploiting
Order and Disorder in Mixing Processes', Science, 257, 754-60 (1992).
I
should appreciate comments from readers pertaining to related articles in
the area of fluid mixing as well as possible extensions or shortcomings of
the ideas presented in this work.
Notes
I
Even the term inology is com plicated. F or exam ple. in chemical engineering the term s mixing,
agi tat ion. and blending are common (Hyman,
1963;
McCabe and Smith,
1956,
Chap.
2. Section
9 ;
Ulbrecht and Pat terson,
1985).
The terms mixing. advection, and
stirring appear in geophysics; e.g.. Eckart.
1948;
Holloway and Kirstmannson.
1984.
Inevitably. different disciplines have created the ir own terminology (e.g., classical reaction
engineering, combustion, polymer processing, etc.).
2 Kinematics appe ars as an integral part of bo ok s in continuum mechanics but m uch less so
in m odern fluid mechanics. There a re exceptions ofco urse : Ch apte rs
V
and VI of the work
of Tietjens based on the lecture notes by Pra nd tl contain and unusually long description
01 deform ation and motion aroun d a point (Pr an dtl and Tiet jens, 1934).
-
7/25/2019 Cambridge Texts in Applied Mathematics J. M. Ottino-The Kinematics of Mixing_ Stretching, Chaos, And Transport -
10/387
-
7/25/2019 Cambridge Texts in Applied Mathematics J. M. Ottino-The Kinematics of Mixing_ Stretching, Chaos, And Transport -
11/387
-
Acknowledgments
/
This hook grew out from a probably unintell igible course given in Santa
EL : .
Arg entina, in July 1985, followed by a sh ort c ou rse given in Am he rst,
M assach usetts, also in 1985. M ost of the ma terial w as conden sed in eight
lectl~rcsgiven at the California Insti tute of Technology in June 1986,
\vhcrc the bulk of the material presented here was written.
. rhc connect ion between s tretching an d folding, an d mixing an d ch ao s,
bcc;~rnc transp aren t after a con vers ation with H . Aref, then at Brown
IJni\c rsity, du rin g a visit t o Providence, in Sep tem ber 1982.
1
would
particularly like to thank him for communicating his results regarding
the 'blinking v ortex' prior t o pub lication (see Secton 7.3), an d als o for
m a n y research discussions an d his friendship du rin g these years. I would
like to thank also the many c om m en ts of P. Holmes of Corn el l University,
on a rather imperfect draft of the manuscript , the comments of J . M.
Grcenc of
G.
A. Technologies, who provided valuable ideas regarding
syrnnictries as well as to the many comments and discussions with S.
Wiggins and
A.
Le on ard , bo th a t the Cal ifornia Ins t i tute of Technology,
during my stay at Pasadena. I am also gra teful to H . Brenner of the
Massachusetts Insti tute of Technology, S. Whitaker of the University of
California at Davis , W. R . Schow alter of Princeton University, C . A.
Trucsdel l of Johns Hopkins Universi ty, W.
E .
Stewart of the University
of
Wisconsin,
R .
E .
Rosensweig and Exxon Research and Engineering,
and J . E:.. M ar sd en , of the University of California a t Berkeley, for various
comments and s uppor t . I am also part icularly thankful to those who
Supplicd photographs or who permit ted reproductions from previous
Pllhlications (G. M . Corcos of the University of California at Berkeley,
R . ( 'hevray of Columbia University. P.
E.
Dimotakis and L. G. Leal.
the California Insti tute of Technology,
R .
W. Metcalfe of the
Universi ty of Houston, D. P. McKenzie, of the Universi ty of Texas at
Allstin. A. E . Perry of the University of Melbourne, 1. Sobey of Oxford
I J n i \ c r s i t y ,and P. Wellnnder of the Universi ty of Washington).
1
a m
particularly indebted to all my former and present students, but
-
7/25/2019 Cambridge Texts in Applied Mathematics J. M. Ottino-The Kinematics of Mixing_ Stretching, Chaos, And Transport -
12/387
x iv A~~kn o w1cd g mc~n t .s
particularly to R. Chella and D. V . K ha kh ar , for work prior to 1986, and
to J .
G .
Franjione, P. D. Swanson , C . W . Leong, T. J . Danielson, and
F. J. Muzzio, who supplied many of the figures and material used in
Chapte rs
7-9.
1
am a l so indebted to
H.
A. Kusch for help with the
proofs and to H. Rising for many discussions during the early stages
of this w ork . Finally,
I
would l ike to express my grat itude to
D.
Tillwick
who helped me with the endless task of typing and proofing, to D.
Tranah, from Cambridge Universi ty Press , for making this project an
enjoyable one , and to my wife , Al ic ia , for he lp and support in many
oth er ways .
Amherst, Massachusetts
-
7/25/2019 Cambridge Texts in Applied Mathematics J. M. Ottino-The Kinematics of Mixing_ Stretching, Chaos, And Transport -
13/387
/
1
Introduction
1 .1 .
Physical picture
I n .;pite of its universality, mixing does not enjoy the reputation of being
:I
very scientific subject an d , generally sp ea kin g, mixing pro blem s in na tu re
;\nd technology are at tacked on a case-by-case basis. From a theoretical
viewpoint the ent i re problem appears to be complex and unwieldy and
thcre is no idealized sta rtin g picture fo r an aly sis; from a n app lied viewpoint
i t is
easy to get lost in the complexities of particular cases without ever
seeing the s tru ctu re of th e en tire subject .
Figure 1 . I
. I ,
which we will use repeatedly throughout this work,
describes the most important physics occurring during mixing. In the
simplest case, during mechanical mixing, an initially designated material
region of fluid stretches and folds throughout the space. This is indeed
thc goal of visualization experiments where a region of fluid marked by a
suitable tracer mo ves w ith the m ea n velocity of the fluid. Th is case is also
closcly approximated by the mixing of two fluids with similar properties
and no interfacial tension (in this case the interfaces are termed passive,
see Aref and Tryggvason (1984)) .
Ob viously, a n exac t description of the m ixing is given by th e location
of thc interfaces as a function of space and time. However, this level of
description is rare because the velocity fields usually found in mixing
Processes are complex. Moreover, relatively simple velocity fields can
Produce very efficient area generation in such a way that the combined
act ion of s t retching and folding produces exponent ial area growth.
Whereas this is a desirab le go al in ach ievin g efficient m ixing it als o implies
that initial er ro rs in the loca tion of the interface a re amplified exp onen tially
fast and numerical tracing becomes hopeless. More significantly. this is
"so
a
signature ofchao t ic f lows an d
i t
is impo rtan t to study the condi t ions
under
which they are produced (more rigorous definit ions of chcios are
glvcn
in Chapters
5
a n d
6).
However. without the action of molecular
diffusion, a n instan tane ous cut of th e fluids reveals a ltrti~c~llcirrrlrctlrrc.
(Figure
1.1.1) .
A measure of the state of mechanical mixing is given by
-
7/25/2019 Cambridge Texts in Applied Mathematics J. M. Ottino-The Kinematics of Mixing_ Stretching, Chaos, And Transport -
14/387
Figurc I . I . I . Basic processes oc curr in g dur in g mixing of f lu ids: ( 0 ) o r r e s p o n d s
to th e case of tw o simi lar fluids with negligible interfiicial tension an d negligible
interdiffusion: an in i t ia l ly designated mater ia l region s tre tches and folds by
the ac t ion o f
a
f low: ( h ) c o r r e s p o n d s t o
a
blob diffusing in the fluid: in this
case the boundaries become diffuse and the extent of the mixing is g iven by
level curves of concentra t ion ( a p ro fi le n o rm a l t o t h e s t r i a t i o n s is sh o wn a t
the r igh t ) : in
( c , )
the b lo b breaks d ue t o in te r fac ial tension fo rces , p roduc ing
smal le r f ragments which migh t in tu rn s t re tch and break produc ing smal le r
f r a g me n t s . C a se
( h )
s an exce l len t approx imat ion to
( t i )
if diffusion is small
dur ing the t ime of the s t re tch ing and fo ld ing . In
( t i )
the blob is pu.s.sii.e, in
( c )
the blob is trc,tii.c..
s t r ia t ion
th ickness
s t re tch ing
@
concen t ra t ion p rofi le
initial
condi t ion
d i f fus ion
-
i s tance
b r e a k u p
-
7/25/2019 Cambridge Texts in Applied Mathematics J. M. Ottino-The Kinematics of Mixing_ Stretching, Chaos, And Transport -
15/387
Scope trtltl eurljs works
3
~ l ~ i ~ . k n e s s e sf the layers, say s, a n d s,, a n d
:(s,
+
s,,) is called the
s l r . ; t ,~ ; o ,~li ick~ios.ssee O t t ino , Ranz , and M acosko , 1979) . The a m ou nt
of intcrfi~cial rea per unit volum e, interpreted as a struc ture d continu um
p rc )p H ty . s ca lle d t he i r l t c ' r r i ~ ~ ~ t ~ r i ~ ~ lIrett dcnsit~,,I,..T h u s ,
i f
S
designates
t1,c area within
a
volume
V
enclosing the point x at t ime
t ,
S
u, . (x ,
t ) = lim .
v - 0
v
~ ~ l l ~ c
f
the abo ve conc epts require m odif icat ion
i f
the fluids a re miscible
o r
immiscible.
I f
the fluids are immiscible, at some point in the mixing
process the st r iat ions o r blobs d o not remain connected an d break into
snlnllcr fra gm ents (F igu re 1.1.1
( c ) ) .
At these length scales the interfaces
arc not prrs,sir.c~ n d inste ad of being convected (pa ssiv ely ) by t he flow ,
they nlodify the surr ou nd in g flow, m akin g the analysis considerably m ore
complicated (in this case the interfaces are termed uc.tioe, see Aref and
Tryg gvu son, 198 4). If the fluids a re m iscible we can still track ma terial
volumes in term s of a (h yp oth etic al) non-diffusive tracer w hich m oves
with the mean mass velocity of the fluid or any other suitable reference
velocity. Designated surfaces of the trac er remain connec ted a n d diffusing
species traverse them in both direc tions . ' Ho we ver , du rin g the mixing
process, connected iso-concentration surfaces might break an d cuts might
rcvcal i slands ra ther than s t r ia t ions (Figu re I . l . l ( h ) ) . In th is case the
specification of the concentration fields of 11-
species constitutes a
com plete desc ription of mixing b ut it is als o clear tha t this is a n elusive go al.
Thus ,
i t
is apparent that the goal of mixing is reduction of length scales
(thinning of material volumes an d dispersion th ro ug ho ut space, possibly
involving br ea ku p) , a n d in th e cas e of m iscible fluids, uniformity of
con cen tration.' W ith this a s a basis, we discuss a few of the idea s used
to describe mixing and then move to examples before returning to the
problem formulation in Section 1.4.
1.2. Scope and early works
A cursory ex am inat ion of a n eclectic an d fairly arbi t ra ry l ist ing of som e
of the earl iest references in the l i terature gives an idea of the scope of
mixing processes and the ways in which mixing problems have been
attacked in the
pas t . "
I .
Tn>l( ) r ( 1 0 3 4 )The
f o r m a t ~ o n f
c m u l s ~ o n s
n
definable fields of
f l ow .
P r o < , .
HI,\,.
Sot... A146.
501-23.
A .
Brc) t l lm: l l l .
-
7/25/2019 Cambridge Texts in Applied Mathematics J. M. Ottino-The Kinematics of Mixing_ Stretching, Chaos, And Transport -
16/387
W . R . Hawtho rne , D. S . Wende l l, and H .
C .
Hot tc l ( 1048 ) M ix ing and combus t i on
in tu rbule nt gas jets. p. 26 6- 88 in ' I 'h~ r t l ymp. on C 'o t~~hu. \ r io t~t~clFltrtnc, uttd E.uplosion
Pltc~t~otnc~t~u.
al t imore: Wil l iams & Wikcns .
C . Eckart (1948) An analys is of the s t i r r ing a n d m ~ x i n g rocesses in incompressible
fluids,
. I .
M u r i t l r
Kc,.\.,
VII,
265-75 .
R .
S . Spencer and
R .
M . Wiley (1951 ) T he mix ing of very v i scous l iqu ids . .I . C o l l .
& i . .
6 ,
133-45.
P.
V .
Dan ckwe r t s (1 952) T he defin it ion and measurement of so me charac te r is t i c s
of mixtures , Appl. S ( , i . Ros., A.3, 279 96.
P.
V . Dan ckwe r t s (1953 ) Co nt i nu ou s f low sys tems-d is t r ibu tion of residence t imes ,
Ch[ ,m. Eng. S ( , i . .
2 , 1 13.
P.
Weland cr (1955 ) Studies on the general development of mo tion in a two-dim ensional ,
ideal fluid,
Tollu.s,
7 . 141-56.
S. Corr sin (195 7) Simple theor y of a n idealized turbu lent mixer. A . 1 . C h . E . J . ,
3.
329-30.
W .
D.
M o h r , R . L. S a x t o n , a n d
C.
H . Jcpson (1957) Mixing in laminar f low systems,
I n d . E n g . C h r m . , 49. 1855-57.
T h .
N.
Zweiter ing (19 59) T h e degree of mixing in c ont inu ous f low systems C h r m .
Eng . Sc i . .
I I , 1-15.
It seems at first strange to start a discussion on mixing with Taylor 's
1934 paper. However, there are several reasons. The first one is that the
problem is of practical importance and was attacked with the best tools
of the t ime, b oth theoret ical ly a nd experimental ly. T he second on e is that
the natural extension of his ideas to mixing remain largely unfulfilled.
Taylor 's concerns a re ob viou s fro m the t i tle of the p ape r. He distil led the
essence of the pro blem , in general a complex o ne, and reduced th e question
to a
local
analysis : the deform ation, s tretching, and bre ak up of a droplet
in tw o prototypica l flows - plan ar hyperbolic flow a n d sim ple shear flow.4
Pres um ably , the long r ang e goal w as to m imic a com plex velocity field in
te rm s of po pu la tio ns of these tw o flows."his is sim ilar, in sp iri t, to the
approach adopted in this work, in two respects : ( i ) analysis of simple
building blocks which give useful powerful insights into the behavior
of
complex problems. and
( i i )
dec om po sition of a problem in local a n d
global
com pone nts ( this idea is reconsidered in Ch ap ter 9 ) .
B ro thman , W ollan, an d Feldman (1945) had m ore prac tical and
pressing needs in mind and tried to attack the problem of mixing in a
general and abstract way. They spoke of fluid deformation and [fluid]
rearrangement, and regarded mixing as a three-dimensional shuminB
process . They worked out probabil is tic argu m ents di rec ted p red or ni na nt l ~
to mixing in closed systems, such as stirred tanks, and obtained
kinetic
expressions for the creation of interfacial a rea . Ecka rt (19 48 ) had in mind
substantially larger length scales and started his analysis with the
con tinu um field eq ua tio ns an d calculated the 'mixing times'
of
thermal and
-
7/25/2019 Cambridge Texts in Applied Mathematics J. M. Ottino-The Kinematics of Mixing_ Stretching, Chaos, And Transport -
17/387
Scope untl earl), wo rks 5
s ; , l inc >POIS i n oceans wi thout resor t ing to any mechani s t i c descr ip t ion
o f
process. A conceptua l ly s imi la r p rob lem in t e rms of scales, mix ing
i n atmospheric
f lows , was addressed by W eland er
(1955) .
R e m a r k a b l y ,
he did
so
by
cons ider ing the poss ib i li ty of app ly ing Ha m i l to n ia n mechanics
to idea l flu ids an d s t ressed the need fo r s tudying the s t re t ch ing an d fo ld ing
of
elemen ts in the flow an d devised for mu las to follow th e process .
The gront I1 of a mater ia l l ine by fracta l const ruct ion is a l so expla ined in
h i s work as wel l as a t rea tment of mot ion of po in t vor t i ces f rom a
~ a m i l l o n i a n i ew p oin t. H e a l so p e rf o rm e d e x p e r im e n t s a n d o n e o f his
visual izat ions i s reproduced in Figure
1.2.1
.h
7 ' 1 1 ~r~tc .r ac rions ~ r t w r r nu r b u l r n c e a n d c h em i ca l
reactions
is o fu t mo s t
i m p o r t a l ~ c c n combus t i on . A l t hough t he approach t o t hese p rob l ems i n
thc
.iO's ; I I I C ~
60's wa s largely s ta t i s t ical th is wa s not the case an d
I t I S c o r n t o r t ~ l ~ qo k n o b t hn t In be l l t houqh t ou t expc r lmen t a l pape r s such
as th r or lr
h \
Hn \ $ t ho rn r . i l ' endc l l, an d Ho t t c l (1948) o n e f in ds d c s c r ~ p -
t l o n s c ~ r n p h a s i / ~ l ~ qhc qcomc. tr lca1 asp ects of the pr ob lem 7'0 qu ot e
I , '~gure
.:.I.
Reproduct ion of one of the earl?
mixing
exper iments of Welander
(10551: evolution of
an
i n ~ t i a l
ondition in a rotating flow.
He
used butanol
floated o n water and the
~ n ~ t i a l
ond i t ion
(squ are) was
ma d e of
methq l - r ed ;
unfortunatelq f e w
a d d i t ~ o n a l etai ls regarding
the experiment were giten in
t h e o r ~ g ~ n a l
aper.
-
7/25/2019 Cambridge Texts in Applied Mathematics J. M. Ottino-The Kinematics of Mixing_ Stretching, Chaos, And Transport -
18/387
6 Introduction
from their paper,
According t o the physical picture of the turb ulen t f lame. eddies . . . are being
draw n in from the sur round ing a tm osph ere an d being broken up in to part ic les
of various sizes
.
.
.
The total area of f lame envelope is many t imes the area
available in a diffusion flame but nevertheless the final intimate mixing of the
gas an d oxygen m ust o ccu r between eddies a s a result of mo lecu lar diffusion.'
Subseqently, statistical theory took over and the geometrical aspects of
the problem were somewhat lost. Of the many possible offsprings of the
statistical theory of turbulence to mixing (e.g., Batchelor, 1953) we might
mention the short but influential paper by Corrsin (1959) which proposed
a simple model to calculate the rate of decrease of concentration
fluctuations in an ideal, yet subsequently widely used mixer.' Another
important concept based on statistical reasoning is that of the mixing
length theory (Prandtl, 1925; Schlichting, 1955, Chap. XIX), which has
found application in an enormous range of problems ranging from
chemical engineering (Bird, Stewart, and Lightfoot, 1960) to astrophysics
(Chan and Sofia, 1987, Wallerstein, 1988). Even though the statistical
treatment does not lend itself easily to visualization, notable exceptions
exist and many of the early works focused on the stretching of material
lines and surfaces, for example, Batchelor (1952; also Corrsin, 1972)
analyzed the problem theoretically, whereas Corrsin and Karweit presented
experimental results (Corrsin and Karweit, 1969). Other work focused on
local deformation to capture the details of the turbulent motion at small
scales. For example, Townsend (195
)
performed an analysis of deformation
and diffusion of small heat spots in order to interpret experimental Eulerian
data in homogeneous decaying turbulence.
Concurrently, at the other end of the spectrum, the stretching of material
lines and volumes was also a concern in the mixing of liquids in low
Reynolds number flows. Spencer and Wiley (1951 focused on the mixing
of very viscous liquids and stressed the idea of being able to describe the
growth of interfacial area between two fluids and the need to relate the
results to the fluid mechanics. However, even though the mathematical
apparatus, largely developed in continuum mechanics, was already in
place for such a program. i t was not until much later that such develop-
ments took place and most of what followed from Spencer and wiley's
work was confined to deformation in shear flows. Two other points worth
mentioning from their work. which have a clear relationship with
dynamicol systems. are the identification of stretching and cutting or
folding as the primary mixing mechanism, the so-called 'baker's trans-
formation' characteristic of chaotic systems (see Chapter 5). and the idea
-
7/25/2019 Cambridge Texts in Applied Mathematics J. M. Ottino-The Kinematics of Mixing_ Stretching, Chaos, And Transport -
19/387
Scope trnd
e u r l y
works
7
of rcp resen t ing mix ing p rocesses in t e rms o f mat r ix t r ans fo rm at ions . wh ich
is
r el at ed t o m ap p i n g s ( see C h ap t e r 5 ) an d t ran si ti o n m a t ri c es
(see
p ,
164 n Reichl , 1980). A c lose ly re lated s tudy by M oh r . Sa x to n , an d
Jepson (195 7) a lso focused on v iscous l iqu ids in t he c on tex t of po lymer
p,.oces~il lghe y considered th e s t re tch ing of a f i lam ent of a f lu id in th e
bulk o f ano the r o ne in a shea r flow us ing s imp le a rg um en t s to acc oun t
for
the
Llscos i tyof the f lu ids bu t wi th out tak in g in to accou nt the in ter fac ial
t ens ion . T h e p rob lem is s imi la r in sp i r i t to the on e t r ea ted by Tay lo r
(193 4) an d m uc h wor k cou ld fo l low a lo ng these l ines . Nevertheless th is
simplif ied t reatm ent form s the basis for mos t of the subsequ ent d evelop men ts
in
the
mixing of viscous fluids."
~ t 1s p rob ab ly fai r to say tha t mo st of th e p rev ious w or ks ha ve th e
g e o m c t n c n l i n t e r p r e t a t ~ o ngiven In Fig ure 1 .1 .1 . Ho we ver , a point of
departu re fro m this p icture of s ign~fic ant onsequen ce in chemical engineer-
Ing took place w ~ t hhe papers by Zwei ter ing (1959) a nd Danckwer ts (1958) .
In th is c ase t h e ap p r o a ch b ecam e m o r e ' l u m p ed ' o r m ac r o sco p i c an d t h e
em p h as i s sh if ted t o c o n t i n u o u s f lo w sy s tem s an d t h e ch a r ac t e r ~ za t i o nof
m ~ x i ng y the t emp ora l d i s t r ibu t ion o f exit t imes .' '
Whereas the ob jec t ive of mos t o f the ab ov e wor ks w as t o re la te the f lu id
m ech an ic s t o t h e m i x ~ n g r so m e k n o w l ed g e of t h e p ro ce ss t o t h e o u t p u t .
Danc kwer t s (19 53) ocused p r imar i ly o n t he charac te r i za t ion of the mixed
state . 1.e .. he devised nu m be rs o r indices t o indic ate t o th e user h ow well
mlxed
a
system is (e .g . , ho w well mixe d is th e system of Fig ure 1 .3 .1?
F igu re 1 .3 .3?F igu re 1 .3 .4? ). Even thou gh we a r e go ing to say l it tl e a bo u t
' the measurement o f mix ing ' . th is i s p robab ly the p lace to s t ress our
opinion o n
a
few points: ( i ) the meas u re shou ld be se lected acco rd ing t o
t he specific ap p l i c a t ~ o n n d
i t
is fut ile to devise a s ingle me as ure t o cover
contingencies. a n d ( i i ) the measu re men t has t o be re la tab le to the f lu id
m e c h a n ~ c s .
Examplc'.s
( 1 )
T h e s t r i a t i o n t h ~ c k n e s s , . Fig ure 1 .1 .1 . is imp or ta nt in processes
in \ ,o lv ing d i ffus ing an d react ing f lu ids and represen ts t he d is tance tha t
t he m o lecu le s m u s t d if fu se in o r d e r t o r e ac t w ~ t h ach o the r . In s imp le
Gases. .s can be calculated exact ly with a knowledge of the veloci ty
field ( s ee C h e ll a a n d O t t i n o ( 1 9 8 5 a ) an d ex am p l e s In C h a p t e r 4 ) .
t i i )
Mo lten po lymers ar e o ften mixed ( a n oper;at ion of ten refer red to a s
b lend ing ) t o p ro d uce mate r ia ls with ~ ~ n ~ q u eropert i e s. F o r example .
In th e mnnuf;acture o f barr ier pol ym er f i lms ~t might be desir able to
Produce s t ruc t l l rcs wi th ef fec tive permeabi l i ty . Th is requ i res tha t
-
7/25/2019 Cambridge Texts in Applied Mathematics J. M. Ottino-The Kinematics of Mixing_ Stretching, Chaos, And Transport -
20/387
Introduction
the clusters of the m ore perm eable ma terials a re disconnected and d o
not form a percolating structure. However, the details of effective
diffusion near the percolation point depend on the ramification of
the clusters (Sevick, M on so n, and O ttino , 1988). Even though such
measurements can be extracted from electron micrographs (eeg.,
Figure 1.3.4, see color plates) via digital image analysis (Sax and
O ttino , 1985), to da te there are no models allowing the com putation
of such details from the fluid mechanics of the process.
( i i i ) A model for the structure of the Earth's upper mantle (Allegre and
Tur co tte , 1986) postulates that the oceanic crust becomes entrained
in the convective mantle where it is subsequently stretched into
filaments by buoyancy induced motions. It follows that, on the
average, the 'oldest' layers are the thinnest and that the diffusion
processes concurrent with the stretching become important over
geological time scales when the striation thickness is of the order of
0.1-1 m (the typical diffusion coefficients are of the order 10-14-
1 0 - l6 cm 2/s, which implies diffusion o n time scales of the order of
10'6-1020 S. By com pariso n th e time scale based on the age of the
E arth is 1.4
x 10'' s) . In this case a model describing the entrainment
of material in the convective mantle coupled to a model describing
the evolution of the striation thickness as function of time is capable
of describing the gross characteristics of the process.
(iv) Co nsider Eulerian conc entration measurements in a tu rbulent m ixing
layer. In principle, the fluctuations can be taken as an indication of
the mixing between the streams and an index such as Danckwerts's
intensity of se gregation 2 can b e co m pu ted . In the ideal case of a
non-invasive probe with an infinitely fast response and vanishingly
small resolution volume we obtain an indication of the thickness of
the striations passing by the point as a function of time. However,
even
i f
this were possible, the statistics o f the fluctu ation s would be
very complicated and hard to connect to the fluid mechanics of the
proces s itself. W ha t is worse, how ever, is th e inability of th e measure-
ments to give a correct global picture of large scale structures, the
so-called coherent structures (Roshko, 1976). In this case a flow
visualization study base d, for exam ple, on shadowgraphs is in f in te l~
more revealing with regard to the structure of the flow (Brown
and Ro shk o, 1974) (see also Figure 1.3.5; see color plates ).
-
7/25/2019 Cambridge Texts in Applied Mathematics J. M. Ottino-The Kinematics of Mixing_ Stretching, Chaos, And Transport -
21/387
A p p l i c t ~ t i o n s
ncl
yeometr ic .n l s t ruc ture
9
1.3. Applications and geometrical structure
I t Is
clear that even restr ict ing our at tention to f luid f luid systems."
miscible o r ~ m m i s c i b l e ,d i f fus ive or non-di f fus ive . reac t ing or not , the
scope
of mixing p roblem s
IS
en or mo us and it is no t possibl e t o deve lop a
complete and useful theory en com pas sing a l l th e ab o v e s i tua t ion s . It is
nevertheless
ev iden t t ha t . in sp i te o f t he eno rm ou s r ange o f l ength an d
t ime scales . the under lying geometr ical s t ructure associated wi th the
process o f red uc tio n of length scales is th at of F igu re 1.1 . I . In this sect ion
we highlight this aspect by means of a few examples.
Mixing is relevant in processes ranging from geological length scales
( ] O h
m ) and exceeding ly l ow Reyno lds num ber s
(10 -20) .
such as in the
mix ing p rocesses occur r ing in t he Ea r th ' s m an t l e , t o Rey no ld s n um be r o f
o rde r 10" co r r e spond ing to mix ing in oceans an d the a tm osph ere . ' * An
example of a s imula t ion of mixing in the E ar th ' s ma nt le is show n in F igure
1.3.1. in which the f low is model led as a two-dimensional layer heated
f rom be low.13 In these cases , ac tua l mixing exper iments a re of course
impossible . However , laboratory models of large scale c i rculat ion in
oceans can be carr ied ou t w i th l iquids in con ta iners p laced o n a ro ta t ing
turn table somet imes involv ing co mb ina t io ns of sources and s inks . F igure
1 .2 .1 , f rom the ear ly pa per by We lander (1955) . sho w s an exam ple ( it is
wor th not ing tha t the output of s imi la r exper iments , were descr ibed as
' ch ao t~c ' , .g. , Veronis , 1973) .
Undoub ted ly .
c.htlotic
i s an apt descr ip t ion of s t re tch ing in t ru ly
turbulent f lows. Figure 1.3.2 shows the s t re tching of mater ia l l ines in a
turbulen t , near ly i so tropic , f low ( Co rrs in an d Karw ei t . 1969) , where th e
e x ~ e c t a t i o n ' ~s tha t of expo nent ia l gro wt h (Batch e lor , 1952) . N ote ,
howevcr . the inherent l imi ta t ion of exper imental techniques in resolving
the smallest scales.
The deform at ion of mater ial l ines has been s tudied a l so in the case o f
chaot ic Stokes ' s f lows . Exper im ents focus ing o n d efo rm at io n of mater ia l
l ines were carr ied out by Chaiken
et t r l .
(1986) an d Chien , Ris ing , and
Ot t i n o ( 1 9 8 6 ). In t h e c as e o f Ch a l k e n
pt (11.
the f low consis ts in a n eccentr ic
Journal bear ing t ime-per iodic tw o-d ime nsio nal f low which is descr ibed in
det ai l in C ha p t e r 7 . F igure 1 .3.3 sho ws the shap e adop ted by a ma te r i a l
line of a t racer by the per iodic d i scont in uous op era t io n of the inn er and
outer cylinders in
;I
coun ter - ro ta t ing sense . N ot e the absence of ' corners '
"d 'branc hes ' in the folded str uc tur e even tho ug h th e f low is essential ly
d i s c o n t i n l l ~ u ~co m pa re wi th Figure 1 .2 .1
).
Figures 1
. ?. I a n d 1 .3 .3 sho w com plex s t re tched a n d fo lded s t ruc tures .
character is t ic
c,f
mixing in two-dimensional f lows. In both cases the
-
7/25/2019 Cambridge Texts in Applied Mathematics J. M. Ottino-The Kinematics of Mixing_ Stretching, Chaos, And Transport -
22/387
10 Introduction
structure formed is lamellar and an indication of the state of mixing is
provided by the striation thickness. How ever, in ot he r cases the structu res
obtained are considerable more complex. If the fluids are immiscible and
sufficiently different, interfacial tension plays a dominant role at small
Flgure 1.3.1. Deformation of a tracer In a numerical experiment of motion In
the Earth mantle The sides of the rectangle are Insulating but the bottom is
subjected to a constant heat flux while the temperature of the top surface is
kept consta nt. Th e motion is produced by buoyancy and internal heating
effects (th e fluid is heate d half from below an d half from w ithin) . Th e Rayleigh
number is 1.4
x
lo h , the time scale of the num erical sim ulation corre spond s
to 155 M year, a nd the thickness of the layer is 700 km . An instantan eous
plcture of the strea ml ~n eseveals five cells. (R eprod uced with pe rm ~s slo n rom
Hoffman and McKenzle (1985).)
-
7/25/2019 Cambridge Texts in Applied Mathematics J. M. Ottino-The Kinematics of Mixing_ Stretching, Chaos, And Transport -
23/387
Applic3utions ~r nd eometricul structure 1 1
k.jfurc
I .3 .2 .
Cirowth of
a
'mater ial l ine ' composed of smal l hydrogen bubbles
produced
by
a
plat inum wirc s t re tched across an decaying turbulent f low
behind
a g r ~ d l aced a t the ex tr eme l ef t. Th e R eynolds number based on the
diam cter is
1.360.
(Reproduced with permission f rom Corrs in a nd Karweit
( I 9 6 l l J . )
Figure
1.3.3.
Mixing in a creeping flow. Th e l igure sho ws th e deformat ion of
a material region in a journ al bearin g flow w hen it op era tes in a t ime-periodic
hishion lexpcriment from Ch aiken rt r r l . (1986)), for a complete desc r ip t ion see
Scct ion 7 .4 . (Re prod uce d wi th permission. )
-
7/25/2019 Cambridge Texts in Applied Mathematics J. M. Ottino-The Kinematics of Mixing_ Stretching, Chaos, And Transport -
24/387
12 Introduction
scales. Figure 1.3.4 (see color plates) shows an image processed two-
dimensional s t ru ctu re produ ced by mixing an d preserved by quenching
o f w o immiscible molte n polyme rs. In this case there has been a com plex
process of breakup and coalescence. As indicated earl ier, the character-
Figure 1.3.6. C on ce ntr atio n of a turb ulent round jet fluid injected into water
at Reynolds number
2,300,
measured by laser induced fluorescence; the cut
is alo ng a plan e including the axis of a sym metry of the jet. (Re produ ced with
permission from Dimotakis. M iake-Lye, an d Papa ntoniou (1983).)
-
7/25/2019 Cambridge Texts in Applied Mathematics J. M. Ottino-The Kinematics of Mixing_ Stretching, Chaos, And Transport -
25/387
Approuch 13
iwtion of this structure depe nds on the intended appl icat ion of the blend
and
a
large nu m be r of mixing measures ar e possible."
interp lay betwecn chemical reactions a nd mixing is now here mo re
ev id en t than in the case of fast reactions (C h a p te r 9) . In m any cases of
i n t e r e s t the flows are turbulent and careful studies have been carried out
to probe the interplay between the f luid mechanics and the t ransport
processes a nd by using prototy pical flows such as tur bu len t sh ea r
flows
an d wakes, perturbed o r not . Th e react ion itself can be used t o m a p
ou t the interface of reaction. F or ex am ple, Koochesfahani a nd Dim otakis
(1986; see also 1985) have used laser induced fluorescence an d high speed
real-time digital im age acq uisition techn iques to visualize the interface of
reaction between two reacting liquids undergoing a diffusion controlled
reaction (see also Ch ap te r 9) . A similar technique c an be used in th e case
of diffusing scalar. Fo r ex am ple, F igure 1.3.5 (see colo r pla tes) was
obtained by measuring the concentration of a fluorescent dye, initially
located in one of the streams of the mixing layer, whereas Figure 1.3.6
shows the mixing of turbu len t jet c on tain ing a fluorescent dy e with a clear
surroun ding f luid. Lam ellar s t ructure s (C ha pte r 9 ) are clearly seen, even
at Kolmogorov length scales (D imo takis , Miake-Lye, an d Pa pan toniou ,
1983).
Ano ther instance of interplay between mixing w ith diffusion a n d
reaction. but at smaller length scales and lower Reynolds numbers
(approximately 200 5 0 0 ) , well kn ow n in polymer engineering, is provided
by the impingement mixing of polymers (Lee
et
al.,
1980)
w h e r e t h e
objective is to mix tw o viscous l iquids (reactive m on om er s) in s hort t ime
scales (o rder 1 0 ~ 2 1 0 - 1) with reaction t ime scales of the o rde r of
10 '~-1o 2 . T he geom etrical picture is similar to the previous cases; in
th is case the mixing requi rement i s to produce s t r ia t ions of the order
of 20-501tm so that the reaction can take place under kinetically
condi t ions. 'h Mixing at even smaller scales might take place
d u e to spontaneous emulsi ficat ion (Fields , Th om as , and Ot t in o , 1987;
W icker t , Macosko , a nd Ranz , 1987).
1.4.
Approach
In spite of its overwhelming diversity, fluid mixing is basically a process
l n ~ o lv i n g reduction of length scales accomplished by st retching an d
folding of material l ines or surfaces. In some cases the material surface
O r line in question is placed in the flow and then subseqently stretched
le.%-
igures 1 2 .1 , 1.3. 3). in others , the surface is continuously fed into
'he flow (cg.. Figures 1.3.5
6) .
Thus. at the most elementary level (i .e. .
-
7/25/2019 Cambridge Texts in Applied Mathematics J. M. Ottino-The Kinematics of Mixing_ Stretching, Chaos, And Transport -
26/387
witho ut avera ging but at the c on tinu um level) rnising c.onsists of' stretching
lrr~d fbldii~~gf ,fluid filtrmc~nts,rrld distribution tllrou~ghoutspuce, accom.
panied by breakup
i f
the fluids ar e sufficiently different, an d sim ultan eou s
diffusion of species and energy (Figure 1.1.1
) .
In the most general case,
various chemical species might be reacting. We seek an understanding of
this process in terms of simple problems which can serve as a 'window'
for more complicated si tuat ions. Our approach is to combine the kine-
matical fou nda tions of fluid mec hanics with dynamical systems concepts,
especially chaotic dynamics. The objective throughout is to gain insight
into the working of mixing flows. The goal is not to construct detailed
mode ls of specific prob lem s but rathe r to p rovide pro totyp es f or a broad
class of problems. Nevertheless. we expect that the insight gained by the
analysis will be important in practical applications such as the design of
mixing devices and understanding of mixing experiments.
Mixing is also inherently related to flow visualization. However,
contrary to popular percept ion the 'unprocessed ' Eulerian velocity field
gives very little information about mixing and the typical ways of
visualizing a flow (streamlines. pathlines, an d to a lesser degree. streaklines)
ar e insufficient to com pletely u nde rstan d the process. As we shall see, our
problem begirls ra th er rharl ends with the specificurion of v (x , r). T h e so lut ion
of
dxidr =
V ( X , I )
with x
=
X at t ime
t
=
0.
x
=
@ , ( X ) ,which is called the flohv or motion,"
provides th e start in g point for ou r analysis. In even the simplest cases this
'solution' might be extremely ha rd to ob tai n. Actually. the impossibili ty of
integrating the velocity field in the conventional sense is the subject of
much of Chapters 5 a n d
6,
where the modern notion of irlteyrability is
introduced. The kinematical foundat ions l ie in an understanding of the
point t ransformation x = @,(X) .
We consider the following sub-problem s:
( 1 ) Within the framework of x
=
@ , ( X ) :mixing o f a single fluid o r similar
fluids.
Th e basic objective here is to c om pu te the length (o r ar ea ) corresponding
to a set of initial conditions. As we shall see only in a few cases can this
be done exactly and in most of these the length stretch is mild. The best
achievable mixing c orresp onds to e xpo nential stretching nearly everywhere
and occurs in some regions of chaotic flows. However. under these
con dit ions the (exa ct) calculation of the length a n d loca tion of lines an d
are as is hopelessly com plic ate d. As we shall see in C h a p te r
5
even extremely
-
7/25/2019 Cambridge Texts in Applied Mathematics J. M. Ottino-The Kinematics of Mixing_ Stretching, Chaos, And Transport -
27/387
simpliljcd
~ O R S
ight be inherently chaotic a n d a
corripleie
char ;~cter izat ion
is n o t
p o ~ i b l c . o r e x am p l e . f r o m a d y n a m i ca l s y st em s vi ewp oin t Re sh all
, ,,that i f th e system possesses horse shoe s we ha ve infinitely m a n y perio dic
and wi th i t the impl ica t ion tha t we cannot poss ib ly ca lcula te
;dl of t h e m For tuna t e ly . a s f a r
;IS
mixing i s concerned we are
intercstcd in low period events, s ince we want to achieve mixing quickly.
~ ~ ~ ~ ~ t h c l e ~ ~here is a lways the int r ins ic l imi ta t ion of being unable to
ca]culatc p ~ c i s enform at ion (m os t prac t ica l p roblem s involve s t re tch ings
of
10'
o r h ighe r ) such a s l eng th s t r et ch an d loca t ion of ma te r i a l
surf ;lccs, No te th at this is t ru e even th ou gh no ne o f the f lows discussed
in Chapters 7 a n d 8 i s turbulent in an Euler ian sense . Rather , the
previous f indings should be used t o es tabl ish the l imits o f wh at might
const i tute reasonable answers in more complicated f lows ( real turbulent
f lows come
immediately
t o m i n d ) .
T h e
prohlcm /irr.cl is /io\c, t o hcsr
~ ~ / 1 1 1 1 . 1 1 c ~ r c ~ 1 ~ ; : 1 ~/ ] o i ~ i - ~ i i ~ q ,t~o\vi t~qgf:)rP/lut1d /iiit u c o t ~ i p l ~ i ~~ / ~ ~ ~ r ~ ~ ~ ~
i,s ; l ~ l p o , ~ , ~ ; l ~ l ~ ~ .
( 2 )
Within the f ramework of
a
family of flows
x,
=
@ , , ( X , ) ,
s
=
1 ,
. .
. .
,1 ;
each of the mot i ons is assu me d t o be topo logical (see Sect ion
2.3) :
mixing
of similar diffusing a n d reactin g f luids.
This case cor respo nds to th e case of mixing of tw o s t ream s. com posed
ofpossibly several species th at ar e rheological ly ident ical , i .e . , they have th e
same dens i ty . v i scos i t y . e t c . , and have no in t e r f ac i a l t ens ion . Concur -
rently wi th th e mecha nica l mixing th er e i s ma ss d i ffus ion , an d poss ib ly .
chemical r eac t ion . H ow ev er . fo r simpli c it y , w e will a s sum e tha t ne i th e r
t h e d i f f u s i o n n o r t h e r e a c t i o n a f f e c t s t h e f l u i d m o t i o n . ' Th i s c a s e i s
d iscussed in C h ap te r 9 a n d c o r r e s p o n d s t o t h e c a s e of
lumellur structures.
( 3 )
Mixing of d i f fe rent f lu ids ; case in w hich the mot io ns a r e n on -
topologica l. i , e , . there is b re ak up a n d o r fus ion of m ater ia l e lemen ts .
In t h i s ca se . t he mix ing o f two o r more f l u ids l eads t o b reakup and
coalescence of mater ia l regions . This problem is complicated and only a
few special cr lses belonging to this category are discussed in Chapter
9.
by d e c o n ~ p o s i n g he p rob lem in to
I o ~ o l
n d
qlobiil
c o m p o n e n t s .
An out l ine of the orga niza t ion o f the res t o f t h e chapte rs i s the fo l lowing:
2
descr ibes t he k inemati cal fo und a t ion s an d C ha p t e r 3 presents
a
b r i e f n i c r v i e , o f flu id m e c h an ic s. W h e r e ; ~ s h e m a te ri al of Ch a p t e r
2
is
' n d i s ~ e n ~ i i b l e .;~rge p;lrts o f C hi lpt er
1
were ; ~ d d e d o provid e b ;l l;~ncc .
4
focuses on
;l
f e ~x; lmp]es which can be solved in deta i l and
ends
in
;I
riit11cr dcfc; lt is t no te t o pr ov ide a bridg e for th e st ud y o f ch ao s.
5 iwcrcnts ;I pener.11 d iscussion o f d y n i~ m ic ;~ lystems ; ~ n d h i ~ p t ~ r
-
7/25/2019 Cambridge Texts in Applied Mathematics J. M. Ottino-The Kinematics of Mixing_ Stretching, Chaos, And Transport -
28/387
16
Introduction
6
focuses on Hamiltonian systems. Similar comments apply in this case.
We use more heavily the 1,laterial of Cha pte r 6 but omission of Ch apter 5
would result in serious imbalance and a misleading rep resen tation of facts.
Chapters
7
and
8,
by far the longest in this work, give examples of chaotic
mixing systems in a n increasing orde r of complexity. C ha pte r 7 discusses
two-dimensional flows, Chapter 8 focuses on three-dimensional flows.
Chapter
9
discusses briefly the case of diffusing and reacting fluids and
active microstructures.
Notes
1 Obviously, on e o f the diffusing species can be tem pera ture, as is the case of mixing
of fluids with different initial tem pera tures o r processes involving exothe rm ic chemical
reactions.
2 A gross, but po pu lar , me asure of the concentration variation is given by th e intensity of
srqreqution, I . If ~ ( x )enotes the concentration at point x and ( . ) denotes a volume
average,
I
is defined as
[(((.-
( c ) ) ~ ) ] ' / '
(Danckwerts, 1952).
3 M any ofthese references inspired addition al wo rk. So me , however, were largely ignored.
4 Subsequently, this problem took a life of its own and much research followed. See for
example Rallison (1984).
5 Th is idea was not widely followed and m ost of the mixing work in the ar ea of d ro p breakup
an d coalescence in complex flow fields resorts to pop ulation balances where break up and
coalescence are taken into account in a probabilistic sense.
6 This paper contains many good ideas, however, it has remained largely ignored
by the mixing community.
7 In the past few years there has been a revival of this idea (e.g., Spaldin g, 1976, 1978b;
Ot tino , 1982).
8 The theory also found use in two-phase mixing. An early reference is Shinnar (1961).
9 The analysis of mixing of viscous fluids has been largely confined to polymers
(Midd leman , 1977; Ta dm or and Gog os, 1979).A follow up pape r, written in the context
of the m ixing of glasses. is C oo pe r (196 6). Even tho ugh C oop er's trea tment of the
kinematics is at the same level as Mohr, Saxton, and Jepson (1957), there is
substantially m or e, since it deals explicitly w ith mass diffusion. This pa pe r, how ever,
has remain ed largely ignore d.
10 This app ro ach wo rks well for pre-mixed reactors with slow reactions, but it is not suited
for diffusion controlled reac ions, such a s in com bus tion. Curiously enough, the
par ticipation of chemical engineers in subjects dealing with non-pre-m ixed reac tors has
been relatively minor a nd in spite of comp lex chem istry , the are a has become largely the
do m ain of mechanical a nd aerospa ce engineering researchers. See the discussion ( pp. 100-
102) following the paper by Dan ckw erts (1958 ). Mu ch w ork followed alo ng these
lines. For a su mm ary, see Nau m an an d Buffham (1983).
I I
W ith the possible exception of the very last exa mple of the last cha pter . we do no t consider
fluid-solid systems.
12 See for example Veronis, 1973: Rhines. 1979; 1983. and the articles by Holland
('Ocean circulation models' . pp. 3 4 5 ) . Veronis ( 'The use of tracer in circulation
studies '. pp. 1 6 91 8 8) , an d Rhines ( 'The dynam ics of unsteady currents ' .
pp. 18%318), in Goldberg er
a l . ,
1977.
-
7/25/2019 Cambridge Texts in Applied Mathematics J. M. Ottino-The Kinematics of Mixing_ Stretching, Chaos, And Transport -
29/387
-
7/25/2019 Cambridge Texts in Applied Mathematics J. M. Ottino-The Kinematics of Mixing_ Stretching, Chaos, And Transport -
30/387
Flow , trajectories, and deformation
In the first part of this ch ap ter we record the basic kinem atical fou nda tions
of fluid mechanics, starting with the primitive concept of particle and
motion, and the classical ways of visualizing a flow. In the second part
we give the bas ic e quat io ns for th e d eform at ion of inf ini tes imal mater ia l
l ines, planes, and volumes, both with respect to spatial , x , and material ,
X, variables, and present equations for deformation of l ines and surfaces
of finite extent.
2.1. Flow
T h e physical idea of
flow
is represented by the m a p or point t ran sforma tion
(Arno ld, 1985, C ha p.
1
)
x
= @ ,(X) with X = @,=,(X), (2.1.1)
i.e., th e initial con diti on of particle X ( a m ean s of identifying a point in
a continuum, in this case labelled by its init ial position vector) occupies
the posi tion x at t ime t (see F ig ur e 2.1 .I ). W e say tha t X is mapped to x
after a time
t . '
In con tinuu m mechanics (2.1 I ) is called the
motion
and is usually
assumed to be invertible an d differentiable. In th e lang uag e of dyn am ical
systems a map ping
@,(XI
+
x
(2.1.2)
is called a
Ck
diffeomorphism
i f
it is
1-1
a n d o n t o , a n d b o th
a,(.)
n d
its inverse are k-times differentiable. If k
=0
the transformation is called
a homeomorphism. In fluid mechanics, k is usually taken equal to three
(see Truesdell , 1954; Serrin, 1959). Also, the transf orm ation (2.1.1) is
required to satisfy
or alternatively,
J = d e t ( D @ , ( X ) )
-
7/25/2019 Cambridge Texts in Applied Mathematics J. M. Ottino-The Kinematics of Mixing_ Stretching, Chaos, And Transport -
31/387
whert D d e n o t e s t h e o p e r a t io n i, ),/ax,,. e . , de r iva tiveswi th r e spec t t o t h e
configurat ion, in this case X. I f the Jacobian
J i s
equal to one
flow is called
isochoric.
~ h cequirement (2.1.3) precludes two par t ic les , X , an d X,, f rom
occup) , ing the sam e posi t ion
x
at a given t ime, or one part icle spl i t t ing
into ~LI 'O :.e. ,
notz-topologicul
mo t ions such as b reak up o r coa lescence a re
not a l lowed (Truesdell an d T ou pin , 1960, p .
510).
I n
the lang ua ge of dynamical systems, the set of diffeomorph isms (2.1.1
for all particles X be lon gin g to th e bo dy Vo is called the flow (i.e.,
a
one-p aram eter se t of di f feomo rphism s) an d is represented by
r 1 - 1
, X I
-
@ , (X I ; = @ , ( X i
where X ) is the set of par t ic les belonging to V (Vo
=
{ X } a n d
V, =
( x ) ) ,
Thus. we say that V,, is m ap pe d in to Vr at tim e t ,
f V 1
- @
rv
1
i ri -
r t
01
or that the materia l l ine L o is mapp ed into L, a t t ime
t ,
( L , ) = @ I L o : .
Figure 2.1.1. Deformation of l ines and volumes by a flow x
= Q , ( X ) .
motion
-
-
7/25/2019 Cambridge Texts in Applied Mathematics J. M. Ottino-The Kinematics of Mixing_ Stretching, Chaos, And Transport -
32/387
20
F l o w , trujrc.tories ur~ tl lt$)rmtrtior~
Note that f lows can be composed according to
@,
+, (XI
= @,(@AX)) ,
i.e ., X is taken to pos i t ion @ ,(X) an d then to @ ,+, (X ) .T he f low can also
be reversed,
@ ,-,(X I = @ , ( @ - , ( X ) ) = @ l ( @ l - l ( x ) ) x
i .e. , X is taken to x and then back to X. '
2.2. Velocity, acceleration, Lagrangian and Eulerian viewpoints
The velocity is defined as
a n d it is the velocity of the particle X . T he a cce lera tion , a , is defined as
a
-
c ' . ' @ , ( ~ ) / i ? t ' ) I ~a ( X , t ).
Any fun ction G (scala r, vector, ten so r) ca n therefore be viewed in two
different ways:
G(X,
t ) =
Lagrangian or mater ia l
i.e., follows the motion of a particular fluid particle, or
G(x , t ) = ~ u l e r i a n % r sp atial
i .e. , the pro per ty of the particle X th at hap pen s t o be at the spatial location
x at t ime t .
Thus , v(X, t ) is the Lagrangian velocity and v(x, t14 is the Eulerian
velocity. In most classical problems in fluid mechanics it is enough
to
ob tain the spat ial desc ript ion. Th e material (o r Lag ran gia n) derivative
is
defined as
DGIDt
-
('G/?t)1,
representing the cha ng e of pro perty G with t ime while following the mo tion
of particle X, whereas the standard time derivative is
?G/?t = (?G/?t)lx,
an d represents the chang e at a fixed position x. T h e relationship between
the two is easily obtained from the chain rule
DG/Dt = ? G / i t
+
V - V G ,
wh ere V is defined a s V =
(?/?.u,)e,
(see App endix). Th e expression allows
the comp uta t ion of the accelerat ion a t (x ,
t )
without compu t ing the mot ion
first.
T h e expression for the m aterial t ime derivat ive of the Jacob ian o f the
flow is known as Euler 's formula (Serrin, 1959,
p.
131
;
Chadwick,
1976,
p.65).
DJIDt = J ( V . v ) = J t r (Vv)
and is the basis of a kinematical result known as the trtrnsport theorem.
-
7/25/2019 Cambridge Texts in Applied Mathematics J. M. Ottino-The Kinematics of Mixing_ Stretching, Chaos, And Transport -
33/387
Veloc,ity,acceleration, Layrangian and Eulerian
consider the integral
1,ix,
t )
d i
where V, represents a m aterial volum e, tha t is. a vo lum e com pose d alw ays
of the sam e particles X belonging to the b ody Vf. M ak in g reference t o
Figure
2.1
I
:V , ]
= 1 : V,}, i.e. . the flow
0,
ransforms { V , } in to
{ V f }
a t
time t . Th e integral ca n be w rit ten, using the definition of the Ja co bia n, as
r P
G ( x ,
)
d c
=
J G ( X , t ) J d V
where d
V
represents a volum e in the reference conf iguration V,,. T he t ime
derivative.
(which is a material derivative since all the particles in
V,
remain there)
can be written as
and since the dom ain V is not a funct ion of t ime, exp and ing the mater ial
derivative we obtain
: I",
(x.
r )
d r
= jv
&
+
G( V.
v)]
dc .
By
means of the divergence theore m.
where n i s the outward normal to the boundary of
V, ,
deno ted ?v , . In
general this result holds fo r an y ar bit rar y co ntr ol volum es V; movin g with
velocity v , ,
Problem
2 . 2
.J
Show that
i f
A (r ) is invert ible, d(det A)/dr
=
idet A ) t r [ ( d A / d r ) ' A ' l .
Problem 2.2.2
that
i f
I
I ; /
=
e l [
~,]
hen
; i v , \
=
@ ,; iV , , ] , .e ., the b ounda ry is
mapped
into the hound;iry. Us ua]]y, this is taken to be that the surfiice
Of
a ma terial bo dy consists o f the siime p;irticles (v on Mises an d
-
7/25/2019 Cambridge Texts in Applied Mathematics J. M. Ottino-The Kinematics of Mixing_ Stretching, Chaos, And Transport -
34/387
22 Flow , trajecto ries and defbrmution
Friedrichs, 1971). Th is result is called 'Lagrang e's theo rem ' by P ra nd tl
and Tietjens
( 1
934, p. 97 ).
2.3. Extension to multicomponent media
In th e case of multicom pone nt media we envision m aterial surfaces m oving
with the m ean ma ss velocity (see C ha pte r 9 ). If th e system h as several
components , s = 1 , . . . , N , we assume the existence of a set of motions
a?),
nd Equation (2.1 I ) is generalized as
X, =@?)(X,) (no su m )
where X, represents a particle of species-s and x, its position at time t.5
Each species is assigned a d ensity p,
=
p,(X,, t ) such that p
=
x ,, where
the sum runs from 1 to
N .
Individual
velocities
are defined as
v, - a@ls)(x,)~at)l ,= v,(x,, t )
and the average mass velocity is defined as
v
-
x
P, /P)~ , .
T h e time derivative of an y function G following the m otion of the species
denoted s, is given by
DG'"/Dt = aG/at + v;VG,
and the relative velocities are defined by
Us= v , - V .
The simplest constitutive equation for
us
s
us = - W, DVO,
(i.e. , dilute solution o r eq uim olecula r counter-diffusion, B ird, Stew art, and
Lightfoot, 1960, p. 502) where
w ,
is the mass fract ion (=p/p,) and D is
the diffusion coefficient.
O th er q uantities, such a s individual deform ation tensors for species-S,
etc., can be defined ana logou sly (B ow en, 1976) but they a re not used in
this work.
2.4. Classical means for visualization of flows
There are several ways of visualizing a flow. In this section we record the
three classical ones.
2 .4 .1 .
Particle path, orbit, or trajectory
Given the Eulerian velocity field v = v(x,
t ) ,
the particle path of X is given
by the solution of
dxldt = v(x, t ) with x =
X
a t t = 0. Physically it
-
7/25/2019 Cambridge Texts in Applied Mathematics J. M. Ottino-The Kinematics of Mixing_ Stretching, Chaos, And Transport -
35/387
Cla.ssicu1 meuns ,for visualization of Jows
23
orrcspo nds to a long time expo sure ph oto gra ph of a n il luminated f luid
p a r t i ~ l ~ .
AS
seen in texts of differential equa t ions, the solut ion t o the a bo ve
x = a , @ ) , is unique a nd co nt inuo us with respect to the in itial
data i f v (x ) has a Lipschi tz con stant ,
K
> 0.' Under these condi t ions, i f
we d en ote x l = @ , ( X I ) a n d x 2 = @ , ( X 2 ) ,we have the trajectories evolve
according to
I x -
x21 G IX1- X21 ex p( K t) ,
K > 0.
AS we shall see there ar e man y systems (C ha pte r
5 )
tha t diverge from the
init ial condit ions at an exponential rate, i .e. , the non-strict inequali ty
becomes a n eq uali ty. '
2 .4 .2 . Stveamlines
The st reamlines corresp ond to the solut ion of the system of equat ions
dxlds = v(x,
t )
where the time r i s t reated as constant and s is a parameter (that is, we
take
a
'picture' of the vector field
v
at t ime
t ) .
Physically, we can mimic
the streamlines by labelling a collection of fluid particles and taking two
successive photographs at t imes t and t + At. Joining the displacements
gives
v
in the neigh borho od of the point x. Th e st reamlines are tangential
to the instanta ne ous velocity a t every p oint , except a t po ints where v = 0.
2.4 .3 .
Stveaklines
The picture at t ime t of the streakline passing through the point x ' is the
curved formed by all the particles X which h app ene d to pass by x ' dur in g
the time 0 < t ' < t . Physically, i t corresponds to the curve traced out by
a non-diffusive tracer (i .e., the particles X of the tracer move according
to
x =
@ ,( X )) njected a t the posi tion x ' .
Example 2.4.1
Com pute the pathlines, s t reamlines, an d st reakl ines corresp onding t o the
unsteady Eulerian velocity field
v 2
=
1. 8
1 =
x l l ( l + t ) ,
To co mp ute the pathlines we solve
d.u,/dt = x 1 / ( l
+
t ) ,
dx2 /d t = 1
with the condition x,
=
X , ,
x, =
X,, at t =0.
.Y ,
= X l ( l
+
t ) ,
s, = X, +
t
and, el iminating t , we obtain
.Y1 - x 1 x 2= X 1 ( l - X2).
-
7/25/2019 Cambridge Texts in Applied Mathematics J. M. Ottino-The Kinematics of Mixing_ Stretching, Chaos, And Transport -
36/387
24
Flo w, tr~ljectories l n d eformrrtion
i .e. , the p articles mo ve in s traigh t l ines. T h e streamlines ar e given by the
solution of
d x , ~ d s
x , / ( l +
t ) , d x , / d s
= 1
with the condit ion
x,
=
x y ,
x,
=
x ; ,
a t
s
=
0,
while holding
r
cons tan t .
Th us, the s treamline passing by
x,
=
xy, x ,
= x;, is given by
x ,
=
xy
exp[s/( l +
t ) ] ,
0
x, = x, +
s,
and e l imina ting the pa ram ete r s,
0
( 1 +
t )
n ( x ,
x y )
=
x ,
- x,,
which show s tha t the s treamline s are
t ime dependent .
T o get the s treakline
pass ing through
x',, x;
we first invert the particle paths at t ime
t'
X I
= x i / ( l +
t ' ) ,
X ,
=
,u;
-
' ,
which indicates that the particlt
X , , X ,
will be foun d a t the p osition xi , x;
a t t ime t ' . T he place occupied by this par ticle at an y t ime t is found again
from the par t icle path as
x , =
x',( l +
t ) J ( l
+
t ' ) ,
x, = .u; -
'
+
t ,
and is interpreted as: the particle which occupied position
x;, x;
at t ime
t'
will be found in position x,,
x ,
at t ime
t .
Eliminating [ 'we get the locus
of the streakline passing by
x',,
x;:
x , x ,
-
x , ( l +
x ;
+
t )
+ x ' , ( I
+
t )
= 0
which shows that the s treaklines ure ulso ,functions
of'
t ime (plots
cor responding to th is example are given by Truesdell a nd To up in, 1960,
p .
333) .
2.5.
Steady and
periodic
flows
A flow is steady i f i t lacks explicit t ime dependence, i .e. , v = v( x ) . N o te
that the concept of steadiness depends on the frame of reference.
An
unsteady f low in o ne f ram e can be s teady in a no the r f rame (mo ving
f rames are s tudied in C ha pt er
3 ) .
W he n the flow is stead y in a given frame
F,
the stream lines an d pathlines coincide when viewed in the frame F .
Fu rthe rm ore , the s treaklines coincide with bo th s treamlines an d pathlines
provided th at th e position o f the dye-injection ap p ar at us is fixed with
respect to the frame
F.
I t is obvious that these statements are a property
of dynamical systems in general and they are not confined to f luid
mechanical systems.
A point x such tha t v(x)
=
0 for all
t
is called a
,fired or ,siri
-
7/25/2019 Cambridge Texts in Applied Mathematics J. M. Ottino-The Kinematics of Mixing_ Stretching, Chaos, And Transport -
37/387
Ste ad y and periodic f l o w s
point in fluid mechanics). A point P is periodic, of period
T,
i f
P = @,.(P)
for ,
= T
but not for any r
t l l
~ i h c r eG is an y sca la r , vec to r . o r t enso r func tion
3.2.
Principle of conservation of linear momentum
/rttcgrul cersion ( o r Euler's u x io m )
This
pr incip le s ta tes tha t fo r
a
m at e r i a l v o l u m e
V ,
l i n ea r m o m en t u m i s
conserved. i .e . .
rate of
c h a n g c
of forces
; ~ c t i n g
n
rnomcnturn
body
jb,
~ v d i .
= jL,
f d i .
+
4
b , ,
t d . s
(3.2.1 )
d t
;~ccclerat~on bod
Corccs
contact
Corcca
\ i llere t (n . x .
t )
is a s yet a n undefined vector cal led the
t r r r c l i o n
which
d ep en d s o n t h e p l acem en t x o n t h e b o u n d a r y ,
CV,,
the in s tan taneous
o r ~ c n t a t i o n .n, and t ime ,
t
(see F igure
3 . 2 . 1 ) .
The vec to r f i s the body
force . which in th is work is assu me d to be indepe nden t of the conf igu ra t ion
o f t l i e b o dy ,
V,.
Using the t r anspo r t theo rcm (Sec t ion
2 . 2 )
w e o b t a i n
k > r a n y e gio n
V, ,
a t a n y t im e ( w h e r e
V,
c a n be a n a r b it r a ry c o n t r o l v o l u m e ) .
Pr.ohlcm 3.2.1
f 'l.o \e tha t t he Pr inc ip le of Co ns er va t io n of L inear M om e nt um implies
the co n t inu um version of 'New ton ' s th i rd law ' : t ( x . n , 1 )
=
t ( x . n . t ) .
3.3 Traction t (n , x , r )
It
can be proved tha t t (n .x . I )
= T'
e n ,
where
T
is a tensor (n o te conv ent io n) .
T h ~ sm p li es t h a t t h c i n f o r m a t i o n ab o u t t r a c t io n s a t t h e p oi n t x a n d an y
' . l l~f i~cc i th no rm al n i s con ta ine d in th e tens or T. Th e proof cons is t s of
t ll rcc s t eps which \ye will briefly repeat here.
( i )
Cuuchy's theorcm
( 'ons ide r a ( smal l ) r eg ion
V,
su r round ing a pa r t i c l e
X .
T h e n , if L r e p r e -
h en ts so m e l en g t h s ca l e
o f V, w e
h a v e :
-
7/25/2019 Cambridge Texts in Applied Mathematics J. M. Ottino-The Kinematics of Mixing_ Stretching, Chaos, And Transport -
56/387
14
C'onserlution rquu tions, chun ge of ir u tn e , untl lwrtic.it.r
V o l u m e (
V,)
= c o n s t . ,
L"
Surface c ?
V , )=
const . , L2
T h e m e a n v a lu e t h e o re m s t a te s th a t f o r a n y c o n t i n u o u s f u n c ti o n G ( X ,
I )
defined over
C
a n d i?V, we can f ind X ' and X" (a t any t ime
t )
s u c h t h a t
I-
J,.
d o
=
c o n s t . ,
L-
G ( X f ,
) ,
where X ' be longs to V,
(; d s = con st . , L2 G(X", t ) ,
where X" be l ongs t o ?V,
F~gure .2.1
( t i )
Mate r ia l
region In
present configuration.
V,,
ind~cating o r m a l
n a n d tri tctlon I;
( h )
cons t ruc t ion for balancc of a n g u l a r momentum.
-
7/25/2019 Cambridge Texts in Applied Mathematics J. M. Ottino-The Kinematics of Mixing_ Stretching, Chaos, And Transport -
57/387
Truction t(n , x , t)
Applying the theorem to E qu at io n (3 .2 .2) we obta in
D
~r
( v ( X 1 , )
cons t . , L")= pf cons t . , L
Thcn, dividing by
L2
and le t t ing
L
go to zero (preserving geometr ica l
similarity) ,
Th at is, the tract io ns are locally in equil ibr ium (see Serr in , 1959, p. 134) .
( i i ) Traction on an arhitvary plane (C au ch y's tetvahedvon constvuction)
By me ans of this co nst ruc tio n, which we will not repeat here (see Se rrin ,
1959, p . 134, for de ta i ls ) , Cau chy was ab le to prove tha t the com po ne nts
of the traction t( n ), [t([n])] =
( r n l ,
rnZ, n3 ) , re re la ted to the n orm al n ,
I n ] =
( n l ,
,, n,) , by the ma trix mu ltip lica tion : '
I " '
= 7 ' . . ) 1 .
11
I
o r equiv alently [ t([n])]
=
[n][T], [t([n])]
=
[T1][n]
whcrc the brackets [ ] represent the display of the components of the
vector n and the matr ix representat ion of the tensor T.
(iii) The third step is to prove that [TI is indeed the matrix
rc~presentutionof a tensor T (see Appendix)
I n
order to prove tha t the T,,s a r e the com po ne nts of a tensor T , we need
to prove tha t the components Ti, transform as a tensor. Since n is just a
frcc vector, which is objective,
n' = Q - n ,
a n d since the tract ion t t ransforms as
t ' = Q . t ,
then. T transforms as '
T ' = Q . T . Q T .
Problem
3.3.1
( Jsc Cauchy 's cons t ruct ion in conjunct ion wi th the 'mass balance ' t o show
that the mass f lux in the direct ion normal to a plane n is given by j-n.
I+
licrc
j
is the mass f lux vector. This allows the definition of the Eulcrian
\clocity a s v
=
/p which can now be regarded a s the primitive qua nti ty