C'ON T E M P O R A R Y D E V E L O P M E N T S IN CONTINUUM MECHANICS A N D PARTIAL DIFFERENTIAL EQUATIONS
This Page Intentionally Left Blank
N0 RTH-H0 LLAND MATHEMATICS STUDIES
30
Contemporary Developments in Continuum Mechanics and Partial Differential Equations Proceedings of the International Symposium o n Continuum Mechanics and Partial Differential Eyuations Rio de Janeiro, August 1977 Edited b y
GUILHERME M . DE LA PENHA Universidade Federal do Rio de Janeiro and Financiadora de Estudos e Projetos (FINEP/SEPLAN) and
LU lZ A D A U T O J. M E D E I R O S Universidade Federal do Rio de Janeiro
1978
NORTH-HOLLAND PUBLISHING COMPANY AMSTERDAM - NEW YORK- OXFORD
0
North-Holland Puhlishing Compan,~ 1978
All rights reserved. No part of this publication may he reproduced, stored in a retrieval s.vstem, or transmitted in any form or h.v any means, electronic, mechanical, photocopying, recording or otherwise, without the prior pcrmission of the copvright owner.
ISBN 0 444 85166 6
PUBLISHERS
NORTH-HOLLAND PUBLISHING COMPANY AMSTERDAM. NEW YORK,OXFORD
SOLE D I S T R I B U T O R S F O R T H E U . S . A . A N D C A N A D A :
ELSEVIER/ N O R T H - H O L L A N D , INC. 5 2 V A N D E R B l L T A V E N U E . NEW Y O R K . N.Y. 10017
Library of Congress Cataloging in Publication Data
International Symposium on Continuum Mechanics a d Partial Differential Equations, Rio de Janeiro, Brazil, 1977. Contemporary developments in continuum mechanics and partial differential equations. (North-Holland mathematics studies ; 30) Includes index. 1. Continuum mechanics--Congresses. 2. Differential equations, Partial--Congresses. I. Penha, Guilherme M. de la. 11. Medeiros, Luis Adauto da Justa. 111. Title.
~~808.2.1585 1977 ISBN 0-444-851666
531
78-5270
P R I N T E D I N THE N E T H E R L A N D S
FOREWORD
An I n t e r n a t i o n a l Symposium on Continuum Mechanics and P a r t i a l D i f f e r e n t i a l E q u a t i o n s was h e l d a t t h e I n s t i t u t o de Matemztica, U n i v e r s i d a d e F e d e r a l do Rio d e . J a n e i r o , Rio de J a n e i r o , R . J , 1977.
Brasil, from August 1 - 5 ,
T h i s meeting had an o b j e c t i v e o f a c q u a i n t i n g m a t h e m a t i c i a n s and
o t h e r s c i e n t i s t s i n p h y s i c s and e n g i n e e r i n g , w i t h t h e p r i n c i p a l l i n e s o f modern r e s e a r c h i n such i n t e r r e l a t e d s u b j e c t s so t o a f f o r d s t i m u l a t i n g d i s c u s s i o n s l e a d i n g t o t h e i n v e s t i g a t i o n o f new problems o r r e v i v i n g o l d o n e s s t i l l open.
Another aim o f t h e Symposium was t o a t t e m p t t o a s s e s s
what t h e r e c e n t r a p i d growth o f i n t e r a c t i o n s between t h e two mathematical s u b j e c t s h a s accomplished and may accomplish i n t h e n e a r f u t u r e . Wide-spread a c t i v e i n t e r e s t was shown i n t h e Symposium, and twelve European, f i f t e e n North-American and n i n e t y - o n e B r a z i l i a n s c i e n t i s t s registered, papers. by J . L .
The c h i e f a c t i v i t y o f t h e c o n f e r e n c e was t h e p r e s e n t a t i o n o f
Two s h o r t c o u r s e s were g i v e n , Lions.
one by M . E .
G u r t i n and t h e o t h e r
C . T r u e s d e l l p r e s e n t e d a c r i t i c a l a p p r a i s a l o f continuum
therniomechanics and t h e k i n e t i c t h e o r y o f g a s e s , a l l o f t h e s e b e i n g reproduced h e r e .
S e v e r a l p a r t i c i p a n t s c o n t r i b u t e d p a p e r s , and s h o r t
communications were p r e s e n t e d .
Duplicated a b s t r a c t s of presented papers
were a v a i l a b l e a t t h e meeting and a l m o s t a l l p a p e r s were followed by a q u e s t i o n and d i s c u s s i o n p e r i o d .
I t i s n a t u r a l l y n o t p o s s i b l e t o reproduce
a l l t h e s e d i s c u s s i o n s , b u t most of t h e p a p e r s which were p r e s e n t e d are included here.
Nearly a l l t h e p a p e r s p r i n t e d were r e t y p e d from t h e o r i g i n a l s
s u p p l i e d by t h e a u t h o r s and s u b s e q u e n t l y e d i t e d .
To speed t h e a v a i l a b i l i t y
o f t h i s volume, t h e f i n a l typed v e r s i o n was n o t s u b m i t t e d t o t h e r e v i s i o n of t h e authors, thus t h e e d i t o r s bear a l l t h e r e s p o n s i b i l i t y f o r anything t h a t may n o t p l e a s e t h e a u t h o r s o r r e a d e r s o f t h e c o n t e n t s o f t h i s volume, I t i s hoped t h a t t h e s e p r o c e e d i n g s w i l l s e r v e b o t h t h o s e who a t t e n d e d t h e Symposium and t h o s e u n a b l e t o a t t e n d , and t h a t t h i s permanent r e c o r d w i l l add t o t h e s t i m u l a t i o n and d i a l o g u e g e n e r a t e d a t t h e m e e t i n g . We s h a l l be most p l e a s e d i f , b e s i d e s t h i s , it w i l l h e l p t o some e x t e n t i n e n c o u r a g i n g m u t u a l l y advantageous c o n t a c t s between t h e p r o d u c e r s and u s e r s o f mathematics.
Many t h a n k s a r e due t o t h e B r a z i l i a n a g e n c i e s f o r r e s e a r c ] ] ,
F i n a n c i a d o r a de Estudos e P r o j e t o s (FINEP) and Conselho Nacional de Desenvolvimento C i e n t f f i c o e Tecnol6gico (CNPq) f o r t h e i r generous f i n a n c i a l V
vi
FOREWORD
s u p p o r t which made t h e Symposium p o s s i b l e .
The Universidade F e d e r a l do
Rio de J a n e i r o , t h r o u g h t h e Centro de C i e n c i a s Matematicas e da Natureza and t h e I n s t i t u t o de Matematica p r o v i d e d f a c i l i t i e s for t h e c o n f e r e n c e , and many p e o p l e o f t h e s e two u n i v e r s i t y i n s t i t u t i o n s gave f r e e l y o f t h e i r h e l p . I t remains t o e x p r e s s a p p r e c i a t i o n t o t h o s e p e r s o n s who have h e l p e d i n v a r i o u s ways t o b r i n g t h e s e p r o c e e d i n g s i n t o b e i n g :
.
Our c o l l e a g u e s on t h e O r g a n i z a t i o n Commission, Gustavo P e r l a Menzala, Luiz C a r l o s M a r t i n s and Rubens Sampaio,
.
The Symposium s p e a k e r s f o r t h e i r c o o p e r a t i o n i n s u b m i t t i n g manuscripts (mostly i n t i m e ) ,
.
Wilson Goes for t h e n e a t n e s s o f t h e t y p i n g ,
.
t o t h e p u b l i s h e r s , North-Holland,
and i n t h e person o f D r . E . Fredriksson.
Rio d e J a n e i r o , December 1977 G.M.
d e La Penha
L.A.
Medeiros
CONTENTS
Foreword
S.S. Antman
V
A family of semi-inverse problems of nonlinear elasticity
1
G.S.S. Avila and D.G. Costa Asymptotic properties of general symmetric hyperbolic systems
25
J.B. Baillon and J.M. Chadam The Cauchy problem for the coupled SchroedingerKlein-Gordon equations
37
T.B. Benjamin
H. Brezis
Applications of generic bifurcation theory in fluid mechanics
45
The Hamilton-Jacobi-Bellman equation for two operators via variational inequalities
74
Nonlinear problems related to the Thomas-Fermi equation
81
H. Brezis
F. Cardoso and J . Hounie Global solvability and hypoellipticity of abstract complexes and equations
90
Michael O'Carroll On the inverse scattering problem for linear evolution equations
102
Nonlinear bifurcation problem and buckling of an elastic plate subjected to unilateral conditions in its plane
112
B.D. Coleman
On the thermodynamics of non-classical systems
135
R.L. Fosdick
On an inequality in thermodynamic stability
143
Claude Do
R. Glowinski and 0. Pironneau Least square solution of non linear problems in fluid dynamics
171
P. Podio Guidugli Elastic bodies in a Signorini-type environment
225
M.E. Gurtin
On the nonlinear theory of elasticity
237
D.D. Joseph
Constitutive equations and free surfaces
254
J.L. Lions
On some questions in boundary value problems of mathematical physics
284
vii
viii
R.C. Maccamy
CONTENTS Memory effects in one-dimensional problems of continuum mechanics
34 7
A general framework for problems in the statics of finite elasticity
363
P. Nowosad
Elliptic metrics on Lorentz manifolds
388
S. Osher
Boundary value problems for equations of mixed type
396
J.P. Puel
A free boundary, nonlinear eigenvalue problem
400
James Serrin
The concepts of thermodynamics
411
W.A. Strauss
The nonlinear Schrddinger equation
452
Walter No11
G. Svetlichny and P. Otterson Derivative dependent infinitesimal deformations of differentiable functions L. Tartar
466
Non linear constitutive relations and homogenization
472
R. Temam
Qualitative properties of Navier-Stokes equations
485
G. Truesdell
Some challenges offered to analysis by rational thermodynamics
495
Some simplified equations from the theory of mixtures
604
W.O. Williams Author Index
613
G.M. de La Penha, L.A. Medeiros ( e d s . ) Contemporary Developments i n Continuum Mechanics and P a r t i a l D i f f e r e n t i a l Equations Worth-Holland Publishing Company (1978)
A FAMILY O F SEM I-INVERSE
PROBLEMS OF N O N L I N E A R ELASTICITY
STUART
s.
ANTMAN*
D e p a r t m e n t o f Ma t,herna t i c s and I n s t i t u t e f o r P h y s i c a l S c i e n c e and T e c h n o l o g y U n i v e r s i t y of Maryland,
C o l l e g e P a r k , Maryland
20742
f ' o r t h e development of
the
1. I n t r o d u c t-i _on Much o f t h e i m p e t u s
theory
o f n o n l i n e a r e l a s t i c i t y f o l l o w i n e ; t h e Second World War was s u p p l i e d by the s t u d y of
c o n c r e t e problems
that
exhibit
i n t e r e s t i n g e f f e c t s due t o n o n l i n e a r m a t e r i a l r e s p o n s e . the study of
i n v e r s e p r o b l e m s R i v l i n f s work
[
111 was
In
paramoiint.
The n a t u r e o f s u c h p r o b l e m s was e x p o s e d by E r i c k s e n l s p r o o f
61 t h a t t h e o n l y d e f o r m a t i o n s p o s s i b l e i n e v e r y homogeneous ( c o m p r e s s i b l e ) e l a s t i c m a t e r i a l a t r e s t u n d e r z e r o body f o r c e a r e a f f i n e and by h i s s t u d y
[51
p o s s i b l e i n e v e r y homogeneous
of n o n - a f f i n e
deformations
incompressible e l a s t i c material
a t r e s t u n d e r z e r o body f o r c e .
(Ericksenls analysis
spawned a s m a l l l i t e r a t u r e d e v o t e d
[51
t o those aspects o f
the
problem t h a t h e l e f t u n r e s o l v e d . )
Semi-inverse
problems,
studied i n t h e 1950's, lead t o
?+
The r e s e a r c h r e p o r t e d h e r e was s u p p o r t e d b y N a t i o n a l S c i e n c e Foundation Grant
MCS77-03760.
2
STUART S.
a r i c h e r c l a s s of d e f o r m a t i o n s , by q u a s i l i n e a r s y s t e m s of (Cf.
G r e e n & Adkins [
for references.)
7,
Most
of solutions, A notable
many of w h i c h a r e d e s c r i b e d
ordinary d i f f e r e n t i a l equations.
Ch 111 a n d T r u e s d c l l & N o 1 1 [ t r e a t m e n t s of
y i e l d i n g a number o f r e s u l t s avoided the questions
ANTMAN
of
t h e s e problems,
o€ p h y s i c a l
interest,
1 3 , Sec.591 while
either
e x i s t e n c e and q u a l i t a t i v e b e h a v i o r
or e l s e t r e a t e d them o n l y f o r s p e c i a l m a t e r i a l s . to this i s the work
exception
radial oscillations elastic material.
of
of Knowles
a c y l i n d r i c a l t u b e of
181 on t h e
incompressible
The manner i n which Knowles employed
constitutive inequalities i s the natural precursor of
the
method u s e d h e r e . I n t h i s p r e s e n t p a p e r we examine a f a m i l y o f s e m i i n v e r s e problems
f o r compressible e l a s t i c m a t e r i a l s s a t i s f y -
ing the strong e L l i p t i c i t y condition.
( F o r s i m p l i c i t y we
assume t h a t t h e m a t e r i a l i s homogeneous and i s o t r o p i c . ) By f u l l y e x p l o i t i n g t h e s t r o n g e l l i p t i c i t y c o n d i t i o n w e r e a d i l y show t h a t
" r e a s o n a b l e " s e m i - i n v e r s e boundary v a l u e problems
a l w a y s h a v e s o l u t i o n s a n d t h a t a number of
their qualitative
f e a t u r e s can be determined. N o-n . - o t a t i.
E',
V e c t o r s , which a r e h e r e d e f i n e d t o b e e l e m e n t s o f
and v e c t o r - v a l u e d
l e t t e r s over t i l d e s .
f u n c t i o n s a r e denoted by lower-case S _e _cond-order
t e n s o r s , which a r e t a k e n
~
t o be elements o f
L ( E 3 ,(E 3) ,
and f u n c t i o n s w i t h v a l u e s
in
L(E3,E3),
a r e d e n o t e d by u p p e r c a s e l e t t e r s o v e r t i l d e s .
The s u b s e t
o f second-order
i s denoted tensors
L+(E3,S3)
i s denoted
tensors with p o s i t i v e determinant
and s u b s p a c e o f symmetric second-order
S(E3,E3).
The d _ o. . t product
of
2
and
b
SEMI- I N V E R S E PROBLEMS O F NONLINEAR ELASTICITY
a.!. -A*
is written The a d j o i n t
-
all
a,?.
-
A
of
-
B,
and
- --
= A.(B-A).
(ab)-c
=
(b*c)a 2
d i f f e r e n t i a l o f t h e mapping
-
i s denoted
t h e mapping
- -
U - I - z F- ( U-)
3.
-2
[af(a)/ay]
- --
b . ( A * a ) = a*(;*.!)
= bu . A- .
The p r o d u c t of
(fi-n)-,a = ah --
The dyad
2.
for all
k+,f(u)
3
at
We s e t
a (fl,.
a (f19f2,f3)
= A
a(Av+Bw,u 2
is
the
The FrGchet
i n the direction
i n the direction
-
B
i s denoted
.. , f n ) ,11
n
3(fl
a
afi
= d e t (-),
) I
a U .
3
f2
9
f3)
(v,u2,u3)
a ( f l 3 f -~ *,f3)
+ B-
a(w,u2,u3) *
The E q u a t i o n s o f E l a s t o s t a t i c s .
' (21,i2,i-3) =
Let
domain
n
in
E3
( i1 ,,I.z ,& 3 )
B3.
orthonormal b a s i s f o r that
( -i , J , E )
I
he a f i x e d
We i d e n t i f y a body ~ i t h the
i t occupies i n i t s r e f e r e n c e
c o n f i g u r a t i o n and w e i d e n t i f y a m a t e r i a l p a r t i c l e
-z
with i t s p o s i t i o n
- z a I. -
-a
-z
for
and t h e F r G c h e t d i f f e r e n t i a l o f
at
a (u,, . . .
2.
Asa.
D i a g o n a l l y r e p e a t e d i n d i c e s a r e summed f r o m
[aF(A)/$U]:B*. 1 to
A*?.
- -
is written
i s d e f i n e d by
As?,
= trace
A::
t e n s o r d e f i n e d by
h
- -
A*.b
-
denoted
-
We s e t
-a
at
i s d e f i n e d by
We a c c o r d i n g l y w r i t e
-A
tensors
A
The v a l u e of
3
= xi + y j
-
-
+
i n this configuration.
-
zk.
Let
~ ( 5 )h e
i n a deformed c o n f i g u r a t i o n .
The d e f o r m a t i o n
F
of
We s e t
t h e p o s i t i o n of
t h e body
-z
particle
W e set
preserves orientation i f
=
and o n l y i f
4
STUART S. ANTMAN
Let at
5.
z(z)
be the first Piola-Kirchhoff stress tensor
Then the equilibrium equations f o r a body subject to
zero body force are
(2.3)
Div
n
The material of
I
-
ia.-
a
aza
AT*= 0,
-
-
is h_omogeneously elastic if there are
T: L + ( E 3 , E 3 )
functions
?-
4
L(!E3,E3)
2:
and
S(IE3,E3)
+ S(E3,E3)
such that
We assume that
T
and
5
are continuously differentiable.
This representation ensures Kirchhoff stress tensor.
_ _ _ _ _ _ . _ _ _ _ ~ -
-S
only if
where
-
I
(2.4).
,S
is the second Piola-
The material is isotropic if and
has the form
is the identity on
E3
and where
depend on the principal invariants of
c.
ao, a l , a 2
Assuming that
T
is continuously differentiable, we require it t o satisfy the strong ellipticity condition
Below we shall impose specific conditions ensuring that large stresses accompany large strains and strains for which det is small.
F
SEMI-INVERSE PROBLEMS OF NONLINEAR ELASTICITY
5
3. Formulation of the Semi-Inverse Problem. We consider deformations defined by
p(z) =
(3.1)
f(x)_el(z)
+
[h(x)+C~+Dz]k,
where
--
(3.2) ~ ~ ( =2 C)O S e(z)i
e(.)
+ sin e ( ? ) i ,
(This is ”Family 1 ” of [ 13, Sec. 591 . )
= g(x)
+ AY + BZ.
O u r problem will be
to determine the existence and properties o f functions and numbers
A, 8 , C, D
f,g,h
such that (3.1) satisfies the
equations (2.3) and (2.5) and certain subsidiary conditions. We set
(3.3)
-
e ( z ) = &~e_~(z), e = k , -2 -3
c?g
x
Denoting derivatives with respect to
where
f’el + fg’e -2
+
h‘
(3.5)
/
f’
0
0
(3.6)
Condition (2.2) reduces to
(3.7)
(AD-BC)ff’ > 0 .
e,c .
by primes, we have
(3.4)
aE2 ax =
=
,e3,
6
ANTMAN
STUART S.
(3.7) as
By making o b v i o u s s i g n c o n v e n t i o n s , we may i n t e r p r e t being e q u i v a l e n t t o t h e requirement be p o s i t i v e .
l i-a i-' j
C
The components o f
that
(3.7)
e a c h f a c t o r of
with r e s p e c t t o t h e b a s i s
are
+
,/(f')2
(cap) =
(3.8)
I\
+
(fg')2
(h')2
A f 2g ' + C h '
Bf 2 g ' + D h ' \
A2f2
2 ABf +CD
+C2
'> 'a
,/
B2f2+D2 /
T
We decompose
as
-
T = ~a L - eL- ia
(3.9) s o that
( 2 . 3 ) and
( 2 . 5 ) imply t h a t
(3.10)
Relations
on
G
( 3 . 5 ) and ( 3 . 8 ) t h u s i m p l y t h a t
- -
{ F a = e L . F . -a i ]
(3.12)
In p a r t i c u l a r , that
h',
g',
f , A,
B,
C,
D.
--
. + * i a ]a r e i n d e p e n d e n t
of
y
and
(3.10) has the componential f o r m 1
(3.13)
(3.14)
From
depend o n l y
and t h e r e f o r e o n l y o n
f',
(:&
(Tta]
(T2')'
+
-
g'Tll
AT2
+
( 3 . 1 5 ) we o b t a i n
(3.16)
T~~ = H ( c o n s t ) .
2
-
AT12
BT2
2
-
+ BT13
BT2'
= 0,
= 0,
z
so
SEMI-INVERSE
Among o t h e r r e l a t i o n s
f[ g ' T l l
(3.17)
implies that
(2.14)
+
+ BT13] =
ATI'
t h e s u b s t i t u t i o n o f which i n t o
D
f'T2
1
y
produces
the integral
(const).
I
Without loss of
C,
(3.14)
T, 1 = G
f
(3.18)
domain
7
PROBLEMS O F NONLINEAR E L A S T I C I T Y
R
i s the unit
t h e reference
g e n e r a l i t y we s u p p o s e t h a t cube
( 0 ~ ) ~ W e .s u p p o s e t h a t AD-BC
are prescribed with
>
0.
I n Section
7
By
A,
we d i s c u s s
how t o r e l a t e t h e s e c o n s t a n t s t o v a r i o u s r e s u l t a n t s a c t i n g over t he m a t e r i a l f a c e s o f any t r a c t i o n s p r e s c r i b e d with the i n t e g r a l s
H
and
(3.16), (3.18).
of
a non-zero
W e accordingly regard
corresponds
d e a d s h e a r l o a d s on t h e f a c e s
p r e s c r i p t i o n of
be c o m p a t i b l e G
e n t e r i n t o a d i s c u s s i o n of
€1
the prescription of
x = O,1
( 3 . 1 6 ) and (3.18), b u t m e r e l y n o t e
c o n d i t i o n s compatible b i t h that
the faces
011
W e do not
as g i v e n .
W e a l s o require that
t h e cube.
G
t o the p r e s c r i p t i o n
x = 0 , 1 , whereas
the
c o r r e s p o n d s n e i t h e r t o a dead
l o a d n o r t o a c o n s t a n t Cauchy s t r e s s .
The __ g o -. v_ e -r.n i n g e q u a t i o n s f o r o u r problem c o n s i s t of
(3.13), ( 3 . 1 6 ) laws
4.
(3.18),
which i n c o r p o r a t e
the s t r e s s - s t r a i n
(2.5).
Conseauences o f C o n s t i t u t i v e R e s t r i c t i o n s .
\\ie now t r a n s f o r m o u r e q u a t i o n s f u r t h e r by b r i n g i n g the strong e l l i p t i c i t y condition choosing
-a
-
= i
in
(2.7) we get
( 2 . 7 ) t o b e a r on them.
By
8
ANTMAN
STUART S .
b = b'
for
e
-L
f u n c t i o n of' Note t h a t
2.
Thus
T *-i
i s a s t r i c t l y monotone
f o r f i x e d v a l u e s of i t s o t h e r a r g u m e n t s .
ag/ax
t h e remarks p r e c e d i n g ( 3 . 1 2 )
i n d e p e n d e n t of g, y,
#
t h e b a s i s u s e d h e r e and t h e r e b y i n d e p e n d e n t of
I n consonance with
Z .
(TLa] are
imply t h a t
(4.1),
we impose t h e growth
conditions
(4.2)
TZ1
4
as
*m
for f i x e d values
of
fg'
+
+m,
T
+
3
t h e o t h e r arguments.
r e s t r i c t t h e arguments o f
-
T
as
fa
h'
+
fm,
Here and below we
t o b e d e f o r m a t i o n s of
t h e form
( 3 . 1 ) ; we t h e r e b y a v o i d t h e i n t e r e s t i n g q u e s t i o n o f p o s i n g r e a l i s t i c growth c o n d i t i o n s f o r e l a s t i c m a t e r i a l s u n d e r a r b i t r a r y deformation.
( T h i s g e n e r a l q u e s t i o n might b e hand-
l e d by combining i d e a s of B a l l Brezis
[4]
31 . ) The m o n o t o n i c i t y c o n d i t i o n
conditions
(4.2)
ensuring t h a t
(4.1)
and t h e growth
j u s t i f y a g l o b a l i m p l i c i t f u n c t i o n theorem
( 3 . 1 6 ) and (3.18), r e g a r d e d as a l g e b r a i c
e q u a t i o n s , can b e uniquely of
w i t h t h o s e of Antman &
t h e o t h e r v a r i a b l e s of
solved
for
fg'
t h e problem.
and
h'
i n terms
W e represent these
s o l u t i o n s by
(4.3)
(4.4)
fg' h'
= y(f',
G/f,
= n(f', G/f,
A,
B,
C, D ) ,
€1, A ,
B,
C,
H,
A l o c a l i m p l i c i t f u n c t i o n theorem e n s u r e s t h a t a r e c o n t i n u o u s l y d i f f e r e n t i a b l e because
T
is.
D).
y
and
q
SEMI - I N V E R S E
We
=ubstitute
a u t on ornous,
9
PROBLEMS O F N O N L I N E A R E L A S T I C I T Y
( b . 3 ) , ( 4 . 4 ) i n t o (3.13) t o get the
q u a s i 1i n e a r ,
s e c ontf
- ord e r
ord inary differentia 1
equation
h = (A,B,C,D,G,H)
where
T2' + A T 2 2 + BTz3 F21
-
[P(f',f,X)I'
(4.5)
and
F31
a ( f ' , f , h ) = 0,
aiid \ % h e r e p
evaluated
at
and
The m o n o t o n i c i t y c o n d i t i o n system
(3.l3),
semi-monotone
(fT2 (for
1
)
1
are
3
of
(3.6)
by d e f i n i t i o n o f
with
for
ensures t h a t the f,
g, h
i s formally
~t c o m e s a s n o s u r p r i s e t h a t
( 4 . 5 ) i s i t s e l f a f o r m a l l y semi-monotone e q u a t i o n f o r Indeed,
and
T1
('t.h).
(4.1)
(3.75)
= 0,
f > 0).
4,
{F a]
the
(4.3)
r e p l a c e d by
and
y, q
and
p,
f.
we f i n d
(4.6)
w h i c h i s s t r i c t l y p o s i t i v e by
a l s o compute
(4.7)
(4.8)
(4.1).
I n a l i k e manner we
10
STUART S.
ANTMAN
(4.9)
The s t r o n g e l l i p t i c i t y c o n d i t i o n i s i n c a p a b l e that the first
t e r m on t h e r i g h t s i d e d o f
s o t h e mapping monotone.
(p(f’,f,k),
(f’ ,f)-
of ensuring
(4.9) i s positive,
~ ( f, f‘ , x ) )
need n o t be
( 4 . 5 ) need n o t be f o r m a l l y
Therefore the equation
i t i s l i k e l y t h a t b o u n d a r y v a l u e problems
monotone
and h e n c e
for ( 4 . 5 )
f a i l t o have unique s o l u t i o n s .
T o e x a m i n e t h i s and r e l a t e d q u e s t i o n s f u r t h e r we G = €1 = 0.
c o n s i d e r t h e s p e c i a l c a s e i n which 1
(2.6) imply t h a t
1
= T3
T2
=
F
when
t h e monotonicity c o n d i t i o n ( k . 1 ) alone, that
2
Now ( 2 . 5 ) and
= F31 = 0.
without
Thus
( 4 . 2 ) , implies
( 3 . 1 6 ) and ( 3 . 1 8 ) h a v e t h e u n i q u e s o l u t i o n = 0,
g’
(4.10)
h’
= 0;
t h i s i n t u r n i m p l i e s t h a t t h e d e f o r m e d c o n f i g u r a t i o n s of t h e faces
y = O,l,
z = 0,1
(2.6)
imply t h a t
(4.11) so that (4.12)
T2
1
= T3
a r e planes.
1
= T1
2
I n t h i s case
= TI3 = 0
(3.13) o r ( 4 . 5 ) r e d u c e s t o p’-a
E
w h e r e t h e a r g u m e n t s of
-
(Tll)’
T1
1
,
T2
AT2 2
2
,
-
T2’
BT23 = 0
are
(2.5) and
SEMI-INVERSE
PRODLEMS O F NONLINEAR E L A S T I C I T Y
O
0
Af
BY
C
D
11
‘.
I n t h i s case
1
-Pa
(4.14)
af7 =-I
a T1
aFll
(4.15)
(4.16)
2
as
--
(4.17)
af
T+AB---aF 2
-
(4.1)
Inequality but
A
2
(p(-,-,h),
a T2 +
aF22
n o w implies that U(-,-,h))
3 AB -+ B
ap/af‘
>
2 aT2 __ 2 ‘ aF
and
0
as/af
> 0,
may n e v e r t h e l e s s f a i l t o be
monotone.
5 . E x i s t e n c e Theory f o r Boundary Value P r o b l e m s . W e impose b o u n d a r y c o n d i t i o n s
1 T1
(5.1 a,b)
I x=o
- qo
or
f(0) = fo
More g e n e r a l c o n d i t i o n s a r e p o s s i b l e . that
A,
y = 0,1,
B, z
C,
D
= 0,1
are prescribed, can e n s u r e t h a t
moment o n t h e body v a n i s h when
> O,
S i n c e we a r e a s s u m i n g
the reactions
on t h e f a c e s
t h e r e s u l t a n t f o r c e and
( 5 . l a ) and ( 5 . 2 a ) a r e prescribed.
STUART S. ANTMAN
12
There are several effective ways to attack the boundary value problem (4.5), ( 5 . 1 ) ,
(5.2).
The first method
is to give it a weak formulation i n a reflexive Sobolev space, the choice o f which is dictated by sharper growth conditions that must be imposed. In this setting the problem can be cast as an equation involving a pseudo-monotone operator.
The difficulty here
lies in the treatment of the strict inequality (3.7) and the
T
growth o f
where
(3.7) is small.
of Antman [2] and Antman & Brezis
Methods similar to those
133
can be used to handle
this difficulty and to yield a full regularity theory.
(The
latter work describes a useful set of growth conditions in Section
4.) This approach shows that there is a weak solution for
each
h
satisfying (3.7) and that each such weak solution is
classical.
Instead of carrying out the details of this theory,
we turn to another that provides somewhat more information about classical solutions. We impose the growth condition
(5.3) for fixed values of
f
and
1.
This condition and the positivity of
(4.6) imply that the
algebraic equation
(5.4)
P
has a unique solution for
(5.5)
(f’,f
1) = 9
f‘, which we denote by
f’ = cp
s,f,X).
13
SEMI-INVERSE PROBLEMS O F N O N L I N E A R ELASTICITY
W e c a n r e w r i t e t h e boundary v a l u e problem
(4.5), ( 5 . 1 ) ,
(5.2)
a s t h e system
(5.9 a , b )
q(l) =
q1
or
f ( l ) = fl.
T o b e s p e c i f i c and t o a v o i d minor problems w i t h Neumann conditions w e r e s t r i c t o u r attention t o conditions (5.8a),
(5.9b). f
>
0
where
a
Let
( T h e s e c o n d i t i o n s do n o t r e n d e r t h e r e q u i r e m e n t of
( 3 . 7 ) c o m p l e t e l y i n n o c u o u s a s would ( 5 . 8 b ) . )
@
i s a c o m p l e t e l y c o n t i n u o u s mapping f r o m
reduces
a
=
;(p),
0 E IR,
a
so that
A
(5.12a)
to
A
If
ExA i n t o E.
be c o n f i n e d t o l i e o n a s i m p l e smooth c u r v e i n
g i v e n p a r a m e t r i c a l l y by
Then
i s continuously d i f f e r e n t i a b l e ,
then
Y
i s completely
14
STUART S. ANTMAN
continuous.
W e assume t h i s .
F o r s u c h problems we h a v e R a b i n o w i t z [ 101
Theorem ( L e r a y & S c h a u d e r [ 9 ] ,
cf.
(uo,e0)
(5.12b).
b e a s o l u t i o n p a i r of
u = Y(u,po)
equation
t h e component of
E x
Suppose t h a t t h e
t h e s e t of
i s e i t h e r unbounded i n
To e x p l o i t t h i s
of
Then
s o l u t i o n p a i r s that E x
and i n
[po,m)
o r e l s e a p p r o a c h e s t h e boundary of
(-m,po]
(u0,$,)
t h e c l o s u r e of
(uo,p0)
contains
Let
as i t s u n i q u e s o l u t i o n .
uo
has
-
)
these s e t s .
theorem we must f i r s t f i n d a s o l u t i o n p a i r One way t o g e t t h i s i s t o u s e t h e weak
(5.12b).
But i t i s more i l l u m i n a t i n g t o
f o r m u l a t i o n mentioned a b o v e .
u s e a d i f f e r e n t approach m o r e i n keeping with t h e continuation methods d e v e l o p e d by R a b i n o w i t z [ 103
.
I n m o s t problems t h e n a t u r a l c h o i c e f o r t r i v i a l s o l u t i o n , which would c o r r e s p o n d s t a t e f o r o u r e l a s t i c i t y problem. s t a t e i s n o t i n t h e form
(uo,Bo)
is a
t o the reference
Unfortunately the t r i v i a l
( 3 . 1 ) of a d m i s s i b l e s o l u t i o n s ;
i s rather a singular l i m i t of
such s o l u t i o n s .
it
The continuation
of s o l u t i o n s from t h i s t r i v a l s o l u t i o n i s o f t e n termed " b i f u r c a t i o n from i n f i n i t y " .
We c a n u s e t h e p h y s i c a l
structure
of o u r problem t o r e n d e r t h i s t e c h n i c a l d i f f i c u l t y i n n o c u o u s . W e s t a r t t h i s p r o c e s s by c o n s i d e r i n g t h e problem i n which
B = C = G = H = 0, case (3.6)
D = 1.
(Cf.
(4.10)-(4.17).)
In this
i s d i a g o n a l and i s t h e s q u a r e r o o t of t h e correspond-
i n g v e r s i o n of
(3.8).
L e t u s assume t h a t t h e r e f e r e n c e s t a t e
is a natural stress-free
state, i.e.,
:(I)
=
2.
We assume
that
(5.13)
-
H:
-
-
aT(I)/aF : H > 0 -
I
V diagonal
I;! f
2.
(This i s a milder r e s t r i c t i o n linear elasticity.) then says t h a t
F‘s
c l a s s of
t h a n t h a t commonly made i n
The c l a s s i c a l i m p l i c i t f u n c t i o n t h e o r e m
the constitutive
relations f o r our special
a r e e q u i v a l e n t t o e q u a t i o n s o f t h e form
(5.14)
f’
1 = V ( T ~,
(5.15)
Af
= V(T1
where
(P,v)
15
PROBLEMS OF NONLINEAR E L A S T I C I T Y
SEMI-INVERSE
i s monotone,
1
,
AT,^), AT2
w(O,O)
2
),
= 1,
v(0,O)
= 1.
Note t h a t t h e c u b i c a l r e f e r e n c e s h a p e i s a t t a i n e d i n t h e l i m i t
+
that
A + 0,
while
( 4 . 12) g i v e s
f
Af
+
1.
(Tll)’
-
m ,
(5.17) The s y s t e m ( 5 . 1 6 ) ,
From ( 5 . 1 4 ) ,
AT2‘
(5.17) replaces
(5.15)
we get
= 0. Equation (5.16)
(4.12).
i s a compatibility condition.
Now t h e b o u n d a r y c o n d i t i o n s (5.18)
T1 1( 0 ) = q 0 ,
(5.8a), ( 5 . 9 b ) become v ( T 1 1(I), A T 2 2 ( 1 ) ) = Afl.
In v i e w of o u r m o n o t o n i c i t y c o n d i t i o n s , t h e u n i q u e s o l u t i o n of
(5.16)-(5.18)
I f we v a r y
when
(A,Afl,qO)
9,
= 0,
Afl
from (O,l,O)
= 1
is
T1
along a curve i n
t h e n a s t a n d a r d i m p l i c i t f u n c t i o n theorem e n s u r e s (5.16)-(5.18) enough t o
= T 2 2 = 0. R3,
that
has a unique s o l u t i o n f o r t h e s e parameters n e a r
(0,1,0). B y o u r c o n s t r u c t i o n ,
such a s o l u t i o n
would c o r r e s p o n d t o a d e f o r m a t i o n i n w h i c h t h e f a c e s
x = 0,l
l i e on c o n c e n t r i c c i r c u l a r c y l i n d e r s o f f i n i t e r a d i u s . A n y such s o l u t i o n i s a s o l u t i o n of
(5.10), (5.11) f o r
16
STUART S .
the se parameters.
L e t one s u c h s o l u t i o n p a i r be d e n o t e d
A condition sufficient
(uo,ao).
u = Y(u,@,), that
ANTMAN
a. = & ( P o ) ,
where
t h e mapping
t o ensure that
the equation
h a s a t m o s t one s o l u t i o n i s
1 2 1 1 2 ( F 1,F * ) A T 1 ( F 1 ,O , O , 0 , F ~,0 , O , O , D )
2 1 2 T2 ( F l , O , O , O , F 2,0,0,0,D)
a.
o u r c o n s t r u c t i o n of
be s t r i c t l y monotone.
,
(Note t h a t
F
e n t a i l s t h a t t h e components o f
have t h e form i n d i c a t e d i n t h e a r g u m e n t s o f
TI1
and
T2
2
.
T h i s r e s t r i c t e d m o n o t o n i c i t y c o n d i t i o n i s i m p l i e d by t h e
[ I 3 1 b u t n o t by t h e s t r o n g e l l i p t i c i t y
Coleman-No11
inequality
condition.)
The c o n t i n u a t i o n theorem o f Leray & S c h a u d e r
a p p l i e s t o parameters
a =
a^(@)
a
confined t o c u r v e s of
t h e form
with The c o n t i n u a t i o n method o f Leray & S c h a u d e r i m p l i e s
that solutions of
(5.10), ( 5 . 1 1 ) a r e c l a s s i c a l u n t i l the
continuum of s o l u t i o n p a i r s becomes unbounded o r e l s e approaches everywhere.
~ ( E x A ) . Such a c l a s s i c a l s o l u t i o n s a t i s f i e s
(3.7)
One c a n g e t a somewhat s t r o n g e r r e s u l t by look-
ing f o r solutions
(5.10),
( 5 . 1 ~ )i n a s m a l l e r c l a s s of
f u n c t i o n s , say L i p s c h i t z continuous f u n c t i o n s .
By s t r e n g -
t h e n i n g o u r growth c o n d i t i o n s o n t h e c o n s t i t u t i v e f u n c t i o n s we c a n show t h a t any- L i p s c h i t z c o n t i n u o u s s o l u t i o n must satisfy
( 3 . 7 ) everywhere provided t h i s
with t h e c h o i c e of
a.
The p r o o f
i s not
incompatible
r e l i e s on t h e o b s e r v a t i o n
t h a t a L i p s c h i t z c o n t i n u o u s f u n c t i o n whose r e c i p r o c a l i s i n t e g r a b l e o n a n i n t e r v a l c a n n o t v a n i s h on t h a t i n t e r v a l . (Cf.
C11.)
SEMI - I N V E R S E
PROBLEMS O F N O N L I N E A R E L A S T I C I T Y
17
6. Qualitative Behavior o f Solutions.
Since
(4.5)
o r the equivalent system ( 5 . 6 ) ,
(5.7) is
autonomous, we can readily determine the qualitative behavior of all solutions of these equations by studying their phase-
plane trajectories.
For simplicity we fix the parameter
at a value for which
aJ/af
> 0
(cf. ( 4 . 1 0 ) - ( 4 . 1 7 ) )
1
and we
assume that
(6.1)
D ( f / , f , ~ )
as
+
f +
{+om].
Then the algebraic equation
o(f',f,h) =
(6.2)
0
has a unique solution
We sketch the curve defined by (6.3) in Fig. 1.
Figure 1 We are now ready to s t u d y the
(5.6),
(5.7).
j-rnplies that
(f,q)
phase-plane diagram o f
W e first note that our construction o f
f'
cp
> 0. Thus there are no singularities f o r
18
Y
STUART S. ANTMAN
>
0.
left f
of
Moreover
(5.7)
t h e image o f
i s t o the right.
says t h a t
the curve
q‘
f =
< 0 (f’
when
,x)
the rurve ( 6 . 2 )
and
i s to the
q’
> 0
when
U s i n g t h e s e i d e a s we s e e t h a t t h e p h a s e -
p l a i i e d i a g r a m h a s t h e c h a r a c t e r shown i n F i g . of
f
or
2.
The image
( 6 . 3 ) , which i s g i v e n by all t h e
(f,q)
satisfying
(6.4a) or
(6.’kb) i s i n d i c a t e d b y t h e dashed l i n e . ‘1
= o
Figure 2
SEMI-INVERSE PROBLEMS O F WONLI'JEAR
Suppose t h a t
t h e boundary ( - 0 n c 1 i t i o n s a r e
c a n d i d a t e s for t h e s o l u t i o n s of a r e the t r a j e c t o r i e s
123, 2 3 , h i 6 ,
a segment o f u n i t
( J. 8 a ) , ( 5.9a)
. Then
t h i s b o u n d a r y v a l u e problem
c a n d i d a t e would be a s o l u t i o n i f traverses
19
ELASTICITY
5 6 , c t c . of F i g . 2 .
A
the independent v a r i a b l e
length as the point
x
(f,y)
traverses the indicaked t r a j e c t o r y . (From t h e pseudo-monotone
operator analysis described i n the
l a s t s e c t i o n w e know t h a t t h e r e i s a t l e a s t one s o l u t i o n f o r
X .)
each
Fig.
2 t e l l s us that
different solutions that
1 2 3 , has t h e s t r e s s
q
there a r e t w o q u a l i t a t i v e l y
arc possible: d e c r e a s e from
The f i r s t , of
x = 0
t o an i n t e r i o r
x = 1.
minimum and t h e n i n c r e a s e t o i t s v a l u e a t of
the f o r m 23, has the s t r e s s increase with
7.
F u r t h e r R e s u l t s and Comments.
t h e form
The second,
x.
-
A l l our r e s u l t s a r e v a l i d f o r non-homogeneous, materials
aeolotropic
with t h e property t h a t t h e r e s u l t i n g equations a r e still
ordinary d i f f e r e n t i a l equations with
independent variable x .
S u i t a b l e a e o l o t r o p i c m a t e r i a l s would even y i e l d autonomous e q u a t i o n s ( c f . [ 7 ] ) . I n place o f prescribing some r e s u l t a n t
A,
B,
C,
D
w e c o u l d prescribe
f o r c e s and moments on t h e f a c e s
y = 0,1, z=O,l.
Then t h e u n s p e c i f i e d c o n s t a n t s a r e t o be d e t e r m i n e d f r o m a system o f e q u a t i o n s o f
(7.1)
t h e form
'I
T ( f ' , f , h ) d x = const.
By u s ng t h e s t r o n g e l l i p t i c i t y c o n d i t i o n or
-
k
( 2 . 7 ) with
2 = J
we o b t a i n a m o n o t o r i i c i t y c o n d i t i o n , which w h e n coupLd
STUART S. ANTMAN
20
with mild growth conditions, enables us to solve appropriate parameters as functionals of ing parameters.
f
(7.1) for
and the remain-
These functionals may be substituted into
(4.5) to convert it into a functional-differential equation. A preliminary analysis indicates that the resulting equakion
together with reasonable boundary conditions generates both a pseudo-monotone operator equation on a suitable Sobolev space and an equation on
C1([O,l])
involving the sum of the
identity plus a compact operator.
Thus these problems can be
readily treated by the methods of analysis described in Section
5. The difficulty with the singular character of the
cubical reference state disappears in "Family 2 " of semi-inverse problems (as categorized in [ 1 3 ] ) .
Here the reference
configuration may be taken as a body described in cylindrical polar coordinates z1 < z < z 2 .
(r,B,z)
by
r
1
< r < r2,
8 , < 8 < €I2,
The deformation is defined by a representation
obtained from (3.1) by replacing
(x,y)
by
(r,B).
Because
the independent variables are polar coordinates, the resulting equations will be singular at
r = 0.
This singularity
is manifested i n a semi-inverse problem when
rl = 0 ,
i.e.,
when the reference configuration contains the material line r = 0.
When
E (0,n)
U (n,m)
this singularity might
cause serious analytical difficulties because the domain has a corner.
n
It is not so obvious that tliis singularity can
cause serious difficulty when a segment o f a solid cylinder.
B2-01 = m , i.e.,
L-2
is
I n this case the analogs
of
when
the various completely continuous operators used i n the
21
SEMI-INVERSE PROBLEMS O F N O N L I N E A R ELASTICITY
analysis o f Section
6 may f a i l t o he c o m p l e t e l y c o n t i n u o u s r = 0.
when t h e m a t e r i a l i s a e o l o t r o p i c n e a r i s o i r o p y f o r s u c h problems t r e a t s t h e h u c k l i n g of When
rl
The r o l e o f
i s examined i n d e t a i l i n
[ 7 1 , which
a circular plate.
> 0 , t h i s problem d o e s n o t a r i s e .
The
5 c a n he a d a p t e d w i t h o u t change t o t r e a t
methods o f S e c t i o n
I n particular,
t h e c o r r e s p o n d i n g s e m i - i n v e r s e problems.
the
p r o c e d u r e b e g i n n i n g w i t h t h e weak e q u a t i o n s p r o d u c e s a soluticn
t o t h e e v e r s i o n problem,
d i s c u s s e d by T r u e s d e l l [ 1 2 ]
elsewhe=
i n t h i s volume.
T h i s problem c a n h e s o l v e d u n d e r t h e
requirement
that
there he zero r e s u l t a n t
t h e ends of
the tube.
equations of t h e problem.)
t h e form
f o r c e and moment a t
(This r e s t r i c t i o n yields
(7.1)
I n general,
by v i r t u e o f
.just t w o
t h e symmetries o f
t h e r e w i l l be n o n - z e r o
tractions
r e q u i r e d o v e r t h e end f a c e s i n o r d e r t o e n s u r e t h a t of
the c i r c u l a r c y l i n d e r s
This semi-inverse
z e r o end t r a c t i o n s .
remain c i r c u l a r cylinders.
r = r 1,r2
p r o c e s s would n o t
t h e images
treat
t h e problem w i t h
The f l a r i n g t h a t T r u e s d e l l d e s c r i b e s ,
which o c c u r s u n d e r z e r o end t r a c t i o n s , singular perturbation of
seems t o r e p r e s e n t a
t h e s e m i - i n v e r s e s o l u t i o n and t o
o f f e r a n e x p e r i m e n t a l v e r i f i c a t i o n of a S t .
Venant P r i n c i p l e
f o r t h i s problem o f n o n - l i n e a r
This e v e r s i o n
problem and o t h e r s o f
elasticity.
Family 2 w i l l b e d i s c u s s e d elsewhere.
( A l t h o u g h t h e g o v e r n i n g e q u a t i o n s of a r e n o t autonomous,
t h e problems
t h e q u a l i t a t i v e b e h a v i o r of
of F a m i l y 2
t h e i r solutions
may b e s t u d i e d by means of P r f l f e r t r a n s f o r m a t i o n s and o t h e r a p p a r a t u s u s e d i n t h e s t u d y of S t u r m - L i o u v i l l e
systems.
STUART S.
22
ANTMAN
The r e s u l t s r e p o r t e d h e r e s u g g e s t t h a t a l l r e a s o n a b l e boundary v a l u e problems f o r q u a s i l i n e a r systems of o r d i n a r y d i f f e r e n t i a l equations describing semi-inverse
problems of
compressible e l a s t i c bodies ha>e c l a s s i c a l solutions.
o n a f o r m u l a t i o n i n m a t e r i a l c o o r d i n a t e s and
analysis relies
on t h e e x p l o i t a t i o n o f Many of
the strong e l l i p t i c i t y condition,
t h e s t u d i e s of semi-inverse
employed s p a t i a l d e s c r i p t i o n s , equations
(cf.
s i m p l i c i t y of
Our
[7,l3]).
problems i n t h e 1950's
which y i e l d o s t e n s i b l y s i m p l e r
Bu1. t h e s e f o r m u l a t i o n s o b s c u r e t h e
the strong e l l i p t i c i t y condition.
At
one s t a g e
o f o u r a n a l y s i s w e made a c o n s t i t u t i v e a s s u m p t i o n t h a t i s i m p l i e d b y t h e Coleman-No11
inequality.
I t s use could be avoided
but a p e r i p h e r a l r o l e i n our bork: by u s i n g a c o n t i n u a t i o n t h e o r e m b a s e d (cf.
[41)
This i n e q u a l i t y plays
on weak er h y p o t h e s e s
o r by u s i n g t h e pseudo-monotone
operator theory.
8. R e f e r e n c e s [l]
S.S.
Antman,
O r d i n a r y D i f f e r e n t i a l E q u a t i o n s of
Non-linear
E l a s t i c i t y 11: E x i s t e n c e a n d R e g u l a r i t y T h e o r y f o r C o n s e r v a t i v e Boundary V a l u e Problems, Mech. [2]
S.A.
Antman, Arch.
[ 3 ] S.S.
Anal.
Arch. R a t i o n a l
6 1 ( 1 9 7 6 ) 353-393.
B u c k l e d S t a t e s o f N o n l i n e a r l y E l a s t i c Plates,
R a t i o n a l Mech.
Antman & H .
BrGzis,
Anal.,
67(1978)
111-149.
The E x i s t e n c e o f O r i e n t a t i o n -
Preserving Deformations i n Nonlinear E l a s t i c i t y , Research Notes i n Mathematics, ed. R. Knops, Pitman, London,
t o appear.
SEMI-INVERSE PROBLEMS O F NONLINEAR ELASTICITY
B a l l , Convexity Conditions
J.M.
i n Non-linear
-63 _ ( 1977) \J.L. E r i c k s e n ,
and E x i s t e n r e Theorems
E l a s t i c i t y , Arch.
337-'ho'3
R a t i o n a l Mech.
Anal.,
-
Deformations P o s s i b l e i n Every I s o t r o p i c ,
Incompressible,
P e r f e c t l y E l a s t i c Body.
Z.A.M.P.
5
( 1 9 5 4 ) , ,466-486. E r i c k s e n , D e f o r m a t i o n s P o s s i b l e i n E v e r y Compressible,
J.L.
Isotropic, Perfectly Elastic Material,
J . Math.
Phys.
-qb _ (195'1)~ 126-128.
A.E.
G r e e n & \J.E.
Non-linear Oxford,
L a r g e E l a s t i c D e f o r m a t i o n s and
Continuum M e c h a n i c s , C l a r e n d o n P r e s s ,
1960.
Knowles,
J.A.
Adkins,
L a r g e AmpLitude O s c i l l a t i o n s
Incompressible E l a s t i c Material, (1960, J.
o f a Tube of
Appl.
Q.
Math.
18 -
71-77.
Leray & J.
Schautler, Topologie e t Gquations
fonctionelles,
Ann.
S c i . E c o l e Norm.
Sup ( 3 ) j l
-
(1934), 45-78. Rabinowitz,
[lo] P . H .
v a l u e Problems
A G l o b a l Theorem f o r N o n l i n e a r E i g e n -
and A p p l i c a t i o n s ,
l i n e a r Functional Analysis, Academic P r e s s , New Y o r k ,
[ll] R . S .
e d . E.H.
Zarantonello,
1971.
R i v l i n , Large E l a s t i c Deformations of I s o t r o p i c
Materials,
Parts I V , V , V I , P h i l . T r a n s . Roy.
London A 2 4 1 (1948) A
C o n t r i b u t i o n t o Non-
379-397,
Proc.
Roy.
SOC.
S O C . London
195 ( 1 9 4 9 ) 4 6 9 - 4 7 3 , P h i l . T r a n s . Roy. S o c . London ( 1 9 4 9 ) 173-195.
A 242
24
[l21
STUART S. ANTMAN
C. Truesdell, Comments on Rational Continuum Mechanics, i n this volume.
[13l
C. Truesdell & W.
Noll, T h e Non-Linear Field Theories o f
Mechanics, Handbuch d e r Physik, III/3, Springer-Verlag, Berlin, 1965.
G.M.
de La Penha. L.A. Medeiros (eds.) Contemporary Developments i n Continuum Mechanics and P a r t i a l D i f f e r e n t i a l Equations QNorth-Holland P u b l i s h i n g Company (1978)
ASYMPTOTIC PROPERTIES OF GENERAL SYMMETRIC HYPERBOLIC SYSTEMS
L V I L A and D.G. COSTA
G.S.S.
Departamento de Matemgtica Universidade de Brasilia Brasilia, DF
-
Brasil
1. Introduction
We consider the Cauchy problem
-a +U at
n C
A . aU --= .I a x . j=1 J
where the A .Is are constant J
u
and
f
x = (xl,
are column vectors o f
... ,xn) E
the time.
n,
R
Hermitian symmetric matrices, k
components,
is tlie space variable and
t E IR
is
We use the Radon transform to study the behavior
of the solution times.
kxk
0,
u
for large (positive as well as negative)
We obtain results on the energy distribution o f
according to the several components of
f
u
(Theorem 3.1 and
Corollary 3 . 2 ) , o n energy decay (Theorem 4.1 through b . h ) , and on the existence o f wave operators, which are defined by comparing equation (1.1) (regarded as unperturbed) with a perturbed equation au
+ at
E(x) -
11
aU C A. j ax = O. j=l j
26
(Theorem 5 . 1 ) .
COSTA
A V I L A and D . G .
G.S.S.
These a r e e x t e n s i o n s of r e s u l t s p r e v i o u s l y
o b t a i n e d by t h e a u t h o r s [
11,
[2] and by Wilcox [ 91.
Proofs
w i l l appear elsewhere. Systems
(1.1) have b e e n d i s c u s s e d i n t h e l i t e r a t u r e
X
u n d e r c e r t a i n a s s u m p t i o n s on t h e r o o t s
= h(p)
o f the
c h a r a c t e r i s t i c equation P ( h , p ) = det(X1
n
-
C
p.A.) = 0
j=1
(see
[43, [ 51, [ 9 ] ) .
symmetric,
A
J
I t should be n o t e d t h a t o u r r e s u l t s a r e
v a l i d f o r g e n e r a l s y s t e m s of on t h e m a t r i c e s
3
type
( l . l ) , w i t h no r e s t r i c t i o n s
o t h e r than t h a t
j’
they a r e Hermitian
t h e r e f o r e w i t h no r e s t r i c t i o n s on t h e r o o t s
X(p).
Of c o u r s e , t h i s i s i m p o r t a n t n o t o n l y from a p u r e l y t h e o r e t i c a l p o i n t of view, but a l s o f o r a p p l i c a t i o n s , s i n c e t h e r e a r e c a s e s , s u c h a s i n magnetogasdynamics the r o o t s
X(p)
[ 6 1 , [S]
,
where
d o n o t s a t i s f y t h e r e s t r i c t i o n s u s u a l l y made
i n the l i t e r a t u r e . I n what f o l l o w s of
i s g i v e n by
f E Ho,
u(t,*) = Uo(t)f,
u n i t a r y group i n
w i l l i n g e n e r a l d e n o t e an e l e m e n t
k H o = L2(Rn)
t h e H i l b e r t space
It i s known t h a t i f
f
Ho
,
w i t h norm d e f i n e d b y
the solution where
(see [9]).
Uo(t)
u
of
(1.1)-(1.2)
i s a one-param*
Such a s o l u t i o n i s c a l l e d a
s o l u t i o n w i t h f i n i t e energy, s i n c e i t s energy, defined as 2
Ilu(t,.)llo,
is finite:
27
SYMMETRIC HYPERBOLIC SYSTEMS
The group
is generated by the operator
Uo(t)
-iA,
where k A = -i"L
A .
j=1
D(A) = { f € H o :
a ~
J ax. J
Af E H o ] .
We observe that under the Fourier transformation,
A
is unitarily equivalent to the operator of multiplication
by the matrix
.
k
A(p) =
Z j=1
p.A. J J'
that is,
Af = 5-l [A( .)5f]
D(A) =
{r
E H ~ :A ( . ) ? : E H ~ ] .
2. The Radon transform and solutions of the Cauchy problem. -.
In this section we shall use the Radon transform in order to derive an expiicity formula for the solution of The Radon transform provides a translation
(1.1)-(1.2).
representation for the free solution group
Uo(t),
and this
is a key fact in establishing the results anounced i n the following sections.
If
f
is a function i n the Schwartz space of rapidly
decreasing functions,
8(Rn),
n
2
3,
its Radon transform
= @ f is defined by the formula (see [ S l y [ 5 1 y
?(S,uJ)
=
f
(xIdsx 7
The following properties are valid:
x E IR,
*i
c71)
E Sn-l.
28
A V I L A and D.G.
G.S.S.
COSTA
Also, we point out the following relation between the A
Radon transform
where
5,
and the Fourier transform
f
-f:
denotes Fourier transformation i n the variable Now, for
n
2
3
odd, let
operators
a
K = a
K
L
and
s.
denote the
n-1
(-)
n as
where
a > 0 n
is a suitable constant.
Also,
for
n > 3
even, let
where
bn
shown ([ 31
is another suitable constant.
, [ 51 ,
[ 71 )
Then, it can be
that the following Inversion Theorem
and Parseval Theorem hold: Theorem 2.1 (2.3)
-
Let
f
E 8 (IRn),
f(x) =
I
Ilu
n
2
3.
Then
(K;)(x-~,a)dw,
I =I
x E Rn.
29
SYMMETRIC HYPERSOLIC SYSTEMS
-
Theorem 2.2
@ = Lm.
Let f
I f(x) I
(2.14)
Then, f o r any
' Rn @
In fact, onto
that
-
K = L
'l*rl=l
2
n
3,
5
l@f(s,J;)I2 dsds.
I-m
extends to a unitary mapping from all of
L2(Rn)
.
L~( [ ~ s xn-')
Remark 2.3
f"
dx =
f E 8(Rn),
I t follows from the definitions (2.2)0 and (2.2)e
.
Ipln-'
Also, since
the definition of
K
= sgn(p) p
in this case differs -1
ki = 5,
o d d by the Hilbert transform
n-1
for
even,
n
from that for
sgn(*)Zl
n
(up to a
constant). Now, let datum
f
u(t,*) 8(Rn)
belongs to
C(t,.,-)
=
?.
k
.
Then, i n view of (2.1)
a; t aC = a + A(&) as
0
n C
with corresponding eigenvalues
A(&)
and setting
we diagonalize (2.5) into the scalar equations
a;.
&+ &.(O,s,u))
J
a;.
x J. ( W ) A as
= gj(s,w),
= 0,
J
J
I S j S k.
w e obtain
C.(t,S,W) = P.(s-x.(W)t,&), hence
(iii),
w .A The initial datum is n o w 6 ( 0 , - , * )= j=l J j' S o , denoting b y [ej(u)] a complete set of orthonormal A(w) =
eigenvectors o f
Since
-
satisfies the eyuation
(2.5) where
be a solution of (1.1) whose initial
J
1 5 j S k ,
xj(d),
30
AVILA
G.S.S.
and D.G.
COSTA
(2.7)
As we see, the Radon transform provides a translation representation f o r the solutions of (1.1). the eigenvalues order.
and eigenvectors
Xj(;u)
First, we are enumerating the
A few words about
e.(a)
are now in
J
h j ( a ) I s
which are not
identically zero in decreasing order (counting multiplicity): X,(m)
(2.8)
hence setting
Xr+l(~)
...
2
... =
=
2
Xr(JJ),
= 0
A,(*)
this way, it is a known fact that the functions (see
[ l O l , Theorem
i n case
1.Is
r < k.
In
are continuous
.J
O n the other hand, it has
1).
been proved by Wilcox ([ 101, Theorem 2) that the corresponding (orthonormalized) eigenvectors measurable functions,
u(t,x) =
Remark
-
2.4
J
can be chosen to be
Assuming that this choice has been made,
(2.7) and obtain
we use Theorem 2.1 to invert
12-91
e.(a)
“i
z
j=1
~
~
( Ki‘ ) ( x - w - X . (a) t ,a ) e . (3) d W ,
l
=
~J
J
The non-zero eigenvalues i n (2.8) can be re-
numbered as w,(m)
...
5
2
pL(w),
L
5
r,
where all the p . ( u J ) ! s are distinct, except f o r J
No
of measure zero i n
1 < j
S
L ,
projection o f =
[ 101
.
We denote by
the orthogonal projection of
space associated with
Xr+l(w)
Sn-1
Ck
pj(w),
and by
iu
Ck
in a set
Pj ( a ) ,
onto the eigen-
P0(w)
the orthogonal
onto the eigenspace associated with
... = Xk(;u)
E
0,
make the convention that
i n case
P0 ( w )
=
0,
r < k Cr0(w)
(if
=
0).
r = k
we
Then, we
SYMMETRIC HYPE RRO L I C SYSTEMS
31
(2.10) and
respectively.
3.
Partition of
energy.
F o r any g i v e n
matrix
kxk
M,
we c o n s i d e r t h e
bounded o p e r a t o r s
Ho = L2(Rn)
mapping
k
we d e f i n e t h e M-energy
where
into f
of
i s t h e norm i n
-
If
u(t,.)
L2(IR~Sn-l)
t h e n t h e e n e r g y of
of
as
It]
E
S
t Ho,
Then, we 'nave: (1.1) w i t h i n i t i a l
t o t h e M-energy
tends
Mu(t,*)
m:
In f a c t , there exists a s e t f
.
i s a s o l u t i o n of
f,
+
f-
by
datum f
Also for
L2(iRxSn-1)k.
k
11) *II[
Theorem 3 . 1
each
1 s j s k,
M . = MP.(*)@, J J
(3.1)
we h a v e
S,
dense i n
2 l l M u ( t , * ) ( j o= EM[f]
HO
,
such t h a t f o r
for a l l
It[
sufficiently large. C r o_ l l a_r y_ 3 .~2 _ o_
-
F o r any f i n i t e e n e r g y s o l u t i o n
(1.1) t h e r e e x i s t s
E . = E .[u(O, *)] J J
,
j=l,
... , k ,
u(t,x)
of
such t h a t
4.
A V I L A and D.G.
G.S.S.
32
COSTA
Energy decav.
A solution
non-static
if
with f i n i t e energy i s c a l l e d
Uo(t)f
E ( k e r A)'.
f
R e c a l l i n g (2.l)(iv) a n d Remark
2.4, i t i s n o t h a r d t o s e e t h a t t h i s is t h e c a s e i f P o ( ~ ) ? ( s , , u )E
0.
i s non-static
if
Thus, i n view o f and o n l y i f
(2.10),
has
u(t,*) = U(t)f
the representation
.e. G(t,s,s) =
(4.1)
T j=1
Theorem 4__ .1 ( k e r A)', (i) (ii)
cone
-
P.(JJ)P(S-V . ( o ) t , w ) . J J So
There e x i s t s a s e t
C
(ker A ) I ,
dense i n
such t h a t Uo(t)So =
so
f o r each
f
1x1 5 a l t l - R ,
for all
E So, It1
t E R;
Uo(t)f
R/a,
2
v a n i s h e s i n some d o u b l e -
where
a = a ( f ) > 0,
R = R ( f ) > 0. T h i s r e s u l t is u s e d t o p r o v e t h e e n e r g y d e c a y s t a t e d i n the following: Theorem 4 . 2
-
G i v e n a bounded m e a s u r a b l e s e t
for any n o n - s t a t i c
solution
More g e n e r a l l y , i f increase "too fast" as any n o n - s t a t i c
t
-b
u.
B(t) m,
i s a s e t t h a t does not
then t h e energy i n
s o l u t i o n also d e c a y s t o z e r o ,
p r e c i s e l y i n the following:
B c Rn,
B(t)
as stated
of
SYMMETRIC 1IYPERBOLTC SYSTEMS
Theorem
'1.3
-
Let
t > O}
(B(t):
e(t)
m e a s u r a b l e s e t s such t h a t
8 ( t ) = sup{ 1x1: x E B ( t ) ) .
37
be a f a m i l y o f hounded = o(t)
t
as
-+
dx = O,
t-+ m
t-+ m f o r any n o n - s t a t i c
sol.utio11
u.
F o r t h e n e x t r e s u l t we n o t e t h a t
g = (I-P
and
)f E
( k e r A)'.
f E Ho
any
f = g+h,
decomposed u n i q u e l y i n t h e form
B
If
C
a s t h e energy o f
On t h e o t h e r h a n d ,
in
h
B,
0
Rn
measurable s e t s converges s l o w l y t o
where t o t h e c h a r a c t e r i s t i c i n this situation, tends t o the value Theorem
4.4
-
If
Eg[ f ]
B(t)
B
B(t)
as
t +
B(t) m,
{B(t): if
i n_ _
B,
t > 0)
e(t)
= o(t)
of and
converges almost cvery-
f u n c t i o n of
t h e energy i n
i s any
that i s ,
we s a y t h a t a f a m i l y
the c h a r a c t e r i s t i c function o f
can b e
h = P f' €
where
m e a s u r a b l e s e t , we d e f i n e t h e _ s t a_t i c. _ e n_ e r_ g y_of f _ Ei[f],
bhere
Then,
l i m Ilu(t,
E ker A
m,
B, of
as
t +
m .
any s o l u t i o n Uo(t)f
that is:
converges slowly t o
Then,
B,
then
5 . The p e r t u r b e d e q u a t i o n and t h e wave o p e r a t o r s . W e now c o n s i d e r t h e p e r t u r b e d Cauchy problem
A V I L A and D . G .
G.S.S.
34
COSTA
u(0,x) = f ( x ) ,
(5.2)
i s a positive definite
E(x)
where
kxk
H e r m i t i a n symmetric
matrix s a t i s f y i n g the following properties: t h e r e e x i s t p o s i t i v e c o n s t a n t s c and c’ s u c h t h a t
i)
(5.3)
CI
ii)
s ~ ( x s) e
there exists
for a l l
>
such t h a t
0
1x1 +
= ~(~xl-l-‘) as I n view o f c o n d i t i o n
11 * ] l o
Letting
(5.3) ( i )i t i s c l e a r that the
/I - 1 1 ,
d e f i n e d by
E(x)f(x)-f(x)dx.
€lo = L2(Rn)
denote the H i l b e r t space
H
1) * I ) ,
norm
b;.
=
IE(x)-II =
m.
norm i s e q u i v a l e n t t o t h e norm
II f l l
x;
C’I
k
we d e f i n e t h e o p e r a t o r A~ = E ( X ) -1 A ,
It follows that
-iAE
= D(A).
D(A,)
i s a self-adjoint
o p e r a t o r on
t h e f i n i t e energy s o l u t i o n o f
(5.1)-(5.2)
u(t,-) = U(t)f,
i s t h e one-parameter
group i n
where
U(t)
generated by
H
-iAE
The wave o p e r a t o r s groups
with t h e
Uo(t)
and
U(t)
W+,
H
and
i s g i v e n by
unitary
(see [9]).
W-
associated w i t h the
a r e d e f i n e d by
(5.4) Po = I - P
where
Theorem 5.1 exist
.
-
i s the projection of
Under t h e h y p o t h e s i s
Ho
onto
(ker A I L .
( 5 . 3 ) , t h e wave o p e r a t o r s
35
SYMMETRIC HYPERBOLIC SYSTEMS
Remark 5 . 2
(5.4)
-
Letting
€*
and the u n i t a r i t y of
l i m t+*m
f E Ho,
= W*f,
that
U(t)
lIU(t)f*
i t f o l l o w s €rom
- U ~ ( ~ ) P ~ ~ =/ I 0.
T h e r e f o r e , by c o n s i d e r i n g s o l u t i o n s o f
(5.1)
of
t h e form
i t i s now e a s y t o s e e t h a t m o s t o f o u r p r e v i o u s
U(t)f*,
results
( n a m e l y Theorem 3 . 1 ,
through
4.4)
C o r o l l a r y 3.2
a n d T h e o r e m s 4.2
are s t i l l valid f o r the perturbed equation
(5.1)
(5.3).
under t h e hypothesis
References -
111
G.S.S.
A v i l a and D.G.
symmetric hyperbolic equations
[z]
D.G.
costa,
Decay o f
Costa,
systems of p a r t i a l d i f f e r e n t i a l
( t o a p p e a r i n Rocky M o u n t a i n J .
of Math.).
O n p a r t i t i o n of energy f o r uniformly
propagative systems, J. M a t h . pp.
solutions of
Anal.
Appl.
52 (1977),
56-62.
[ 3 ] S. H e l g a s o n ,
T h e Radon t r a n s f o r m o n E u c l i d e a n s p a c e s . . . ,
A c t a Math.
113 ( 1 9 6 5 ) , p p . 93-106.
[ b ] T. I k e b e , S c a t t e r i n g f o r u n i f o r m l y p r o p a g a t i v e s y s t e m s ,
[5]
P.D.
Proc.
Int.
Tokio
( 1 9 6 9 ) , p p . 225-230.
Lax a n d R . S . Press,
C6] D .
Conf'.
Ludwig,
on Func.
Phillips,
Anal.
and Related T o p i c s ,
S c a t t e r i n g T h e o r y , Academic
1967. The S i n g u l a r i t i e s o f
NYU R e s a r c h R e p o r t
(1961).
t h e Riemann F u n c t i o n ,
G.S.S.
36
AVILA and D.G.
[71 D. Ludwig, The Radon transform Cornm.
[8] S.I. Pai,
COSTA
on Euclidean spaces,
Pure Appl. Math. 19 (1966), pp. 49-81. Magnetogasdynamics and Plasma Dynamics,
Springer, 1962.
[9]
C.H.
Wilcox,
Wave operators and asymptotic solutions o f
wave propagation problems of classical physics, Arch. Rat. Mech. Anal. 22 (1966) PP. 37-78. [lo] C.H. Wilcox, Measurable eigenvectors f o r Hermitian matrix-valued polynomials, J. Ma h. Anal. Appl. 4 0
( 1 9 7 2 1 , ~p.12-19.
G.M. de La Penha, L.A. Medeiros (eds.) Contemporary Developments i n Continuum Mechanics and P a r t i a l D i f f e r e n t i a l Equations @North-Holland Publishing Company (1978)
THE CAUCHY PROBLEM FOR THE COUPLED
SCHROEDINGE R-KLE IN- GORDON E QUATI0N S
and
J.B. Baillon
John M .
Department of Mathematics
Dept. de Math6matiques &ole
Chadam*f
Pontificia Universidade
Polytechnique
Cat6lica do Rio de Janeiro Paris, France
Rio de Janeiro, Brasil
Introduction. The classical versions of the non-linear equations o f relativistic quantum physics have for a long time been a subject of great interest.
One o f the major unsolved problems
in the area has been the global existence o f the solution t h e Cauchy problem in
R3
to
o f certain important equations
-
having quadratic non-linearities
the coupled Dirac-Klein-
Gordon equations,
P (-iY au + ~ ) =t g $ @ ,
-
( A - a tt
2
m ) @=
6t1
and the coupled Maxwell-Dirac (-i?
)r
+
a,,
+
O n leave from Mathematics Department, Indiana University, Bloomington, Indiana 47401.
Supported in part by NSF grant MPS 73-08567
J.B. BAILLON and J.M. CHADAM
38
Here we consider the Cauchy problem for the closely related coupled Schoredinger-Klein-Gordon ( S K G ) equations which are a semi-relativistic version o f equations (1):
with nucleon field
8 : R 3X R
-b
C
and meson field
@:
R'xR
4
R
I n previous works the Cauchy problem has been solved for equations (1) and (2) i n one space dimension [l] and for special cases in higher dimensions [2,31.
Equations ( 3 ) have
recently been treated i n bounded regions of
R3[
4,5]
.
The Results. We begin by writing equations ( 3 ) i n vector form,
from
R
JI
(g
) are considered as maps t to complex and R2-valued functions respectively. The
where the components
and
conventional solution spaces in which one wishes to solve equations
(4) for the components
"escalated energy" spaces
C m = Hm
( $ y(:t)) @
(HmCBHm'')
are the so-called [l].
Indeed,
following the example of quantum mechanics, it is really the exponentiated o r , in the non-linear case, the integrated form of equation
(k),
-
SC HR O E D IN GE R K LE IPJ - GO R D 0N E QUA T I0 N S
which a r e of
interest.
S(t) = e
Here
i A t
and
a r e t h e f r e e S c h r o e d i n g e r and Klein-Gordon respectively.
(+ ,(:
Specifically then,
39
propagators
))
is a
Cm solution
t of
t h e SKG e q u a t i o n s o v e r t h e i n t e r v a l
($,(:
)):
(0,T)
4
Cm
i s continuous
(0,T)
i f t h e map
and s a t i s f i e s e q u a t i o n s
t
( 5 ) where t h e i n t e g r a l i s i n t e r p r e t e d i n t h e s t r o n g Riemann
zm.
sense i n
-
Theorem
solution i n Proof:
zm
if
problem.
of
+
S ( t ) : tim
a r e continuous, uniformly
( a , (m 0 12 ) )
rn 2 2 .
-+
boundedness,
[ 6 1 c a n be a p p l i e d t o this
and
-+
K( t ):
bounded o n e - p a r a m e t e r
, ( O
IJI
I
2)): Zm + Cm
H
~
~
g r o u p s and
i s l o c a l l y Lipschitz
t h i s f o l l o w s from t h e l o c a l
o f t h e map v i a some t r i v i a l a l g e b r a .
The l o c a l
on t h e o t h e r h a n d , f o l l o w s f r o m t h e f a c t t h a t
i s an a l g e b r a i f
dimension
(@,$
H~
One c a n s h o w t h a t
boundedness
Hm
Segal
The l o c a l e x i s t e n c e f o l l o w s i n t h i s way b e c a u s e t h e
f r e e propagators
if
m)
2.
m 2
The a b s t r a c t r e s u l t
t h e map
T =
The SKG e q u a t i o n s h a v e a u n i q u e g l o b a l ( i . e .
m > d/2
where
d
i s the s p a t i a l
"71.
The main c o n t r i b u t i o n o f t h i s n o t e i s t o show t h a t this s o l u t i o n can b e e x t e n d e d t o
T =
p r e v i o u s l y mentioned r e s u l t
of
+m.
This o b t a i n s f r o m t h e
Segal i f t h e
Zm
-
norm of
s o l u t i o n c a n be s h o w n t o be f i n i t e f o r e a c h f i n i t e t i m e .
the We
H
~
-
J.B.
40
B A I L L O N and J . M .
CHADAM
b e g i n with t h e p h y s i c a l l y r e l e v a n t conserved q u a n t i t i e s o f c h a r g e and e n e r g y and t h e n s y s t e m a t i c a l l y b o o t - s t r a p o u r way
t o the xm-norm.
Specifically
[4,
p.4031
one h a s
4
3x = c o n s t a n t . I $ ( x , t ) 1 2 @ -b( x , t ) d F r o m t h i s one h a s
t h e second l i n e f o l l o w i n g f r o m H s l d e r ' s
i n e q u a l i t y and t h e
l a s t l i n e from S o b o l e v i n e q u a l i t i e s and t h e f a c t t h a t li$(t)ii 2 i s conserved.
Now
Combining t h e above t w o e s t i m a t e s , one o b t a i n s
Adding e q u a t i o n ( 6 a ) t o i n e q u a l i t y
(7),
one o b t a i n s on t h e
l e f t h a n d s i d e t h e s q u a r e o f a norm which i s e q u i v a l e n t t o t h e
Cl-norm
t h u s p r o v i n g i t i s u n i f o r m l y bounded i n s p i t e o f
non-definiteness
of
t h e i n t e r a c t i o n energy.
the
SCHROEDINGER-K LE I N - GORDON EQUATIONS
and m o s t c r u c i a l ,
The n e x t ,
Proof:
s t e p i s t o show
Here we f o l l o w c l o s e l y t h e t e c h n i q u e of B a i l l o n e t .
C 8 , 9 I by showing t h a t
a1
~ ~ $ ( t ) ~i s~ fl i,n4i t e and t h e n u s i n g
111
t h e Sobolev i n e q u a l i t y
( t ) l l mS
( t ) l ( z / 4 119 (t)ll
C11 vp
to
To b e g i n
establish the desired result.
A l l t h a t r e m a i n s t h e n i s t o e s t a b l i s h t h e f i n i t e n e s s of
IlVh(t)114.
Let
D
d e n o t e a n a r b i t r a r y weak s p a t i a l d e r i v a t i v e .
D i f f e r e n t i a t i n g e q u a t i o n ( 5 a ) , u s i n g t h e L e i b n i t z f o r m u l a and t a k i n g norms
D
since
one o b t a i n s
commutes w i t h
S(t).
U s i n g t h e well-known
e s t i m a t e f o r t h e Schroedinger p r o p a g a t o r [lo,
decay
p.601
1 1 -d(-- -) IlS(t)fIlp s
for a l l
f
E Lq(Rd)
where
C't
+
l/p
l/q
= 1
and
1
2;
q S
2,
one o b t a i n s f r o m i n e q u a l i t y ( 8 )
rt IIS(t)$0112,2 +
IID$(t)l14 s
Igl C i ,
( t - s ) - 3 / 4 I I D 0 ' ~ + @ D $ l l ~ / 3d s
t s
co
+
lglc
S
Co
+
Igl
(t-s)-3/4 c I I D 0 ~ s ~ l 1 2 1 1 ~ ~ s ~ l 1 4 + l l s ~ s ~ l l , I l ~ d s~ ~ s ~ l l , 1 C{C,
c
(t-s)-3/4 ds
+
C2
c
( t - ~ ) - ' / I~I D + ( S ) ( ~d~s ] .
The r e s u l t n o w f o l l o w s f r o m a v e r s i o n o f G r o n w a l l ' s sketched i n reference Lemma
3
-
[91.
The C2-norm of
f i n i t e f o r each
t <
lemma as
m.
the solution
($(t),(st(t)
42
J.B.
Proof:
But,
B A I L L O N and J . M .
CHADAM
From e q u a t i o n ( 5 b ) one h a s
o f Lemma 2 ,
f r o m t h e proof
one h a s t h a t
/IIII(S)//~,~ is
a
l o c a l l y i n t e g r a b l e f u n c t i o n g i v i n g t h e f i n i t e n e s s o f t h e meson p a r t of
t h e norm.
The r e s u l t now f o l l o w s from t h e s t a n d a r d v e r s i o n o f Gronwall’s lemma b e c a u s e e v e r y t h i n g e x c e p t
/ID2$
(.)I\
has previously
b e e n shown t o b e l o c a l l y i n t e g r a b l e .
To t h i s p o i n t we h a v e shown t h e e x i s t e n c e of a u n i q u e g l o b a l s o l u t i o n t o t h e Cauchy problem f o r t h e S K G e q u a t i o n i n t h e space
Z2.
A l l t h a t remains i s t o prove t h e r e g u l a r i t y
p a r t of t h e r e s u l t .
-
Lemma
4
than
2,
have
114 ( t ) l l m , 2 11@(t)11m,2 and
Suppose
m
i s an a r b i t r a r y p o s i t i v e i n t e g e r l a r g e r
t h e n t h e a b o v e , g l o b a l s o l u t i o n of e q u a t i o n s
II@,(t)ll
m-l,2
locally
(5))
-
S C H R O E D I N G E R - K LE I N GORDON E Q U A T I O N S
f o l l o w s f r o m L e m m a 3 by g i v i n g t h e i n d u c t i v e
Proof:
The p r o o f
step.
Suppose then that
some
But
m 2
H
3,
m- 1
the r e s u l t
is known in
Then
m 2
for
the left-hand
47
~
m- 1
for
i
3
i s an algebra g i v i n g t h e f i n i t e n e s s of
side f r o m the inductive hypothesis.
and t h e r e s u l t f o l l o w s
from Gronwall's
Similarly
l e m m a b e c a u s e of t h e
11 a ( s)II m , 2 .
l o c a l i n t e g r a b i l i t y of
References
[l] C h a d a m , J . M .
-
G l o b a l Solutions of
the ( C l a s s i c a l )
Coupled Maxwell-Dirac
Space D i m e n s i o n ,
[2]
Chadam,
J.M.
J. Functional A n a l . ,
and G l a s s e y ,
Solutions of
-
R.T.
Equations
3,1 7 3
in One
(1973).
On C e r t a i n G l o b a l
the C a u c h y P r o b l e m for the ( C l a s s i c a l ) E q u a t i o n s i n O n e and T h r e e
Coupled Klein-Gordon-Dirac Space D i m e n s i o n s ,
[ 3 ] C h a d a m , J.M.
the C a u c h y P r o b l e m f o r
Arch.
and G l a s s e y ,
Rat. R.T.
Mech.
-
Anal.,
54
(1974).
O n the M a x w e l l - D i r a c
E q u a t i o n s w i t h Z e r o M a g n e t i c F i e l d and T h e i r S o l u t i o n
i n T w o Space D i m e n s i o n s ,
53,
495 ( 1 9 7 6 ) .
J. M a t h .
Anal.
and A p p l i c .
44
C 41
J.B. BAILLON and J.M.
Fukuda, I. and Tsutsumi, M.
-
CHADAM
On the Yukawa-Coupled
Klein-Gordon-Schroedinger Equations i n Three Space
Dimensions, Proc. Japan Acad., 5 1 ,
C 51
Fukuda, I. and Tsutsumi, M.
-
Schr8dinger Equations, 11,
C 61
Segal, I . E .
-
402, (1975).
O n Coupled Klein-Gordonto appear.
Non-linear Semi-groups, Ann. Math. 7 7 ,
339 (1963).
C 73
Grisvard P,
-
Contribution i n Proceedings o f Evolution
Equation Seminar, Nice, France
1974-5.
[ 81 Baillon, J.B., Cazenave, T. and Figueira, M.
-
Equation
de Schr8dinger Non-lingaire I, t o appear C.R. Acad. Sci
[ 91
.
Baillon, J.B., Cazenave, T. and Figueira, M.
-
Equations
de Schradinger Non-lin&aire, 11, to appear CR. Acad. Sci. [lo] Reed, M. and Simon, B.
-
Methods o f Modern Mathematical
Physics 11, Academic Press,
1975.
G.M. de La Penha, L . A . Medeiros (eds.) Contemporary Developwnts i n Continuum Mechanics and P a r t i a l D i f f e r e n t i a l Equations @North-Holland Pub1i s h i n g Company (1978)
APPLICATIONS OF GENERIC BIFURCATION THl3ORY IN FLUID MECHANICS
T.B. BENJAMIN Fluid Mechanics Research Institute University of Essex, England
1. Introduction
In thiq lecture
some
abstract theoretical results
are reviewed with the particular aim of illustrating their application to practical problems in the mechanics of fluids. From the standpoint of the practical-minded investigator into flow problems the modern theory o f partial differential equations seemingly offers much relevant information, and the complexities of the problems in question give weight to the seeming generality of this iiifsrmation.
Its precise bearing
on observed phenomena is sometimes elusive, however, and
physical insights guided by experimental information may then appear to be the more dependable, albeit limited, means of understanding.
Thus, in this subject, there are distinct and
sometime,: ill-matched stvles o f scientific inquiry, and their more effective alliance is still an outstanding objective. T h e present discussion begins by recalling J.Leray’s classic demonstration, elaborated by E . Hopf, O.A.
Ladyzhens-
kaya and others, that the boundary-value problem determining steady motions of a viscous incompressible fluid in any finite domain has at least one solution f o r all value o f the Reynolds number
R
(the dimensionless velocity parameter).
46
T.B.
By i t s e l f
this result
p r e t a t i o n s of
BENJAMIN
i s hardly helpful as regards i n t e r large R
observed s t e a d y f l o w s , p a r t i c u l a r l y a t
when t h e problem can h a v e m u l t i p l e s o l u t i o n s , may r e p r e s e n t
some o f which
s t a b l e f l o w s and o t h e r s r e p r e s e n t u n s t a b l e f l o w s .
O n t h e same b a s i s , however,
a corpus o f
additional inferences
can h e b u i l t up which p r o v i d e s f o r c o m p l e t e i n t e r p r e t a t i o n s of
v a r i o u s c o m p l i c a t e d phenomena.
To t h i s e n d , that are typical of
i t i s essential t o identify properties
e x p e r i m e n t a l l y observed f l o w s , e x c e p t i n g
s i n g u l a r b e h a v i o u r t h a t m a y h e i n c l u d e d w i t h i n t h e s c o p e of the
Comments a b o u t t h i s r e q u i r e m e n t
abstract theory.
made i n
$3.
Then,
$4,
in
a s e t of
are
typical properties
a s s o c i a t e d w i t h hounded s t e a d y f l o w s i s p r e s e n t e d .
In
$5
a
concomitant i d e a with p a r t i c u l a r p r a c t i c a l i n t e r e s t i s introduced, being explained i n t h e f i r s t p l a c e with r e f e r e n c e
t o a problem i n e l a s t i c i t y .
Finally,
in
$ 6 , some r e l e v a n t
experimental o b s e r v a t i o n s a r e reviewed.
2.
The t h e o r y of Leray-Hopf-Ladyzhenskaya
a c c o u n t of
The best-known
L a d y z h e n s k a y a ' s monograph ( 1 9 6 9 ) . recalled.
Here t h e g i s t i s m e r e l y
W e c o n s i d e r s t e a d y motions o f
a v i s c o u s incompres-
s i b l e f l u i d t h a t f i l l s a bounded smooth domain
aD
d e n o t e s t h e boundary,
i s prescribed,
and
Eulerian velocity,
5 and
D
on which t h e v e l o c i t y of
= D U aD. P:
fi +
IR
pressure corrected f o r the effect conservative force f i e l d s ) .
5 of
t h e t h e o r y i s Ch.
Let
U:
5
+ R3
in
the f l u i d he t h e
t h e gauge p r e s s u r e of
El3.
(i.e.
g r a v i t y and o t h e r
The N a v i e r - S t o k e s
equation of
47
GENERIC BIFURCATION THEORY I N FLUID MECIIANICS
motion i n dimensionless f o r m ,
t h e c o n d i t i o n of
incompressibil-
i t y and t h e boundary c o n d i t i o n a r e
-nu
+
+ R u.vU
v-U = 0
VP = 0 U = q
D,
i n
R,
The p o s i t i v e p a r a m e t e r
that
q-n = 0
aD.
on
the Rey n o ld s number,
p r o p o r t i o n a l t o t h e v i s c o s i t y of assume
D,
in
all,
U = W
i s t o put
which i s so i n t h e
later.
s p e c i f i c applications t o be considered The f i r s t
of t h e problem (2.1)
s t e p i n the treatment
-
+ q,
g
%-here
i s a solenoidal extension of
t h e boundary data t o t h e i n t e r i o r of n o i d a l and v a n i s h e s on
aD.
It i s h e l p f u l t o
the fluid.
everywhere on
is inversely
Hence,
Thus
D.
assuming
i s sole-
W
W,
;,
P
and
t h e boundary t o be s u f f i c i e n t l y smooth, m u l t i p l y i n g t h e f i r s t o f
(2.1)
compact
Let
H
by any smooth s o l e n o i d a l f u n c t i o n
D,
support i n
I
[v$:vW
-
D + kJ
with
a n d i n t e g r a t i n g b y p a r t s , we o b t a i n
+
R{{*(W*Vt) =
W-(~*V$) + W.(W*W)]ldx
-
b.(Ag W
d e n o t e t h e subspace of
192
i n t h e Hilbert space
H
(u,v) =
I
(2.2)
Rq-Vq)dx.
(D)3
c o n s i s t i n g of
aD.
s o l e n o i d a l e l e m e n t s w i t h z e r o t r a c e on product
@:
the
The s c a l a r
i s taken t o be VU:VV
dx,
D
and i n view of
t h e Poincar6 i n e q u a l i t y ,
c l o s e d u n d e r t h e norm
I/uII = ( u , u )
for arbitrary
i s evidently
A weak s o l u t i o n o f
t h e hydrodynamic p r o b l e m i s d e f i n e d as a (2.2)
H
W E H
satisfying
$ E H.
The f i r s t f u n c t i o n a l of
(I
i n (2.2)
is just
(W,a),
48
T.B.
and a n i m p o r t a n t
BENJAMIN
s t e p i n t h e t h e o r y i s t o show t h a t
remaining l i n e a r f u n c t i o n a l s a r e d e f i n e d i n
H
the
and have
needed p r o p e r t i e s a c c o r d i n g t o t h e r e p r e s e n t a t i o n theorem of Riesz.
-
q
A c h o i c e of
i s shown t o b e p o s s i b l e s u c h t h a t t h e
f u n c t i o n a l on t h e right-hand t w o involving
s i d e i s d e f i n e d , and t h a t t h e
l i n e a r l y on t h e l e f t - h a n d
W
completely continuous o p e r a t o r s
-+
H
H.
side define
It also appears that
a c o m p l e t e l y c o n t i n u o u s q u a d r a t i c o p e r a t o r i s d e f i n e d by t h e
l a s t term on t h e l e f t - h a n d
c(R) =
where
@:
H
+
H
+
side.
Thus
(2.2)
i s equivalent t o
i s a p r e s c r i b e d e l e m e n t of
RC2
H
and
i s completely continuous. The most t a x i n g s t e p i n t h e t h e o r y i s t o e s t a b l i s h
t h e f o l l o w i n g a p r i o r i e s t i m a t e depending on a s u i t a b l e c h o i c e of
p
(p >
G.
\lclll)
s E [0,1]
F o r any f i n i t e
R,
t h e r e e x i s t f i n i t e numbers
s u c h t h a t t h e r e i s no
w E H,
IIwII
= p ,
and
satisfying
w = sR@w
+
C(sR).
A c c o r d i n g t o t h e homotopy i n v a r i a n c e of
L e r a y - S c h a u d e r degree,
i t f o l l o w s that deg(1 where
S = ( u E H: P
p r o p e r t y of d e g r e e ,
-
RH,c(R),S
IIu/l < p } .
P
) = 1,
(2.41
Hence, by t h e f i x e d - p o i n t
( 2 . 3 ) h a s a t l e a s t one s o l u t i o n i n
The a v a i l a b l e r e g u l a r i t y t h e o r y shows t h a t , j e c t t o m o d e r a t e c o n d i t i o n s on t h e smoothness of equivalence c l a s s o f
aD,
S
P
.
subthe
t h i s s o l u t i o n h a s a member c o n s t i t u t i n g
a classical solution of
(2.1).
GENERIC BIFURCATION THEORY IN FLUID MECHANICS
3.
The n o t i o n of ~
"typical" o r
~
~
49
'generic' ~
With r e g a r d t o s p e c i f i c p r a c t i c a l a p p l i c a t i o n s ,
the
a n a l y s i s o u t l i n e d above i s i n c o n v e n i e n t l y s w e e p i n g i n i t s g e n e r a l i t y , and t h e p o s s i b i l i t i e s i n a b s t r a c t need rowed i n o r d e r t h a t solution set
a u s e f u l account
{W(R)]
c l a s s i f i c a t i o n s of
in
R+xH.
If
t o be n a r -
c a n b e made of
the
one a c c e p t s p l a u s i b l e
the r e a l p o s s i b i l i t i e s f o r steady flows,
however, v a r i o u s p r a c t i c a l l y s i g n i f i c a n t p r o p e r t i e s may be deduced from t h e c e n t r a l r e s u l t framework.
I n o t h e r words,
(2.4)
and i c s t h e o r e t i c a l
s i m p l i f y i n g h y p o t h e s e s a r e needed
t o d i s c r i m i n a t e between t y p i c a l b e h a v i o u r o f
t h e system ( 2 . 1 )
and e x t r a o r d i n a r y b e h a v i o u r t h a t may b e i n c l u d e d
i n the
i n f i n i t e s e t s of a d m i s s i b l e domains and boundary d a t a
q
but
i s impossible t o capture experimentally.
I t i s p e r t i n e n t h e r e t o acknowledge t h e r e c e n t s t u d y by F o i a s & Temam
( 1 9 7 7 ) , i n which p r o g r e s s h a s b e e n
made t o w a r d s a r a t i o n a l c l a s s i f i c a t i o n of s t e a d y Navier-Stokes
flows.
problem w i d e r t h a n ( 2 . 1 ) , non-conservative
field of
the properties o f
T h i s s t u d y was c o n c e r n e d w i t h a
which i n c l u d e s t h e e f f e c t s o f a e x t e r n a l f o r c e s a c t i n g on t h e f l u i d .
Considering a c e r t a i n space
F1
boundary f u n c t i o n s , F o i a s & Temam showed by means o f Sard-Smale value o f
theorem t h a t , f o r a g i v e n domain
R,
F 2 of
of f o r c e s and a s p a c e
D
t h e r e i s a n open d e n s e s u b s e t of
the
and a g i v e n FlxF2
for
which t h e number o f
solutions i s f i n i t e .
number i s c o n s t a n t ,
and e a c h s o l u t i o n i s a f u n c t i o n of p o s i -
tion,
i n e v e r y c o n n e c t e d component
Moreover,
o f t h i s subset.
this
This
50
BENJAMIN
T.B.
p r o p e r t y , which i s g e n e r i c i n t h e s p e c i f i e d s e n s e , h a s sub-~ s e q u e n t l y b e e n shown by Tromba & Marsden
(1977) t o e x t e n d t o
a s p a c e o f domains. While t h e s e r e s u l t s g o some w a y t o w a r d s d e t e r m i n i n g t h e o v e r a l l c h a r a c t e r of
t h e s o l u t i o n s e t , many i n t e r e s t i n g
q u e s t i o n s r e m a i n which s o f a r h a v e o n l y b e e n answered speculatively.
For instance, i s the finiteness property nearly
universal i n the sense t h a t
i t e x h a u s t s some a p p r o p r i a t e
measure p u t on t h e s p a c e s o f p o s s i b i l i t i e s ( s e e n o t e b e l o w ) ? Are t h e e x c e p t i o n s c o n f i n e d t o a s e t of t h e f i r s t B a i r e category?
I n view of
t h e e m p i r i c a l f a c t t h a t t h e number of
s o l u t i o n s c a n change w i t h c h a n g e s of
D
(cf.
$ 6 ) , what i s
t h e t y p i c a l form of t h e b i f u r c a t i o n s from which t h e new solutions arise?
What p r o p e r t i e s
of
such b i f u r c a t i o n s and o f
o t h e r s t e a d y - f l o w phenomena a r e s t r u c t u r a-~ l l y s t_a_b le, so that t h e y may b e o b s e r v a b l e i n p r a c t i c e ? Adopting t h e p r a c t i c a l s t a n d p o i n t a l r e a d y proposed, we may c l a s s i f y p r o p e r t i e s as t y p i c a l when t h e weight a v a i l a b l e evidence and i n t u i t i v e
-
-
mathematical,
suggests that
of
such a s i t i s , e m p i r i c a l
t h e s e p r o p e r t i e s a r e predominant
and t h e r e i s no o b v i o u s r e a s o n t o t h e c o n t r a r y .
Such p r o -
p e r t i e s may be u s e f u l l y j u x t a p o s e d w i t h o t h e r p r o p e r t i e s t h a t have been p r o v e d t o h o l d i n g e n e r a l ( e . g . P r o p e r t y 3 i n A s regards bifurcations,
$4).
we s h a l l i d e n t i f y a s t y p i c a l t h e t w o
s i m p l e s t s t r u c t u r a l l y s t a b l e p r o c e s s e s t h a t comply w i t h d e g r e e t h e o r y and a c c o u n t p l a u s i b l y f o r known phenomena.
In
f a c t , we t h u s r e c o v e r t h e t w o s i m p l e s t c a t a s t r o p h e s a c c o r d i n g t o R.
Thorn's t h e o r y o f
c a t a s t r o p h e s , even t h o u g h t h e p r e s e n t
problem d o e s n o t have a ' g e n e r a t i n g p o t e n t i a l '
and s o i s n o t
G E N E R I C BIFURCATION T H E O R Y I N F L U I D M E C H A N I C S
d i r e c t l y amenable t o t h e s t a n d a r d v e r s i o n of
that
theory.
(Note:
It i s w o r t h remembering i n t h e p r e s e n t c o n n e x i o n t h a t
a
S
set
that
i s open and d e n s e i n a s p a c e
n e c e s s a r i l y h a v e t h e p r o p e r t y of
covering
X
does n o t
'almost all'
of
X,
a p r o p e r t y t h a t may be d e f i n e d i n t e r m s o f a measure p u t on X
For instance,
xn
c o n s i d e r t h e d e n u m e r a b l e s e t o f r a t i o n a l numbers
belonging t o
a n open i n t e r v a l x
n
]O,l[,
and f o r e a c h
Sn c ] O , l [
and which h a s l e n g t h
n
= 1,2
,...
define
which i n c l u d e s t h e r e s p e c t i v e
6/2n,
where
0 < 6 < 1.
]O,l[,
t h e s e t o f r a t i o n a l numbers i s d e n s e i n
Because
so also is
m
t h e open s e t
S =
not g r e a t e r than
u
Sn b u t t h e m e a s u r e of S i s evidently n=1 6 , which we can make a s s m a l l as w e p l e a s e
without invalidating the property t h a t
4.
S
i s open and d e n s e . )
T y p i c a l p r o p e r t i e s o f bounded s t e-~ a d y f_ l.-. ows -. ~ ~ ~
_
The f i v e p r o p e r t i e s l i s t e d a s f o l l o w s h a v e b e e n d i s c u s s e d i n previous papers
(Benjamin
1976,
1977a), i n the
l a t t e r of which t h e i r p r a c t i c a l i m p l i c a t i o n s were emphasized. Here v a r i o u s p o i n t s o f of
and i n
remarks,
$5
i n t e r p r e t a t i o n a r e covered i n a s e r i e s
a r e l a t e d i d e a w i l l be d e v e l o p e d which
has p a r t i c u l a r i m p o r t a n c e r e g a r d i n g t h e e x p e r i m e n t s r e p o r t e d
in
96. P r o p e r t i e s 1 and 2 a r e g e n e r i c i n t h e d e f i n i t e sense
e s t a b l i s h e d by Foias-Temam
and Tromba-Marsden,
l i k e l y t o be v i r t u a l l y u n i v e r s a l t h e c l a s s of Property
4
observable f l o w s .
and t h e y seem
('measure-exhausting') i n Property
3 i n p a r t and
p e r t a i n t o t h e c o r r e s p o n d i n g time-dependent
problem.
T.B. BENJAMIN
52
Property
3 i s c e r t a i n , h a v i n g b e e n p r o v e d q u i t e g e n e r a l l y by
S e r r i n ( 1 9 5 9 ) , and P r o p e r t y
Thm 3 ) .
Property
4
i s a l s o c e r t a i n (Benjamin
1976,
5 i s t h e m o s t d e p e n d e n t on p l a u s i b l e reason-
ing f o r i t s c l a s s i f i c a t i o n , f i e d by t h e d i s c u s s i o n i n
and i t s meaning w i l l be e x e m p l i -
56.
(2.3) are isolated.
1. The s o l u t i o n s of
Being t h e f i x e d
p o i n t s o f a completely continuous o p e r a t o r , the s o l u t i o n s e t i s therefore f i n i t e . (i,,
m = 1,2,.
degree,
An i n d e x i s d e f i n a b l e f o r e a c h s o l u t i o n
.., k <
a n d , by t h e a d d i t i v e p r o p e r t y o f
a);
(2.4) implies that k
C i =I. m m= 1 Remark .( i f :The t y p i c a l p r o p e r t y t h a t s o l u t i o n s a r e i s o l a t e d
i s n o t u n i v e r s a l , a s t h e f o l l o w i n g couter-example s h o w s . C o n s i d e r t h e T a y l o r e x p e r i m e n t on f l o w b e t w e e n c o n c e n t r i c rotating cylinders,
and s u p p o s e t h a t t h e l e n g t h o f t h e f l u i d -
f i l l e d a n n u l u s i s s m a l l enough f o r a f l o w i n t h e form of j u s t t w o a x i s y m m e t r i c c e l l s t o b e p o s s i b l e a t m o d e r a t e v a l u e s of
R
(se
56).
I t i s w e l l known t h a t a l l s u c h c e l l u l a r f l o w s
i n t h e T a y l o r e x p e r i m e n t d e v e l o p a z i m u t h a l waves when
R
is
r a i s e d above a c r i t i c a l v a l u e , and i t a p p e a r s c e r t a i n t h a t
if
t h e i n n e r and o u t e r c y l i n d e r s a r e r o t a t e d i n o p p o s i t e d i r e c t i o n s , t h e i r v e l o c i t i e s c a n b e a d j u s t e d t o make t h e wave motions s t e a d y ( i . e .
t h e p h a s e of
t h e waves, which i s u s u a l l y
moving a z i m u t h a l l y , i s b r o u g h t t o r e s t a t p a r t i c u l a r combinations o f t h e boundary v e l o c i t i e s ) . flow i s p o s s i b l e t h a t i s n o n - t r i v i a l l y muthal c o - o r d i n a t e .
If
Thus, a s t e a d y
periodic i n the azi-
t h e a p p a r a t u s were p e r f e c t l y symmetric,
GENERIC B I F U R C A T I O N THEORY IN F L U I D MECHANICS
53
all translations of this flow would be equally possible. Therefore the solution of the hydrodynamic problem is not isolated.
It is likely, however, that such exceptions to
Property 1 are extremely rare. 2. The solutions
of (2.3)
Wm(R)
possibly intersecting) curves i n
Cm
are
IR+xH,
(though i = rtl m
and
except
at points o f intersection (two-sided bifurcation points) and turning points (one-sided bifurcation points). Remark (ii): In view of (4.1) this obviously implies that, except at isolated critic-a2values of
1
2
solutions is odd and
3. F o r
R
(k-1)
Rc
of
R,
the number
of them have index
sufficiently small, the solution
W1
k
-1.
is unique
and the f l o w that it represents is uncondition lly stable (i.e.
W1
is an attractor for the whole of
H
Remark .. - - flow may accordingly be defined. . I n - _ _ _ (iii): A primary R+XH,
the corresponding solution curve
one extending to
R = 0.
W1(R)
Uniqueness for small
is the only R
is obvious-
ly to be expected, because (2.1) becomes a linear elliptic problem in the limit
R
4
0.
Stability is to be expected
because the time-dependent, parabolic problem also becomes linear i n this limit.
4. At any non-critical value of by a solution with index
i = -1
R,
the flow represented
is unstable.
Remark (iv): Combined with Property 2, this implies that
i=l
is a necessary condition for stability of a steady flow away
from a bifurcation point.
But it is generally not a
sufficient condition for stability.
The most significant
54
T.B. BENJAMIN
p r a c t i c a l i n f e r e n c e t o be drawn i s t h a t any o b s e.r ~ v eId_ s t e a d y . f l o w must have i n d e x 1.
5. A t b i f u r c a t i o n p o i n t s , t h e d i m e n s i o n o f t h e c e n t r e manifold i n
R+xH
is
2,
and t h e i n d e x of t h e r e s p e c t i v e
solution i s zero. Remark-
(v):
This property i s a t t r i b u t a b l e t o bifurcations
that are realizable i n practice,
f o r the exceptions t o i t
depend on s i n g u l a r c o n d i t i o n s of
symmetry or o t h e r s p e c i a l
f a c t o r s which c a n n o t b e produced e x a c t l y i n a n y e x p e r i m e n t , b e i n g a n n u l l e d by any p e r t u r b a t i o n ( e x p e r i m e n t a l ' n o i s e ' however small. Property
Thus,
5 a r e not
f o r t h e problem (l), t h e e x c e p t i o n s t o
' s t r u c t u r a l l y s t a b l e ' , a l t h o u g h t h e y may
sometimes need c o n s i d e r a t i o n a s f e a t u r e s o f i d e a l i z e d t h e o r e t i c a l models.
< i= 0
r
)
R
f
Cr --iIO
- --
55
G E N E R I C BIFURCATION THEORY I N FLUID M E C H A N I C S
Consider the forms o f b i f u r c a t i o n i l l u s t r a t e d i n Fig.
1, where
mernhers of ( b ) and
r
i s any l i n e a r f u n c t i o n a l d i s c r i m i n a t i n g
the solution s e t
[Wn(R)].
I n t h e diagrams
( c ) t h e a x i s r e p r e s e n t s t h e p r i m a r y s o l u t i o n , and
a d d i t i o n a l solutions a r e represented off c r i t i c a l form o f
Property
Fig.
c’
il = 1
The s u p e p
l ( a ) , being
i s e x c l u d e d by
a t the bifurc-
But t h e t r a n s c r i t i c a l ( a s y m m e t r i c ) form i n
l ( b ) i s admissible,
p a i r of
R = R
5 s i n c e (4.1) r e q u i r e s t h a t
ation point.
the axis.
bifurcation i l l u s t r a t e d i n Fig.
symmetric a t t h e b i f h r c a t i o n p o i n t
Fig.
as a l s o i s t h a t i n Fig.
l ( c ) where a
s o l u t i o n s a r i s e s s e p a r a t e l y from t h e p r i m a r y s o l u t i o n .
l ( d ) shes a n o t h e r c a s e c o n s i s t e i i t w i t h P r o p e r t y
primary where
5: t h e
s o l u t i o n curve i s f o l d e d , having t w o t u r n i n g p o i n t s il = 0 .
single-valued
The t y p i c a l p r o c e s s whereby a n o r i g i n a l l y s o l u t i o n curve i n
IR+
x H
becomes f o l d e d a s a
s u p p l e m e n t a r y p a r a m e t e r i s v a r i e d w i l l be d i s c u s s e d i n Remark
(a),
$5.
( v i ) : I n t h e i d e a l i z e d t h e o r e t i c a l m o d e l for t h e Taylor
experiment,
t o infinity.
t h e e n d s of
t h e f l u i d - f i l l e d a n n u l u s a r e removed
A c c o r d i n g l y , a n a r b i t r a r y c o n d i t i o n of p e r i o d -
i c i t y i s imposed i n p l a c e of
t h e r e a l i s t i c end c o n d i t i o n s
i n c l u d e d i n t h e problem ( 2 . 1 ) . p r o v i d e s examples of
A s i s w e l l known,
t h i s model
s u p e r c r i t i c a l b i f u r c a t i o n , but
no r e a s o n t o e x p e c t P r o p e r t y
there is
5 t o be evaded by a bounded
model o f t h e T a y l o r f l o w . P r o p e r t y 5 does not
carry
over w i t h t h e s a m e f o r c e
t o t h e i n t e r p r e t a t i o n o f t h e Bdnard e x p e r i m e n t on i n c i p i e n t c o n v e c t i v e m o t i o n , f o r which t h e t h e o r e t i c a l problem i s a k i n
to
( 2 . 1 ) b u t somewhat more c o m p l i c a t e d .
Among v a r i o u s bounded __
56
T.B.
BENJAMIN
models that may be considered, the simplest o f them features supercritical bifurcations and s o Property 5 does not apply.
In this model the thermal boundary conditions are taken to be commensurate with a motionless state of the fluid for all values of the temperature parameter (the Rayleigh number), the dependence of viscosity on temperature is ignored, and
If any one of these three
internal heating is also ignored.
simplifications is relaxed, however, Property 5 becomes typical.
For a discussion o f these facts, see the Appendix to the paper by Benjamin (1977a).
5.
Morphogenesis T o introduce a far-reaching idea whose importance
in the hydrodynamic problem will be demonstrated presently we
turn to a familiar problem i n elasticity.
Its qualitative
aspects being intuitively obvious, this problem serves to secure the idea i n question, whose generality then becomes plain i n terms of Leray-Schauder degree.
The problem is
considered i n three progressive stages of elaboration, the first two of which are well known but the last of which presents a new aspect. (i) We first consider Euler's elementary model of buckling (cf. Love 1927, $263; Reiss 1969).
A thin inextensible but
flexible wire (elastica), constrained to lie i n a plane and straight when unstressed, is subjected to a compressive end load
P.
The bending stiffness
only of arc length s = 0
and
s = ,f.
s
$
of the wire is a function
along its centreline, and its ends at
are free to turn.
If
y
is the displace-
57
G E N E R I C B I F U R C A T I O N THEORY I N F L U I D M E C H A N I C S
ment of then
t h e c e n t r e l i n e from t h e s t r a i g h t l i n e b e t w e e n t h e ends,
dy/ds
= sin
where
@,
i s angle t h a t
In t e r m s of
makes w i t h t h i s line. of
0
t h e tangent
= @(sf,
@
the equation
e q u i l i b r i u m and t h e end conditions a r e
(5.1) and i n p h y s i c a l r e s p e c t s n o g e n e r a l i t y i s l o s t by t h e
< n.
assumption t h a t
P,
For a l l
( j . 1 ) has the
r e p r e s e n t i n g t h e unflexed
s t a t e of
additional solutions a r i s e , exceeds t h e
@
the wire.
0+
and
t h e l i n e a r i z e d form of
i s a constant).
@-
@ z 0
But t w o
= -@+,
bhen
P
which i s t h e f i r s t
Pc
‘Euler buckling load’
eigenvalue o f if
say
t r i v i a l solution
( 5 . 1 ) thus
The m u l t i p l i c i t y o f
Pc = O n 2 / L
2
solutions further
i n c r e a s e s by t w o a t e a c h of a n i n f i n i t e s e q u e n c e o f h i g h e r c r i t i c a l l o a d s , b u t we c o n f i n e a t t e n t i o n t o t h e f i r 5 t p a i r o f non-trivial
solutions.
f a c t a case o f
The g e n e s i s of
these solutions i s i n
s u p e r c r i t i c a l b i f u r c a t i o n from t h e n u l l
solution, as i l l u s t r a t e d by Fig.
l(a).
t h e f i g u r e c o u l d b e , f o r example
0(0)
The f u n c t i o n a l
or
$I(&),
f
in
although
f o r analysis the preferable choice i s f
where
C(s)
value
Pc
=
i s t h e e i g e n f u n c t i o n corresponding t o the eigen (thus
5 =
Questions
cos(l~s/t) is
i s a constant).
o€ s t a b i l i t y c a n b e answered by s t u d y i n g
t h e p o t e n t i a l energy o f
P < Pc
@
t h e system.
the n u l l function realizes
It a p p e a r s t h a t when a minimum o f
t h e energy
58
T.B.
BENJAMIN
and so represents a stable state, but when
P
7
Pc
the un-
flexed state is unstable and then the non-trivial solutions @+
and
represent stablc states.
@-
The problem (5.1) can
be reformulated so that degree theory is applicable (rf. Krasnosel’skii 1964, pp. 181-183), and the disposition o f indices f o r the first bifurcation is found to b e a s shown in Fig. l(a). A perturbation analysis shows that, respective to
p(s),
every given positive function
(P,f)-plane is locally parabolic.
the branching in the
Thus the bifurcation is
described by the cubic equation
-
f3
(P-P )f = 0,
the real roots of which are f = &(P-P
)
f = 0
P > Pc.
for
(5.2) P
for all
and
O n the understanding that the
appropriate normalizing factor is included i n the definition of
f,
this representation holds for all choices of the
linear functional such that
f(@+) = -f(@-)
#
0.
(ii) W e next consider the more general problem where the unstressed wire need not b e perfectly straight, that is, where
Q, = @ o ( s ) f o r
letting the number flexure.
G
P = 0.
We write
(p@g)’ =
EF(s),
represent the magnitude of the residual
While the end conditions remain the same as before,
the equation o f equilibrium becomes
+ P sin
(PO’)‘
@ = SF(S).
(5.3)
The solution set is now qualitatively different from the previous case
E
= 0.
T h e symmetric bifurcation of
before is unfolded i n the manner that was first made clear
59
G E N E R I C B I F U R C A T I O N T W O R Y IU F L U I D M E C H A N I C S
i n the hritings of W.T.
Koiter about
It can be shown that provided
problems of elastic stability. Q0(s)
‘imperfection-sensitive’
satisfies the undemanding condition
‘0
I€\,
the situation for small
P-P
with
positive and small
o r negative, is described by the cubic equation
f7
-
(l’-Pc)f
-
6
= 0
(cf. Chillingworth 1 9 7 5 , Thompson & Hunt
(5.k)
1975).
P
FIGURE 2.
FIGURE 3 .
This situation is illustrated in Fig. 2 , where the real roots of
(5.4)
are plotted against
curves are for
E
= 0.05,
P
with
P c = 1.
The continurns
and the dashed curves f o r
E
= 0.
Note also the disposition in indicesshownin the figure. The discriminant of the cubic values o f
P
and
E
satisfying
(5.4)
vanishes for
60
T.B.
BENJAMIN
which d e s c r i b e s a c u s p i n t h e
(P,E)-,~lane(Fig.
7).
For
p o i n t s i n s i d e t h e c u s p t h e c u b i c has t h r e e r e a l r o o t s , f o r p o i n t s o u t s i d e i t h a s only one. d e s c r i b e d by
(5.4) i s a
and
The u n f o l d i n g p r o c e s s
p a r t i c u l a r i n s t a n c e of t h e a l g e b r a i c
c h a r a c t e r i z a t i o n commonly known by t h e t e r m
'cusp c a t a s t r o p h e ' .
The s i t u a t i o n i s i n t u i t i v e l y o b v i o u s
f r o m Fig.
4.
The p r i m a r y mode of
the
f l e x i n g upwards i n t h e f i g u r e ,
-system,
P
P d e v e l o p s s m o o t h l y and r e m a i n s
i- L(;LRF: ' t .
with increasing
P.
stable
P
But f o r
s u f f i c i e n t l y l a r g e , a n o t h e r s t a b l e mode becomes p o s s i b l e , f l e x i n g downwards evident that
there i s an intervening s t a t e o f
which i s u n s t a b l e .
and i t i s
i n the figure,
equilibrium
Having b e e n r e a l i z e d by p u s h i n g t h e column
dohnwards u n d e r a s u f f i c i e n t l y l a r g e mode w i l l a b r u p t l y d i s a p p e a r when
P,
t h e s t a b l e secondary
P i s reduced g r a d u a l l y t o
the c r i t i c a l v a l u e corresponding t o t h e one-sided p o i n t i n t h e lower h a l f mined a p p r o x i m a t e l y by
of F i g .
(5.5)
( i i i ) Suppose now t h a t
2
when
(i.e.
c
bifurcation
t o the value deter-
i s small).
t h e p e r t u r b e d system j u s t c o n s i d e r -
ed i s s u b j e c t e d t o a n a d d i t i o n a l p h y s i c a l f a c t o r t e n d i n g t o c a n c e l t h e p r i o r i t y o f t h e p r i m a r y mode.
F o r example, a
b e n d i n g moment may b e a p p l i e d a t t h e end
s = 0
column,
s o t h a t t h e end c o n d i t i o n becomes
It i s c l e a r from F i g . the
left-hand
4
of
the
Q ' ( 0 ) = m.
t h a t a c l o c k w i s e moment a p p l i e d a t
end w i l l have t h e r e q u i r e d e f f e c t , and i f
large
61
G E N E R I C BIFURCATION TIIEORY I N FLUID M E C € I A N I C S
e n o u g h i t w i l l c a u s e t h e column t o b u c k l e downwards r a t h e r t h a n upwards u n d e r a g r a d u a l l y i n c r e a s e d l o a d
P.
mode i s t h u s made t o u n d e r g o a m o r p h o g e n e s i s :
t h e continuum
~
The primary
~~~
s t a t e s g e n e r a t e d by i n c r e a s i n g t h e l o a d g r a d u a l l y f r o m z e r o
of
c h a n g e s f r o m d e f i n i t e l y u p w a r d s - t o d e f i n i t e l y downwards, l e a v i n g t h e u p w a r d s c o n t i n u u m a s a s e c o n d a r y mode r e a l i z a b l e
It i s c l e a r , moreover, that
only a t s u f f i c i e n t l y l a r g e loads. if
i s small,
E
then a comparatively s m a l l
m
w i l l
switch
Clie p r i m a r y mode f r o m u p w a r d s t o downwards. I n view of p r o p e r t i e s a p p l i c a b i l i t y of §I+),
degree theory
the nature of
that are inherent i n the (like the properties
t h e morphogenesis can be p r e d i c t e d i n
geometrical terms without
explicit
analysis.
d i a g r a m s u c h a s i s m o d e l l e d by F i g .
a
P = 0
to the f a r
t h e d i a g r a m , and a s e c c i n d a r y mode by o n e b r a n c h o f
l o o p s u c h as i n t h e l o w e r h a l f
of' F i g .
i n d i c e s i s a l w a y s 1, t h e c u r v e s a r e points,
In a state
2 , a p r i m a r y mode i s
r e p r e s e n t e d by a c u r v e e x t e n d i n g from right of
listed in
Cm
2.
The sum of' t h e
except a t bifurcatim
and d i s c o n n e c t e d s e c o n d a r y l o o p s a r e s m o o t h a t t h e i r
extremities
(i.e.
locally p a r a b o l i c ) .
W e seek a process
whereby a n o r i g i n a l p r i m a r y modc a n d a s e c o n d a r y mode e x change r o l e s as a supplementary parameter continuously.
is varied
Hence w e e x p e r t t h e p r i m a r y and s e c o n d a r y
c u r v e s t o b e deformed c o n t i n u o u s l y s o t h a t i n t o c o n t a c t a t one p o i n t . order,
m
they a r e brought
They t h e n d i v i d e i n a d i f f e r e n t
f o r m i n g a new p r i m a r y c u r v e a n d new s e c o n d a r y l o o p . One p o s s i b i l i t y f o r s u c h a p r o c e s s i s t h e u n f o l d i n g
o f a symmetric s u p e r c r i t i c a l b i f u r c a t i o n ,
a s d e s c r i b e d by
62
T.B.
(5.4),
but
BENJAMIN
t h i s i s e v i d e n t l y not ge n e r i c i n t h e considered
c l a s s of p h y s i c a l p o s s i b i l i t i e s . singular specifications
(i.e.
w i l l e n f o r c e symmetry o f
of
I n o t h e r words,
~ ( s )i n ( 5 . 2 ) )
the function
t h e m u t a t i o n when t h e p r i m a r y and
secondary c u r v e s i n a s t a t e diagram a r e brought The s i n g u l a r e v e n t u a l i t y would c o r r e s p o n d the generic Property
(c F I G U R E 5.
only
into contact.
t o a n e v a s i o n of
5 l i s t e d i n 93.
1
(a)
S u c c e s s i v e f o r m s of
l o c i i n a s t a t e diagram f v s P.
T h e r e r e m a i n s o n l y one p o s s i b l e form o f e v e n t s , which i s i l l u s t r a t e d i n F i g . sense (a)
( d ) of
4
as t he parameter
5.
The p r o c e s s m a y g o i n t h e
t h e f o u r diagrams,
m
i s i n c r e a s e d , but
o r i n t h e r e v e r s e sense, i n e i t h e r case the
p r i m a r y mode m u t a t e s a t a t r a n s c r i t i c a l b i f u r c a t i o n p o i n t a s shown i n t h e d i a g r a m ( c ) . by t h e f i g u r e i s t h a t
An i m p o r t a n t c o n c l u s i o n made c l e a r
the continuous process
through t h i s b i -
f u r c a t i o n n e c e s s a r i l y causes t h e primary curve t o be f o l d e d (i.e.
a s i l l u s t r a t e d by F i g .
parameter
m
(P,m)-plane,
1 ( b ) ) when t h e s u p p l e m e n t a r y
l i e s i n some i n t e r v a l .
P r o j e c t e d onto t h e
t h e f o l d a p p e a r s a s a c u s p , whose a p e x
63
G E N E R I C BIFURCATLON THEORY I N FLUID M E C H A N I C S
f = f(P)
c o r r e s p o n d s t o t h e p o i n t where t h e p r i m a r y c u r v e
T h i s i s a n o t h e r cusp c a t a s t r o -
f i r s t has a v e r t i c a l tangent.
p h e , b u t w i t h o r i e n t a t i o n d i f f e r e n t from t h e one c o n s i d e r e d previously.
The f o l d i n g of
i n practice,
a h y s t e r e s i s w i l l be m a n i f e s t e d i n t h e dependence
of
t h e p r i m a r y s t a t e on
t h e primary curve i m p l i e s t h a t ,
P.
A p p r e c i a t i n g t h e g e n e r a l i t y of
t h i s argument,
we
may p r e s e n t t h e c o n c l u s i o n a s f o l l o w s , where no p a r t i c u l a r p h y s i c a l meaning i s a t t a c h e d t o
and
P
m:
P R I N C I P L E O F MORPHOGENESIS W I T H RESPECT T O TWO PARAMETERS.
u E H
Suppose t h a t t h e s t a t e s a primary
'loading'
parameter
m. ~
-a b_-l e- (A.e.--a
P E R+
parameter
F o r every
s. u f f i c i e n t l y s mall,
o f a s y s t e m a r e d e t e r m i n e d by
m,
u(P)
and a s u p p l e m e n t a r y
i s u n i q u e when -
s o t h -~ a t - a p r i. m a r y s e t o f s t a t e s i s d e fin~
R+XH).
curve i n
~
-
Suppose ___ _. t h a t a s
c o n t i n u o u s l y t h r o - u g p z c_r_i t i.___ c a l -~ value u n d e r g-o__ e s a m o r p h o g e n e s i__ s (i.e. ~
mc,
m
i s _va r ied _
t h e p r i m a r y ___ set ~
a r a d i c a l q.u a l i t a t i v e change -
i n t.. h e__ s.-t . a. t e .s d.e v e l o p e d a t s u f f-i c i e n t l y _l ~a r_g_e_ . -- -
TbE,
P).
~
g_e_n e r i c a l l y~, t!le P
when - __
m
is
P
~s y s. t e_ m-
man -i f e s t s h y s t e r-____ e s i s w i t h res-pect
l_i_e-s -i n a n i n t e r v a l above - o- r below -
B "c
C
FIGURE 6 .
m
.
to
64
T.B.
BENJAMIN
This principle is illustrated in Fig. 6, which shows the typical form of the 'catastrophe set' i n the (P,m)-plane. Except f o r
B
and
C,
every point i n the set is the project-
ion o f a one-sided bifurcation point, the extremity of a secondary loop, i n R+xH
.
T h e turning point
B
corresponds
to the transcritical bifurcation point shown i n F i g . and the cusp between
C
and
fold in the primary curve
m = m
C
u = u(P).
5(c),
is the projection of the
T h e generic attribute
being asserted is simply that the cusp points either upwards or downwards relative to the P-axis.
The singular case where
it is aligned with the P-axis recovers the case o f mutation through a symmetric bifurcation, which we have earlier dismissed as a practical possibility. It is noteworthy that closely parallel reasoning has been used by Zeeman (1975) a s the basis of a speculation about biological processes.
6.
Experimeptal results The qiialitative predictions reviewed in 994 and
5
have been borne out by some recent experiments on steady flows.
In particular, morphogenetic processes have been studied, and they have been found to accord with the general principle explained above.
We summarize as follows the outcome o f one
set of experiments which has already been reported (Benjamin 1977b). Fig. 7 is a diagram o f the apparatus, which is basically the same as that used i n the famous experiments o f Taylor (1923) mentioned earlier.
The length
4,
of the fluid
G E N E R I C R I F U R C A T L O N F - I E O R Y I N FLUID M E C H A N I C S
f i l l e d annulus i s comparatively continuously adjustable.
s m a l l , however,
The o u t e r c y l i n d e r made of a
62
i s r o t a t e d a t c o n s t a n t speed
a continuous range o f speeds. arid o u t e r r a d i i i s 0 . 6 2 ,
( r a d / s ) by a s e r v o m o t o r w i t h
2
R = 62r,/V,
the fluid.
hhere
R.
wide r a n g e s o f
R
cells.
(mm2/s)
R,
but
a l l of
i t was
? t a b l e s t e a d y f l o w s e x i s t s over
The p r i m a r y f l o w was p r o d u c e d by g r a d u a l -
R
f r o m s m a l l v a l u e s , and f o r
a b o u t 100 i t s s t r u c t u r e became q u i t e c l e a r , number of
V
flow v i s u a l i z a t i o n .
which became u n s t e a d y a t s u f f i c i e n t l y l a r g e of
- 2 3 mm. -
a pearly
V a r i o u s a x i s y m m e t r i c f l o w s were o b s e r v e d ,
found t h a t a m u l t i p l i c i t y
the inner
An aqueous s o l u t i o n
i n which a s m a l l amount o f
s u b s t a n c e w a s mixed f o r t h e p u r p o s e o f
of
d = r '-rl
and t l i c . gap w i d t h
i s t h e k i n e m a t i c v i s c o s i t y of
g l y c e r i n bas u s e d ,
rl/r2
The r a t i o
The Reynolds number i s g i v e n b y
l y increasing
trans-
( P e r s p e x ) i s f i x e d , and t h e i n n e r c y l i n d e r
parent material
of
and i s
g r e a t e r than
f e a t u r i n g a n even
c e l l s w i t h c o n t r a r y s p i r a l l i n g motions i n a d j a c e n t
S e c o n d a r y modes were p r o d u c i b l e by s e v e r a l means,
s u c h a s by s t a r t i n g t h e m o t i o n of
L
f r o m r e s t , or by v a r y i n g
t h e i n n e r c y l i n d e r suddenly
g r a d u a l l y t o change a w e l l -
d e v e l o p e d p r i m a r y f l o w i n t o a s t a t e where i t i s no l o n g e r primary. A n example of
t h e p o s s i b i l i t i e s i s g i v e n with photo-
graphs i n t h e c i t e d paper. demonstrated a t
L/d
= 3.3,
t w o c e l l s f o r t h i s v a l u e of modes h a v e f o u r c e l l s .
F i v e d i s t i n c t s t e a d y flows a r e
R = 650. L/d,
I n the
The p r i m a r y f l o i q h a s
and t w o of
'normal'
t h e secondary
f o u r - c e l l mode,
c a l l e d because it occurs a s t h e primary flow f o r l a r g e r
so
66
T.B.
values o f
L/d,
BENJAMIN
the motion is inwards near the end walls;
and i n the other, abnormal four-cell mode, the motion is outwards there.
The remaining two secondary modes have three
cells, with a n abnormal cell at the t o p of the annulus in one case and at the bottom in the other. 2 and
4 explained i n $4, this evidence
Because of Properties o f five different
stable flows implies the existence o f four other flows which, having index
-1, are necessarily unstable.
FIGUIILI 7 .
-dh 1.01
FIGURE 8.
1.
a:
0
I
400
I
600
R*
G E N E R r C n L F U R C A T I O N TYEORY I N F L U I D M E C H A N I C S
4.01-
-1
d -
3.8
-
3.6 -
3.4 -
> -
ioo
160 IZ
?GO
140
120
FIGURE 9. The s t r u c t u r e o f
t h e primary f l o w ,
d i s t i n c t l y a t moderately l a r g e
L/d,
and i t i s o f
R,
as manifest
d e p e n d s d i s c o n t i n u o u s l y on
obvious i n t e r e s t t o i n q u i r e i n t o t h e
n a t u r e of' t h e m u t a t i o n t h a t
takes place a t c r i t i c a l values of
k/d.
i n v e s t i g a t i o n was t o measure
The a d o p t e d method of
properties
of
t h e s e c o n d a r y modes t h a t e v o l v e i n t o t h e
p r i m a r y flow a s t h e c r i t i c a l v a l u e i s c r o s s e d r e s p e c t i v e l y from a b o v e or below. Fig.
A s a n example o f s u c h m e a s u r e m e n t s ,
8 shows e x p e r i m e n t a l v a l u e s of t h e h e i g h t
c e l l i n t h e normal f o u r - c e l l
d e c r e a s i n g f u n c t i o n of
R
significant feature of Fig.
s i d e of
R
end,
and h
is a
The most
8 i s t h a t t h e curve
has a v e r t i c a l tangent a t i t s left-hand
t h e end
where t h e
the f i g u r e ,
up t o t h i s l i m i t .
of
L/d = 3 . 3 .
s e c o n d a r y mode a t
T h i s f l o w becomes u n s t e a d y a t t h e v a l u e of c u r v e e n d s on t h e r i g h t - h a n d
h
h = h(R)
indicating the
68
T.B.
c r i t i c a l v a l u e of (with
R
f o r t h e one-sided
bifurcation point
from which t h e s e c o n d a r y mode a r i s e s .
i=O)
t o the p r o p e r t i e s explained
i s a view of
R xII.
BENJAMIN
According
e a r l i e r , t h e experimental curve
one b r a n c h ( w i t h
i = l ) of a s e c o n d a r y l o o p i n
The complementary b r a n c h ( w i t h i = - l )r e p r e s e n t s a n
u n s t a b l e f l o w and t h e r e f o r e c a n n o t be r e a l i z e d e x p e r i m e n t a l l y . The r e s u l t s of many s u c h measurements a r e shown i n
9.
Fig.
For
L/d
B
below
i n the figure,
h a s t w o c e l l s , and t h e l i n e generating the four-cell t h e l o c u s of flow.
two-cell
R
flow i s primary,
is
b u t below
B
a new e f f e c t
i s gradually increased f r o m small values, a
f l o w w i t h embryonic f o u r - c e l l
i s then g r a d u a l l y reduced,
f e a t u r e s i s evolved
a t which a n a b r u p t
CB,
t r a n s i t i o n t o a c l e a r two-cell
DC,
DC
T h i s f l o w i s a s e c o n d a r y mode
B,
c o n t i n u o u s l y up t o t h e l i n e
line
The l i n e
c o r r e s p o n d i n g b i f u r c a t i o n s i n v o l v i n g t h e two-cell
and t h e f o u r - c e l l If
i s the locus o f bifurcations
s e c o n d a r y mode.
Above t h e l e v e l of
appears.
RC
t h e primary f l o w
s t r u c t u r e takes place.
If
R
t h i s s t r u c t u r e remains u n t i l t h e
a t which t h e f l o w r e t u r n s a b r u p t l y t o i t s o r i g i n a l .
form. The i n t e r p r e t a t i o n o f l y a s e x p l a i n e d a t t h e end of
ween F i g .
t h e s e o b s e r v a t i o n s i s precise-
$ 5 , and t h e e q u i v a l e n c e b e t -
6 and t h e e x p e r i m e n t a l F i g . 9 c a n b e a p p r e c i a t e d .
H y s t e r e s i s of t h e p r i m a r y t w o - c e l l
CB
flow t a k e s p l a c e between
and t h e o p p o s i t e s i d e o f t h e c u s p , and
B
corresponds
t o t h e t r a n s c r i t i c a l b i f u r c a t i o n p o i n t a t which t h e p r i m a r y f l o w m u t a t e s b e t w e e n t h e t w o - c e l l and f o u r - c e l l forms. O t h e r e x p e r i m e n t s have b e e n made o n s t e a d y f l o w s i n
GENERIC B I F U R C A T I O N THEORY I N F L U I D YECIIANICS
a domain w i t h s e m i - c i r c u l a r
c y l i n d r i c a l boundary,
69
t h e diametric
f a c e and e n d s o f w h i c h a r e s t a t i o n a r y a n d t h e c i r c u l a r portion of w h i c h i s r o t a t e d a t c o n 5 t a r r t s p e e d . llake c o m p l i c a t e d c e l l u l a r
f’orms, a n d ,
of‘ t h e p r i m a r y f l o w e v o l v e d a t liigh
l y o n t h e l e n g t h o f t h e domain. and t h e m u t a t i o n s o f
T h e o b s e r v e d €lobs
as before, R
the structure
depends d is c o n tin u o u s -
P r o p e r t i e s of
s e c o n d a r y modes
t h e p r i m a r y f l o w ha\-e a p p e a r e d t o be t h e
5ame q u a l i t a t i v e l y as i n t l i e f i r s t e x p e r i m e n t s ,
thus a g a i n
confirming t h e t h e o r e t i c a l p r e d i c t i o n s .
It should be noted f i n a l l y t h a t t h e n o n l i n e a r boundary-value
problems r e - p e c t i v e
t o t h e s e two s e t s o f
e x p e r i i n e n t s a r e beyond t l i e r e a c h o f Moreover, no numerical r c . + u l t
15
constructive theories.
yet
available.
The
clualitative theory i s therefore particularly useful i n c l a r i f y i n g s u c h experitrierital
fxndings.
REFERENCE S B e n j a m i n , T.R.
1976
Applications
o f Leray-Schauder d e g r e e
t h e o r y t o problems o f hydrodynamic s t a b i l i t y .
Benjamin, T.B. of
1977a
Benjamin, T.B.
I. Theory.
Proc.
Roy.
S O C . Land. -.A ~
~
[ F l u i d Meclianics R e s e a r c h I n s t i t u t e , U n i v .
of E s s e x , Rep.
of
B i f u r c a t i o n phenomena i n s t e a d y f l o w s
a viscous f l u i d .
359, 1-26.
Math. __ __
no 83.1
197713
a viscous
B i f u r c a t i o n phenomena i n s t e a d y f l o w s
fluid.
11. E x p e r i m e n t s .
~
~ Roy-._Soc. ~ c A.
359, 2 7 - 4 3 . Chillingworth,
D.
1975
The c a t a s t r o p h e o f a b u c k l e d beam.
I n Dyriamical S y s t e m s Mathematics v o l . -~
468,
-
Warwick pp.
1974.
86-91.
L e c t u r e s N o t e s i -_n
Berlin:
Springer-Verlag.
70
T.B. BENJAMIN
Foias, C.
Temam, R .
&
1977
Structure of the set of station-
ary solutions of the Navier-Stokes equations.
Communs -
Pure Appl. Math., 30, 149-164.
1964
Krasnosel'skii, M.A.
Topological Methods i n the Theory -
o f Nonlinear Integral Equations.
1929
Love, A.E.H.
London: Pergammon.
TA-Mathema&icaLTheor_y of Elasticity,
4th ed. Cambridge U n i J ersity Press. (Dover edition l9hh.)
1969
Reiss, E.L.
Column buckling
Bifurcation.
Xar-ue
-
A n elementary example o f
In Bifurcation Theory and - -Nonlinear _____ Eigen_ _ __ Problems (ed. J.B. Keller & S. Antman). New Y o r k : ~-
B en,jam in. Serrin, J.
1959
O n the stability o f viscous fluid motions.
Arch. Ration. Mech.- _ Analysis _ --
1923
Taylor, G.I.
1,1-17.
Stability o f a viscous liquid contained
between rotating cylinders.
Phil. Trans. Roy.~SOC. A-
2213, 2 8 9 - 3 4 3 . Thompson, J.M.T.
& Junt, G.W.
furcation theory. Tromba, A . J .
1975
Towards a unified bi-
z.
Math. Phys. 26, _. angew. ~- 581-604.
& Marsden, J.E.
1977 Generic finiteness of
the stationary solutions of the Navier-Stokes equations. (Preprint.) Zeeman, E.C.
1975
Levels of structure i n catastrophe theory
illustrated by applications in the social and biological sciences.
Proc. Int. Cong. Math., Vancouver 1974,
pp. 533-546.
Canadian Mathematical Congress.
G E N E R I C BIFURCATION THEORY I N F L U I D MECHANICS
L e c t u r e hy T . R .
71
Benjamin
S U P P LEME NT
A t the conclusion of
the lecture,
a practical
d e m o n s t r a t i o n w a s made i l l u s t r a t i n g t h e p o s s i b i l i t y of m u l t i p l e s t a b l e s t a t e s i n a s y s t e m w i t h i n f i n i t e freedom. A c o i l e d wire capable o€ l a r g e f l e x u r e s without
permanent s t r a i n ( ' c u r t a i n w i r e ' ) i s f i x e d i n a h o r i z o n t a l board, forming a n a r c h i n a v e r t i c a l plane.
The w i r e p a s s e s
t h r o u g h t h e b o a r d a t one e n d , g i v i n g a s p a r e l e n g t h b e n e a t h , and s o t h e l e n g t h varied.
L
of w i r e i n t h e a r c h c a n r e a d i l y be
[ T h i s kind o f wire roughly s i m u l a t e s E u l e r ' s e l a s t i c a
w i t h c o n s t a n t bending s t i f f n e s s
8,
moment t o c u r v a t u r e i n a p l a n e ( c f .
the r a t i o of bending
$5(i)).
But i n f a c t
p
i s a m i l d l y d e c r e a s i n g f u n c t i o n of c u r v a t u r e . ]
The p r o p e r t i e s s t a t e d i a g r a m , where
f
p o s i t i o n o f equilibrium. about t he L-axis,
of
t h i s system a r e i n d i c a t e d i n t h e
i s d e f l e x i o n from t h e u p r i g h t Ideally,
t h e diagram i s symmetrical
and t h e u n f o l d i n g d u e t o r e s i d u a l i m p e r f e c t -
72
T.B.
BENJAMIN
i o n s i s shown by t h e d a s h e d l i n e s .
The f o l l o w i n g p r o p e r t i e s
a r e noteworthy:
( i )F o r stable.
L < L’
,
the upright position i s unconditionally
T h i s i s d e m o n s t r a t e d by d e f o r m i n g t h e w i r e g r o s s l y
and t h e n r e l e a s i n g i t , whereupon t h e e x t r a e n e r g y i s q u i c k l y d i s s i p a t e d and t h e w i r e comes t o r e s t u p r i g h t .
( i i )F o r
L
L’,
i n a n i n t e r v a l above
position i s s t i l l stable
( d e m o n s t r a t e d by g i v i n g t h e w i r e a
s m a l l p u s h from t h i s p o s i t i o n ,
t o which i t t h e n r e t u r n s ) , b u t
t w o other s t a b l e e q u i l i b r i a e x i s t .
Given a s u f f i c i e n t l y
vigorous push f r o m t h e u p r i g h t p o s i t i o n , one o r o t h e r o f
the upright
t h e wire f a l l s i n t o
t h e s e lower p o s i t i o n s .
( i i i ) The e x i s t e n c e of
three stable
e q u i l i b r i a implies the existence of
t w o a d d i t i o n a l e q u i l i b r i a which a r e unstable. theory,
T h i s f o l l o w s by d e g r e e
or by t h e Morse i n e q u a l i t i e s
r e f e r r e d t o t h e energy f u n c t i o n a l ( e x p r e s s i n g s t r a i n energy p l u s g r a v i t y p o t e n t i a l ) , common s e n s e i n t h i s example.
E i t h e r one of
or by
these unstable
e q u i l i b r i a i s d e m o n s t r a t e d by g e n t l y g u i d i n g t h e w i r e i n t o t h i s position.
On b e i n g r e l e a s e d ,
t h e w i r e t h e n s l o w l y moves
away f r o m t h e p o s i t i o n e i t h e r upwards or downwards. ( i v ) The t u r n i n g p o i n t
P
i n t h e s t a t e diagram i s
d e m o n s t r a t e d by p u t t i n g t h e w i r e i n t o one o f positions of
s t a b l e equilibrium with
gradually decreasing
L
towards
L’.
L
i t s lower
> L’ ,
and t h e n
Thus t h e a r c
QP
is
G E N E R I C D I F U R C A T L O N THEORY I N F L U I D M E C H A N I C S
f o l l o w e d , and a s t h e upright
(v) For unstable.
P
i s a p p r o a c h e d t h e w i r e a b r u p t l y jumps t o
position.
L
sufficiently large,
t h e uprigllt
position is
The w i r e t h e n falls s p o n t a n e o u s l y if r e l e a s e d i n
t h i s position.
71
de La Penha, L.A. Medeiros (eds.) Contenporary Developments in Continuum Mechanics and Partial Differential Equations @North-Holland Publishing Company (1978) G.M.
THE HAMILTON-JACOBI-BELLMAN EQUATION FOR TWO OPERATORS VIA VARIATIONAL INEQUALITIES
HAIM BREZIS DGpartement de Mathgmatiques Paris VI
Universit;
4 pl. Jussieu, 75230 Paris Cedex 0 5
Let
(Au]
denote a family of second order elliptic
0c
operators on a bounded domain
where aa. 1J
p > 0
!,isj
f'(x),
is a constant, 2
2
N
R :
aa
(a > 0) 4
and
:a
RN,
+a.
Ij
5 E
are smooth and Given functions
we consider the problem: find
(1)
u(x)
n
on
and
such that u = 0
on
Sup {A'u(x) U
-
fa(x)]
= 0
an.
Problem (1) occurs in the theory of optimal stochastic control (see e.g.
In case
C21).
N 62 = R ,
assumptions on $L
p <
m.
difficult
Krylov [ 3 ] proves
f&)
His proof
-
(under reasonable
that (1) has a unique solution
-
u E
which is extremely technical and
relies on an explicit construction of
u
by use
of probability tools.
In what follows, I report on a joint work with L.C. Evans [l]. O u r results concern o n l y the case where the family
of two operators.
A'
consists
It is not clear how to extendour proofs to
the case of more than two operators.
O n the other hand our
THE H A M I L T O N - J A C O B I - B E L L M A N
EQUATION
75
method h a s some a d v a n t a g e s : a ) i t i s q u i t e s i m p l e and c o n s t r u c t i v e , b) i t leads t o c l a s s i c a l c)
solutions
(u E C2’a(n)),
i t i s v e r y f l e x i b l e and c a n be a d a p t e d : ( i ) t o t h e c a s e where
A1
and
( i i ) t o t h e c a s e where
A1
i s e l l i p t i c and
parabolic
are parabolic
A*
( t h i s h a s been d o n e by P . L .
is
A2
-
Lions)
s u c h a problem o c c u r s i n a w o r k of B e n s o u s s a n and L e s ourne
.
Our main r e s u l t s a r e t h e f o l l o w i n g Theorem 1 then
V
f
1
Theorem 2 of
Po > 0
There i s a c o n s t a n t
,f
2
E L‘(n)
1 Max { A u
(2)
u
-
-
If
3
-
11
E H2((n)
2 fl, A u
Theorem 3
-
When
solution
u
of
to
f ]
H:(n) = 0
> p0,
@
unique s o l v i n g a.e.
f 1 , f 2 E H1(n)
i n addition
( 2 ) belongs
2
-
n
such t h a t i f
in
R.
then the s o l u t i o n
H3(n).
f1,f2
E
W1
”(n)
p
>
N,
then the
~ ~ , ~ f( o nr some )
o
( 2 ) belongs
with
o < a < 1.
Remarks and open p r o b l e m s . f1,f2 E C”(n),
1) I n g e n e r a l , e v e n f o r
( 2 ) does n o t belong t o = A,
(2)
continuous
reduces
prove a t b e s t t h a t a s k i s whether
C1
i n general.
u E
C2’l.
a
A’
of
2 = A =
which i s o n l y l i p s c h i t z T h e r e f o r e we may hope t o
A more r e a s o n a b l e q u e s t i o n t o
v a < 1
u E C29a
s h a l l see l a t e r the
F o r example when
Au = M i n ( f 1 , f 2 }
to
and n o t
C3(n).
u
the s o l u t i o n
when
f 1 , f 2 E Cm.
A s we
which o c c u r s i n Theorem 3 i s found by
u s i n g D e G i o r g i s theorem,
t h e r e f o r e i t might be very s m a l l .
76
H A I M BREZIS
A related question i s :
( o r f o r what
p ’ s ?)
does
belong t o
u
W39p(n)
that
f1,f2 E Lm(n)
3
m
f1,f2 E LP(n),
2 ) It would b e o f i n t e r e s t t o p r o v e t h a t i f
u
p <
f1,f2 E Cm.
when
then the solution
+
of
u E W2”
(2) satisfies
u E W2”
p > N).
for some
( o r just A positive
answer would be v e r y u s e f u l i n d e a l i n g w i t h t h e problem: find
u
such t h a t
1 Max(A u
-
which o c c u r s i n q u e s t i o n s o f s t o c h a s t i c systems ( J . L .
f
1
,
2 A u
-
f
2
,
u]
= 0
on
R,
o p t i m a l s t o p p i n g t i m e for
L i o n s , p r i v a t e communication).
3 ) Many q u e s t i o n s c a n be a s k e d c o n c e r n i n g f o r example t h e ” f r e e boundary” s e p a r a t i n g the s e t s [A’u
= f
1
3,
1 [A1u < f ]
i t s s h a p e , smoothness e t c . . .
and
Many r e s u l t s h a v e
been o b t a i n e d r e c e n t l y o n v a r i o u s o t h e r f r e e boundary v a l u e problems
( s e e t h e works
Nirenberg,
o f C a f f a r e l l i , Friedman, K i n d e r l e h r e r ,
Riviere, e t c ...)
Sketch o f proof
f o r Theorem 1
-
B y an e a s y change o f unknown
I
we c a n a l w a y s assume t h a t 1 Max ( A u
(3) Set
1 2 -1 T = A (A ) ,
Dirichlet condition. operator i n Set
2 cp = A u
(A2)
-1
f2 = 0
that is
2 f , A u]
= 0
in
0.
being understood with zero
Clearly,
T
i s a l i n e a r and bounded
~ ~ ( 0 ) . a s new unknown,
By a w e l l known theorem of
and s o
( 3 ) becomes
Stampacchia ( s e e
141)
problem
(5)
THE H A M I L T O N - J A C O R I - B E L L M A N
tlas a u n i q u e s o l u t i o n p r o v i d e d '
'n
T W ' Y dx
2
alp)
T
i s coercive i n
a > 0.
E L2,
V $
1 \\here
=
v
-1
fI
Tq * $ d x =
n
A1(A2)
w*$
P
=
allvl
'hl
'
-1 (A')
$
2
T y * v dx
Therefore
al$l2
2
l7
'.
< Cllvll
i.e.
A v - A v dx
( b y a n i n e q u a l i t y o€ S o b o l e v s k i , H2 Appendix i n [ 11 ) .
3
L'
But
L2
f
77
EQUATION
see e.g.
2
u; = A v
since
the
I(! 1
and
S
L
L2
I1
o f p r o o f f o r Theorem 2
Sketch
-
We a r i t c ( 3 ) a s a m u l t i v a l u e d
equation
(6)
A
P
where
(6)
1 A u
(7)
P,
+ p(a2u)
U
3 f
=
r
y
by
+ 6 ( A 2u e )
E
= f
0
> 0
for
r > 0.
'2
in
y : R + IR
( f o r example) and
nondecreasing f u n c t i o n siirh t h a t
Y(r)
R,
in
i s t h e maximal monotone g r a p h d e f i n e d a s
a n d we a p p r o x i m a t e
where
~
Y(r)
= 0
i s a n y smooth for
r < 0,
It f o l l o w s f r o m t h e s t a n d a r d t h e o r y o f
monotone o p e r a t o r s ( a n d f r o m S o h o l e v s k i Is i n e q u a l i t y )
u
h a s a s o l u t i o n and t h a t
(7) is
G
+ u
s t r o n g l y nonlinear, but
smooth p r o v i d e d
f
as
0
in
11'.
F o r s i m p l i c i t y we d r o p now
is.
1 (A u
+
)
+~ 8 '
x
2 2 ( A u ) ( A u),
we find
= fx.
(7)
Equation
e l l i p t i c and t h e r e f o r e
Differentiating (7) with respect t o
(8)
E
that
uF 6 .
i s
78 Let
HAIM B R E Z I S
;
b e a smooth f u n c t i o n w i t h compact s u p p o r t i n
M u l t i p l y i n g ( 8 ) by
2
C2(A u
)
a.
~y i e l d s
(9) Note t h a t
R1,
where
R2
so that (10) 0 ) and S o b o l e v s k i ‘ s i n e q u a l i t y t h a t
I n e q u a l i t y (11) l e a d s independent of
c
To e s t a b l i s h t h a t
E H’
-
in
Hloc
Hloc’
u p t o t h e b o u n d a r y one u s e s f i r s t
l o c a l c h a r t s t o s t r a i g h t e n t h e boundary. a s above
3
c
3
and t h e r e f o r e LI
u
t o an estimate f o r
used i n t a n g e n t i a l d i r e c t i o n s
t h i r d order d e r i v a t i v e s , except
u
The same a r g u m e n t
-
x x x ’
shows t h a t a l l belong t o
N N N
(xN
d e n o t e s t h e normal v a r i a b l e ) .
F i n a l l y we g o b a c k t o e q u a t i o n
with
S1,S2
E
H1,
But U
x x
( 1 2 ) says t h a t
= min(-
N N
and t h e r e f o r e
1 E H .
u X~ X~
( 3 ) w h i c h we w r i t e as
s1 , 7s2 1 1
a
NN
L2
THE HANILTON- JACOBI-BELLMAN E Q U A T I O N
-
Sketch of proof for Theorem 3 ~
(13)
M E A1
Set -
> 0
6
79
s o small that
A2
E
is still uniformly elliptic with top order coefficients
ak d *
We may rewrite ( 3 ) as
(14)
+
max{Mu
E A
2
-
u
f, A ~ U ]= 0
0.
in
Set v = M u - f
(15) and s o
(14) becomes ~2 u
E
(16)
We already know that
v
+
v
+
= i n~
6 M
2
A u
M
let us apply
HI-;
resulting expressions makes sense i n
(17)
a.
+
H")
to (16) (the
:
M v+ = 0.
The commutator 2 2 K u z M A u - A M u
involves at most third order derivates from (15) and
(18)
of
u.
(17) that C[A
2
(v+f)
+
Ku]
+
M V+ = 0.
We rewrite (18) as
(gkL
vXk ) x
= Rl + R 2
4,
where
(XE
denotes the characteristic function of
E)
It follows
80
HAIM B R E Z I S
where
a
ap
E
(note f o r example that
Lm
can be included i n
R2
The coefficients
a
term like
by ( 1 5 ) ) . are bounded measurable and uniformly
elliptic; thus the interior estimates o f De Giorgi-MoserStampacchia apply to (19). u E H3,
Since
- -1
it f o l l o w s that
u
I
(and the usual modification when
therefore apply Theorem
,E ~L2” ,
~ J
N=l
w2 y 2 * * ( n 2 )
and so
J
A n easy bootstrap argument shows that
some u E
Lp
ux.x . E L2**(n2) 1
p > N.
a < 1.
1
F -
W e may
u E
estimates with
:w:
Then Theorem 6 . 2 i n [ 4 1 implies that 0 <
2=
5.4 i n [41 to ( 1 9 ) to conclude that
Then by equation (15) and the standard
u E
1
where or 2).
n2 r c
R,.
f o r some
v E CoYa f o r
Schauder estimates applied to (15) lead to
c2ya. References
[l] H . Brezis [2]
W.H.
-
L.C.
Fleming
-
Evans,
to appear.
R.W. Rishel, Deterministic and stochastic
optimal control, Springer ( 1 9 7 5 ) .
[33 N.V. Krylov, Control of a solution o f a stochastic integral equation, Th. Proba. Appl. 1_7 (1972) p.114-13L [4]
G. Stampacchia, Equations elliptiques d u second ordre & coefficients discontinus, Presses Univ. Montreal(l966).
G.M. de La Penha, L.A. Medeiros (eds.) Contemporary Developments in Continuum Mechanics and Partial Differential Equations @North-Holl and Pub1 ishing Company (1978)
N O N L I N E A R PROBLEMS R E L A T E D T O T I I E THOMAS-FERMI
EQUATION
HAIM H R E Z I S
D 6 p a r t e m e n t d e MathGmatiques U n i v e r s i t G P a r i s VI pl.
’I
Jussieu
75230 P a r i s Cedex 05
Consider t h e following f u n c t i o n a l
p : a Z 3 + R+
where
and
a n i m p o r t a n t example i s
V:
R3
V(X)
+ R.
V(x)
I<
m.
c
=
i s a g i v e n function:
m
where
i=t Ix-ail ai
E ~3
e(p)
where
> 0
and
a r e given.
i s c a l l e d t h e Thomas-Fermi
f u n c t i o n a l and t h e f o l l o w -
i n g p r o b l e m o c c u r s i n quantum m e c h a n i c s (1)
i
Min e ( p ) w h e r e K = ( p PEK I > 0
[41 )
E L 1 (IR 3) , p 2 O a n d
i s a f i x e d number.
The f u n c t i o n electrons.
(see
p
t o be determined r e p r e s e n t s a d e n s i t y of
The s y s t e m c o n s i s t s o f e l e c t r o n s and o f
positive nuclei of The f u n c t i o n a l
e
t h e k i n e t i c energy,
charge
mi
has 3 terms,
placed a t p o i n t s
a.
k i n space.
corresponding r e s p e c t i v e l y to
t h e a t t r a c t i v e p o t e n t i a l energy
( i n t e r a c t i o n b e t w e e n e l e c t r o n s and n u c l e i ) and t h e r e p u l s i v e p o t e n t i a l energy ( i n t e r a c t i o n between e l e c t r o n s ) . Even t h o u g h p r o b l e m (1) seems t o b e a s i m p l e c o n v e x m i n i m i z a t i o n problem, i t need n o t have a s o l u t i o n .
The
82
HAIM BREZIS
pn
difficulty lies in the fact that if sequence for (l), then 1< p
5 / 3 and
that /p(x)dx K
S
converges weakly to
pn
e(pn)
is a minimizing
7
e(p),
in Lp,
but we can only assert
In fact if we replace in (1) the convex
I.
by = (p E Ll,
and
p 2 0
/p(x)dx
then the problem becomes much simpler.
< I]
To convince the reader
that there is a serious difficulty in solving (l), we begin with a non-existence result. 1
Proposition 1
-
as
Then, n o solution of problem (1) exists (no
1x1 -+ m.
matter what Proof:
I
Clearly
Assume
V(x)
E Lloc,
V
S
0
a.e. and V ( x) -+ 0
is). e(p) 1 0
p
for
E K.
On the otherhand we
have
Indeed, we choose ball of radius
R,
I ' = ' R = -
lBRl
where
BR
denotes the
XBR
centered at
0
in
R3
and
X
is the BR
characteristic function.
It is not hard to check (using
properties of the Marcinkiewicz spaces, see e.g. 131 Appendix)
and therefore
C (p,)
-+ 0
as
R
.$
-.
Thus (1) can not have a solution since it would have to be p
E
0.
We recall first remarkable resultsdue to E. Lieb and
B. Simon [41.
NONLINEAR PROBLEMS RELATED T O THE T H O M A S - F E R M I
Theorem 2
-
k m C -i=l lx-ail
V(x) =
Assume
83
EQUATION
k
=
I.
and s e t
C mi. i=l
Then ( a ) If
0 < I < Io,
(b) If
I > Io,
( c ) If
I < I .
problem (1) h a s a unique s o l u t i o n .
( 1 ) h a s no s o l u t i o n .
problem
( 1 ) h a s compact
t h e n t h e s o l u t i o n of
support.
The proof
-
then
in
-
[ 4 ] i s i n d i r e c t : f i r s t one s o l v e s
Min E ( p ) K one shows t h a t i f 0 < I S
-
by a c l e v e r a r g u m e n t
the solution l i e s i n f a c t i n
K.
a j o i n t w o r k w i t h Ph.
( t o a p p e a r ) which e x t e n d s
Renilan
I n what f o l l o w s
I
and
Io,
describe
Theorem 2 i n v a r i o u s d i r e c t i o n s : 1)
p5”
i s r e p l a c e d by
j(p)
where
i s any
j
C1
convex
j ( 0 ) = j ’ ( 0 ) = 0.
f u n c t i o n such t h a t
m. 2)
general function d e t e r m i n e what
--I- i s r e p l a c e d by a
V(x) =
The s p e c i a l p o t e n t i a l
V(x).
I
I
x-ai O f course w e w i l l have t o
plays t h e r o l e of
Io.
3 ) The v i e w p o i n t i n s o l v i n g ( 1 ) i s t o t a l l y d i f f e r e n t f r o m t h e one i n
C41.
d e r i v e d f r o m (1) directly. ,j(p)
= pp,
some
p’s
Inf E K
=
One of
W e c o n s i d e r the Euler-Lagrange
-
which i s a n o n l i n e a r p . d . e . t h e advantages i s
that,
equation
-
and s o l v e i t
for exemple when
we c a n h a n d l e a l a r g e r r a n g e o f
p’s,
including
f o r which t h e v a r i a t i o n a l a p p r o a c h f a i l s s i n c e
-m.
Consider now
where
c
i s a normalization c o n s t a n t such t h a t
f u n d a m e n t a l s o l u t i o n of
-A
i.e.
IXI
is the
84
HAIM B R E Z I S
T h e Euler-Lagrange
e q u a t i o n d e r i v e d ( f o r m a l l y ) from Min K
(3)
e(p)
is
Here
E
p
Find
and f i n d a c o n s t a n t
K
an
[p >
03
j’(p)
-v+
an
[p
=
01
j’(p)
-
h
v +
p l a y s t h e r o l e of
with t he c o n s t r a i n t
I
B
B
1
so that
=
P
A.
2
P
a Lagrange m u l t i p l i e r a s s o c i a t e d
pdx = I .
Our n e x t r e s u l t d e s c r i b e s t h e
p r e c i s e l i n k between ( 3 ) and ( i t ) . Proposition 3
-
Assume t h a t
there i s a constant
j*(v+c)
(5)
d e n o t e s t h e c o n j u g a t e convex f u n c t i o n o f
-
j*(r) = s u p ( r s
such t h a t
E L ~ ( R ~ ) .
( 3 ) and ( 4 ) a r e e q u i v a l e n t .
Then t h e problems
C
j
Here
j*
i.e.
j(s)].
s2 0
The p r o o f t h a t when
j(r) =
o n l y when
p >
of P r o p o s i t i o n 3 w i l l be found i n [ 2 ] . and
3
V(x) = C
(in particular
mi
F5q ,
p =
-3
is
then
Note
(5) holds
O.K.)
From now o n , we w i l l c o n c e n t r a t e on ( 4 ) .
Onemain
r e s u l t i s the following Theorem k
-
Assume
(6)
V
1 E * L1
(7)
V(X) >
1.1
o
a.e.
(i.e.
AV E L1
and
V(m)
= 0)
on some s e t o f p o s i t i v e measure.
85
NOWLINEAR PROBLEMS R ELATED T O THE THOMAS-FERMI E Q U A T I O N
A)
€3)
C)
(a)
If
0 < I
(h)
If
I > Io,
When
I < I .
then
p
such t h a t
t h e r e e x i s t s a u n i q u e s o l u t i o n o€ ('t)
Io,
t h e r e e x i s t s no s o l u t i o n o f
V ( x ) -+ 0
and
as
1x1 -+
(4).
(uniformly),
a
h a s compact s u p p o r t .
S(r) = rp
If
> 0
I .
Then t h e r e e x i s t s
p 2
and
4/3
( t h i s i s j u s t a n example
-
t h e g e n e r a l assumption a p p e a r s l a t e r ) , t h e n
s I .
\(-Av)~x
r0
In p a r t i c u l a r D)
If
V E
I d e a of A)
1
dx.
z 0.
4
p > -
3
B)
= bounded m e a s u r e s o n
and C ) hold.
the proof.
(4)
we f i n d
-
v +
13
j'(p)
1x1 -+
t o transform
2
A
a.e.
on
we s e e t h a t n e c e s s a r i l y
m
(4)
P
IR
u = V - B
P
s o that
o
= u + 2
. I n order
i n t o a nonlinear p a r t i a l d i f f e r e n t i a l
(9)
j'(p)
3
A < O.
e q u a t i o n we i n t r o d u c e t h e new unknown
and
R3)
( f o r the general conditions see
F i r s t n o t e t h a t by
Letting
(-Av)' -:iV
when
(h
1.1
A),
l a t e r ) then
SVdx
-* h
j ( r ) = rP,
and
-
J
( 6 ) we assume t h a t
i n s t e a d of
(8)
=
<
u
X
on the s e t
[p >
+ X
on the s e t
[P
=
01
01.
86
HAIM B R E Z I S
In other words we find
-
u + x < o
o
i u + h >
p = o
=1
p = (j')
-1
(u+X).
We can summarize (11) i n a single equation
Y: IR
where
-b
[R
is defined by (j/)-'(r)
(13)
Problem
r >
for
r < 0
(4) becomes equivalent to Finding a function
-nu +
and a constant
u
1
=
Y(u+X)dx
the following.
(4).
First we
0 such that
o
= =
I.
(14) has been solved, the function
provides a solution of
X s
-zv
y(u+A) =
u(=
Once
o
for
~ ( r=)
p(x) = y(u(x)+X)
O u r approach i n solving
fix x
5;
0
and find
ux
(14) is
unique
solution of: -nux
+ y(ux+X) = - i v
(15)
such that
Y(ux+h) E L1.
The existence and uniqueness of
ux
follow f r o m the next
Lemma Lemma 1 ( [ 3 1 )
-
Let
function such that unique
u
8 : R + IR e(0)
solution of
= -3u
0.
+
b e a continuous nondecreasing Given
f
@ ( u )= f
E
Ll,
in
R3
there exists a and
u(=) =O.
87
N O N L I N E A R PROBLEMS RELATED T O THE TIIOMAS-FERMI E Q U A T I O N
X
Next we s e t f o r e a c h
I(X)
=
0
5;
[ Y ( u , ( x ) + X ) dx. i
I(X)
The f o l l o w i n g p r o p e r t i e s of
-
Lemma 2
I(h):
The f u n c t i o n
I ( , ) = 0.
l i m
n o n d e c r e a s i n g and
+
(-m,O]
play a c r u c i a l role: i s continuous,
[O,+m)
I ( X )
I n addition
is
X+-m
s t r i c t l y i n c r e a s i n g a s s o o n a s i t becomes p o s i t i v e and
>
I(0)
0.
I t f o l l o w s immediately t h a t w i t h
< I0
0 < I
h a s a unique s o l u t i o n f o r
= I(0)
I.
(14)
problem
and n o s o l u t i o n when
I > Io.
The f a c t t h a t c o n s e q u e n c e of
X +
as
I
i s nondecreasing i s an easy
t h e maximum p r i n c i p l e .
I ( X )
To s e e t h a t
+
0
i t i s s u f f i c i e n t t o n o t e t h a t ( b y maximum
-m
principle )
(17) and s o Next
y(uX+X) < y(V+X) + 0 I(0)
solution
> u
satisfies
X +
as
-03.
would imply t h a t t h e
of
y(uo)
E
0
+ Y (u,)
i.e.
u
S
= -AV
0
and
= V
u
-
acontradih
(7). I(x)
I n proving t h a t becomes p o s i t i v e , B) For a given
1
a.e.
I(0) = 0
otherwise
0;
I -nuo with
v
ux s
such t h a t
I
i s s t r i c t l y i n c r e a s i n g as s o o n a s i t
one u s e s Lemma such t h a t
I(X)
= I
3.5 f r o m [ 3 ] .
0 < I < Io,
satisfies
the corresponding
X < 0.
Since
88
HAIM B R E Z I S
p ( x ) = Y(u,(x)+X)
Y(V(x)+x)
p(x) = 0
follows that
when
Under t h e a s s u m p t i o n s of
C)
<
5
,/
/ p(u)' (-All)+
dx <
if'
dx.
and 1x1
V(x)
4
0
as
1x1
it
m
4
is sufficiently large.
Lemma 1 , o n e a l w a y s h a s
= /Y(uo)dx =/Y(uo)+dx
I .
Therefore
[31)
(see
dx.
The i n e q u a l i t y -AV d x < I .
(18)
i s more d e l i c a t e t o p r o v e .
I n f a c t one uses t h e following:
3 - Under t h e a s s u m p t i o n s of Lemma 1 a n d i n a d d i t i o n
Lemma
dx =
S k e t c h of
t h e proof
assume t h a t = 0
-
-
f a r away.
By a d e n s i t y a r g u m e n t w e c a n a l w a y s
E (r)
4
0
r
as
4
m
-
which shows t h a t
u(x) < c ( l x l ) 1x1 and t h e c o n c l u s i o n f o l l o w s .
-
When
holds provided
1
j(r) = - r p D p 2
1x1 F i n a l l y one
Y(u) =
2 + P-1 (u ) and
so
(19)
4 -
3 .
hold for a g e n e r a l
proof
u)
t h e n we must h a v e
then
D ) The a n a l o g u e o f Lemma 1 when
Ournext r e s u l t
u(x) S
(we u s e h e r e (19)).
proves t h a t i f
Remark
+ p(u)=
-Au
i t i s easy t o construct a supersolution
Next
h a v i n g s p h e r i c a l symmetry
where
m .
h a s compact s u p p o r t a n d t h e r e f o r e
f
if
f
i s a measure does
not
@.
r e p l a c e s Lemma 1 when
i s e s s e n t i a l l y unchanged.
f
E h,
the r e s t o f the
NONLThmAR PROBLEMS RELATED TO THE TI-IO!IAS-FER?II
Lemma
-
4
p:
Assume
R + R !
b,
f E
Then f o r e v e r y
i s a continuous nondecreasirig
= 0
@(O)
function such t h a t
89
EQUATION
and s u c h t h a t
t h e r e e x i s t s a unique f u n c t i o n
u
such t h a t
-nu + p ( u )
= f
(21)
=
u ( m )
p(u) E
and i n a d d i t i o n
o
L1.
1) A r e s u l t s i m i l a r t o Lemma h
Remarks:
(when
was p r o v e d i n d e p e n d e n t l y by B a m b e r g e r [ 11
p(u) E
such t h a t
L1
f =
and
f
6
i s a power)
.
2 ) The a s s u m p t i o n (20) i s a l s o n e c e s s a r y ; i f
u
@
( 2 1 ) h a s a solutica?
( D i r a c mass a t t h e
o r i g i n ) , then (20) holds.
1 + P-1 Y(u) = ( u ) ~
3) I n c a s e
1 j ( r ) = - rp, P p >
provided
4 3
then
( 2 0 ) holds w i t h
.
Ll i b 1i o g r a phy
[11 A .
B a m b e r g e r , E t u d e tle d e u x e q u a t i o n s n o n l i n 6 a i r e s a v e c u n e m a s s e d e D i r a c a u s e c o n d membre Benilan
[ 2 ] Ph.
-
H.
B r e z i s , D e t a i l e d p a p e r on t h e Thomas
Fermi e q u a t i o n ,
[3I
Benilan
l'h.
-
equation i n
[41
H.
Brezis
L1(IRN),
-
M. Ann.
Crandall, A semilinear Sc.
Norm.
P.
523-555.
-
B.
Simon, The Thomas-Fermi
Lieb
M o l e c u l e s and S o l i d s , Adv.
p.
22-116.
-
( t o appear).
2 (1975) E.H.
( t o appear).
i n Math.
Sup.
Pisa,
T h e o r y o f Atoms,
29 ( 1 9 7 7 )
G.M. de La Penha, L.A. Medeiros (eds.) Contemporary Developments in Continuum Mechanics and Partial Differential Equations Worth-Holland Publishing Company (1978)
GLOBAL SOLVABILITY AND HYPOELLIPTICITY
OF ABSTRACT COMPLEXES AND EQUATIONS
FERNANDO CARDOSO
and
JORGE HOUNIE
Universidade Federal de Pernambuco
1. Introduction_. -.
n
Let
be a V-aimeiisiorial
ed, orientable by
A
Cm
l),
compact, connectWe will denote
a linear selfadjoint operator, densely defined in a
has bounded inverse instance,
IDXI ) .
H,
Am'. on
(l-AX)'
which is Enbounded, positivo and (We may think of
Rn,
A
defined by
elements
u
)I UII
H
of
= ))ASullo ,
s < 0, HS
A:
if
such that where
s
2
0, HS
ASu E H,
1) 1)
s E
R,
m E R,
( f o r the Hilbert space structures) of
By
H+m
HS
is the space of equipped with the
H Am HS
is a n isomorphism onto
HSmm.
we denote the intersection of the spaces
union, with the inductive limit topology. H'and
so can
H+=
H = Ho; IIulIs =
for the norm
equipped with the projective limit topology, and by
sER,
(for
denotes the n o r m in
is the completion of
= )IASullo. Whatever
as being, for
o r a selfadjoint extension of
We will use the scale of Sobolev spaces
s E R),
if
2
manifold without boundary.
complex Hilbert space
norm
(V
H-'
and
HS,
H-m their
Since for each
can be regarded as the dual of each other, Hmm:
with their topologies, they are the
91
GLOBAL SOLVABIL I T Y AND H Y P O E L L I P T I C ITY
s t r o n g d u a l o f each o t h e r . role of
space o f
I n concrete cases
H+m
plays
the
H - ~ of distributions.
test-functions,
When s t u d y i n g a n a l y t i c i t y p r o p e r t i e s i t i s advantageous
t o introduce another s c a l e of spaces: let
denote t h e subspace o f
ES
functions
u
ES
let
the
Eo = H
s'
< s, I
embedded, w i t h a g r e a t e r norm,
s , ES
i n the
i s a Hilbert space.
ES
and f o r any t w o r e a l numbers
r e a l number
H
denote t h e completion of
Equipped w i t h t h i s norm, have
consisting of
satis€ying
s < 0,
f o r any
H
s 2 0,
f o r any r e a l
ES
arid d e n s e i n
ES
.
We
is
Given a n y
E-'.
i s t h e d u a l of
A s w i t h S o b o l e v s p a c e s we f o r m t h e u n i o n and i n t e r s e c t i o n of
t h e spaces
than going t o (1.2)
or
-m
E+O
=
ES, +m):
u
ES,
E-O
s> 0 The s p a c e s
E+O
and
E-O
> 0,
=
n S<
E ~ .
0
w i l l be e q u i p p e d w i t h t h e i r n a t u -
r a l l o c a l l y convex t o p o l o g i e s , s
going t o z e r o ( r a t h e r
s
b u t for
t h e i n d u c t i v e l i m i t of
on t h e f i r s t , t h e p r o j e c t i v e l i m i t
of the
on t h e s e c o n d , which i s t h e n a FrGchet s p a c e .
of
each o t h e r .
t h e r o l e of funct io n s
From o u r p r e s e n t s t a n d p o i n t
"analytic functions1',
62
CJ(62;Hk)
that
s < 0,
They a r e d u a l s
Efo
w i l l play
of t h e
"hyper-
.
We d e n o t e by in
E-O
ES,
t h e ES,
valued i n H+m.
Cm(n;Hfm)
t h e space of
Cm
functions
It i s t h e i n t e r s e c t i o n o f t h e spaces
(of the j-continuously
d i f f e r e n t i a b l e functions
92
FERNANDO CARDOSO and JORGE H O U N I E
R
defined i n
j, k ,
tend t o
Hmm.
a s t h e nonnegative i n t e g e r s with i t s n a t u r a l
W e equip Cm(R;H+")
+m.
8' ( R ; H - m )
W e w i l l d e n o t e by
topology. and r e f e r
Hk)
and v a l u e d in
Cm(R;H+T,
t h e d u a l of
t o i t as t h e space o f d i s t r i b u t i o n s i n
R
Cm
valuedin
A c t u a l l y , we s h a l l need f o r m s and c u r r e n t s w i t h v a l u e s
E
i n some t o p o l o g i c a l v e c t o r s p a c e
I?"
either
E*O).
or
t h e space o f E-valued
(@; tl,
I;
p G V ,
p-currents
simply distributions i n charts
0
If
...,t v )
n
in
R,6
C
0
w i l l always be
we d e n o t e by
n
p=O
(if
valued i n
of
I?
(here
A'
fl'(n;E)
they a r e
I n terms o f l o c a l
E).
a n open s e t , s u c h a
c u r r e n t i s a l i n e a r combination
(1.3) where
J
i s an ordered m u l t i - i n d e x
such t h a t
to
p,
and
1S
<.
j,
. .<
j P
dtJ = d t .
5
A...A
AP
Cm(n;E)
J
-b
of E - v a l u e d
P
)
of i n t e g e r s
i t s l e n g t h , here equal The c o e f f i c i e n t s
uJ
in
P
distributions i n
Cz(8)
l i n e a r mappings
.
dt.
1 (1.1) a r e E - v a l u e d
[ J[
V ,
(jl,. . . , j
6,
i.e.,
continuous
6.
A n a l o g o u s l y one d e f i n e s t h e space
Cm
p-forms
in
0.
( 1 . 3 ) c a n he made g l o b a l by i n t r o -
Of c o u r s e , f o r m u l a
d u c i n g a p a r t i t i o n of u n i t y s u b o r d i n a t e d t o a l o c a l l y f i n i t e covering o f
R
by l o c a l c h a r t s
s h a l l always assume t h a t in
2.
(in fact,
from now o n , w e
such formulas a r e g l o b a l l y d e f i n e d
0). G lobal S o l v a b i l i t y . ____
O u r b a s i c datum i s a on t he parameter
h E o(A)
Cm
one-form i n
( t h e s p e c t r u m of
R, A):
depending
GLOBAL SOLVAHI L I T Y A N D F I Y P O E L L I PTICITY
V
b(t,k) =
(3.1)
b.(t,h)dt
J
j=1
We s h a l l assume t h a t t h e 12
.(t,A),
jzl,..
J
folloLbs:
.,v,
of
%A(")
A-I,
iihich converge i n
x(EI,H)
o p e r a t o r s on
a s w e l l a s e a c h one o f
t
iD = d
on
t h e i r t-derivatives,
n.
" d i f f e r e n t i a l operator" i n
n
+ h(t,h)AA.
t
p = O,...,v-l,
F o r each
C"(n),
in
( t h e Banach s p a c e o € bounded l i n e a r
We s h a l l t h e n f o r m t h e (2.2)
defined as
w i t h complex c o e f f i c i e n t s
to
uniformly w i t h r e s p e c t
%,(n)
a r e t h e s e r i e s i n t h e non-
n e g a t i v e powers of
H),
..
J
"convolution operators"
belong t o t h e r i n g
t h e elements
93
i t defines a l i n e a r operator
(2.3) D"
W e set
= 0.
Our f i r s t r e q u i r e m e n t i s ( t h e complex g e n e r a t e d by t h e
= 1,.
..
i.e.
,V),
that
( 2 . 3 ) form a complex
operators
V
I m IDp c Ker
__ a
atj
+
b.(t,A)A,j= J
rihich means t h a t
the
s t a n d a r d "Frobenius i n t e g r a b i l i t y c o n d i t i o n s " a r e s a t i s f i e d , and t h i s i s w e l l known t o be e q u i v a l e n t w i t h t h e r e q u i r e m e n t t h a t t h e 1-form We
be c l u s d e d f o r a . e .
f o r a.e.
where
a.e.
XEu(A),
t h e one-form b ( t , X ) i_____ s exact
i s t o b e u n d e r s t o o d i n t h e s e n s e of
A.
We t h e n i n t r o d u c e a p r i m i t i v e of
t (2.5)
1.
s h a l l assume more:
(2.4)
measure o f
b(t,X)
B(t,X)
=
b:
(in
n),
the spectral
FERNANDO CARDOSO and J O R G E H O U N I E
94
O u r purpose is to study the equation (2.6)
IDU = f,
where form
is a E-valued p-current,
u
(p s
V
-
l),
n.
in
f
a H+m-valued, Cm (p+1)-
Let us consider a spectral
resolution of the operator
A
(considered i n
(2.7)
jX
dE(X).
A =
H):
F r o m it one easily derives that a compatibility c o r r l i t h
for
(2.6) is
Eq
the - (p+l)-form ___- e B(t,A)AE f I -
(2.8) for a.e. XEu(A), T h e set o f forms will be denoted by
f
E'+'A
is exact.
- -
satisfying (2.8)
Cm(n;H+")
gp+I Cm(h2;H+m). D
I t is our goal to complement (2.8) with conditions
B
bearing on the function
so as to obtain conditions which
a r e necessary as well sufficient for the Ttglobalsolvabilitytt of ( 2 . 6 ) . We state now o u r fundamental hypothesis about the primitive
B( t
,X)
or rather aoout its "principal part".
Let
(2.9)
B(t,X)
= Bo(t)
F o r any.real number
(2.10)
nr
=
The natural injection of
nr
+
B,(t)X-'
... .
w e write
r
t E
+
n;
Re Bo(t) < r]
into
h2,
.
gives rise to homo-
morphisms of the De Rham homology spaces:
apPr) -
(2.11)
i
P.'DP(").
We also have the transpose homomorphisms (restriction of cohornology classes):
GLOBAL SOLVABILITY AND HYPOELLIPTICITY
95
i*
4 " @ ) P4 P ( n r ) .
(2.12)
We d e f i n e t h e f o l l o w i n g " g l o b a l " ( Q )
-
Definition 2 . 1 i n d .i m e nsio - _ . __n
p
(0 5
Ker i
(2.13)
D
W e say t h a t t h e system
p
F
s
= {O],
i.e. _-
(see
has Property
(O),
0 , ______ i f f o r every r e a l
in
V - l ) l
condition
i
is injective
P
--
o r equivalently I m i* = @ P ( n r ) , P
(2.14)
i* p
i.e. ~
i s s ur j e______ ctive.
($)
We o b s e r v e t h a t a l t h o u g h P r o p e r t y i n t e r m s of on
the particular primitive
(2.5),
i s formulated
i t depends only
b.
We s h a l l s t a t e now a r e s u l t p r o v e d i n
-
Theorem 2 . 1
The f o l l o w i n g c o n d i t i o n s
( I ) : T h e system
ID
(11): G i v e n a n y e l e m e n t ..-
element -~
u
of
A'
f
are equivalent: ~
(e),
has Property
of
121 :
i n dimension
C"(O;H+")
- OaD
p,
in
t h e r.e _.___ i s an
c ~ ( ~ ; H + " ) ______ s o l u t i o n of
~u = f
in
n. (111): G i v e n a n y e l e m e n t element
u
of
f
of -
6dP+l
AP 8' (o;H-")
C"(n
ID
H'")
s o l u t on of
t h e r e i s an
iDu = f
in ~~
Remark
2.1
-
When
V
= 1,
deals with a single operator o n l y one c a s e :
p = 0 = V-1.
= S1 L =
(the unit circle),
a + at
b(t,A)A
one
and t h e r e i s
I t i s s e e n at once t h a t t h e
96
FERNANDO CARDOSO and JORGE H O U N I E
(,I,)
v a l i d i t y of
in
( p e r f o r c e i n dimension z e r o ) i s
S1
equivalent t o the following property: Re b o
-
Remark 2 . 2 model.
vanishes -identically i n
i t i s worth t o mention t h a t t h e r e i s o n e c a s e
where t h i s model i s m o r e t h a n t h a t :
t h i s i s when t h e Ley&
f o r m of t h e complex i s n o n d e g e n e r a t e .
m a t r i x of
.
1
T h e complex u n d e r s t u d y p r o v i d e s a v e r y s i m p l e
However,
means t h a t ,
S
t
whatever
B(t,h)
n
X
and
A =
(1-n X )l!'
on
E a(A),
notation i t t h e Hessian
a '€3
t,
with respect t o
I t i s shown i n [l]
i s nonsingular. situation that
in
our
In
...,
( -a-t-j-a b . ] ~. , k = 1 ,
( i n the
"concrete"
H = L2(Rn))
IRn,
V '
that
u n d e r s u c h a h y p o t h e s i s , o u r complex i s t h e t r u e m i c r o l o c a l f o r m ( p o s s i b l y a f t e r a s i m i l i t u d e by a n e l l i p t i c F o u r i e r i n t e g r a l o p e r a t o r ) o f a n y g e n e r a l complex w i t h t h e same property,
and i n f a c t c a n be f u r t h e r s i m p l i f i e d ( l o c a l l y or
m i c r o l o c a l l y ) i n t o t h e complex g e n e r a t e d o p e r a t o r s ) p r e c i s e l y b y o p e r a t o r s of
a a t 3. +
L . =-J
Remark 2 . 3
-
c . tJA, J
the type:
j=l,.
..,
V
E . = kl. J
A n i n t e r e s t i n g a l t h o u g h more d i f f i c u l t q u e s t i o n
t o investigate i s the study of if
(modulo r e g u l a r i z i n g
one r e p l a c e s
the global solvability of
(2.6)
( 2 . 4 ) by t h e following l e s s r e s t r i c t i v e
condition:
(2.15) f o r a.e.
XEa(A),
t h e one-form b ( t , X ) i s c l o s e d
(in 0).
3. G l o b a l H y p o e l l i p t i c i t y a n d H y p o a n a l y t i c i t y I n t h i s s e c t i o n , we o n l y c o n s i d e r t h e c a s e
V
= 1,
GLOBAL SOLVABILITY A K D HPPOELLIPTICITY
L-2
= Sp
T1 = [R/2nr&).
( w h i c h we i d e n t i f y w i t h Let
a
(3.1) IQhere
at
thought
of
at +
s1
b ( t , ~ ) ~ ,t E
gt
- and t h e c o e f f i c i e n t
means
aA(S1)
the r i n g
=
Sectioii 2 .
defined i n
as an oper ator oil
b(t,A) Thus,
belongs
!L
to
may be
with periodic coe€ficients
lR
and we o n l y l o o k a t p e r i o d i c s o l u t i o n s , when p e r i o d i c d a t a are prescribed. We o b s e r v e t h a t
(2.15)
holds automatically, although
i n general
( 2 . 4 ) may b e f a l s e ( s e e C o r o l l a r y 3 . 3 I s e l o h ) .
Definition
3.1
s1
We s a y t h a t
i s globally hypoelliptic i n
[L,
u E c"(s';H-")
_______ i f given
Lu E C"(Sl;H+")
(3.2) D e f i n i t i o n 3.2
s1
-
if
given
-
W e say that
u E
C"(SI;E-O)
Remark
3.1
-
Cm(S1;H+m).
i~-s g l o b a l l y h y p.o a n a l y t i-~ c in
L
~
LU E Cm(S1;E+o)
(3.3)
u E
=
u E Cm(S1;E+O).
A c c o r d i n g t o P r o p o s i t i o n 111.2.2 i r i
i s no g a i n i n g e n e r a l i t y i n D e f i n i t i o n
3.1,
o u r " s p a c e of
Crn(S1;E-O)
solutions"
C"(S';H-"),
bigger spaces o f d i s t r i b u t i o n s Remark 3 . 2
-
to
!Lu
with r e s p e c t
the coefficients i n
t)
b(t,A)
to
t
,
if
we replace
there
by t h e
Q' ( S1;E-").
Q' ( S 1 ; H - " ) ,
Traditionally i n Definition 3.2,
a n a l y t i c i t y of that
3.2,
[ 61
one assumes also
(under the hypothesis
a r e analytic with respect
and d e r i v e s t h e same p r o p e r t y f o r
u.
But
in fact
t h e r e i s no g a i n i n g e n e r a l i t y i n d o i n g t h i s h e r e :
if
the
coefficients i n
to
t,
b(t,A)
are a n a l y t i c with respect
98
FERNANDO CARDOSO and JORGE HOUNIE
u E 8' ( S 1 ; E ) ,
and if in
1 S ,
and
Lu
E,
with values in
Proposition 111.2.2 i n [ 6 ] , of
t
S1
in
t
is an analytic function of
then by virtue of a variant of will be an analytic function
u
with values in
(E
E
is one of the four
dm, E*').
spaces
In [ 5 ] , a theorem characterizing globally hypoelliptic operators in of
u(A)
of the form (3.1), w a s proved; the behavior
S1,
(the spectrum of
A)
in a neighborhood of infinity
played an important role.
In this section we shall present a n analogous theorem (see
[31) characterizing globally hypoanalytic operators in
S1,
of the form (3.1). According to the hypotheses made on
C
IR, for certain positive
(3.4)
r(z) =
m liQ m
which is analytic on
p.
> p,
= r z+rl+r2z
A
is clear that constant.
C
where
-1
+...,
z
E
C
and either has a pole of order
one or a removable singularity at
be the subset of
C
We introduce a function
zb(t,z)dt IzI
A,u(A) c [p,+-)
r(z)
z
=
m.
Let
takes integral values.
is a discrete set when
r(z)
It
is not a
We denote b y
Theorem 3.1
-
defined on
S1,
The evolution operator
L,
given by (3.1)
is globally hypoanalytic i n
if the following conditions hold:
S1
and
if and only
GLOBAL SOLVABILITY AND HYPOELLIPTICITY
(p1
Re bo
(GG)
If
Corollary 3.1
-
S1;
~
Re b o
5 E a(A)
for a l l
does -__ not change sign i n
99
0, for any positive real number
I
k
sufficiently large. ~
If -
L
~~~
is _ _ globally hypoelliptic, then it is
globally hypoanalytic. Proof:
Condition
Remark 3.3 -
-
(6)
(see [ 51 ) implies (Cis).
The converse of Corollary 3.1 is not true in
general, as follows from Corollary 3.2, b e l o w , and We recall that if
P
s m o o t h orientable manifold
hypoanalytic if
Pu E A(M) u E
is a differential operator on a M,
we say that
P
is globally
(the distributions on
(the real analytic functions on
M)
and
M)
imply that
A(M).
Theorem 3.2 on -
u E r9'(M)
[4].
S1
-
Let
b(t)
be an analytic complex valued function
and consider the vector field
(3.8) defined on the 2-torus
T'
= S1xS1 = {eit]x{eix].
Thep
P
is globally hypoanalytic if and only if the following
conditions hold:
(e ) (GLJJ)
I m b(t)
If
Im b(t)
t
d o e s n o t change sign
0, y =
211
("
Re b(t)dt
is an
10
irrational number satisfying: (A)
Given
K > 0, there is
Q > 0
s o that for
q
2.
Q,
FERNANDO CARDOSO and JORGE H O U N I E
100
lp-yql
e-Kq,
2
When
p, q
b(t)
Lntegers.
is constant in ( 7 . 8 1 ,
Theorem 3.2 yields a
result due t o Greenfield (see [ 4 ] ) namely: Corollary 3.2
-
The complex vector field
is globally hypoanalytic in Re b
if and only if
T2
-
hax, b E C,
Im b f 0
pr
is an irrational number satisfying (A).
Corollary 3.3 L
P = at
-
If -
(2.4) holds, that is if
b(t,A)
is exact,
is hypoanalytic (in fact the same is true - not globally --_ ~~
concerning global hypoellipticity). This Corollary is also a consequence of the proof of the following necessary and sufficient condition for hypoellipticity at the first step 1 E U(A),
p = 0
B(t,X)
the primitive
(see [ 6 ] ) :
for each
should not have any local
t.
minimum with respect to
References [l] Boutet de Monvel, L.
-
Hypoelliptic operators with double
characteristics and related pseudodifferential operators, Comm. Pure Applied Math., vol. 2 7 ( 1 9 7 4 ) . [2]
Cardoso, F. and Hounie, J.
-
Global solvability of an
abstract complex, to appear in the Proc. Amer. Math. SOC.
[3] Cardoso, F . and Hounie, J.
-
Global hypoanalytic first
order evolution equations,
[4]
Greenfield, S . J .
-
to appear.
Hypoelliptic vector fields and
continued fractions, Proc. Amer. Math. SOC., vol. 3 1
( 1 9 7 2 ) , 115-118.
GLOBAL SOLVABILITY A N D HYPOELLIPTICITY
[ 5 ] H o u n i e , J.
-
equations,
[ 6 ] Treves, F.
-
101
Global hypoelliptic first order evolution t o appear.
Study of a m o d e l in the theory of complexes
o f pseudodifferential operators, Ann. Math.,
vol.
( 1 9 7 6 ) , 269-324.
T h i s w o r k o f F. Cardoso w a s partially supported b y CNPq (Brasil).
104
G.M. de La Penha, L.A. Medeiros (eds.) Contemporary Developments i n Continuum Mechanics and P a r t i a l D i f f e r e n t i a l Equations @North-Holland Pub1 i s h i n g Company (1978)
ON THE INVERSE SCATTERING PROBLEM
FOR L I N E A R EVOLUTION E Q U A T I O N S
M I C H A E L 0 1 CARROLL
bepartment
of M a t h e m a t i c s
P o n t i f i c i a Universidade C a t 6 l i c a do R i o d e J a n e i r o , B r a s i l
I t i s o u r aim t o p r e s e n t r e s u l t s w i t h s i m p l e p r o o f s o f some a s p e c t s o f dimension
> 1.
n
t h e i n v e r s e s c a t t e r i n g problem i n s p a c e T h e s e r e s u l t s were o b t a i n e d j o i n t l y w i t h
Paul O t t e r s o n o f PUC/RJ. s c a t t e r i n g problem.
Let us f i r s t r e c a l l t h e d i r e c t
W e c o n s i d e r a system undergoing l i n e a r
e v o l u t i o n i n a H i l b e r t space U(t)
= e
-it H
evolution.
61
g i v e n by t h e u n i t a r y g r o u p
and want t o a n a l y s e t h e o r b i t s of
The e v o l u t i o n may r e p r e s e n t t h a t
this
o f an e l e c t r o -
m a g n e t i c f i e l d i n a d i e l e c t r i c and m a g n e t i c medium, a c o u s t i c wave i n a l i q u i d o r s o l i d , mechanical system o f p a r t i c l e s
(take
o f an
o f a quantum or c l a s s i c a l
W
t o be
L2
o f phase
space i n t h e c l a s s i c a l mechanical c a s e ) i n t e r a c t i n g through p a i r p o t e n t i a l s and w i t h a s t a t i c e l e c t r o m a g n e t i c f i e l d , e t c . One can a l s o c o n s i d e r o b s t a c l e s i n t h i s framework by d e f i n i n g H
w i t h a p p r o p r i a t e b o u n d a r y c o n d i t i o n s , f o r example, Dirichlet,
Neuman or mixed.
I n many c a s e s f o r l a r g e p o s i t i v e and n e g a t i v e t i m e s and s p a t i a l d i s t a n c e s t h e e v o l u t i o n of n o n - s p a t i a l l y ed o r b i t s of
U(t)
localiz-
may b e a p p r o x i m a t e d by a s i m p l e r e v o l u t i o n
O N THE INVERSE SCATTERING
u
(t) = e
-itHo
where s a y
Ho
103
represents a self-adjoint
p a r t i a l d i f f e r e n t i a l operator with constant c o e f f i c i e n t s , T h i s can be made p r e c i s e by r e q u i r i n g t h a t
s-lim t-b*
e
i H t e-iHot
= w*
(la)
m
e x i s t s and
R ( Y ) = Pc# where
R = r a n g e and
denotes t h e orthogonal p r o j e c t i o n
Pc
on t h e c o n t i n u o u s s p e c t r u m of f o r every
f E
If
H.
Pc#
there exist
l i m
-iHot (e
fit
( l a ) and ( l b ) h o l d t h e n
f* E #
-
e
-iHt
such t h a t
fJ = 0
t+*m
and i t i s i n t h i s s e n s e t h a t e v e r y o r b i t
(t +
asymptotically Also if
Uo(t).
H
f
behaves
( l a ) and ( l b ) h o l d t h e n t h e l i n e a r o p e r a t o r
f
ween
-iHt
l i k e t h e o r b i t s under t h e e v o l u t i o n
km)
S : # + # ,
i s u n i t a r y and
e
S-I
-
s = w ?++ w -
-+ f +
g i v e s a measure o f t h e d i f f e r e n c e b e t -
Ho.
and
Now l e t u s c o n s i d e r t h e i n v e r s e s c a t t e r i n g problem. W e consider the non-linear
function
s: c c C
w i t h domain
-b
L(#)
-t
s(c)
some c l a s s of i n t e r a c t i o n s and r a n g e t h e
unitary operators i n
#.
We p o s e t h e f o l l o w i n g q u e s t i o n s :
1.
I s t h e r e an a l g o r i t h m f o r f i n d i n g
2.
Is
S
injective?
3.
If
S
i s i n j e c t i v e and we i n t r o d u c e a Banach s p a c e
C
from
S(C)?
104
MICHAEL O'CARROLL
topology on
@,
such as
(Clg
S-I
is
continuous and
differentiable?
!I.
Can we completely characterize the image
S(@)
C
of
L(#)?
in
We now respond to questions 1-4 for some systems of interest in mathematical physics: Case I:
C =
{V:
id = L 2(R),
n=l, R + R
1
2
H = -d /dx
+
2 2 H o = -d /dx ,
V,
Faddeev El1 has given a
E L1].
(l+lxl)V(x)
2
solution to 1 by generalizing the Gelfand-Levitan equation
S
and to b e
is not 1-1. However if the scattering data is taken
and the normalization constants of' bound states
S
C
then there is a 1-1 correspondence between
and the
scattering data and 4 has a n affirmative answer. Case 11:
@ =
In V: R
-A,
8 = L ~ ( R ~ ) , H~ =
n=I,
{vE
L'
n LP
n
H = H +V
~ q , p < n/2 < q j , 3 L3j2) n L1],
+ R: {VER(Rollnik class
{v
-1/2-0
= (1+x2)
W for some $ > O
n > 3 n = 3 1
and W € L m n L
1,
n = 2
Mochizuki [ 2 ] has answered 1 and 2 affirmatively for a smaller class t h a n
@
for
n=3.
We now give a simple proof showing
that the response to 1, 2 and 3 is yes. space
S
has the kernel
-
S(k/ ,k) = 6(k/-k) where
k,k'
E Rn
lim Oig-bO
f
e
2
m i 6(kf2-k )T(k' ,k)
and for large
T(k/ ,k) =
-
In Fourier transform
- ik' - x
k2
~ ( x eikSx ) dx
-ik/.x V ( x ) [ (H-k2-io)
-1
Ve
ik.
'1
(x) d x
(2.1)
O N THE I N V E R S l i : S C A T T E R I N G
x E Rn,
where
i n t e g r a l of
k2 = kj2
V1j2
(2.1)
-1
V1’*(H-k2-is)
and
L2 -+ L
2
.
( s e e Simon
[31).
= M(II,c ,k2)
( n > 1) and G i n i b r e and Moulin [
k
31 ( n = 3 ) ,
Agmon [
41
51
(M(Ho,k
SUP
L2
as a n o p e r a t o r f r o m
From t h e e s t i m a t e s of Simon [
2
I n t,he second
i s c o n s i d e r e d a s a n e l e m e n t of
V1/’
l i m
105
2
,c)l =
0
O
-+m
s o using the resolvent equation l i m sup /M(H,k 2 k -+a W c S 1 The i d e a of
2
k‘-k = q
T(k’ ,k)
each
we i n t r o d u c e a u n i t v e c t o r
cl
and w i t h
so that
keeping
0 < a E
k(a)’
R
= k’(a)
To do t h i s f o r
fixed. n(q)
perpendicular t o k ‘ ( a ) = an+q/2
k ( a ) = an-q/2,
write 2
0.
1 and 2 i s t o i s o l a t e t h e f i r s t
the proofs o f
term o f q
, ~ ) j=
2
+
= q /I1
a
2
,
k‘(a)-k(a) = q.
For
n o t a t i o n a l c o n v e n i e n c e we d r o p t h e a d e p e n d e n c e i n what We g i v e t h e p r o o f s t o
follows.
1) W e find
1 , 2 and
( A denotes the F o u r i e r transform)
l i m IT(k‘,k)-G(q)\
l i m IVIL/(
a3 m
a+ a
2 2 IM(h,a cq / 4 , 6 ) (
sup O<ES
= ?2(q)
T ( k ‘ ,k,V1)
so that
= T ( k / ,k,V2) V1
= V2
+ lvlLP
for
n >
3
T
on
as
lvlg
B,
=
T(k’,k,V) G1(q)
=
#.
as operators i n
or
V
t h e n a s i n 1 above
3 ) I n t r o d u c i n g t h e Banach s p a c e
= 0.
1
2 ) W r i t i n g t h e e x p l i c i t dependence of if
3:
= (VILl
l v l R + lvlLl
+ for
lVILq
f
n = 3
and u s i n g t h e f a c t t h a t
s UP
SUP
2 k E[O,m) by t h e e s t i m a t e s of [
31
2
lM(Ho,k
,El
<
m
0<~<1
and [ 51 f o r
IVIB
small t h e Fr6chet
106
MICHAEL O'CARROLL
derivative of
I
e-ik'.x
T(V)
is
-1
( ( I-V(H-k2-is )
V, W
and is continuous where a +
rn
fixed
W ( I-( H-k2-ia
belong to
V) e
B.
ik..
) (x) d x
Taking the
limit gives
q
w
dV Therefore
)-'
dT(V)/dV
= ij(q).
is injective.
inverse function theorem
B
sional subspaces of
T(V) is a
By a version of the
.restricted to finite dimenC1
diffeomorphism.
Case 111: The same results as in Case I1 hold for the wave equation with a linear multiplicative perturbation, i.e.
Ott + (For
n=3
and small
V
o
(3.1)
see also Perla [ 6 ] ) .
A s in Birman [ 7 ]
the free equation with
2
( - ~ + m+v)@ =
we transform equation (3.1) and
V = 0
to the Schroedinger form
(3.2) where =
r
x = (xl,x2),
(au(O),
v(0)).
HPac(H = -A+V)
x1 = q B u, B
x2 = du/dt = V
x(0) =
and
is the positive self-adjoint operator
for equation (3.1) where
Pac
is the ortho-
gonal projection on the subspace of absolute continuity of H; for the free equation we use the operator place of
B.
A = -A
= H
in 0
T h w n i t a r p group which gives the time evolution
of ( 3 . 2 ) i n the Hilbert space
L2@L2
is
ON THE I N V E R S E SCATTERING
107
The wave operators associated with the unitary group e -iat
and the free particle unitary group = s-lim
W,(b,a)
e
ibt
e
-ibt
are
-iat e
t + f m
Combining the results o f A g m o n [ b ] on the existence and completeness o f the wave operator
W*(B,A)
= s-lim e
iBt .-iAt
t+*m and the invariance principle of Birman 161 i.e.
shows that
W*(b,a) =
W*(b,a)
where
exist and are complete.
/I (w* + w*)t)/2
(w*
-
w*)/2i
(-W* + W*)/2i
(w*
+
W*)/Z
W* = W*(B,A).
Also
s = W (b,a)*W
- (b,a),
scattering operator associated with (3.2),
where
S = W+(B,A)*
and
S*-I;
,’
the unitary
is
-
W (B,A).
From Equation (3.3) we s e e that
S-I
Furthermore
s-I
only involves
the restriction to the energy shell of
S-I,
as we have seen, is given by ( 2 . 1 ) . Rather than transforming (3.1) to (3.2) we could take
T
as given by the time independent approach to
scattering where we let
@(x,t)
= $(x,k) e -iwt
By imposing the outgoing wave condition o n
(w=(k 2+m 2 )1/2),
$(x,k)
-
ik-x e 9
108
MICHAEL 0 ' C A R R O L L
$(x,k)
e i k - x -[ ( H - k 2 - i C ) - l
i s g i v e n by
i s g i v e n by 2 . 1 .
T(k',k)
dependence
x ( x , k ) eiwt
x
condition
= e
i k -x
-ik. x
(x)
and
We a l s o have a s o l u t i o n w i t h t i m e
and by i m p o s i n g t h e o u t g o i n g wave
-
[ ( H-k2+ie )
V ( s ) x(x,k)ds.
-1
Ve
ik.
T-(k' , k )
i n g t o t h i s t i m e dependence f e
Veik"]
Both
T
'1 ( x )
.
Correspond-
i s g i v e n by
and
T
a r e considered as
-
g i v e n s c a t t e r i n g d a t a b u t a s we h a v e s e e n i n Case I1 i t i s
T
enough t o c o n s i d e r Case I V :
3.
i n responding t o q u e s t i o n s 1, 2 ,
We c o n s i d e r t h e S c h r o e d i n g e r e q u a t i o n i n a n e x t e r n a l
s t a t i c electromagnetic f i e l d , i . e . i
where
a@/at
= [ ( - i a/axi
2 A ~ )+
-
i s a s t a t i c r e a l p o t e n t i a l and
V
vlo
A = (A1,A2,
...,An)
i s t h e r e a l m a g n e t i c v e c t o r p o t e n t i a l i n t h e Coulonb gauge which we r e q u i r e t o s a t i s f y a r e t o b e summed o v e r i = 1 , 2 , We assume
V,
2 Ai
and
Ai
C = (Q:
and
A
t o be s m a l l .
U(x) = 2 i A
i
lim a+ m and a s mines
i3Ai/axi A.
= 0
Knowing
or A
i n the distributionalsense.
2 -1/2-c
Q = (l+x
=
e
+
A:
)
W E Lm
and
For large
a/axi
(repeated indices
t o belong t o t h e c l a s s
c > 0
T(k',k) where
...,n )
Rn + R :
some
= 0
i)Ai/axi
n
W
for
L']
k2
- i k ' ax
+
T(k',k)/a
V.
We f i n d t h a t
= i(q)-n(q)
q-i(q) = 0 we o b t a i n
u(x) $ (x,k)dx
2 Ai
equation
+
(4.1)
(4.1)
V and h e n c e
s u b t r a c t i n g from t h e s e r i e s e x p a n s i o n of
T
deter-
?(q)
by
a l l ternis involv-
ON T H E I N V E R S E SCATTERING
A
ing
a +
and t a k i n g t h e
limit.
m
/Q1"
have a l s o used t h e e s t i m a t e where
i s a c o n s t a n t for a l l
C
b u t d e p e n d i n g on
2 ko.
(4.1)
(Ho-k
.
2
-1
-1s)
c
2
[ko,m)
only t h e f i r s t
remains i n t h e l i m i t o f
T
we a p p r o x i m a t e d a f i n i t e number o f t e r m s i n and
V
b y f i n i t e sums o f
Ai
a r b i t r a r y mean and v a r i a n c e , r e p r e s e n t e d i ( I I o - k 2- i s ) t by i e dt f o r E > 0 and u s e d
the e x p l i c i t expression f o r the kernel
Case V:
2 - i.c ) Q1 / 2 / <
i n the interval
A l s o i n showing t h a t
t h e s e r i e s by a p p r o x i m a t i n g G a u s s i a n s of
I n t h i s a n a l y s i s we
a/axi(-4-k k2
term i n t h e s e r i e s expansion o f equation
109
e
iHot
Here we c o n s i d e r t h e K l e i n - G o r d o n e q u a t i o n m i n i m a l l y
coupled w i t h t h e s t a t i c e l e c t r o m a g n e t i c f i e l d , i . e .
-a2Q/at2 V
and
Ai
Letting
2
= (-k+m - 2 i A i
@ ( x , t )= $ ( x , k ) e
U = V2
-
2 A.
-
-iwt
ik-x
-
2Vw
-
i
T(k',k) = The s o l u t i o n o f e q u a t i o n n o t needed t o r e c o v e r -2w
2 2 A ~ - V
2iAi
and t a k i n g t h e
V
e
,
V
+
2vi
and
Ai
w = ( k +m
)
2
-1
[ (-A+U-k - i e )
s c a t t e r i n g d a t a corresponding
by
+
a r e a s i n Case I V w i t h
$(x,k) = e with
a/axi
a/axi
t o the
-ik' .x
aG/at. small. where
U eik"]
a 4
m
Ai.
(x)
and we t a k e as o u r e
-iwt
t i m e dependence
U(x) 1 ( x , k ) d x .
( 5 . 1 ) w i t h time dependence and
(5.1)
(5.2) e
iwt
Dividing equation (5.2)
l i m i t we f i n d
is
MICHAEL O'CARROLL
110
Witti
fixed take normal directions
q
determine
?(q)
and
i(q)
n(q)
and
-n(q)
to
separately.
Case VI: We consider the Dirac equation minimally coupled to a static electromagnetic field, i.e. (-iyo a/at where
yo, y
spinor.
i
+
iyi a/axi
+
m
+ vy0
are Dirac matrices and
V,
We assume
av/aXi,
and
satisfy the conditions of Case IV.
-
aiyi)@
=
o
is a multi-component
@
2 Ai
are small and
We omit the details but
by assuming a solution of equation (6.1) with time dependence -iwt and
this case is similar to Case V and we can determine
A
V
separately. In conclusion we think i t would be interesting to
extend these results to a wider class o f interaction (i.e. the
L1
condition can probably b e dropped and a distribution
type p r o o f given) and to find applications of the inverse method to solve non-linear evolution equations i n more than one space dimension.
References
111 L.D. Faddeev,
AMS Translations, Series 2, Vol.
65,
139-166 (1967).
[2] Mochizuki,
[3] B. Simon,
Proc. Japan Acad.
43 (1967).
Quantum___.___ Mechanics for Hamiltonians Defined as
Quadratic Forms, Princeton University Press, 1971.
[ k ] Agmon, S.. Ann. Scu. Norm. Sup. Pisa 2 (1975), 151-218.
ON THE INVERSE SCATTERING
[51
J . G i n i b r e a n d M.
Moulin,
Ann.
111
Inst. Henri Poincar6 21,
8 (1974).
[6]
G.
Perla,
J o u r n a l o f D i f f e r e n t i a l E q u a t i o n s 20 ( 1 9 7 6 )
NO 1, 2 3 3 - 2 4 7 .
[ 7 ] B i r m a n , M.S., 91-117
Am.
Math.
(1966).
SOC.
Transl.,
Ser.
2, Vol.
54,
G.M. de La Penha. L.A. Medeiros (eds.) Contenporary Developments i n Continuum Mechanics and P a r t i a l D i f f e r e n t i a l Equations @North-Holland Publishing Conpany (1978)
NONLINEAR BIFURCATION PROBLEM AND BUCKLING OF AN ELASTIC PLATE SUBJECTED TO UNILATERAL CONDITIONS IN ITS PLANE
DO
CLAUDE
Universit6 de Nantes (ENSM) et
Laboratoire de Mbcanique ThGorique associ6 au
CNRS, Universitg Paris VI
I. --Introduction. Many concrete situations of buckling o f elastic structures arise in engineering Sciences. The corregondingproblems have been studied since a long time, when the prescribed conditions are bilateral; they are described by an eigenvalue problem for a system of partial differential equations and boundary conditions.
One can find a general discussion i n
[161 and the corresponding mathematical analysis i n [121. However, in many situations, the structures are subjected to unilateral conditions; in these cases, we still have an eigenvalue problem, but there are some inequalities amongst the relations which describe the mechanical system. The case of elastic plates, when the vertical displacement is subjected to unilateral conditions, has been recently studied. It has been formulated as an eigenvalue problem for a variational inequality on a Hilbert space (Cf. [ 2 ]
[41 ,
[ 181
,
[ 201
,
[ 211
,
[ 221 )
.
,
[3I,
The corresponding bending
problems have been studied i n [ 61 , [ 111
, [ 171 .
A new situation with unilateral conditions, this time
113
NONLINEAR r3IFURCATION
i m p o s e d on t h e h o r i z o n t a l d i s p l a c e m e n t ,
h a s b e e n s u g g e s t e d by
e n g i n e e r s , b h o h a v e t o coiisicler a c r a c k i n s i d e t h e p l a t e . The p r e s e n t p a p e r i s r e l a t e d elastic plate, of
t o d i s c u s s t h e b u c k l i n g of
an
when a c r a c k i s e x p a n d e d a t t h e b o u n d a r y ( c r a c k
t h e clamp, f o r example); a l o n g t h e c r a c k ,
d i s p l a c e m e n t i s l i m i t e d by t h e s u p p o r t .
the horizontal
Under some assumptions,
we c a n d e s c r i b e t h e c r i t i c a l l o a d a n d e x h i b i t b i f u r c a t e d solutions; but
t h e number of
t h e s e s o l u t i o n s r e m a i n s unknown.
U s i n g v o n Karman's
theory,
we h a v e t o b e g i n t o
s o l v e a problem o f l i n e a r e l a s t i c i t y i n t h e p la n e o f plate,
w i t h u n i l a t e r a l boiindary c o n d i t i o n s
so
t h e horizontal
d i s p l a c e m e n t i s g i v e n b)- a v a r i a t i o n a l i n e q u a l i t y . ing t h i s displacement,
where
1: + c(1,;)
(I-XL), I
Eliminat-
obtain a functional equation f o r
hnic
the v e r t i c a l displacement
(1.1)
the
i s the identity,
L
i z i n g the given l a t e r a l load,
= 0,
i s a l i n e a r o p e r a t o r character-
X
i s a p o s i t i v e r e a l parameter
measuring t h e i n t e n s i t y of
t h e s e l o a d s and G i s a n o n l i n e a r
o p e r a t o r i n t r o d u c e d by
von K a r m a n ' s
The r e l a t i o n
(1.1) h a s
p r o b l e m s a n d we c a n hope
that
theory.
t h e u s u a l f o r m f o r t h i s k i n d of t h e l i n e a r i z e d problem
(1-XL)L
w i l l s t i l l g i v e t h e c r i t i c a l l o a d and t h e b u c k l i n g mode.
The
f o l l o w i n g s t u d y only g i v e s a p a r t i a l answer t o t h i s q u e s t i o n . We h a v e n o t situations
the very p r e c i s e r e s u l t s as obtained i n c l a s s i c a l
([ 1 2 1 , [ 131
of
C
i s v e r y weak:
to
I?
and
z
, C
[: 141 , [ 151 )
,
because t h e r e g u l a r i t y
i s o n l y Lipschitxian with respect
( t h a t comes f r o m t h e f a c t t h e r e i s a
v a r i a t i o n a l i n e q u a l i t y i n our problem).
Then we c a n n o t u s e
CLAUDE DO
114
methods requiring differentiability of
C
(implicit function
theorem, for example). A s s u m i n g that the first characteristic value o f
L
is
simple, we still use the classical Lyapunov-Schmidt method (Cf. [lg]),
but only
Brouwer's fixed point theorem allows us
to solve the bifurcation equation and from this alone we cannot know the number o f solutions.
We give another example (without
physical application) for which it is possible to use Picard's fixed point theorem in place of Brouwer's theorem, and so, to recover the fact that there exists only one bifurcating
branch of solutions. The plan is as follows 2. The Physical Problem.
Some results are summarized at the
end of this section. 7 . Variational formulation, and functional formulation for
the vertical displacement.
4. Bifurcation.
The Theorem 4.2 is the main result.
11. The Physical Problem.
I n its natural state, the plate occupies a bounded
open
0 in
R2,
generic point of
with sufficiently smooth boundary
n is
denoted by
the outward unitary normal and deduced from
n
by a
n/2
x = (x )
a
T=(T
n = (n )
a
A
is
) is the tangent vector,
a
rotation.
characterized by the thickness
,
r.
h,
The plate is
the flexural rigidity
D,
Greek indices take values 1 and 2; we use the summation
af
convention on repeated indices, and the notation-=
axa
-
f ,a
N O N L I N E A R BIFURCATION
v
the Poisson's r a t i o n
aa$Y6
and
115
LamGIs e l a s t i c i t y c o n s t a n t s
with
'
1
a
(2.1)
a
agY6
-
a B ~ 6- a a p b ~
A
aPAY6
' a.
--
a
~6aS
*aBAaO'
symmetric t e n s o r ;
(ao
-U A
= (A
a$
),
' 0).
H o r i z o n t a l and v e r t i c a l d i s p l a c e m e n t s a r e r e s p e c t i v e l y c a l l e d u = (u
a
)
and
C.
E q u i l i b r i u m s t a t e s , when body f o r c e s a r e
z e r o , a r e d e s c r i b e d by t h e f o l l o w i n g von Karman's r e l a t i o n s ( s e e [11):
U
Y6
i s t h e n o n l i n e a r t e n s o r of deformations used i n
von Karman's t h e o r y ; stresses
0
aP
i s the tensor of
tension
( i n the plane o f the p l a t e ) .
Boundary c o n d i t i o n s . The theorem o f v i r t u a l w o r k s p r o v i d e s t h e v a r i o u s b o u n d a r y problems a s s o c i a t e d w i t h t h e t h e o r y o f plates.
L e t us i n t r o d u c e some n o t a t i o n s :
116
CLAUDE DO
a bs
i s the tangent d e r i v a t i v e ;
a an
i s t h e normal d e r i v a t i v e .
T h e n , when body f o r c e s a r e z e r o ,
[5 ] ) ,
works can be w r i t t e n ( s e e [ 3 ] , displacement
( v ) dx =
R
-
From
a t t h e boundary i s
(2.9).
for a l l virtual
(v,z) as:
(2.11)
Remark 2 . 1
t h e theorem o f v i r t u a l
1r
aUp nB vu du
.
( 2 . 1 0 ) , w e observe t h a t the v e r t i c a l load
,a
hUuD
na
+
F,
where
F
i s g i v e n by
T h i s p r o p e r t y a l s o r e s u l t s from t h e s t u d y i n
[SI.
The boundary conditions u s e d i v the f o l l o w i n g a r e w r i t t e n i n (2.12),
On
rl
. . . , (2 . 2 0 1 .
(crack of
the clamp), the v e r t i c a l
a n d t h e r o t a t i o n a r e f r e e ( ( 2 . 1 6 ) and ( 2 . 1 7 ) ;
(2.8)
i s t h e b e n d i n g moment;
by t h e p r e v i o u s Remark 2 . 1 ) ; l i m i t e d o n a s i d e ((2.12),
t h e m e a n i n g of
displacement
M,
g i v e n by
( 2 . 1 7 ) i s provided
the h o r i z o n t a l displacement i s
(2.13),
( 2 . 1 4 ) ) and t h e p o s s i b l e
contact with t h e support i s without f r i c t i o n ( ( 2 . 1 5 ) ) .
On
r2
and
r3,
the p l a t e i s clamped
the h o r i z o n t a l displacment i s f i x e d ( ( 2 . 2 0 ) ) . the plate
( ( 2 . 1 8 ) ) ; on Finally,
i s s u b j e c t e d t o g i v e n boundary t r a c t i o n :
i s a given l i n e a r d e n s i t y of f o r c e s i n t h e p la n e o f and
a r e a l parameter
on
r3 r2,
t = (t,) the plate,
( p o s i t i v e i n t h e f o l l o w i n g ) measur-
NONLINEAR BIFURCATlON
i n g the intensity of the boundary traction ((2.20)). u
(2.12)
n
= u
a
S O ;
a
n n
R = u
(2.13)
n
S O ;
a B
R un = 0 ;
(2.1k)
,
(2.15)
0
i-1
UP
a
T
on
rl.
= 0;
13
M = 0;
(2.16)
h o
(2.17)
c
(2.18)
aB
0
aP
(2.20)
'la
+
F = 0.
? L =0 an
= 0,
(2.19)
,p
n
B
= Xt,.
r2 u r q ;
on
r2;
on
r
on
U = O
3'
In the following, we assume (2.21) and (2.22).
Then
rigid displacements of the plate are n o t possible; meas
(2.21)
r2 u
(2.22)
Remark 2.2 (2.12),
- 5
= 0
. . . , (2 . 2 0 ) ,
r3
r3
> 0;
is nonrectilinear.
,...,( 2 . 6 ) ,
is a solution of the problem (2.2)
for all real
X,
i.e. for all intensity
o f the lateral load; the horizontal displacement is the
solution of a classical elasticity problem (with unilateral boundary conditions), The question is to k n o w i f , for some real
h ,
load
X.
other solutions
exist
and to find the critical
I t is natural to introduce the Problem 2.3 (linearized problem) _ _ ~
_
-
To find
5
and
uo
such
CLAUDE DO
118
in
with the boundary conditions (for
<
and
on
r2.
n,
uo) (2.12), ...,(2.18),
(2.20) and a,;
(2.26)
nB = t,
It is an eigenvalue problem. The r e s u l t s obtained in this paper are as follows: Results -
2.4
-
Let ?,
be the first positive characteristic
value of Problem 2.3.
Then, for the problem (2.2),,..,(2.6),
(2.12), ...,(2.20), (i) (ii)
5 =
0
if
l o is simple,
lo
is the critical load).
is the unique solution;
l o is a bifurcation point (i.e.
111. Variational formulation and functional formulation.
Using the theorem of virtual works, it is easy t o obtain the variational formulation of the previous problem (2.2), ...,(2 . 6 ) ,
(2.12),...,( 2.20).
First, we have to fix
the admissible displacements, the selection of functional spaces being clear and classic.
(3.1)
V =
{V
=
(v,) E
1
(H
Let us introduce
( n ) )2 1
v = 0
on
r33 ,
119
NONLINEAR B I F U R C A T I O N
By Korn inequality,
V
is a Hilbert space.
Now, the set of
kinematically admissible horizontal displacements is the closed convex cone of K = (v
(3.3)
V:
E v
I
o
vn = v n s a a
r,}.
on
The set o f vertical admissible displacements is defined by (3.4)
2 Z = ( z E H ( n )
I
Because of(2.21) and ( 2 . 2 2 ) , a(z1,z2)
z = O ,
az-
an -
o
on
r2 u r33.
the symmetric bilinear form
introduced by ( 2 . 7 )
is a scalar product on
Z,
equivalent to the natural one; in the following, we will use this scalar product, and we will write, more simply:
(3.5) Now, we assume that
(3.6) The variational formulation is obtained from the theorem o f virtual works, using admissible virtualCKSplacemen0j; then, we obtain Problem 3.1
-
T o find
u E K
(3.7)
and z
5
E Z
d x = O
such that ' & z € Z ,
(3.9) Elimination _ ___
of
u.
We have already observed that
a solution f o r all positive real solutions, we eliminate
u
1.
5 =
0
is
To find other possible
between the relations
(3.7), ( 3 . 8 ) ,
120
CLAUDE DO
(3.9):
i t remains a f u n c t i o n a l f o r m u l a t i o n o n l y f o r
t h i s , we f i r s t s o l v c Lemma 3 . 2
(3.8), ( 3 . 9 ) ,
when
gives the existence o f
<
and
1
j.
For
a r e given.
u = ~ ( 1 , s )and u = ~(1,;).
I n o r d e r t o e x h i b i t a l i n e a r i z e d p r o b l e m , we may i n t r o d u c e the following n o t a t i o n s
(3.10)
II
(because
li
i s a cone):
u ( h , < ) = xuo
+
..(A,<),
where
u
= xuo
+
s(x,c),
where
0
O ( h , ; )
0
= u(1,O);
0
= u(1,O).
2
0.
O bser v e t h a t
Lemma 3 . 2
-
exists a
unique
(L2(62))‘,
Let
z
be given i n
u = u(X,z)
in
with t h e decomposition
Z, V,
and
h
10) T h e r e
u = a(1,z)
and
in
(3.10), s u c h t h a t
c
2 0 ) W e have t h e f o l l o w i n g e s t i m a t e s C )
(with various constants
:
(3.14)
(3.16) 30)
I n addition,
and compact f r o m
the a p p l i c a t i o n
Z
into
L2(61).
z+sag
(X,z) i s c o n t i n u o u s
NONT,TUFAR H I F ' T J R C A T I O N
Proof: results
about v a r i a t i o n a l i n e q u a l i t i e s
particular,
results
e l a s t i c i t y (cf. (and in
(7.13)
Tlle problem ( 3 . 1 2 ) )
1 0 )
L2("),
i/-
7
theorem ( c f . [
i s coercive.
( c f . [ 71
General
) axid, i n
a b o u t \ - a r i a t i o n a l i n e c l u a l i t i e s of
C 51 ) by
Uao(h,<)
121
give
(3.13)),
E Z.
u(X ,:)
e x i s t e n c e arid u n i c i t y of
i f t$e o b s e r v e t h a t
7
,Y
3 8
T h i s p r o p e r t y comes f r o m t h c imbedding
91 )
(3.17)
H1(n)
t
12"(0),
tL.ith a compact i n j e c t i o n , v e r y u s e d i n what f o l l o w s . particular,
is
z
In
we have
w i t h a compact i n , j e c t i o n . 2 0 )
U s i n g t h e e s t i m a t e , for all
- -
f , g, f, g
in
we d e d u c e from t h e g e n e r a l r e s u l t 5 c o n c e r n i n g t h e
of
L
4( n ) :
continuity
the solution o f a v a r i a t i o n a l i n e q u a l i t y with respect t o
the datas
(cf.
101 )
the rollowing, with various constants:
3 g ) F r o m t h e s e e s t i m a t e s , and (3.18), we d e d u c e t h e compactness o r the statement. Remark and
z
3.3
-
s
i s L i p s c h i t x continuous b-ith r e s p e c t t o
(at least locally i n
z ) .
W e cannot h o p e f o r a b e t t e r
result and, i n p a r t i c u l a r , h c c a n n o t hope f o r any d i f f e r e n t i a t l i l i t y properties,
because
s
comes f r o m a v a r i a t i o n a l i n e q u a l i t > - .
122
CLAUDE DO
Now,
because
3.2) a r e i n
L2(n),
uo
and
( g i v e n by ( 3 . 1 0 )
s
and because ( 3 . 1 8 ) , we have
( L
Then, by R i e s z theorem, t h e r e e x i s t in
Z
and Lemma
and
C(k,<)
such t h a t
Theorem 3.4
(Properties of
L
-
L
i s a compact s e l f a d j o i n t
Z.
l i n e a r o p e r a t o r on Proof:
L)
i s s e l f a d j o i n t because
' 0
is symmetric.
c o m p a c t n e s s comes from t h e f i r s t i n e q u a l i t y ( 3 . 1 9 )
The
and
(3.17),
(3.18). Theorem 3.5
(Properties of
C)
i s a n o n l i n e a r o p e r a t o r on
Z,
(i)
c(X,O)
(ii) z
c
5
clIzI13,
z bC(k,z)
0,
w i t h the f o l l o w i n g p r o p e r t i e s .
c
1
(c(h,z),z)
2
and 0
( i i )L e t u s w r i t e
1;
c1
constant with
z;
v z E Z; w E W
Z ;
c o n s t a n t , independent of
~ ~ C ( ~ , z ) - C ( ~5, zc)l II1~- i I I I z l l ,
there exists Proof:
2
i s c o n t i n u o u s and compact on
C(1,z)
respect t o (V)
1
For a l l
= 0;
(iii) IlC(x,z)II (iv)
-
when t h e e q u a l i t y h o l d s ,
such t h a t
e
UB
(w)
+
1 2
-z
z
,Y ,6
= 0.
N O N L I N E A R BIFURACTION
123
and use the decomposition, with clear notations:
1I;I
then, i f
S
1,
we have
Lemma 3 . 2 ,
Now,
2 0 ) , gives
and the conclusion foflows,xqith (3.18), (iii)
Comes from ( 3 . 2 3 ) and (ii).
(iv)
We write
-
IlC(h,z)
like in ( 3 . 2 2 )
C(5,z)II
and tho
conclusion comes from ( 3 . 1 6 ) . (v)
First, we know solution, f o r
(3.12),
(3.13)
v = Xuo
+ w
that ?,
(with
0,
5
z=O),
Xu
0
,
defined in ( 3 . 1 0 ) , is
of the variational inequality and we choose the test function
to obtain the inequality:
Afterwards, we choose the test function
(3.12),
(3.13)
0
to obtain + s
a@) cap
Using
v = Xu
(w) d x s
[
t w
n a a
dU.
( 3 . 2 4 ) , we deduce: /
(3.25) Computing obtain:
(C(X,z),z)
with ( 3 . 2 1 ) , and using ( 3 . 2 5 ) ,
we
in
CLAUDE DO
124
Now,
we u s e t h e c o n s i t i t u t i v e r e l a t i o n s , i n e q u a l i t i e s
and ( 3 . 1 1 ) .
(2.1),
Then, /
(C(X,z),z)
c
5
(3.26)
( C ( x , z ) , z ) 5 0;
W e get
f o r a suitable
€
ap
a@
aB
i n addition, if
then there
z,
IJ
LJ
n
(w)
+
is
w
1 5 Z,aZ
,P
dx,
the equality i s true
V such t h a t
in
= 0.
T h i s c o n d i t i o n shows t h a t t h e p l a t e can bend w i t h o u t s t r e t c h i n g d i t w a s a l r e a d y found,
M.
POTIER-FERRY
[ 113
.
S o , t h e Theorem Remark
-
3.6
i n c l a s s i c a l s i t u a t i o n s , by
3.5 i s proved.
The above d i s c u s s i o n shows t h a t ,
i s s u f f i c i e n t l y f i x e d a t i t s boundary,
if
the p l a t e
s o t h a t i t c a n n o t bend
w i t h o u t s t r e t c h i n g , t h e n we have
Functional formulation. problem f o r
5
3.7
-
Problem
U s i n g t h e above n o t a t i o n s , we g e t a
only:
Let
1 2 0
be g i v e n .
Find
5 E
Z
such
that
(3.27) where
(I-XL)< I
Remark 3 . 8
i s t h e i d e n t i t y on
-
+ C(X,C) 2.
5
A s a l r e a d y observed,
t h e previous problem,
for all
= 0,
X
2
0.
= 0
i s a solution of
Now,
we have t o look
12 5
NONLINEAR BIFURCATION
h.
€or nontrivial solutions with suitable of
Existence results
nontrivial solutions describe the buckling.
By (iii), Theorem 3.5, it is clear that the linearized problem is
We may hope that this linearized problem gives (as in classical situations) the critical load and the buckling mode. while, the regularity of
C
C
with respect to
is o n l y Lipschitz-continuous ((iv),
1
Mean-
is very weak,
Theorem 3 . 5 )
S O
we
cannot hops f o r a better resu1t.C comes f r o m the solution of a variational inequality (see Remark 3 . 3 ) . techniques of H . B . FERRY [ 1 2 1
,
The results and
KELLER and W.F. LANGFORD [ls], M. P O T I E R -
D.H. SATTINGER [14] cannot be used here, because
they require the differentiability of
C.
Meanwhile, the
classical Lyapunov-Schmidt method (already used i n these papers) will be used i n Section I V to study the Problem 3 . 7 . Let of
L,
Eo
Lo
be the smallest positive characteristic value
the associated proper subspace (of finite
dimension, because the Theorem 3 . 9 ) .
Concerning the unique-
ness, we have the Theorem 3.8
-
If
o f the Problem
h = h
0
0
S
h < h
3.7 and
c=O
tknc=0
is the unique solution
still is the unique solution f o r
if we have the following condition.
126
CLAUDE DO
I f we h a v e
5
e q u a l i t y , then
E Eo,
equality i s
but
n o t p o s s i b l e when ( 3 . 2 9 ) h o l d s .
IV.
B i f u r c a t i o n a t a Simple C h a r a c t e r i s t i c Value. An a b s t r a c t v e r s i o n of
t h e problem d e s c r i b e d i n t h e
above s e c t i o n c a n b e f o r m u l a t e d
w i t h p o s s i b l e new applicaticns.
I n t h e f o l l o w i n g , we c o n s i d e r t h i s
abstract
version,
and t h e
a s s u m p t i o n s which a r e done a r e t h e p r o p e r t i e s o b t a i n e d i n t h e concrete case o f
the previous sections.
i s a r e a l H i l b e r t space, w i t h s c a l a r product
Z
11 1 1 .
and r e l a t e d norm
(4.1)
(4.2)
(4.3)
r
L
simple;
C
L,
Eo
is
t h e a s s o c i a t e d p r o p e r subspace
Go,
with
i s d e f i n e d on
R+XZ,
(4.4)
c(x,0) = 0,
(4.5)
zctC(h,z)
c
w
x
2
Il$oll
,
)
a s follows:
the f i r s t characteristic value o f
g e n e r a t e d by
and
and
L
i s a compact s e l f a d j o i n t l i n e a r o p e r a t o r o n
l o > 0,
C
The d a t a s a r e
(
Z; is
= 1.
w i t h t h e f o l l o w i n g properties:
0;
i s c o n t i n u o u s and compact on
constant with respect t o
z,;
and
Z;
x;
N O N L I N E A R BIFURCATION
4.1
Problem
-
x
Find
(4.11)
2
and
0
<
E Z
+ ~(x,;)
(I-xL),:
127
such t h a t
= 0.
We a r e l o o k i n g f o r n o n t r i v i a l s o l u t i o n s o f problem,
t h e above
using the
Lyapunov-Schmidt
method.
It c o n s i s t s i n looking f o r
s o l u t i o n s with the f o r m
<
(4.12)
I
+ w,
= EOo
I h - X o l < rl.
w E Eo'
I
Let
P
Eo.
be t h e o r t h o g o n a l p r o j e c t i o n on
Then
( 4 . 1 1 ) i s e q u i v a l e n t t o t h e s y s t e m ( 4 . 1 3 ) , (4.14), i f sufficiently s m a l l so that
w
+
(I-IL)-l
F i r s t , we s o l v e
w e have t o f i x ; f o r this.
(4.14)
e x i s t s on
Pc(x,e@o+ w) = for a l l
is
*O'
0.
and
i n a domain
E
P i c a r d ' s f i x e d p o i n t theorem can b e used
Afterwards,
of bifurcation)
(I-XL)-'
q
we p u t t h i s s o l u t i o n i n
(4.13)
wich we h a v e t o s o l v e w i t h r e s p e c t t o
Implicit function
theorem c a n n o t b e used
sufficiently regular;
Picard's fixed point
because
C
(equation
X. i s not
theorem c a n n o t b e
u s e d b e c a u s e we h a v e n o t a s t r i c t c o n t r a c t i o n .
Only
R r o u w e r ' s f i x e d p o i n t theorem a l l o w s u s k s o l v e t h e e q u a t i o n o f b i f u r c a t i o n ; f r o m them, i n f o r m a t i o n
s o l u t i o n s c a n n o t be o b t a i n e d .
on t h e number of
128
CLAUDE DO
The m a i n
-
Theorem b . 2
result i s the
eo >
There e x i s t s
such t h a t ,
0
h,
t h e Problem 11.1 p o s s e s s e s a s o l u t i o n
all
for
; (,; #
leI<EOY
if
0
6
#
0)
with t he e s t i m a t e s :
1 = A0(1
€ao
j = 0 s p
<
C l E l
+
Then,
u), w E El,,
W)
P-1 I'WII
+
s
(and
p
clc
I P.
> 0
if
E
f 0),
t h i s s o l u t i o n b i f u r c a t e s from t h e t r i v i a l one a t t h e
lo.
point
-
Remark j1.3 By ('1.9) ~-
,
t h e u n i q u e n e s s Theorem 3 . 8 h o l d s i n
i t s s t r o n g e s t form.
-
4.4
Remark
l o a d if it is
Theorem k . 2 shows t h a t
-
Lemma j1.5
IIwI1 5
There e x i s t s and
<
IE
w = w(1,S)
solution
4.5,
14.6,
4.7
L.
s e t up t h e p r o o f
of
4.2.
< no
Ih-kol
i s the c r i t i c a l
a s i m p l e c h a r a c t e r i s t i c v a l u e of
The f o l l o w i n g Lemmas Theorem
lo
clgl',
qo > 0,
where
(4.14)
C o y
with
Eo
IIwI S
> 0
such t h a t ,
possesses a IC
I.
if
unique
I n a d d i t i o n , w e have
c
i s a c o n s t a n t with r e s p e c t
A = A(h,€)
be t h e o p e r a t o r d e f i n e d on
to
X
and
E. Proof:
Let
2
if
IEl
(4.15)
and A
Ix-hol
*(I-lL)-l
PC(h,E~o+z);
are s u f f i c i e n t l y small,
then:
i s a s t r i c t contraction on the b a l l
R(0,IeI).
NONLINEAR B I F U R C A T I O N
T o s e e t h i s , we c h o o s e
(4.16) so,
when
(4.6), we
IG
if
Ih-hol
I I ~ z l ls
cIc
EL
a n i n v e r s e on
1
< Eo,
there c
are
d
for a l l
i s proved.
c:
IS\
6
<
if
0'
z
1.
and
tliat i s necessary,
we o b t a i n
the statement
w(X,E)
c
I G I P. of
(4.14),
IX-X~
(c
16
1
we h a v e t h e
< Co):
constant i n
But
so
and t h e r e s u l t f o l l o w s f o r
c0
Then,
we h a v e
= IIAwll S
C ~ E I
Now, b y
B(O,ISl),
L i p s c h i t z c o n d i t i o n (uniform with r e s p e c t t o l l w ( ~ , ~ ) - w ( i , ~ ) I 5I
I).
11 z - ~ l l;
c c p-l
(4.16),
For the solution
1;
IS
The f i r s t p a r t o f
Eo.
But now, by
-
in
1 1
IE
B(0,
a constants, ~ independent of
l!lyll
Lemma 4.6
5
possesses
i t i s e a s y toohtajn
Then, IIzII 5
if
and
z
AZ-AZII
a f t e r a reduction o f
(4.15)
no.
o p e r a t e s on
A
obtain,for a l l
I)
Ip,
(I-hL)
such t h a t
qo
c
129
sufficiently s m a l l .
x
and c ) .
CLAUDE DO
130
Now, let us study the bifurcation equation (4.13) after w = w(X,E),
substitution of Lemma
4.7
-
I
IE
For all
<
C
solution of o,
(4.14). = X (E )
there exists
such
that:
o
(4.18) (4.19)
(1
x -
< -
x
-
1 s
*
if
CJEIP-',
E
# o
( c ( 1 , E $ o + w(X,E)),Qo)
= 0-
0
Proof: We use Brouwer's
(4.7) and
have by
Lemma
I~C(X,E:G~
fixed point theorem.
First, we
4.5
+ w(a,e)>ll
4
~ I I +~ wG( x~, d P
c ~ ( E ( ~ ;
and it follows that P- 1
I n addition, (4.18) and Lemma I I C ( X , E ~+ ~
-
w(~,c))
That is enough to use
4.6 give + ~(x,E)II
c(X,CQ,
Brouwer's
c3lc
I 1x-X I
fixed point theorem.
The
conclusion follows. Remark 4.8 Lemma
-
Theorem
4.2 (and
4.7) shows that the solutionsof
(4.19) are inside the dark region the annexed picture.
of
But, without
uniqueness, a local representation
X = X(P)
is not possible.
I n fact,
when the assumptions o f the implicit function
theorem are not valid, all the dark region can be
h
131
NONLINEAR BIFURCATION
f ( h ,F:)= 0 .
s o l u t i o n o f a problem l i k e i n g c l a s s i c a l e x a m p l e , where
f(A,e) =
r
I
f
0
-E -h -2E
2
2
W e r e c a l l t h e follow-
i s d e f i n e d by: 2
,
if
h
,
if
-E2
,
if
0 s h s
,
if
E 2 S
,
if
h
-E
5
s
< 0 6
S
29
2
x
2
29
2
2
ExoEpt f o r t h e c o n t i n u i t y of p a r t i a l d e r i v a t i v e s a t
the origin,
e v e r y a s s u m p t i o n o f t h e i m p l i c i t f u n c t i o n theorem h o l d s . we check t h a t
is Lipschitzian, but
f
t h e c o n c l u s i o n s of
i m p l i c i t f u n c t i o n s theorem a r e n o t v a l i d : representation
-
Remark '+.9
member o f
Picard's
On t h e c o n t r a r y , i f
replaces
r e p l a c e d by
a local
f i x e d p o i n t theorem t o
b e c a u s e we h a v e no i n f o r m a t i o n on t h e c o n s t a n t
i n (4.20).
below,
(4.10),
I
cl(e 9
( 4 . 2 0 ) by
1h-A
t h e assumption ( l k . 1 0
t h e r i g h t member i n
I +
c4l?
1'
clP
I
IX-il,
1) W-GI~ with
4 . 7 and we o b t a i n T h i s assumption
C ( h , z ) = XC(2).
l o c a l e x i s t e n c e of a
bis),
( 4 . 1 7 ) can be and t h e r i g h t
a >
Then,
0.
Picard's f i x e d p o i n t t h e o r e m c a n b e u s e d i n t h e p r o o f
this case.
the
i s not p o s s i b l e for t h e s e t
X(C)
We c a n n o t u s e
solve (4.19) c4
h =
Also
unique
of Lemma
h = X(o).
( 4 . 1 0 b i s ) i s , f o r example, s a t i s f i e d i f Theorem 4 . 1 0
But, p r o b a b l y ,
is the result
obtained i n
(4.10 b i s ) i s not true f o r the
problem d e s c r i b e d i n s e c t i o n s I1 and 111.
CLATJDE DO
132
Then, there exists one, unique, family o f nontrivial solutions
4.1 and i t possesses the representation
for the Problem
Besides, the estimates given i n the statement of Theorem 4.2 hold good.
References [11 L. Landau et E. Lifschitz, ThGorie de l'Glasticit6, Physique th6orique , t. V I I , Editions M i r y lloscou (1967). [2] C1. D o ,
~'robli?mes de valeurs propres pour une in6quation
variationnelle sur un ccne et application auflambement unilateral d'une plaque mince, C.R. Acad.
Sc.,
A , 280,
P. 45-48 (1975). c3]
C1. Do, The Buckling Plate Subjected ~- o f a Thin Elastic . __ ~~~~
~
to Unilateral Conditions; Proceedings of the IUTAM/IMU Symposium -
on Applications of Methods of Functional
~
Analysis to Problems in Mechanics, Lecture Notes i n Math.
[4] C1.
Do,
,
5 0 3 , p. 307-316, Springer ( 1 9 7 6 ) .
Bifurcation Theory for Elastic Plates Subjected
to Unilateral Conditions, -~ 6 0 , no 2 ,
[51
J. M a t h . A n a l . A p p l . ,
p. 435. 448, 1977.
G. Duvaut et J . L .
Lions, Les in6quations - - . _en _ MGcanique - . et _ _
en Physique, Dunod, Paris (1972).
133
N O N L I N E A R BIFURCATION
Duvaut e t J . L .
G.
Lions,
Prob1;rnes --
u n i l a t 6 r a u x d a n s ~la
t h e o r i e d e l a f l e x i o n f o r t e d e s p l a q u e s , J . MGca., no 1 J.L.
(1974).
Lions,
Q u e l q u e s rn6thodes d e - r 6 s o l u__t i o n .d-e s p r o b l & m e s ~~~
aux -~ A.
~
Necas,
Sur- l e f l a m b e m e n t d e s p l a q u e s
rniE"s,
Cycle, U n i v e r s i t g de P o i t i e r s
(1976).
Les m6thodes d i r e c t e s d a n s l a t h e o r i e d e s __.
6 q u a t i o n s e l l i p t i q u _e s_, ~
Prague,
(1969).
Dunod, P a r i s
~
Cirnetiere,
G.
~
l i m i t e s rion l i n 6 a i r e s ,
T h e s e d e 3;me J.
13,
e t Masson,
Acad.
T c h g c o s l o v a q u e d e s Science,
(1967).
Paris
Stampacchia, V a r i a t__ i on a l I n e q u a l i t i e s ; T-heory t i o n s o f Monotone O p e r a t o r s , _A -p p l i c a.~
and
Proceedings of a
NATO Advanced S t u d y I n s t i t u t e , V e n i c e ( 1 9 6 8 ) . M.
Potier-Ferry,
C harge s -l _i r_ n-i t e s - e t Th6orerne d ' e x i s t e n c e ~~
~
e n t h 6 o r i e d e s plaqles 6 l a s t i q u e s , C . R .
280, p. M.
Sc.,
A,
1317-1319 e t 1385-1387 ( 1 9 7 5 ) .
Potier-Ferry,
B -i f u r c a t i o n e t s t a b i l i t 6 p o u r d e s
s-y s t 6 m e s d 6 r i v a n t d ' u n . p o.t_ entiel, ~
H.B.
Acad.
K e l l e r e t W.F.
& paraqtre.
L a n g f o r d , I--___ terations , -P e r_ t u_ rb_ a t i_ o n_ s ~
and M u l t i p o r N o n l i__ n e ar.-B _ l_i c -i t.i.e s f-~ - i f u r c-a t i o n P r o b l e m s , Arch. D.H.
Rat.
Mech.
Anal.,
48,2, p . 83-108 ( 1 9 7 2 ) .
S a t t i n g e r , S t a_b_i l i t y o f B i f u r c a t i n g S o l u t i o n s b y
L e r a y - S c h a.___--uder Degree, A r c h .
- .
no 2 ,
D.H.
p.
Rat.
Mech.
Anal.,
43,
154-166 (1971).
S a t t i n g e r , Topics i n S t a b i l i t y and B i f u r c a t i o n
T h e o r y , L e c t u r e Notes i n Math.,
309, Springer ( 1 9 7 3 ) .
CLAUDE D O
134
[16] B. Budiansky, Theory _ _ _ of _ ~Buckling and Post-Buckling Behavior ____-- of Elastic Structures, Advance in Appl. Mech.,
14, Chia Shyn Yih ed., Academic Press (1974). El71 J. Naumann, On Value Problems __Some Unilateral - -__- Boundary _in Nonlinear Plate Theory, A paraztre dans Beitrdge zur Analysis.
[I81 J. Naumann et H.U.
Wenk, On Problems for ._ _Eigenvalue _ _ . -.. ______
Variational Inequalities; an Application to Nonlinear __.___ ___---__--___ -Plate Buckling, Rendiconti di Matematica ( 3 ) , Vol. 9, s6rie VI, p. [l9] M.S.
439-463 (1976).
Berger, A Bifurcation Theory for Nonlinear Elliptic
Partial Differential Equations and Related Systems; Bifurcation Theory and Nonlinear Eigenvalue Problems;
J.B. Keller et S. Antman ed., Benjamin (1969). [ Z O ] E.
Miersemann, Uber Nichtlineare .~ Eigenwertaufgaben in konvexen Mengen, Preprint, Karl Marx Universit3t, Leipzig (1976).
1211
R.C. Riddell, Eigenvalue Problems for Nonlinear ___ - - Elliptic Variational on a Cone, J. Functional Anal., . -. Inequalities - - ..._._ to appear.
[22]
R.C. Riddell, Eigenvalue Problems for ___Nonlinear Elliptic ~
-_
Variational Inequalities, to appear. -._
G.M. de La Penha. L.A. Medeiros (eds.) Contemporary Developments i n Continuum Mechanics and P a r t i a l D i f f e r e n t i a l Equations @North-Holland Publishing Company (1978)
OY Tflk’
THERMODYNAMICS OF NON-CLASSICAL SYSTEMS
BERNARD D. COLEMAN Department of Mathematics Carnegie-Mellon University Pittsburgh, Pennsylvania 15213
In recent years David R. Owen and I have been developing a mathematical theory of thermodynamics in which the existence and regularity of caloric functions o f state, such as entropy and free energy, are deduced, rather than assumed [ 11 -[ e.g.
51 .
When applied to
several classes of materials,
elastic and viscous materials, materials with internal
state variables, and materials with fading memory, our new formulation o f thermodynamics yields restrictions on the constitutive relations for the stress, temperature, and heat flux agreeing with those obtained i n treatments [ 61 -[ 103 which start with a differentiable entropy function and employ the Clausius-Duhem inequality. to hypoelastic materials [ 3 ]
The theory has been applied also and elastic-plastic materials
[Z] [4] for which the thermodynamical restrictions had not previously been fully understood.
The new formulation o f the
principles of thermodynamics has two advantages:
(1) It
permits one to determine the class of entropy functions compatible with given constitutive relations for stress, temperature, and heat flux.
(2) It permits one, in principle,
to find thermodynamical restrictions on such constitutive
136
BERNARD D.
r r l - i t i o n s without
e v e r m e n t i o n i n g e n t r o p y or f r e e e n e r g y .
T h i s s e c o n d a s p e c t of
t h e t h e o r y was i l l u s t r a t e d i n a n o t e
i n t h e Reridiconti d e l l ’ b c c a d e m i a d e i Lincei
r e c e n t l y published
[41.
COLEXAN
I t i s t h e m a t t e r of t h e e x i s t e n c e of e n t r o p y f u n c t i o n s
which I s h o u l d l i k e t o d i s c u s s t o d a y .
L 51
Owen and I g i v e p r o o f s
of
I n r e f e r e n c e s [l]
t h e theorems s t a t e d below.
(C , V ) i n
I n the present theory, a system i s a p a i r
Z
which
i s a t o p o l o g i c a l space shose elements
II
s t a t es , and
a n open s u b s e t called form
g(P)
Pp:
C + R(P) ’6
P-valued f u n c t i o n t
Z
of
Z;
pp
P
)
H
#
Q Y
P”P‘
i s called
pP’,
ll &(P”)),
Pp
is
of
R(P).
Thus
I t i s assumed
I”’
processes
and
there i s specified a with t h e p r o p e r t y t h a t
PNP’,
i.e.,
is the
for e a c h
t h e r e holds
J
pPNplD
in
= pp,,p’a;
t h e p r o c e s s r e s u l t i n g from t h e s u c c e s s i v e
P‘
a plication of __
with
of
and i s w r i t t e n i n t h e
i s d e n o t e d by
(P” yP‘)
pairs
With
Pp
t h i s function
and i s c o n t i n u o u s .
(I)” , P ’
pPN
-1 f i ( p ” p ’ ) = pp’(6?(P’)
if
into
t h e transformation induced b y
“composition” of
P”P’
C
R(P’) n a S ( P ” )
f o r which
P P// p‘
Z
of
The r a n g e o f
on t l i e s e t
that
I”
Q(P)
p,a. C
a r e called
i s a s s o c i a t e d a c o n t i n u o u s mapping
t h e t r a n s f o r m a t i o n inducg-d by 0 I--,
u
i s a s e t o f e l e m e n t s c a l l e d __ p r o-c___ esses.
each process t h e r e
and
and P”. -
(z,n)
The p a i r
f o r m s a system
t h e s t r u c t u r e ,just described has t h e following p r o p e r t y ,
which w e c a l l p r o p e r t y
I:
F o r eacli
states accessible f r o m
uY
i.e.,
i.s d e n s e i n
0
in
c,
the set of
t h e set
C.
We have r e c e n t l y observed [ 5 ] t h a t s e v e r a l o€ our r e s u l t s f o r s y s t e m s h o l d a l s o f o r more g e n e r a l o b j e c t s c a l l e d
ON THE THERMODYNAMICS O F N O N - C L A S S I C A L
F o r a semi-system,
semi-systems.
n
the s e t
assumed t o h a v e p r o p e r t y
-
I;
o f s t a t e s and
a s t r u c t u r e of' t h e t y p e
o f processes possess
employed t o d e f i n e s y s t p m s , b u t
c
the s e t
137
SYSTEYS
t h i s s t r u c t u r e 1s no l o n g e r
w e assume i n s t e a d t h a t t h e
s t r u c t u r e h a s a l e s s r e s t r ' i c t l v e p r o p e r t y , which I s h a l l h e r e
I t :T h e r e
c a l l property
I
n u o := { p p a o i s dense i n
C
in
C.
A state
such t h a t
the
set
E l9(P)3
Do
such t h a t t h e s e t
Do
1%
E n,
p
is c a l l e d a b a s e s t a t e .
b a s e s t a t e , and a s y s t e m
oo
is a state
70
i s dense
A semi-system h a s a t
l e a s t one
a s e m i - s y s t e m f o r which e v e r y
s t a t e i s a base s t a t e . Let
noc
fi
An action
nOc
-4
IR
(i)
19 (P"P' )
:=
n ~ cI
u E ~J(P)}.
(C,F)
i s a function
with the following two properties: Additivi-ty
,
(ii)
(X,y).
E
f o r a semi-system
-
if
Continuity IR
(P",P')
is in '6
and
C7
is in
then ,cL(l-"'P'
$(P) +
((r.,~)
-
+
,a) = p(P' ,a) f o r each
p(p" , p p ' " ) ; 9,
in
P
the function
&(P,.):
i s continuous.
Let
A
b e a n a c t i o n and
If
@
i s a subset of
s t a t e o f a semi-system
0
C,
.A{O + S]
we w r i t e
t h e s e t o f numbers o b t a i n e d by l e t t i n g
P
vary over t h e
p r o c e s s e s whose i n d u c e d t r a n s f o r m a t i o n s t a k e A{.
If € o r e a c h
-4
c > 0
(43
:=
(,A(P,U)
1
P
E
n,
for
u
into
@,
i.e.
of
a
p p a E 81.
t h e r e i s a n open n e i g h b o r h o o d
8
U
BERNARD D.
138
COLEMAN
such t h a t
u.
t h e n P i s s a i d t o have t h e C -l a u s i u s p r o p e r t y a t
( C ,ll)
i s a b a s e s t a t e of
8
neighborhood
and i f
each s t a t e
u
uo
If
h a s an open
such t h a t
D
t h e n we s a y t h a t
i s u p p e r l i m i t e d from
,&
F o r systems,
uo.
t h e f o l l o w i n g t w o theorems a r e e a s i l y
p r o v e d [ 11 : Theorem 1
-
a system,
then
from
i.e.,at
0,
I f A -
,a, h a s t h a t p r o p e r t y a t a l l s t a t e s a c c e s s i b l e a l l s t a t e s i n the s e t
h y p o t h e s i s , dense i n
-
Theorem 2
uo
which i s , by
An a c t i o n kL h a s t h e C l a u s i u s p r o p e r t y a t a s t a t e
F o r g e n e r a l semi-systems, t o base s t a t e s ,
i s upper l i m i t e d f r o m
Do.
when a t t e n t i o n i s c o n f i n e d
t h e p r o p e r t y of b e i n g u p p e r l i m i t e d i s a t
l e a s t as strong as the Clausius property,
-
nu,
C.
of a s y s t e m i f a n d o n l y i f ,&
Theorem 3
of
u
has t h e C l a u s i u s property a t a s t a t e
f o r we have C51:
If a n a c t i o n ,A f o r a s e m i - s y s t e m i s u p p e r
from a base s t a t e . -
00'
thenp -
limited
has t h e C l a u s i u s p r o p e r t y a t
Do* There a r e semi-systems
( b u t n o t s y s t e m s ) , f o r which
a n a c t i o n which h a s t h e C l a u s i u s p r o p e r t y a t a b a s e s t a t e n e e d ' n o t be upper l i m i t e d f r o m t h e base s t a t e . I f p. i s a n a c t i o n f o r a s e m i - s y s t e m A:
G
C
C
4
lR
i s c a l l e d an upper p o t e n t i a l f o r
(C
,n), ,&
i f
a functim
O N THE THERMODYNAMICS O F NON-CLASSICAL SYSTEMS
(1)
t h e domain
(2)
whenever
E > 0,
n
Note: -
If
Pa
of
u1
i s dense i n
A
a2
and
are i n
8
a n open n e i g h b o r h o o d
in
P
G
with
Pal
8,
in
G ,
u2
of
and
t h e r e i s , f o r each such t h a t ,
f o r every
there holds
i s a n u p p e r p o t e n t i a l f o r p-
A
C,
are b o t h i n t h e domain of
,
u
and i f
4
-
If
f o r a semi-system h a s an upper
then p has the C l a u s i u s p r o p e r t y a t every
s. t a t e i n t h e domain ~
[1][5]: ~~
A,
potential
an a c t i o n A ,
and
then
A,
The f o l l o w i n g theorem h a s a v e r y s h o r t p r o o f Theorem
1 39
G
f A.
~~~~~~
A far less t r i v i a l result
i s t h e f o l l o w i n g theoremC51,
w h i c h , w i t h i t s c o r o l l a r y , Theorem 6 ,
i s the principal result
[l] of t h e p r e s e n t t h e o r y : Theorem 5 if
-
I f p. i s a n a c t i o n f o r a s e m i - s y s t e m -
t h e r e i s a base s t a t e
limited,
then
L%
uo
in Z
(C,n)
from & , ____ which --
has an upper p o t e n t i a l
A
and
i s upper
which i s upper __
semicontinuous. I n v i e w of Theorem 2 ,
t h e theorem j u s t s t a t e d has t h e
following corollary: Theorem 6
-
I f p. i s a n a c t i o n f o r a s y s t e m -
there is a s t a t e i n th e n L?,
C
a t which U ,
(C,n),
and i f
h a s t he.-~ Clausius property, ___
has an upper semicontinuous upper p o t e n t i a l .
~
BERNARD D .
140
COLEMAN
Without w r i t i n g o u t h e r e o u r p r o o f s h o u l d l i k e t o show you t h e form of t h e proof y i e l d s .
(C,n),
l e t ,&
Let
oo
5, I
of Theorem
t h e u p p e r p o t e n t i a l which
be a base s t a t e o f
b e u p p e r l i m i t e d from
no,
t h e semi-system
and c o n s i d e r t h e
(oo,o):
f o l l o w i n g f a m i l y o f open s e t s a s s o c i a t e d w i t h t h e p a i r
6 ( a o , u ) := { Q c C
I
O o p e n , UEO, s u p p / { o o + S] <
m } .
Let
Here t h e infimum may o r m a y n o t b e f i n i t e ; set states
0
We s h o w t h a t in C.
for which i t i s f i n i t e , i . e .
nuo
i s a s u b s e t of
be t h e
let
C o and h e n c e
We a l s o s h o w t h a t t h e f u n c t i o n A o :
Co
let
C o + R,
co
i s dense
d e f i n e d by
i s u p p e r s e m i c o n t i n u o u s and i s a n u p p e r p o t e n t i a l f o r A . If
the s e t
m(oo,uo) = 0 P
Ao,
in
n
Co
uo,
contains the base s t a t e
and i f
( f o r t h i s t o b e s o i t s u f f i c e s t h a t t h e r e be a
and s u c h t h a t
Pao =
uo
and
,&(P,ff
) =
0),
then
d e f i n e d a b o v e , i s t h e minimum e l e m e n t o f
t h e s e t of a l l
Co
and v a n i s h a t
u p p e r p o t e n t i a l s f o r ,& which have domain 00
T h i s r e s e a r c h w a s s u p p o r t e d by t h e U . S . Science Foundation.
I am v e r y g r a t e f u l
National
t o our B r a z i l i a n
c o l l e a g u e s w h o have s p e n t time a t Carnegie-Mellon namely P r o f e s s o r s Guilherme M . F i l h o , and L u i z C a r l o s M a r t i n s ,
University,
d e L a P e n h a , Rubens Sampaio f o r making t h i s v i s i t t o B r a d
ON THE THERMODYNAYICS OF NON-CLASSICAL SYSTEMS
141
possible and f o r their friendship and kindness to me here.
References [l] Coleman, B . D . ,
D.R. Owen:
&
Arch. Rational Mech. Anal. 54,
€or thermodynamics.
1-104
A mathematical foundation
(1974).
[ 2 1 Coleman, B.D.,
D.R. Owen:
&
On thermodynamics and e l a s 6
59, Arch. Rational Mech. Anal. -
plastic materials.
25-51 ( 1 9 7 5 ) .
[3] Coleman, B.D.,
& D.R,
Owen:
On thermodynamics and
intrinsically equilibrated materials.
108,189-199
Pura Applicata (LV),
[4] Coleman, B.D.,
&
D.R. Owen:
plastic materials.
&
(1976).
Thermodynamics of elastic-
Rend. Accad. Lincei (Cl. Scienze
fisiche, Serie VIII)
[ 5 ] Coleman, B.D.,
Annali Mat.
61,
D.R. O w e n :
77-81 ( 1 9 7 7 ) . On the thermodynamics of
semi-systems with restrictions on the accessibility 6 6 , 173-181 ( 1 9 7 7 ) . of states. Arch. R a t i o n a l blech.Ana1. -
[6] Coleman, B.D.,
& W. N o l l :
The thermodynamics o f elastic
materials with heat conductionand viscosity. Rational Mech. Anal.
[ 7 1 Coleman, B.D.,
&
3,167-178
V.J. Mizel:
(1963).
Existence of caloric
equations of state i n thermodynamics.
9,1116-1125 [8] Coleman, B.D.:
Arch.
J. Chem. Phys.
(1964).
Thermodynamics o f materials with memory.
Arch. Rational Mech. 1 7 , 1-46 (1964).
BERNARD D .
142
[9]
Coleman, B.D.:
O n thermodynamics, s t r a i n impulses, and
viscoelasticity.
230-254
COLEMAN
Arch.
R a t i o n a l Mech. Anal. 17, -
(1964).
[lo] Coleman, B.D.,
& M.E.
Gurtin:
internal state variables,
(1967).
Thermodynamics with
J. Chem. Phys.
47, -
597-613
G.M. de La Penha, L.A. Medeiros (eds.) Contemporary Developments i n Continuum Mechanics and P a r t i a l D i f f e r e n t i a l Equations @North-Holland P u b l i s h i n g Company (1978)
OK AN INEQUALITY IN THERMODYNAMIC STABILITY
ROGER L. FOSDICK
1
Department of Aerospace Engineering and Mechanics University o f Minnesota Minneapolis, Minnesota 55455
1. Introduction. -.
In recent years significant progress has been reported on the relationship between the thermodynamics and
A key step in much of
the stability of a general continua.
this work is the demonstration that a functional o f the form
is, i n a broad class of processes
P,
a non-increasing
function of time, i.e.,
61dt for all
t > 0. Here,
Be >
@(t) s 0 0
is the prescribed constant
equilibrium absolute temperature;
and
E(.,.)
n ( - , * ) are,
respectively, response functions f o r the densities of internal energy and entropy;
e(X,t) > 0
and
K(X,t)
are,
w
respectively, the fields o f absolute temperature and a deformation tensor (e.g., an appropriate strain tensor for an 'This article is based o n work that was done jointly with J. Ernest Dunn, a more detailed version o f which will appear shortly in the Archive for Rational Mechanics and Analysis. See, f o r example, [: 1-71 and, for a general theory [: 81
.
144
ROGER L. F O S D I C K
elastic material,
rl-31,
c.f.
or t h e s t r e t c h i n g t e n s o r i n
c e r t a i n incompressible f l u i d s , c . f . velocity field. d e n o t e s t i m e and
the processes
course, i n the foregoingdescription,
t
B
i s t h e body m a n i f o l d w i t h p a r t i c l e s
XEB
1
$ ( t ) is, for
t h e i n e q u a l i t y ( 2 ) shows t h a t
IP,
o f
bounded ____ above by i t s i n i t i a l v a l u e @ ( O ) ,
a f i n i t e lower bound f o r
t h e e x i s t e n c e of
i s the
v(X,t)
Of
and mass measure m ( . ) . While
[9]);
@ ( t ) i s provided
by t h e i n e q u a l i t y
(e,E) f ( B e , K e ) ,
for all
justified [1,2] w i t h ( 2 ) and result
under
an i n e q u a l i t y which E r i c k s e n h a s
f a i r l y natural conditions.
Moreover,
( 3 ) t h e f o l l o w i n g r a t h e r weak i m p l i c i t
stability
i s immediate:
i
B
f(e,K)dm
@ ( o )-
where t h e p o s i t i v e d e f i n i t e f u n c t i o n
f(.,-)
C+,+eed-,.>l
me,
f(.,.)
(4) i s d e f i n e d by
- C ~ ( e e l ~ e ) - e e ~ ( e e , ~, e ) l( 5 )
The f o l l o w i n g q u e s t i o n may b e r a i s e d :
Since i n
1 F o r s i m p l i c i t y , we assume t h r o u g h o u t t h i s w o r k t h a t B is regular a t a l l X E B. T h i s n o t i o n o f r e g u l a r i t y i s made p r e c i s e i n t h e forthcoming paper r e f e r r e d t o i n f o o t n o t e 1 on page 1. I n words, i t means r o u g h l y t h a t B h a s a smooth c o n n e c t i n g s t r u c t u r e s o t h a t a t any X E B any n e i g h b o r h o o d can b e decomposed i n t o d i s j o i n t s u b b o d i e s e a c h h a v i n g a p r e s c r i b e d f r a c t i o n a l mass t o t a l i n g t h e mass of t h e neighborhood.
ON AN INEQUALITY
practice neither known,
nor
€ ( a , * )
the function
I N THERMODYNAMIC
I](.,-)
i s ever e x p l i c i t l y
-
i s a l s o n o t known
f(.,.)
145
STABILITY
u n d e r what
c o n d i t i o n s i s i t n e v e r t h e l e s s p o s s i b l e t o c o n c l u d e from
more e x p l i c i t bounds f o r
8(
a
,
t)
c o n n e c t i o n Coleman and G r e e n b e r g
-
K(
and
[4]
a
,
(4)
t)? In this
and Coleman
~ s t a b l i s h e dt h e i n t e r e s t i n g r e s u l t t h a t i f
[5]
f(.,-)
llave
i s not
o n l y p o s i t i v e d e f i n i t e b u t a l s o convex o u t s i d e a compact s e t
>
then f o r every
where
II*II
points
0
there exists a
6 > 0
d e n o t e s a norm on Lhe s p a c e o f n - t u p l e s
(e,g).
The a p p l i c a t i o n of
we s e e t h a t i f
Ie
enough" t o LI
sense
P
by
( 4 ) and ( 7 )
i s convex o u t s i d e a compact s e t
f(-,.)
f o r any p r o c e s s i n
containing
s u c h a lemma t o t h e
s t a b i l i t y a n a l y s i s o u t l i n e d above i s c l e a r :
an
such that
with the i n i t i a l value
it follows that close" t o
Be
e(.,t)
and
Ke.
and
I(0)
hut broadly generalized, i s the subject o f
"close
K(-,t)
A r e s u l t of
then
are i n t h i s type,
this work.
In
f a c t , i n o r d e r t o b e a p p l i c a b l e t o a w i d e r c l a s s of m a t e r i a l s , t h e requirement w i l l b e dropped,
that
f(.,.)
b e convex o u t s i d e a compact s e t
and t h e s p e c i f i c norm
11 - 1 1 ,
as introduced
a b o v e , w i l l n o t b e assumed. F o l l o w i n g a d i s c u s s i o n of problem and r e l a t e d results,
t h e a b s t r a c t fundamental
theorems i n S e c t i o n 2 ,
i n S e c t i o n 3,
we t h e n a p p l y o u r
t o show how c e r t a i n m i l d c o n d i t i o n s on
t h e s t r a i n energy f u n c t i o n i n h y p e r e l a s t i c i t y t h e o r y can l e a d to a result
similar t o
(7),
and t h e r e b y b r o a d l y g e n e r a l i z e a
s t a b i l i t y t h e o r e m o f G u r t i r i [ 101.
146
ROGER L. F O S D I C K
2. The Fundamental Problem and Related Theorems.
Following upon the question that was raised in the introduction, we state, somewhat more precisely, the associated fundamental problem:
Given a continuous function _____.
on a connected subset
functions
R2,
on D -
N(*)
of a complete metric space
D
the non-negative reals
f(.) defined
into
V
into
determine those continuous
R2
which share with
f(-) 2
common non-empty compact proper subset of zero's, i.e., 1 = f- (0) E
N-'(O)
for every -
E 7 0
P, C
D,
an? which have the ;>roperty that
there is a
wise continuous function
6 > 0
g(*) on B
such that for any pieceinto - D
with ___
one has
More generally, for
where
C
P
A > 0 and for the set of functions given
denotes the class of piecewise continuous €unctions,
the fundamental problem is to determine, in terms of the general structure of N(.)
f(.),
those conditions which any function
must satisfy i n order to ensure that the number
is finite and tends to zero with main point, if
f(.)
A.
Thus, to emphasize the
is known only in a vague, qualitative
way, then it is not likely that a bound on
( B
f(g)dm
will
ON AN INEQUALITY I N THERMODYNAMIC STABILITY
g ( * ) . Never-
be at once tractable regarding information about f(.)
theless, qualitative information on construction o f an explicit
may suffice for the
and "nice" function
which
N(.)
satisfies the conditions of the fundamental problem statement
g(.).
and so yield concrete information on t o this
proceQure as
f-N
sbitching.
associated with the possibility of when is
A?
G(A)
We shall refer
Thus the basic questims
f-N
switching are, then,
finite valued, and how does its value depend on
i(* E S,)
and when does there exist G(A)
=
i
N(g)dm
such that
?
B
We shall refer to any such function
gc.t
E
s,
as a
Clearly the study o f C; -maximizers is relevant A to the questions associated with the possibility of f-N
% -maximizer. switching.
As we shall now see, the local characterization,
at its points of continuity in
B,
of any
A
-maximizer
requires that a certain relative concavity property must hold between
f(*)
and
N(.).
This, in turn, will be given a
useful geometric interpretation in terms of the notion of f - N touch points.
First, then, we begin with a theorem which is
concerned with necessary conditions for the existence o f a
GA -maximizer. Theorem 1
-
Let
-g ( . )
be a SA-maximizer and let
a point of continuity of and g 2 -
in D -
E- ( - ) .
are such that
Suppose
x o E B
a E (O,l),
and
__
be El
148
ROGER L.
g(Xo)
Proof: Clearly, if trivial.
is connected and
gl
and
g2
f(-)
not in
E E
"X0)
Thus, take
Z
FOSDICK
then both (10) and (11) are
E D 4
D
and observe that since
is continuous there are many points which satisfy (10) for any
a E (Oil).
Now, with these three quantities fixed, suppose initially that
where
f(gl) f
(S,]
f ( _ g 2 ) and define
is a monotone decreasing sequence of open
spheres about
Xo
of continiii.ty o f
with limit
-(.)
Xo.
a E (O,l),
neighborhood of for large erioiigh
subbodies,
Sn 1
= (l-an)m(Sn)
an
the value of
ri
n
and
is a point
it follows that the numerator in the
above fraction vanishes in the limit as large eiioiigh
Xo
Because
so that also
1
with
a.
Thus, for
is within any prescribed
Sn
we m a y decompose
Sn,
n +
m(Sn)
=
an E (0,l). Hence,
into two disjoint
a n m(Sn)
and
2
m(S,)
and define the following function: X E B X €
sIl
1
sn
1
This function has the property that piecewise continuous and
gn(*)
E
S,
since it is
=
ON AN I N E Q U A L I T Y I N T H E R M O D Y N A M I C S T A B I L I T Y
of a Q - m a x i m i z e r ,
A
or,
we s e e t h a t
f o r l a r g e enough
n
gn(-),
w i t h the d e f i n i t i o n of
T h u s , by d i v i d i n g t h r u b y
n
and l e t t i n g
m(Sn)
4
m
we
o b t a i n (11) t o c o m p l e t e t h e p r o o f u n d e r t h e a s s u m p t i o n t h a t
f(gJ
f f(g2)To remove t h i s
f(g(Xo))
that since
f(g)
such t h a t Then,
S
>
f(g(XO))
and by what
q u a l i t y must h o l d f o r see that
then there e x i s t s
0
c l e a r l y for e v e r y
< f(g(XO)),
l a t t e r restriction, first notice
f(2)
a E (0,l)
f f(g') <
we hove
in
D
f(i(XO)).
af(g)
+
-
(1-a)f(g')
we h a v e a l r e a d y p r o v e n t h e same i n -
N(.).
N(g) S N(g(XO))
and
g'
and
Thus,
for a l l
letting
E
D
a +
1
such t h a t
we r e a d i l y
150
f(g)
ROGER L. F O S D I C K
f(C(XO)).
5
Now,
returning to the problem of removing
the restriction noted above, we suppose that
gl
g in D -2 it s o happens that
and
that
a E
and
(0,l)
have been chosen to satisfy (lo), and that
f(gl) = f(g2).
fki) s f(g(Xo))
as we have just noted.
Then, from (10) we have and s o
(i=1,-2,),
Nki)
I:
N(.'(XO)),
I
But then clearly (11) holds.
Those points
E D
which satisfy
are called points of f-concavity f o r
N.
By Theorem 1, it is
in this set that the value of a 6 -maximizer, g ( X ) , 3 whenever X is a point of continuity for g ( * ) .
must lie
-
G
While every point in €-concavity for
N,
is trivially a point of
it is perhaps worth noting that
for GA-maximizer could have all o f its points o f f-concavity __
N
& L,
since
?Z
is a proper subset o f
D.
A s a first step toward a geometric characterization
o f the points of f-concavity for
N
through the notion of
touch points we have the Theorem 2
-
A point
g E
w
D-C
if and_ only ~ _ _ _ i f there exists
g E D.
for all D
then
Proof:
-
If
-X
is a point of f-concavity for N
> 0
such that
is not a global maximum for
N(.)
on -
X > 0.
To seethat (13) with
-
1
2
0
is sufficient for
6
to
151
ON A N I N E Q U A L I T Y I N TKERMODYNAMIC S T A B I L I T Y
be a p o i n t
of
f-concavity
g
respectively, f o r
l-a,
g2
and
b e any
(13) is written, and
g2,
a E (0,l)
and when t h e
and t h e s e c o n d
and t h e n a d d e d , we o b t a i n
a E (O,l),
T h u s , if holds,
gl
r e p l a c e d by
f i r s t i n e q u a l i t y i s multiplied by by
g1
let
N,
and o b s e r v e t h a t when
D
two p o i n t s i n
for
-
1 z
and i f
0,
gl
and
we l e t
a r e chosen such t h a t
(12)2
$
t h e n (12)1 h o l d s which i m p l i e s t h a t
i s a point o f f-concavity f o r Now,
g2
g
N.
be a p o i n t o f f - c o n c a v i t y f o r
N
a E (0,l) by
and d e f i n e
where
p E (0,l)
where
-
g2 E D
gl E
and where
meets
f(g2) <
D
meets
f(g).
Thus,
f(gl) since
2
f(g)
a
is
and
s t r i c t l y l e s s t h a n t h e r a t i o i n b r a c e s and s i n c e t h e denominator i s s t r i c t l y p o s i t i v e , f o r a l l such
N
for of
a
and
g2.
Since
we may t h e n i n v o k e ( 1 2 ) 1 ,
-g
-
i s a point
of f-concavity
which w i t h t h e d e f i n i t i o n
may b e w r i t t e n as
f o r a l l such ting
g1
(12)2 h o l d s
then c e r t a i n l y
p
+
0
g1
and
E2
and any
E
inf N ( E )
g
-
f(E) -
N(g)
Clearly, l e t -
we s e e t h a t
2_
0,
-g
E D
(15)
f(g)
where t h e infimum i s t a k e n o v e r a l l
f(g),
(0,l).
i n t h i s i n e q u a l i t y and d e f i n i n g
x’ f(g) <
p E
(13) holds f o r a l l
which a l s o meet
g E
D
of
this
FOSDICK
ROGER L.
152
kind. F o r all
g E
D
such that
f(g)
= f(E)
we showed
N(g)
i n the closing remarks of the proof of Theorem 1 that
< N(g).
Thus, ( 1 3 ) holds for such
g
establish (13) for all
E D
Toward this end, n o w , define
gl E
f(_g2)
again meets and
g2
f(g) >
such that a E (0,l)
f(gJ f(&) - f(gJ
is such that
< f(E).
’
f ( g l ) > f(E)
Then clearly ( 1 2 )
is a point of f-concavity for
gl
with
g2
and
f(g).
by
interchanged, holds f o r all such
since (12)1,
D
and we have left to
f(g) -
a s where, here,
g,
N
51
2’
and
-g2
with and
El
g21
and
we may invoke a
interchanged, and for this
write
Finally, since choices o f
g1
(14) and (15) are still valid under the present and
g2 *
it follows that by letting
p
-t
1
i n (14), by using this result to continue the latter inequality and then by applying (15) we reach
for any such
g
E D
gl.
such that
Thus, we find that ( 1 3 ) holds for all
f(g) >
f(Z).
M
The condition ( 1 3 ) has an interesting and helpful geometric interpretation which I shall now describe.
To do
so, requires that the set of values o f the functions
f(.)
and
N(.)
defined on
D
be interpreted as surfaces, as is
illustrated, for example, i n Figure 1 .
(Here, 8:
is the
ON A N INEQUALITY I N THERMODYNAMIC STABILITY
-
singleton set
the origin,
a simple "cone".)
(.)
a ,b
5
-
(:,6)
b 2
0,
s u r f ac es
happens t h a t
,.,
but
g E D - i s a l w a y s a t l e a s t a s l o w as
(16)
0.
F- - ( * ) ,just a,b otherwise f o r a l l other
j t
Y
such s u r f a c e i s i l l u s t r a t e d .
>
a
13,
a t some p o i n t
f ( * )
downward t r a ns 1a t e d
and
M
aN(*)
Suppose t h a t f o r some touches
h a s been c h o s e n t o h e
N(.):
associated with
F
N(.)
C o n s i d e r t h e f o l l o w i n g two p a r a m e t e r famiLy
u n i f o r m l y c omp r e s s ed
of
and
153
f(g).
I n F i g u r e 1, o n e
Thus,
which g i v e s
-
- -
h = aN(z)
f(5) 2
0,
and
Accordingly, .g . E D. ( 1 7 ) h o l d s f o r some a E
for all
we c a l l a n y p o i n t
that
[O,m]
f-N
a proper
is clear that
(17) i s
touch point
-
E D
if
impossible.
N
on
D
E D
then -
if and o n l y if
iZ
f-N
i s a touch point
8
5
then
g l o b a l maximum on
N
in
D,
= 0
in
touch p o i n t s ,
i s n o t a g l o b a l maximum f o r
- ,Z
i s a point o f f-concavity
i s a proper
f-N
touch p o i n t .
A
for
It
(0,m).
l i n g t h a t n o ( $ -maximizer can have a l l o f i t s p o i n t s f-concavity
such
touch point
a t
t r u e f o r some
Thus, i n t h e terms of
Theorem 2 a s s e r t s t h a t if
N( . )
if
an
E D
Z,
we s e e t h a t
if N(-)
-
Recalof
has no
t h e n t h e e x i s t e n c e of p r o p e r
for
f-N
154
ROGER L. FOSDICK
touch points is necessary of a QA -maximizer. __ -~ for the existence . -. - One could now ask, to what extent is the existence of proper touch points -~ sufficient for the existence o f a maximizer?
SA -
Before turning to this important question let us
introduce the notion of a touch manifold associated with -~ -
g,
i.e.,
(in Figure 1, 3(g)
has two elements as shown) and consider
the following independent criterion concerning the existence
of touch points.
-
3 Theorem -
M
_Let _
E
D
a compact .set containing -be -~
C
such __
that
f@some~-2 > 0.
Then _ _ _ -proper __
- _.
in particular,
5(2)
point then ^a __ - - ____
f(i)/N(g)
where
q
E
Eg E
I
touch points exist and, ____
is not null.
I
D
f-N
> 0
f(g) s
_If _ ..2 E D
D
C
is a touch
and
f(g)3
*
It is worth noting that if of
-
is a compact subset
M;'
then the necessary condition (18)* is a complete
converse to the sufficient condition (18)1.
However, recal-
ling that a central motive in much of this work is that in
f(.) is
not usually known explicitly, we see
that neither is
M;.
A l l is not s o bad though in that for
any touch point
g
applications
I
{g E D
I
N(g)
S
known, s o that if
(17) requires that
N(g)],
M;
and since
Mh
N(-)
E
MG
E
is known
is a compact subset of
D
M i
is
then s o too
155
O N AN I N E Q U A L I T Y I N THERMODYNAMIC STABILITY
w i l l be
M.:
T h i s theorem w i l l be u s e d l a t e r i n t h e d e v e l o p -
ment of m i l d c o n d i t i o n s f o r non-linear merely
It
elasticity.
f ( * )
i s compact and
-
f(g) 5
$N(g)
t o u c h p o i n t s when
t h i s theorem,
-
ZN(g)
f(g)
f ( g ) < $N(g)
g E
a proper
t o u c h p o i n t and
D
on
0
2
h o l d s f o r all f-N
outside
N(.)
-
5
observe t h a t s i n c e
E M
there e x i s t s
- f(g)
GN(g)
-
of
Z E M
(18)1 t e l l s u s t h a t we have
guarantees
D.
Proof: F o r t h e f i r s t p a r t -__
ZN(g)
switching i n the theory o f
l i e above some c o m p r e s s i o n of
some compact s e t i n -
M
f-N
M.
On t h e o t h e r h a n d ,
D-M.
on
0
on
f($)
such t h a t
D-M,
and f o r t h e g i v e n number
5(2)
A fortiori,
(17)
and t h u s
2 >
-
0
g*
is
i s n o t empty.
F o r t h e s e c o n d p a r t of t h i s t h e o r e m , r e c a l l t h a t since
2
E
E
impossible i n -
8
.
g = g E
D
-
E
D
-
(17).
Further, if
M;
(where
(17)
then
N(.)
g l o b a l l y maximizes
-
f(g) >
which s h o w s t h a t
aN(g)
on
f(8))
D.
5
0
and
= 0
is
for t h e t o u c h p o i n t
i be
a p o i n t which
I n t h i s c a s e f o r any
g E
D-
we may w r i t e
(18)* h o l d s .
- f(8)
5 =
requires that
i n ( 1 7 ) and n o t e t h a t s i n c e then
-
a
i s a touch p o i n t then
We n o w assume t h a t E
iE
D
-
g
a
E
( 0 , m )
i s a touch p o i n t
156
ROGER L.
-
g E D
for any
MP.
FOSDICK
Thus, appropriately defining
2
we
establish (18)
2'
N(.)
We saw earlier that if on
D, then the existence of proper
has n o global maximum f-N
touch points is
necessary for the existence of a Q -maximizer. i\
is a proper
touch point s o that (17) holds for a l l
f-N
and for some
a
E
Then, if
(0,m).
g ( * )E
s,
ED
Suppose
g E D
(recall (8)),
it follows from (17) that
\
s m(B)N(t.)
N(_g(X))dm
+
[!ho)f(ij)l/a
B G(A) <
and we see that
m
exists, even though it is not at
In fact, if
all certain that a Q*-maximizer exists. L!
.-- E
then not only is the constant function with
f-'(A/m(B))
value
in
but it is clearly a @, -maximizer, i.e.,
s,
D
3
f
/
N(g(X))dm
f o r all
g(.)
E 9,.
given
>
the set
*
C,
0
m(B)N(g)
= G(A)
B
While it is hardly likely that for a f-'(A/m(B))
will contain a proper
touch point, o u r observations here indicate that the mere existence
of a proper touch point may be the single most
important sufficient requirement for the existence of a
A
-
maximizer even in this more general situation. The here',
following theorem, which we shall not prove
gives a sufficient condition for both the existence of
a QA-maximizer as well as for the validity o f the procedure
of
f-N
switching.
'The proof o f a similar b u t stronger result w i l l appear shortly in a paper by Dsnn and myself i n the Archive for Rational Mechanics and Analysis.
ON AN I N E Q U A L I T Y
Theorem
4 -
Suppose t h e r e ...e-x.i s t s
mizer e x i s t s . -i
nto
and
D
Moreover, --______ E
> 0,
g i v e n any
> 0
a*
-
such t h a t
g(-)E
Cp
3(a*)$ G
such t h a t
F i g u r e l g i v e s a n example i l l u s t r a t i o n of
a*
f o r a s i n g l e c h o i c e of
-
t h e r e a r e many o t h e r s ) aiid f o r
For this particular
"cone".
t h e touch manifold consists of x(a*)
$
a
(taken equal t o
a*,
-)
P(a*)
i n the f i g u r e
assumed t o b e a sirnple
there are t w o points i n
3(a)).
(=
5(a*)
N(
Clearly, since, here,
the s i n g l e point a t the coordinate o r i g i n ,
E.
Since
i s compact,
P(a*)
B
mapping
6(c) > 0
t h e r e e_. xists
~~
157
I N THERMODYNAMIC STABILITY
iZ
wehave
i t i s b o t h c l o s e d and
t o t a l l y bounded i n t h e c o m p l e t e m e t r i c s p a c e
V
3
D.
Roughly,
t h i s c o n d i t i o n avoids c e r t a i n d i f f i c u l t i e s t h a t could g e n e r a l l y e x i s t a t t h e boundary p o i n t s o f
i t requires that
for f i n i t e dimensional V)
a*N(*),
o s c i l l a t e f o r e v e r about above i t .
I t i s worth n o t i n g t h a t
o u t s i d e of
a compact s e t ,
f-N
f o r the
switching
D,
f(-)
( a t least not
but rather ultimately r i s e need n o t be convex
f(.)
a s Coleman [ 5 ] procedure
and
required,
i n order
t o be p o s s i b l e .
B e f o r e l e a v i n g t h i s s e c t i o n we r e c o r d a s u f f i c i e n t condition o f practical v a l i d i t y of
t h e procedure
g e n e r a l l y the
5
-
{g E
D
Theorem
M;
P
s i g n i f i c a n c e which g u a r a n t e e s t h e
of
existence o f a
Suppose
N(g) <
E D
N(i)]
f-N
S
s w i t c h i n g , though n o t
a -maximizer.
i s a proper i s compact i n
f-N
D.
t o u c h p o i n t and Then,
the
ROGER L.
158
procedure Proof:
of
FOSDICK
s w i t c h i n g-.i s v a l i d .
f-N
-
E D
Supposing f i r s t t h a t
C
i s a p r o p e r touch
p o i n t , we may conclude from t h e second p a r t of Theorem
;E
where
f(g)/N(g) > 0
MA f
B u t , a s noted e a r l i e r ,
M;'
Thus, s i n c e
and where C
G(A) <
M;
f(i)].
i s compact i n D.
3(;)
i s n o t n u l l which, a s r e -
o f Theorem 3 , i s s u f f i c i e n t f o r t h e
marked a f t e r t h e proof e x i s t e n c e of
and s o
f(g) <
we may a p p l y t h e f i r s t p a r t o f
C,
contains
3 t.o conclude t h a t
Theorem
Mi,
= {g€D I
MA f
3 that
m.
Moreover,
f o r any
a
E
we s e e
(0,;)
that
which g u a r a n t e e s n o t o n l y t h a t t h e s e t s t a t e m e n t of Theorem
4 ) satisfies
P(a)
MT,
P(a) G
(defined i n the and i s t h e r e -
f o r e compact, b u t a l s o , f r o m t h e f i r s t p a r t of Theorem 3 , that one
i s not n u l l .
5(a) a
E
3(a)
$
Z
f-N
f(g) for
all
g
E D.
a E
for a l l
5 ( a ) E L!
t h e n f o r any one of t h e s e v a l u e s we s e e from I;
the
switching i s v a l i d .
On t h e o t h e r hand, i f
aN(g)
for a t least
we may a p p l y Theorem 4 t o s e e t h a t
(0,;)
procedure of
Thus, i f
Thus, f o r any
(17)
(0,;)
that
g ( * )E
s,
we
obtain
'B which shows t h a t procedure
of
f-N
Finally,
'B
G(A) I
A/a
and, t h e r e f o r e ,
t h a t again the
switching i s v a l i d . supposing t h a t
E
Z
i s a proper touch
159
O N AN I N E Q U A L I T Y I N THERMODYNAMIC STABILITY
p o i n t , we s e e f r o m ( 1 7 ) t h a t
a
and f o r some
E
a N(g)
f(g)
5
D
t h e l a s t paragraph
The argument of
(0,m).
g E
for all
1
may t h e n be r e p e a t e d t o c o m p l e t e t h i s p r o o f .
T h i s theorem w i l l b e u s e f u l i n t h e n e x t s e c t i o n where we e s t a b l i s h a c e r t a i n s w i t c h i n g specific function
3.
Application:
Let of
procedure
for a
i n t h e t h e o r y of n o n - l i n e a r elasticiiy.
N(.)
Hyperelastic Materials.
Lin
d e n o t e t h e s e t of
l i n e a r transformation
a t h r e e dimensional Euclidean v e c t o r s p a c e , i n t o i t -
E’,
self-the
elements o f
two v e c t o r s
2,
2
in
and
F
are tensors.
E Lin
The ____ i n n e r product o f
i s d e n o t e d by
E’
a 4 2,
t e n s o r p r o d u c t by any e l e m e n t
Lin
2-2
and t h e i r
which i s a n e l e m e n t of
-
we u s e t h e n o t a t i o n s
det F,
t o denote the determinant, t r a c e ,
For
Lin.
F, FT
tr
transpose,
and
inner t r a c e norm, t h e l a t t e r of which i s b a s e d upon t h e ____ product
E
--
t r UWT
f o r any
denote t h e u n i t tensor .
u,
W
in
Lin.
F i n a l l y , we s h a l l u s e
and O r t h t o d e n o t e , r e s p e c t i v e l y ,
We l e t
1
E Lin
I n v , Syrn, Sym+
the s e t o f i n v e r t i b l e ,
s y m m e t r i c , p o s i t i v e d e f i n i t e s y m m e t r i c , and orthoe;orial t e n s o r s .
It i s clearthat I n v , and
Sym+
Inv
i s open i n
i s open i n
Sym
L i n , Sym
i s a s u b s p a c e of
and convex.
A s t r a i n energy f u n c t i o n f o r a h y p e r e l a s t i c material
i s a smooth ( i . e . ,
1 c l a s s C ) mapping
a(.)
of
Inv
into
IR
w h i c h , b e c a u s e of t h e p r i n c i p l e o f m a t e r i a l f r a m e i n d i f f e r e n c e
[ll],
s a t i s f i e s the invariance condition
160
ROGER L.
F E
for all
Inv
FOSDICK
E Orth.
and f o r a l l
I n addition, we
(19) holds i f
know by t h e p o l a r d e c o m p o s i t i o n theorem t h a t and o n l y i f
dF) where,
-
--
F = RU ed on
8(*):
F
f o r any with
;E
E Inv,
f? E O r t h
=
a(;)
i s u n i y u e and i s s u c h t h a t
Sym'
and u n i q u e .
Sym+ + !R
a(*)
Thus,
by i t s v a l u e s o n l y on
Inv
(20)
Sym+ C Inv
u(.)
be t h e r e s t r i c t i o n of
i s determin-
and i f we l e t
to Sym'
we
have
-
a ( F ) = a(!)
= a(;).
(21)
I t i s straightforward t o s h o w t h a t the gradients
- -)
and
E Sym
;E
F E
Inv
1
=
-
c!
T
'F(F) -
+
F
i s a r b i t r a r y and
r a F ( E ) I T 133
-
(22)
9
f? E O r t h
= f?; with
and
Sym+. If
i n a d d i t i o n t o ( 2 0 ) i t a l s o happens t h a t e i t h e r
f o r some s m o o t h symmetric f u n c t i o n of positive reals
R> x
IR'
>
x IR
d e n o t e s t h e p r o p e r numbers o f
a(?) f o r some smooth f u n c t i o n that
E Lin
a r e r e l a t e d by
):(,' where
a,(.)
a(*)
6(.,.,.)
into
u
IR,
E Sym+,
of
where
the t r i p l e (u,}
or
-
(24)
= E(det U)
;(*):
R'
+ R,
t h e n we s h a l l s a y
i s t h e s t r a i n e n e r g y f u n c t i o n f o r a i s o t r o p i c hy-
p e_ r e_ la_ s t i~ c material or a e l a s t i c f l u i d , respectively.
.
T h e r e a r e c e r t a i n i n e q u a l i t i e s t h a t w i l l be mentioned i n t h e r e m a i n d e r
of
t h i s s e c t i o n i n connection with a
O N AY I V E Q U A L I T Y I N TTII~R"40DYNAMIC! S T A B I L I T Y
specific notion of For t h i s reason,
f-N
s h i t r h i n g i n h y p e r e l a s t i c i t y theory.
i t i s c o n v e n i e n t t o f i r s t i n t r o d u c e aird t o
discuss the various relations
t h a t e x i s t hetween t h e s e i n -
and t h e n t u r n t o t h e q u e s t i o n of
equalities,
W e say t h a t
obeys the
U ( * )
-
--
u ( F ) + (GF-F)-UF(E) <
E
for a l l
Inv
-
I
-
G E
and f o r a l l
Sym+
C-N
f-N
switching.
inequality
if
u(GF)
(25)
{i}.
-
In
contrast
,
f o r a much weaker i n e q u a l i t y o f t h e same k i n d , w e sa3 t h a t
G
E
Inv
i s a point of
C-N
u(*)
growth for
I
-
f or all
-
E
G
Sym+.
--
+
a(?)
(G6
-
Clearly,
-
f).OF(E)
U(
i f and o n l y i f e v e r y member o f
for
.)
S
--
obeys t h e
Inv
(26)
o(G$) C-N
i s a point
of
(26) i s a s t r i c t inequality T o r
a t which
U ( . )
O-
We s a y t h a t
ti(-)
i f merely ~
inequality
C-N
growth
2 # 3.
E ~ y m + i__s a p o i-n t o f c o n v e x i t y f o r ~
if
-
i(ir)+ ( -u --G ) G u ( ~ )s ; ( u )
-
E Syrn'.
for a l l
a x i a l convexity f o r that,
b(*)
if
g
Sym+
i___._s a p o i n t of co-
(27) holds f o r a l l
E Sym'
s h a r e a c o m m o n orthonormal b a s i s o € p r o p e r
i n addition,
-
$
A n clerneilt
(27)
A
vectors ~ t i t h U,
i.e.,
g ( g ) + ( c & - ~ ,-. q i js) 8(.,&) for a l l
2
E
Sym+
such t h a t
@
E Sym+. C l e a r l y , c o a x i a l
c o n v e x i t y i s weaker t h a n c o n v e x i t y . p o i n t of
C-N
growth f o r
a x i a l convexity f o r generally true.
; ( a ) ;
U(.)
(28)
Also,
if
E Sym+
is a
then i t i s a p o i n t o f c v -
however,
t h e converse i s not
To emphasize t h e d i f f e r e n c e between t h e s e
ROGER L. F O S D I C K
162
E Sym+,
two notions of convexity at
5 2
common, that any two tensors
-A
stretch if
E
is a point of Inv
-_
GB C-N
for some
growth for
B
E Sym'. a(*)
differ by a pure Then, if
fi
the graph of
E Sym+
a(*)
on
$
is restricted at all those tensors that differ from
by a pure stretch.
Sym'
iE
On the other hand, if
point of coaxial convexity for on
and
let us say, as is
;( * )
Sym+
is a
then the graph of
;(
)
has the same formal restriction but only at those A
tensors that in addition to differing from
';I
by a pure
1
by a pure
n
stretch also differ, like
U
itself, from
w
stretch
I
A
-
for
C-N
growth
U
is composed with any pure
stretch, while for coaxial convexity
.-.
is composed only with
those pure stretches that yield a pure stretch composition. 1 For isotropic elastic materials we have the following Theorem 6
2
-
a(.)
be the strain energy function -______-- for an
isotropic elastic material.
~~
Then the following three state-
are equivalent: -ments -. - _____ ~
-1
fi
One argument [121 for motivating that E S y m + be a point o f C-N growth for a ( . ) is based on the requirement that the actual wor$ done at a particle in a deformation which differs from by a pure stretch be greater than or equal to the work that would be done if the same deformation was executed under locally dead loading. It would seem at least as reasonable to require, additionally, that the deformation a pure stretch. The restriction of coaxial itself convexity would then follow. 2
The proof of this theorem is given by Dunn and myself in an article to appear shortly in the Archive for Rational Mechanics and Analysis,
ON AN INEQUALITY IN THERMODYNAMIC STABILITY
;(.).
(i)
E Sym+
is a point of convexity . for
(ii)
E Sym+
is a point of coaxial convexity ~- for
(iii)
(G1,G2,G3) of -
The __ proper values ___
E(*,.,.),
point of convexity for
q,G2,G3)
;(*I.
~
3
,.
+ c
p
E Sym+
is a
i.e., -
- (Gl,G2,G3)
(ui-u.) 1 aui
+l1,U2,U3)
i= 1 __ and
ul, u 2 ,
for all
u
R'.
in -
3
Now, since, f o r such materials, Coleman and No11
have
shown that if E Sym+
(iv)
C-N
is a point o f
growth for
a ( .)
,
then (iii) must hold, and since it is known [ll, p.323, and
141 that there exists a symmetric smooth function E Sym+
which satisfies (iii) for all all
p
in a non-empty subset of
Sym',
; ( - , a , - )
while (iv) fails for we see that within
the class of isotropic elastic materials any one of (i), (ii)
or (iii) is generally milder than (iv).
However, for elastic
fluids, where ( 2 4 ) holds, this difference breaks down and all four statements are, indeed, equivalent.
We turn now to the subject of touch points and switching within the context of hyperelasticity theory. Here, Gurtin 1101 has noted, essentially, that if
a(-),
the
strain energy function for an isotropic elastic material, has a strict global minimum at E Sym'
for all
-
then the procedure E
;(-)
ed o n
-
a(;)
Sym'
and into
{_1) ) , of N(.) R.'
-1
on
Syh+
and obeys the
f-N
(i.e.
a ( ? ) < a(!)
C-N
inequality,
switching is valid for
1-
- -11,
f(*) c
both of which are defin-
Our purpose in the remainder o f this
164
ROGER L.
FOSDICK
article is to show that the same conclusion holds for any
,1
hyperelastic material, not necessarily isotropic, provided a(-)
is a strict global minimum of
on
Sym'
and provided
there is a "properly distributed" set of points o f coaxial convexity for
6(')
"arouiid" 1.
In order to more clearly
state the precise result, it is convenient to have available
-
the set theoretic notions of a star effectively containing 1 ~ _ _ _ _ _ and a radiant shell of the star. We say that
--
containing 1 if
S E Sym
-
R
-1 is
not a limit point o f
-
the line
S
S,
is a star effectively
is compact and if there exists some sub-
S
set s -
E
+
called a radiant shell about 1, such that (i)
-
UU
+ ( 1 - a )-l
empty intersection with effectively containing
f o r all
RS.
-1
-
and (ii) -~ for any E Sym'
RS,
u E [ 0,1] has a non-
Clearly,
since
Sym'
admits of stars
is open in
Sym'
Sym.
For
example, any small enough closed ball in Sym centered at
,1
will suffice if its boundary is chosen as a radiant shell. This brings us to the following main Theorem
7
-
a(.)
Let
be the strain energy function for an
~
elastic material and suppose that minimum- at
1 c
ly containing -_
____.
RS,
on --
Sym'.
Let __
O(*)
S E Sym
f
has a strict global be a star effective-
,1 and suppose that one of its radiant shells,
is composed entirely of points of coaxial convexity for
;(-).
Then, f o r
Sym+,
proper
-
f(*) E ;(.)
f-N
U(;)
__ and
N(-)
T
1.
-
,I[ on
touch points exist.
Before proving this theorem, it is perhaps worth noting that it's conclusion procedure
of
f-N
guarantees the validity of the
switching.
That is, since at least one
165
OK AN I N E Q U A L I T Y I N TFTERMODYNANIC STABILITY
f-N
proper
touch point exists,
we see that the set
compact in
then f o r it
i n the hypotheses of Theorem 5 , i.e.,
M;
in the present context
E Sym+,
say
(u E
I ];-?I
Sym+
s
1$-11}
,
is
Whence, Theorem 5 is applicable. Of course,
Sym'.
as a prerequisite consideration it is clear that the and the
of the present
N(-)
f(.)
heorem satisfy all of the
conditions that we required of such functions throughout this work. Proof:
A s a first step we shal
point o f coaxial convexity for ;( .. -
for all
> 1
CI
a
Y
s a
Sym'
then ___
)
-1 + I J ( ~ - 1E ) Sym'.
such that
E
show that if
To see this,
-
n
let and
G1
C2
and
G2i
be two tensors in
are in
G1 and G2. containing El by a E G2
containing
by
G 1;
and note that (28) can be written
Sym+,
f o r each
such that
Sym'
Multiplying the resulting inequality
1-a
and the second inequality
(O,l),
and adding the two together, we find
that
f o r all
a
E
G1
and
(O,l),
and
G -2
as described above
such that
+ Now,
let
LI
> 1 be such that n
since
U
-CI
and
-
E
6-l
-Up -
and
-CI
E
-1 + p ( i - & ) E
Syrn'.
Then,
share a common orthonormal basis of proper
U
vectors it follows that G1
U
G2
s
U
-I-I
fi-' -
E Sym+.
Finally, if we set
it follows that
El, G 2 ,
6
G -1- = -Up '
166
ROGER L.
G26 =
and
a
,1
are a l l i n
FOSDICK
and
a € (0,l) p r o v i d e d we s e t
from
(30). Now, A
-
that
U
E
-
aU
RS
convexity f o r f(-)
This,
aGl + (l-a)G2 = 1,
u
Thus,
RS
that
l/p.
E
E Sym+
-
E RS.
T h a t s u c h a number
6 E
and l e t
S
b e i n g a r a d i a n t s h e l l of
,1 + 'E(g-,1),
=
C l e a r l y , we have
and s i n c e
of
u
let
+ (1-E):
a p r o p e r t y of
-1.
and t h a t
Sym',
-
( 2 9 ) follows
be s u c h
(0,l)
6
exists is
the s t a r
where
I
l/a' >
1,
i s composed e n t i r e l y o f p o i n t s o f c o a x i a l
i t f o l l o w s f r o m ( 2 9 ) and t h e d e f i n i t i o n
; ( a )
-
fi f ( 6 )
s f(;),
so that
-
E Sym'
of c o u r s e , h o l d s f o r e v e r y c h o i c e o f
and f o r e a c h s u c h c h o i c e t h e r e e x i s t s a c o r r e s p o n d i n g
N(*),
Thus, u s i n g t h e d e f i n i t i o n of
f!~)
2
2
inf
E
-
where t h e s t r i c t i n e q u a l i t y
be a l i m i t p o i n t of
.fLY)2 -
0,
2 >
0
h o l d s s i n c e a p r o p e r t y of
RS.
S
be compact and t h a t
To c o m p l e t e t h i s p r o o f ,
we f i n a l l y a p p e a l t o t h e f i r s t p a r t of Theorem g u a r a n t e e s t h e e x i s t e n c e of
proper
A s mentioned p r e v i o u s l y , theorem e n s u r e s t h e v a l i d i t y of
hyperelasticity theory, t h a t the p o i n t s of
0 E RS.
we have
s t a r s and r a d i a n t s h e l l s demands t h a t
switching, f o r a s p e c i f i c
S,
UERS N(U)
N@)
-1 n o t
about
S
f-N
this
touch points.
p r i o r t o i t s proof,
t h e procedure
f ( * ) and
3;
then,
N(.),
of
this
f-N
i n non-linear
It i s , I t h i n k , worth ernphasising
c o a x i a l convexity i n the hypotheses a r e
n o t n e c e s s a r i l y t h e same a s t h e p r o p e r t o u c h p o i n t s which a r e
167
ON AN INEQUALITY IN THERMODYNAMIC STABILITY
shown to exist.
However, since it, indeed, can be shown that
any proper touch point which is a point of convexity for N(.) is also a point of convexity f o r has the special form Sym+,
1. -
f ( * ) , and since here
N(.)
,I1 which is convex everywhere on
then in the present context we see that every proper
touch point must be a point of convexity for
b(-)
on
Sym'.
Thus, the existence of a "properly distributed" set of points of coaxial convexity for
'Taround'T ,1,
; ( a )
its point o f
strict minimum, implies, through "touching", the existence o f points of convexity for
;(-)
somewhere on
Syrn'.
The results of this and the previous section suggest that the condition o f having proper
f-N
touch points is a
regularity condition that may well be intimately connected to the development of mild a priori inequalities in the theory of hyperelasticity.
Acknowledgement.
Support o f the U . S .
tion is gratefully acknowledged.
National Science Founda-
168
\
--
A
h
la
v
u \ \
ROGER L.
\
I
FOSDICK
. .. WI
3
a
\
-1-
1
I-
I
Yr i
‘1 N
”
169
O N A N INEQUALITY I N THERMODYNAXIC STABILITY
R e f erenc es -
[ 11 E r i c k s e n , J . L . , theory.
A thermo-kinetic
J. S o l i d s S t r u c t u r e s
Int.
[Z] E r i c k s e n , J . L ,
2,
Appl.
(1966).
573-580
Thermoelastic s t a b i l i t y , Proc.
N a t i o n a l Congr.
[ 3 ] K o i t c r , W.T.,
view o f e l a s t i c s t a b i l i t y
5 t h U.S.
187-193 (1966).
Mech.,
Thermodynamics o f e l a s t i c s t a b i l i t y .
of A p p l .
3 r d C a n a d i a n Congr.
Mech.,
Calgary,
Proc.
29-37
(1971).
[ b 1 Coleman, B . D .
Bc J . M .
G r e e n b e r g , Thermodynamics and t h e
s t a b i l i t y o f f l u i d motion. Anal.
25, 3 2 1 - 3 4 1
[ 5 1 Coleman, B . D . ,
[61
R a t i o n a l Mech.
(1967).
O n the s t a b i l i t y of equilibrium s t a t e s o f
general fluids.
1-32
Arch.
Arch.
R a t i o n a l Mech.
36,
Anal.
(1970).
Coleman, B . D .
& E.H.
D i l l ,
O n thermodynamics
and t h e
s t a b i l i t y o f m o t i o n s o f m a t e r i a l s w i t h memory. Arch.
R a t i o n a l Mech.
,
[ 71 G u r t i n , M.E.
Anal.
51, 1-53 ( 1 9 7 3 ) .
Thermodynamics a n d t h e e n e r g y c r i t e r i o n f o r
stability.
Arch.
R a t i o n a l Mech.
Anal.
52,
93-10?
(1973). 181 G u r t i n , M . E . ,
Thermodynamics a n d s t a b i l i t y .
R a t i o n a l Mech.
[ 9 ] Dunn, J . E .
& R.L.
Anal.
Fosdick,
and boundedness o f
o f second grade. 191-252
(1974).
2,63-96
Arch.
(1975).
Thermodynamics, s t a b i l i t y ,
f l u i d s o f c o m p l e x i t y 2 and f l u i d s Arch.
R a t i o n a l Mech.
Anal.
56,
ROGER L. FOSDICK
[lo] Gurtin, M.E., Moden continuum thermodynamics. Mechanics Today.
New York: Pergamon 1973.
[ll] Truesdell, C. & W. Noll, The Non-Linear Field Theories Flbgge's Handbuch der Physik, III/'j,
o f Mechanics.
Berlin-Heidelberg-New York: Springer 1965. [12] Coleman, B.D., Mechanical and thermodynamical admissibility of stress-strain functions. Mech. Anal.
4, 172-186
Arch. Rational
(1962).
[l3] Coleman, D.B. & W. Noll, On the thermostatics tinuous media,
Arch. Rational Mech. Anal.
of con-
k,
97-128 (1959)*
[14] Bragg, L.E.
& B.D.
Coleman, On strain energy functions
for isotropic elastic materials.
4, 424-426
(1963).
J. Math. Physics
G.M. de La Penha. L.A. Medeiros (eds.) Contemporary Developments i n Continuum Mechanics and P a r t i a l D i f f e r e n t i a l Equations @North-Holland P u b l i s h i n g Company (1978)
LEA.I;T SQUARE SOLUTION O F NON L I N E A R PROBLEMS
I N F L U I D DYNAMICS
R.
GLOWINSKI* and
Universitb de P a r i s Iria-Laboria,
0.
PIRONNEAU
6 and
Iria-Laboria, 78150 le Chesnay
France
France
1. I n t r o d u c t i o n .
Many problems i n Mathematical P h y s i c s a r e of
the
following forms, e i t h e r A(u)u = f,
+ B(u)
Au
In
(l.l), A ( v )
i s in (1.2)
v
i s for
=
f.
g i v e n a l_i _ n e_a_ r . o p e r a t o r and
B
a non l i n___ e a r operator.
This suggests obviously the following algorithms: F o r s o l v i n g (1.1)
(1.3)
uo
then f o r
n
2
0
we o b t a i n
u n+l
given from
u
n
by
I
A(un)un+' = f ,
For s o l v i n g ( 1 . 2 )
we may use an a l g o r i t h m s i m i l a r t o
(1.5)
(1.4)1 i s r e p l a c e d by
*
except t h a t
(1.3)-
~.
V i s i t i n g P r o f e s s o r IJF'RJ,
R i o d e J a n e i r o , July-August
1977.
R. GLOWINSKI and 0. P I R O N N E A U
172
There are some applications in which (l.3), (resp. ( 1 . 3 ) , ( 1 . 4 ) 2 , (resp. (1.2)).
-
(1.5))
( 1 . 4 ) 1’ (1.5)
converges to a solution o f (1.1)
Unfortunately these are important applications
mathematically modelled by equations like (1.1) or (1.2)
-
f o r which the above algorithms can not be used, therefore one
has to use more sophisticated methods.
In some cases (cf., e.g., Glowinski-Marocco Ell) these problems may be reduced to a saddle-point problem by introduc-. ~
ing an artificia-1 constraint and the corresponding Lagran* multiplier. - __ -..~
In some other situation such an approach is not suitable (by lack o f monotonicity usually) and it may be convenient to use aleast square method.
A typical least square formulation o f problem (1.1) is Min J(v)
(1.6)
V
where
(1.7) In
(1.7),
]I*]]
denotes a suitable norm (for more details
about this last point, see Remark 1.2 below) and
J(v)
the
solution o f the linear problem
We obtain similarly a least square formulation of (1.2) by replacing (1.8)1 by
(1*8)2
Ay = f-B(v).
LEAST SQUARE S O L U T I O N O F NON L I N E A R PROBLEMS
Remark _ _ 1~ .1:
I t a p p e a r s c l e a r l y from t h e above r e l a t i o n s t h a t
(1.6), (l.7), s t r u c t u r e of
(l.7),
( 1 . 8 ) l and ( 1 . 6 ) ,
( l . 8 ) 2 have t h e
an o p t i m a l c o n t r o l problem i n which t h e c ontrol ._ ~~
variable
(i.e.
t h e independent v a r i a b l e i n t h e problem) i s
the s t a t e variable i s
y
and t h e c o-s t-f-u n-c.t_ i -o. n is -
v,
J.
T h i s p a r t i c u l a r s t r u c t u r e l e a d s v e r y n a t u r a l l y t o compute t h e derivative o f
111 , [ 2 ] )
J
v i a an a d j o i n t s t a t e e q u a t i o n ( s e e Lions
f o r t h e o p t i m a l c o n t r o l o f s y s t e m s m o d e l l e d by
p a r t i a l d i f f e r e n t i a l equations). T h i s k i n d o f methodology h a s b e e n a d v o c a t e d by Cea-Geymonat
[l] f o r s o l v i n g non monotone, non l i n e a r t w o - p o i n t b o u n d a r y v a l u e problems l i k e
1
(1.9)
+
-u”
cp(u) = f
10,1[
over
u ( 0 ) = u ( 1 ) = 0.
I n B e g i s - G l o w i n s k i [1] a s i m i l a r a p p r o a c h h a s b e e n u s e d f o r s o l v i n g f r e e boundary problems
l i k e those r e l o t e d t o flows o f
l i q u i d s i n porous media. I n a n u n p u b l i s h e d work Derviauc-Glowinski h a v e a l s o u s e d t h i s t y p e o f methods t o compute t h e maximal s o l u t i o n of
I
(1.10)
-u
= Xe
U
over
10,1[
”
u(0) = u ( 1 ) = 0 where
>
0.
In t h e above r e f e r e n c e s t h e o p t i m i z a t i o n i s a c h i e v e d u s i n g e s s e n t i a l l y -~ descent o r steepest descent algorithms. ~
~~
~~~
More
~~~~
r e c e n t l y G l o w i n s k i and P i r o n n e a u o b s e r v e d t h a t t h e u s e of a conjugate gradient algorithm,
t o achieve the minimization,
c a n d r a s t i c a l l y improve t h e s p e e d o f c o n v e r g e n c e ,
specially
17 4
R. GLOWINSKI and 0. PIRONNEAU
in the solution of some hard problems like those to be considered in the following sections.
J
Remark 1.2: The choice of a convenient cost function
is a
very important matter if one wishes to use the above rnethodology.
Let us mention some possibilities we have found very
attractive since they simplify the adjoint state equation and lead to good convergence properties.
In the case of the problem (1.1) assume that for A(v)
E e(V,V’)
dual space. if
A(v)
where
(
where
is an Hilbert space and
V
Assume also thak
A(v)
v
given, V’
is self-adjoint. .
is _ strongly - a natural choice for _ _ _ V-elliptic
,*)
denotes the cu-ality -~_ pairing ~ _ . _o _ f _V’
Similarly i f in (1.2),
A
and
its Then
J
is
V.
is strongly V-elliptic it is
convenient to use
In the following sections we shall see how the above principles can be applied to the numerical solution of two important problems i n Fluid Mechanics, namely: *
The transonic flow of an isentropic, compressible perfect ~
fluid. ’ Navier-Stokes equations for incompressible fluids.
Numerical results will be presented in order to illustrate the possibilities of the above methods.
To conclude this report some indication will be given on the minimization of non quadratic functionals by conjugate gradient algorithms.
LEAST S Q U A R E S O L U T I O N OF NON L I N E A R PRODLEMS
175
2. Numerical solution-of transonic flow problems. . __ ~
2.1
Generalities on transonic flows.
The theoretical and numerical studies of transonic flows for -_ perfect fluids have always been very important,
But these
problems have become even more important these last years in relation with the design and development of large subsonic aircrafts. These transonic problem are very difficult, theoretically and mumerically, for the following reasons: They are n o n linear. Shocks may exist in the flow. An Entropy Condition is needed, in a way o r another, to avoid non physical solutions. They are mixed problems i n the sense that they are el_l.ipt-icin the Eubsgsic_ part of the flow, LyEerbolic i n the supersonic part of the flow. From a theoretical point of view let us mention the work of Morawetz [l].
At the present moment the numerical methods
which are the more commonly used have originated from Murman-Cole [l]
and we shall mention B a u e r - G a r a b e d i a n - K o r n r l ] ,
Bauer-Garabedian-Korn,Jameson
11
,
Jameson [ 11 ,[ 21 ,[ 31 ,[ 41
and the bibliographies there i n (see also Hewitt-Illingworth and Co-editors [ 13 ) . These above numerical methods use the key idea of Murman and Cole which consists to use a finite difference scheme, centered i n the subsonic-part of the flow, kackward (in the direction of the flow) in the supersonic part.
The switching
1 76
R.
GLOWINSRI a n d 0 .
between t h e s e two-schemes
PIRONNEAU
i s a u t o m a t i c a l l y done v i a a
t r u n c a t i o n o p e r a t o r only a c t i v e i n the supersanic p a r t of flow ( s e e Jameson,
loc.
cit.,
f o r more d e t a i l s ) .
We s h a l l d e s c r i b e h e r e a d i f f e r e n t a p p r o a c h ,
which seems v e r y
c o n v e n i e n t f o r comput i n g f l o w s p a s t p r o f i l e s , infinity,
or f l o w s i n n o z z l e s .
the
subsonic a t
I n o u r method t h e t r a n s o n i c
f l o w problem i s f o r m u l a t e d a s a non l i n e a r l e a s t s q u a r e problem,
w h i c h i s i n t u r n i n t e r p r e t e d a s a n o p t i m a l~control
p r o b l -e m .
Then f i n i t e e l e m e n t s a r e u s e d t o a p p r o x i m a t e t h e S i n c e t h e .____ e n t r o p y~-c o n d i t i o n i s f o r m u l a t e d by
above problems.
~
a l i_ n e_ ar _ ine q u a l i t y c~-o n s t . ra i n t , a c o n v e n i e n t method t o h a n d l e .
i t i s t o u s e p e n a l t y a n d / o r d u a l i t y methods ( u s i n g a n augmented l a g r a n g i a n i f p e n a l t y and d u a l i t y a r e co mb in ed ) . Duality techniques a r e not
c o n s i d e r e d i n t h i s r e p o r t and w e
r e f e r t o Glowinski-Pironneau [ d e s c r i p t i o n of
2.2 2.2.1
11,
Glowinski
[13 for a
them.
M a t h e m a t i c a l m o d e l s f o r t r a n s__ o n __ i c f.l_ o ws. . .. B..-a s i c a s s u m p t i o n s .
We a s s u m e t h a t t h e f l u i d u n d e r c o n s i d e r a t i o n i s p e r f e c t a n d compressible,
and t h a t t h e flow of
a n d i r r o t a t i o-n a l ( i . e . ,
potential).
-
such a f l u i d is isentropic These assumptions a r e n o t
t r u e i n g e n e r a l s i n c e t h r o u g h a s h o c k wave t h e r e i s a v a r i a t i o n of e n t r o p y , and a n i r r o t a t i o n a l f l o w becomes rotational
(i.e.
t h e v a l i d i t y of i n t h e c a s e of
t h e r e i s c r e a t i o n o f v o_ r t_ i c i_ ty).
Therefore
t h e model t o f o l l o w i s assumed t o b e c o r r e c t "weak s h o cks". -.
177
LEAST SQUARE SOLUTION O F N O N LINEAR PROBLEMS
I n t h e c a s e o f a f l o w s p a s t a s- h a_ r p~ p_r o_f ._ i l e w e s h a l l suppose t h a t t h e r e i s no wake b e h i n d
2.2.2
E--q u a -~ t i o n s of
the tr a- _ i l ~i n g edge.
t h e flow.
~
n
Let
be t h e domain o f
r
t h e f l o w and
i t s boundary;
then
we s h a l l assume t h a t t h e f l o w i s m o d e l l e d by 1
,
-
(2.1)
, 2
y-1
v-((1 -
where i n ( 2 . 1 ) .
- 0 -
09
i s the flow p otential, _____
the flow v e l o c i t y ,
i s t h e c.r i t i c a l v e l o c i t y ,
C,
~~
-
y
i s t h e r. __ a t i o of
W e have *
(y = 1 - 4 f o r
specific heats
4.
_ _ I
t o add t o ( 2 . 1 )
Boundary c o n d i t i o n s ( o f D i r i c h l e t a n d / o r
Neumann t y p e ,
for
instance), *
A K u t t a - J o u k o v-k y
condition i n the case of
a l i f t i n g body ( s e e L a n d a u - L i f c h i t z
-
the flow past
[ l , Sec.
46]),
An e n t r o p y c o n d i t i o_ n . i n o r d e r t o e l i m i n a t e t h e non p h y s i c a l .s o l u t i o n s of
(1.1); t h i s p o i n t
w i l l be d i s c u s s e d i n S e c .
2.2.3. Remark 2 . 1 :
I t c a n h a p p e n t h a t on some p a r t of
0
have
:
and
ness; -
t o be g i v e n s _ i m_ u l t a n e o u s- __ ly t o i n s u r e unique-
i t is, f o r i n s t a n c e ,
of F i g .
2.1 i f
~~
t h e c a s e of
'4
g i v e n on
t h e divergent nozzle
the velocity a t the entrance i s supersonic.
T y p i c a l boundary c o n d i t i o n s a r e 5-n
t h e boundary,
r4, rl, r 2 .
If
@
g i v e n on
r l , Fg
and
the velocity a t the entrance
178
(i.e.
R.
I-,)
GLOWINSKI and 0. P I R O N N E A U
is subsonic we require ___ less boundary conditions.
Figure 2.1
Figure 2.2
Remark 2.2:
In the case o f t h e f l o w past a multibody profile
(like i n Fig. 2.2)
each
body requires a Kutta-Jourkowsky
condition.
2.2.3
Formulation of the entropy condition.
It follows f r o m Landau-Lifchitz [ l , ch.91 condition c a n be formulated as follows:
that the entropy
1 79
LEAST SQUARE SOLUTION OF NON LINEAR PROBLEMS
1 In the direction
o f the flow, One- cannot have a - ___ ~~
-!
I
(2.2)
subsonic-supersonic transition through
a_;t?ck.
F o r a one-dimensional . . flow, (2.2) implies ~
*<
2 d
(2.3)
dx
3-
2
+m,
is a measure bounded from_ above; _ ~ _ _weak (and more dx rigorous) formulation of (2.3) are: i.e.
-~
There exists a constant ~~
~
M,
such that-either -
where
(We recall that fJ
(R) =
{cp E Cm ( f i )
,
cp
has a compact support in
n] . )
In the case o f a t ~ oo r three dimensional flow we shall suppose that (2.2) can be formulated by
(2.9)
190
R.
GLOWINSKI a n d 0 . P I R O N N E A U
T h e n u m e r i c a l r e s u l t s o b t a i n e d f o r t__ wo and _ t r e_e -_ d i_ m e~n s_ i o_ n a_ l flows u s i n g d i s c r e t e analogous of above formulation Remark 2 2 :
(2.7) seanto justify the
of t h e entropy condition.
The c o n d i t i o n ( 2 . 7 )
i s a c t u a l l y a p a r t i c u l a r case
O f
(w)+ E LPP)
(2.10) where,
assuming t h a t
is a m a s_ u_ re, _e_
( ~ c p ) + = positive p a r t of
(2.11) Then
Acp
(2.7)
corresponds t o
f r o m a computational point
p =
+m;
~cp.
another useful choice i s ,
of view,
p=2
i.c.
(2.12)
To c o n c l u d e t h i s S e c .
2.2
a t t h e moment,
l e t u s mention t h a t ,
t h e t r a n s o n i c f l o w problem i s s t i l l open from a t h e o r e t i c a l ~
~~
of view, even i n t h e simpler c a s e of t h e so-called
point
s m a l l d i s t u r b a n c_e _ c c l u a t i o n ~~
details).
( s e e J a m e s o n [ 11 ,[
41
f o r more
T h i s l a c k o f e x i s t e n c e and u n i q u e n e s s r e s u l t s h a s
v e r y l i k e l y i t s c o u n t e r p a r t i n t h e n u m e r i c a l methodology.
2.3
L - east
s q u a r e f o r m u l a t i on. ~
~~
R edu5;ion _ _t o~ a n o p t i m a l -~ ~~
~
~
~~
c o n t r o l problem.
If
w e s u p p o s e t h a t t h e d e n s i t y o f t h e f l u i d i s o n e when t h e
velocity is zero,
then the coefficient
a s the density of
the
fluid.
of
i n (2.1)
We s h a l l u s e t h e n o t a t i o n 1
2 Y - l (2.13)
V@
lvcpl
)
.
appm
181
LEAST SQUARE SOLUTION OF NON L I N E A R PROBLEXS
The i d e a o f t h e method t o f o l l o h i s t o d ecoup_ le t h e _ density ~ and t h e p o t e n t i a l those described
4
potential
0
and
-
via a _ l e_a s -~ t s_ q u_ a r_ e~ f o_ rm u_ l a.t_____ i o_ n like _ _
@
To d o s o we i n t r o d u c e a new
1.
i n Sec.
-
the c o. n t r o l p o t e n t i a l ~
by m i n i m i z i n g a c o n v e n i e n t
least
square c o s t function.
W e may u s e f o r i n s t a n c e t h e f o l l o w i n g f o r m u l a t i o n formulations s e e Glowinski-Pironneau pironneau
5
and t r y t o r e c o u p l e
(for other
[ 2 1 , Gowinski-Periaux-
[ 11 )
(2.14) where
5
i s a f u n c t i o n of
in
via the state
(2.15) p l u s boundary c o n d i t i o n s f o r
I-.
on
d,
i
In
(2.14
,
a convex s e t Since
a
t h e parameter
'' c o n v e n i e n t l y ''
p ep(x)) = 0
is either cho s e n
or
0
1
.
1 Y+1 2
I Vcp(x) 1
i f and o n l y i f
X
and
= (Fl)
is
and
C,
1
y =
that for a i r , appears t h a t
i n the
1.4
implies
y+1 2
(--)
Y-1
t r a n s o n i c range
=
--
,,/6 = 2 . 4 5 , =
(say
it
lVCpl
1.5
c*)
w e have
(2.16)
o
< 6 s p(cp(x)) s 1
I t f o l l o w s from ( 2 . 1 6 )
that
a.e.
on
0.
( 2 . 1 5 ) i s an e l l i p t i c p r o b l e m f o r
a p p r o p r i a t e boundary c o n d i t i o n s . l i f t i n g b o d i e s , Kutta-Joukowsky
I n t h e case of flows p a s t conditions are a l s o required
t o h a v e , w i t h t h e o t h e r b o u n d a r y c o n d i t i o n s , a -__ physical s o l u t i o n o f problem ( 2 . 1 5 )
(modulo a c o n s t a n t i f
Neumann c o n d i t i o n s on t h e b o u n d a r y ) .
one o n l y h a s
R.
182
2.4:
Remark
If
GLOWINSKI and 0 . P I R O N N E A U
i n t h e o r i g i n a l problem,
0
r,
be s i m u l t a n e o u s l y p r e s c r i b e d on some p a r t of approach,
5
w i t h t h e two p o t e n t i a l
and
'@ have t o an
and
a,
t h e above
i s very
convenient s i n c e t h e boundary c o n d i t i o n s can be s p l i t t e d b e t ween
@
4.
and
However i f
boundary c o n d i t i o n s f o r t a k e a c c o u n t of
0
one w i s h e s t o u s e t h e same and
4
i t i s always p o s s i b l e t o
t h e e x t r a boundary c o n d i t i o n s
(assumed t o be
(2.14) a
o f D i r i c h l e t t y p e ) by a d d i n g t o t h e c o s t f u n c t i o n I
quantity proportional t o e i t h e r
( I
15-@,1
(9-0,1
2
br,
or
'd (or t o a l i n e a r c o m b i n a t i o n o f b o t h ) , where
2
' rd
i s the part
of
on which one r e q u i r e s
A s i m i l a r i d e a i s used i n Begis-Glowinski
@Ird
=
rd
Od.
[l] t o s o l v e some
f r e e boundary p r o b l e m s . Remark 2.5: variant
To s t a t e t h e e n t r o p y c__ o n d-i t_i_ o n ( 2 . 7 )
4
( 2 . 1 2 ) ) we h a v e t h e c h o i c e between
(or its
and
F
(actud-
l y we can a l s o u s e t h e s e two p o t e n t i a l s ~ i m y l t a n e o u ~ l y ) I. f one u s e s
E
( r e s p . @ ) we have a c o n t r o l c o-n s t r a_i_n t
(resp.
a s t a t e c-onstr-int). Remark 2 . 6 :
W e observe t h a t for t h e c l a s s o f flows under
c o n s i d e r a t i o n we h a v e
T h i s o b s e r v a t i o n s u g g e s t s t h a t t h e convex
X
( 2 . 1 4 ) h a s t o be t a k e n a s a convex s u b s e t of
occuring i n
Wl'"(S2).
We
o b s e r v e a l s o t h a t t o s t a y i n t h e t r a n s o n i c r a n g e i t may b e convenient t o intr oduce t h e following c o n s t r a i n t s
or
183
LEAST SQUARE S O L U T I O N OF NON LINEAR PROBLEMS
(2.18) Actually the computations we have done prove that for a physically well-posed transonic problem it is not necessary ~~
to introduce
Remark
~
(2.17), (2.18).
2.7: If the transonic problem has a solution and if
X
is "large enough" then the least square problem has a solution such that the cost function is equal to zero.
This last
property gives us means to check the quality of the computed solution.
2.4
Finite element approximations. __
We assume that
0c
_. Three-dimensional calculations will
2
!R
be reparted i n a forthcoming paper.
2.4.1
Generalities.
The above control problem will be approximated by finite elements.
Compared to finite differences, finite elements
give us the possibility o f handling problems posed on rather complicated geometrical domains.
Moreover the variational . .~
framework of finite element formulations is very appropriate to the problem under consideration and to the methodology we use to solve it.
It will be in particular fairly easy to
approximate the weak formulations of the entropy condition (2.7) (and also the alternate entropy condition (2.12)).
If
R
is unbounded, it is replaced by a bounded domain
still denoted by
-
as large as possible.
-
To approximate
the above continuous problem, we introduce a standard tri-
184
R. GLOWINSKI and 0 . PIRONNEAU
Zh
angulat ion elements).
of
61
(we can also use quadrilateral finite -
7
Then we approximate the functions
and
$I
by
piecewise polynomial functions belonging to the following subspace of
with
H1(n)
(and
Wl’”(n))
Pk = the space o f polynomials of degree
Sk.
Remark 2.8: In the case of a lifting body, to take into account the Kutta-Joukowsky condition, one usually introduces -~ .. (see Fig. 2.3) an arc
y
between the trailing edge o f the
body and the external boundary.
Thus arc
constant jump, a priori unknown of
a@
requires the continuity of
Y.
Since
@
-
an
@
(and
is discontinuous along
and
y
supports a
<,
ae z) at y,
but one the crossing o f
Hl(61)
c a n not be
used any more, but introducing
.
.
0
\
n = n - y one can use
A.
HL(n)
and define
Vh
over a triangulation of
Results of computations for such profiles are given in
Sec. 2.6;
however the method f o r t r e a t i n g t h e Kutta-Joukowsky
condition will not be treated in this report since they are not specific of transonic flows. In order to simplify the presentation o f our method we shall assume i n the sequel that one works directly over
61.
LEAST SQUARF S O L U T I O \
@rU O U
Figure
Remark 2 . 9 :
If
@
and
?
2.3
have t o s a t i s f y o n l y N e u m-a r i r i
b-o u n d a r y_ _c o n d i t -~ i o n s , t h e s u b s p a c e of ~
Vh
itself.
If
on t h e c o n t r a r )
i$
Vh
t o h e u s e d w i l l be
and/or
haxe to satisfi
F
D i r i c h l e t boundary c o n d i t i o n s over s o m e p a r t of s h a l l have t o use subspaccs o f Remark 2 . 1 0 : domain o f
R
Vh,
,
",
t h e n we
s t r i c t l y included _-___
We have t a c i t l y assumed t h a t 2
185
LIhFAR l'ROI3LEVS
R
in V
h'
i s a polygonal
or has b e e n a p p r o x i m a t e d b y s u c h a d o m a i n .
Hoizever i n t h e c a s e o f a curx-ed b o u n d a r y i t i s a l w a y s r > o s s i b l e t o use
( a t s o m e e x t r a computational c o s t ) curved f i n i t e
elements
t-11).
(see,
for i n s t a n c e , S t r a r i g - F i x [I], C i a r l e t - R a v i a r t
R. GLOWINSKI and 0. PIRONNEAU
185
2.4.2
o f the state equation . and . . o f_ _the cost Approximation __. ~~
function. T o simplify the presention, we shall assume that we only have
Neumann .boundary cond>t-ions, i.e.
aa@ n
(2.20)
= g
on
7
(it is the case for the important application of flows subsonic at infinity, past profiles). F
sections that
and
0
It follows from the above
will satisfy the same boundary ~
conditions.
We shall also assume that if
then the corresponding value of (2.21)
p g
p
dr =
g(x)
f 0, x E
r,
is known, and that
0.
'iThen the state equation (2.15) h a s the following variational
(2.22)
which is approximated by
r
i
@
and
f
(pg)h q h
dr
qhevh,
'I-
[
Oh vh' ( ~ g )is~ a Convenient approximation of @h
pg.
Since
are only defined modulo a constant we shall
prescribe the values of some point of ed by
dx =
4
(2.23)
where
P(ch)v@h'v"h
r.
@
and
cph
(and
5
and
5,)
at
The cost function in (2.14) is approximat-
187
LEAST SQUARE S Q L U T I O N O F PJON L I N E A R PROULEFIS
If one
R e m a r k 2.11:
(k=l),
uses
the piecewise linear approximation
then the integrals occuring i n ( 2 . 2 3 ) ,
to compute since property for
05,
p(ch)
(2.24)
piecewise constant implies the same (from (2.1'3)).
If
k=2
a ~ numerical _
integration procedure has to be used in ( 2 . 2 3 ) ,
a
are easy
_
and also in
= 1.
(2.24)
if
2.4.3
Approximation of .the entropy condition. -~ _. ~~
~
~~~
~
T o avoid non physical shocks (i.e. shocks for which the
thermodynamical entropy condition is n o t satisfied) we use the formulations o f Sec. 2 . 2 . 3 .
We still assume that w e o n l y
have Neumanii b o u n d a r y conditions.
We use the notation of the continuous oroblem. the condition ( 2 . 1 2 ) (2.14)
If
one uses
we may "regularize" the cost function
by adding to it, with
C
> 0, either
(2.25)
or (2.26)
E
[ '*
l ( h E ; ) + 1 2 dx
( o r a linear combination of both).
In
(2.25),
(2.26),
is
a "small" parameter.
We can make this approach more sophisticated by using instead of ( 2 . 2 5 ) ,
(2.26),
regularization functionals like
188
R.
GLOWINSKI and 0 . PIRONNEAU
0.
p o s s i b l y e q u a l t o z e r o o v e r some p a r t of
To u s e t h e above methodology f o r t h e a p p r o x i m a t e problem i t i s n e c e s s a r y t o have a n a p p r o x i m a t i o n o f W e may u s e a n a p p r o x i m a t i o n a p p r o x i m a t i o n of
C i a r l e t - R a v i a r t [ 21
,
[
(2.28)
73 ) :
i s s u f f i c i e n t l y smooth, t h e n from
we h a v e
Ai/lcpdx =
'n
(2
i
epdP
-
(2.29)
I
{R
an a p p r o x i m a t i o n of
E H1(n).
-
i
'R
Ah
$I,
0Bh'VCPh dx
A@
of
as
cph E Vh,
bL
I n (2.29),
t o define
the function
g
of
gh
is
(2.20).
we a l s o h a v e D i r i c h l e t boundary c o n d i t i o n s
If
o v e r some p a r t
cp
E Vh*
We u s e t h e same method
Remark 2 . 1 2 :
ghCPhc
=
'h'hcphdx
AhOh
+
V w - v c p dx
F
U s i n g t h i s i d e a we d e f i n e a n a p p r o x i m a t i o n €allows
[21,
e q u a t i o n ( s e e Glowinski
-
Glowinski-Pironneau [
$
L e t u s assume t h a t Grgeq-for5qJ-a-
s u g g e s t e d by F i x e d -f i n i t e e_l e_m_e n t
t h e b i h a r m o n~ic
~
AS).
(resp.
of
r,
t h e above i d e a i s s l i g h t l y more
c o m p l i c a t e d t o a p p l y , b u t t h i s can be d o n e . Remark 2.13: To o b t a i n
from ( 2 . 2 9 ) we have t o s o l v e a
Ah@,
l-~ i n e a r s ys t e m t h e m a t r i x o f which i s _s_yrnmetric, p o s i t i v e ~~
definite,
sparse, but p o t d i a g o n a l
approximation of t h e o p e r a t o r I
approximate
n
4 0 cp
h h h
dx
I).
( t h i s m a t r i x i s an If
k=l
one c a n
u s i n g t h e two-dimensional,
189
LEAST SQUARE SOLUTION O F N O N L I N E A R I'RODLEMS
t____ r a p e-z o i d a l n u m e r i c a l i n t e g r a t i o - n m e t h o d .
D o i n g s o we o b t a i n
~
s o l v i n g a l i n e a r s y s t e m w i t h a _d_i a g o n a l m a t r i x .
by
ahQh
If
t h e a b o v e r e g u l a r i z a t i o n method i s t e c h n i c a l l y more
k=2
complicated t o use.
A
Once <Jh
h a s b e e n a p p r o x i m a t e d we add t o t h e c o s t f u n c t i o n
(see ( 2 . 2 4 ) ) the functional
I
2
I (Ahah)+]
('. 3')
dx
(rcsp'
Eh(x)
I ('h
2
dx)
'R
ch
with (2.30)
" s m a l l " and
0.
2
W e use i n f a c t approximations o f
o b t a i n e d v i a a n u m e r i c a l i n t e g r a t i o n p r o c e.d u r e .
W e have t o m e n t i o n t h a t t h e o p t i m a l c h o i c e f o r
is still
Fh
a n open q u e s t i o n , n e v e r t h e l e s s t h e r e e x i s t s o m e e m p i r i c a l methods t h a t w e d o n ' t d i s c u s s h e r e .
Numerical r e s u l t s
ed w i t h t h e p i e c e w i s e l i n e a r a p p r o x i m a t i o n i n Sec. Eemark
(k=l)
obtairr
a r e given
2.6.
2.14:
T h e r e g u l a r i z a t i o n by
€
(or i t s
l ( h @ ) + 1 2 dx
v a r i a n t s ) m a y be s e e n a s a t e c h n i q u e f o r i n t r o d u c i n g s o m e pseudo-viscosity
B
-
i n t h e problem under c o n s i d e r a t i o n .
A method u s i n g ( 2 . 7 ) .
L e t u s d e s c r i b e t h i s method f o r t h e c o n t i n u o u s p r o b l e m ,
first:
~~
We assume t h a t t h e e n t r o p y c o n d i t i o n i s f o r m u l a t e d by ( 2 . 7 ) . Let
M(x)
b e a s u f f i c i e n t l y s m o o t h u p p e r bound o f
e s t i m a t e d or g u e s s e d ) .
a weak f o r m u l a t i o n o f
R e p l a c i n g ( 2 . 7 ) by
(2.31)
is
hb,
(M i s
GLOWINSKI and 0. P I R O N N E A U
R.
190
-
(2.32)
I
1
[email protected] dx
S
I
n
M ( x ) c p dx
V
cp E 8 + ( n ) .
R i t i s v e r y convenient t o i n t r o d u c e t h e
M
I n s t e a d of u s i n g
s o l u t i o n ( d e f i n e d modulo a c o n s t a n t ) o f
Ago = M
over
0,
= g
over
r,
(2.33)
then
i""'o.vv
(2.33) has t h e f ollow ing v a r i a t i o n a l formulation dx =
-
(
M(x)cp dx +
f
v cp t H 1 ( n ) ,
mdr
r
R
(2.34)
I t f o l l o w s from ( 2 . 3 4 )
that
-i
(2.35)
W e observe t h a t
(2.32)
v ( O - @ O ) ' ~ c p dx
5
a (@-a0) = an
0.
c a n a l s o be w r i t t e n
v cp E
0
rs+(n).
F o r t h e d i s c r e t e problem t h e o b v i o u s s t r a t e g y seems t o be t h e following: F i r s t we a p p r o x i m a t e
vaoh-vvh
dx =
Q0
-
by
\ R
where
Mh
Go,,
solution of
Mh(x)vhdx +
f ghcphdr r
i s a c o n v e n i e n t a p p r o x i m a t i o n of
M.
+ e p , c vh,
191
LEAST SQUARE SOLUTION O F N O N L I N E A R PROBLEMS
( 2 . 3 5 ) by
F i n a l l y we a p p r o x i m a t e
1
-
(2.38)
'
v(@h-@o,)'vqh dx
qh
n
I n f a c t t h e above " o b v i o u s ' l
s t r a t e g y f o r t h e d i s c r e t e problem
h a s t o be m o d i f i e d f o r t h e f o l l o w i n g r e a s o n s : 1)
V:h
k=l,
If
can be g e n e r a t e d by t h e c a n o n i c a l b a s i s
2)
B u t i t i s n o t the c a s e f o r
Voh.
f u n c t i o n s of
k=l
C o m p u t a t i o n s u s i n g ( 2 . 3 8 ) and done w i t h t h a t t h e approximation o f
close t o
k=2. have shown
t h e s o l u t i o n i s n o t good
t h e p o i n t s a t which s o n i c l i n e s t o u c h
r.
To overcome t h e s e d i f f i c u l t i e s we may p r o c e e d a s f o l l o w s : If
k=l:
ed from 1
C,
Let
be t h e s e t o f
Nh,
t o
where
Vh,
c a n o n i c a l b a s i s of
Nh
t h e v e r t i c e s of
= dim(Vh).
Sh,number-
Wh
Let
be t h e
i.e.
(2.39) with
(2.40)
wi
E Vh,
W e observe t h a t
po - .s -i t i v e Z of
e
w
VL
i
2
of
Wi(Pj)
0
onver
Vh
v P j E c.,
= 6 1. J.
R,
+k
i,
i s g e n e r a t e d by
and t h a t
Bh.
the
Then i n s t e a d
t o formulate t h e d i s c r e t e entropy c o n d i t i o n
using (2.38)
one t a k e s
(2.41)
-
V(@h-@oh)*vp dh x
which i s e q u i v a l e n t t o t h e
s e t of
the
'
' Nh
ph
Vi
following l i n e a r
192
R.
GLOWINSKI a n d 0.
T-'IRO"EALl
i z c i u a 1i t y c o n s t r a i n t s
f
( 2 .I42 ) 5,
If
v(ql-cboh)
-1
-vwi
V
dx < 0
i = 1,
...,N h .
i s u s e d , i n s t e a d o€ ( 2 . 4 2 ) x e h a v e I
-
(2.'43)
v(~tl-@oll)'v\\idx
0
4
V
...,N h .
i = 1,
62 Computations done with
and u s i n g ( 2 . 4 2 )
k=l
have p r o d u c e d
good r e s u l t s . k=2: The t e c h n i q u e s
I.f ~
~~
t o be u s e d a r e m o r e c o m p l i c a t e d a n d
a r e d e t a i l e d i n G l o ~ i n s k i - P i r o n n c a u[ 1 1 ,
-Remark ..2.15:
(It holds f o r
If
k=1,2).
some D i r i c h l e t
r,
b o u n d a r y c o n d i t i o n s a r c ' p r e s c r i h e d somewhere o v e r p o s i t i v e cone used
the
t o de€irie t h e d i s c r e t e entropy c o n d i t i o n
w i l l b e r e l a t e d t o t h e subspace of
Vtl
c o n s i s t i n g of
those
f u n c t i o n s v a n i s h i n g a t t h e boundary nodes corresponding t o the
(discrete) Dirichlet condition.
2.26:
Remark
(or
Mh)
may b e a n u n e a s y
computations. p a r t A,
Houever,
t~e c h n i q_u_e i s u s e d
e s p e c i a l l y for p r o f i l e
l i k e i n t h e r e g u l a r i z a t i o n method o f
t o handle
2.'+./+ A p p r o x.i m~-~ ation of ~
(2.'+2) ( o r
if
Part A,
~_
X.
take i n t o account
(2.17),
(2.18),
Xh
then
Xh
is
the d i s c r e t e entropy conditions.
2.4.3 e i t h e r
one u s e s t h e r e g u l a r i z a t i o n methods
or
a penalty
(2.43)).
one u s e s t h e methods d e s c r i b e d i n S ec.
= Vh
if
~~
e s s e n t i a l l y determined b y If
task,
some e m p i r i c a l r u l e s h a v e b e e n o b t a i n e d ,
If we d o n ' t
M
The o p t i m a l c h o i c e f o r the hound f u n c t i o n
of
Sec.
xh =
2.4.3,
i s d e f i n e d by t h e l i n e a r i n e q u a l i t y constraints
LEAST S Q U A R E SOLUTIOV O F NON L I N E A R PKOI)LE,PIS
(2.42)
2.5
( o r (2.43))
ICe-rative
~-
2.5.1
so
i f w e proceed
lution of
l i k e i n Sec.
193
2.4.3,
P a r t B.
t h e a p p~r o x i m a t e p r~. oblems.
-~
~~~
Preliminary statements.
G eneralitie - s.
We s h a l l r e s t r i c t o u r a t t e n t i o n t o t h e f o l l o w i n g s i t u a t i o n : *
R
i s hounded anti w e o n l y have Neumann boiindary
conditions.
~
We assume a l s o t h a t K u t t a - J o u k o w s k y required
i s not
since t h e i r treatment
conditions are not s p e c i f i c of
transonic
flows. '
We d o n ' t
take
i n t o account
*
We s u p p o s e t h a t
a
= 1
in
(2.17), ( 2 . 1 8 ) .
(2.14)
and t h a t an e n t r o p y
c o n d i t i o n i s f o r m u l a t e d u s i n g t h e method o f Part B,
with the formulation (2.43)
Sec.
2.4.3,
(control's constraints).
The c a s e where t h e r e g u l a r i z a t i o n method o f P a r t A i s u s e d i s s i m p l e r , and i n f a c t may b e s e e n a s a p a r t i c u l a r c a s e o f other ( a t least i f '
one u s e s p e n a l t y t o h a n d l e
the
(2.43)).
We f i n a l l y assume t h a t we u s e p i e c e w i s e l i n e a r f i n i t e elements
(k=l).
S i n c e we d o n ' t closed
take i n t o account
convex s u b s e t of
Vh
(2.17), ( 2 . 1 8 ) ,
d e f i n e d by
t h e a p p r o x i m a t e c o n t r o l problem
(2.43).
xh
i s the
Therefore
i s a non l i n e a r ( a n d non ~
c o n v e x ) programming p r o b l e m i n w h i c h t h e i n d e p e n d e n t v a r i a b l e
is
qh.
method.
~~
We s h a l l s o l v e t h i s p r o b l e m by a c o n j u g a t e g r a d i e n t ~~~
T o take i n t o account t h e l i n e a r c o n s t r a i n t s ( 2 . 4 3 )
we h a v e u s e d a p e n a l t y method
( f o r t h e s o l u t i o n by p e n a l t y -
d u a l i t y m e t h o d s v i a ____ augmented ._______ Lagrangians see GlowinskiP i r o n n e a u [ 11 )
.
194
R.
2.5.2
GLOWINSKI and 0 . PIRONNEAU
Penalty approximation of the discrete ____ problem. ~
~
We use the notation o f Sec. 2.4.3.
(Uh9Vh)h
-- _
vh
2
L (n)-scalar product: ~-
the following _ approximate (2.44)
Let us define over
Nh
c
miuh(Pi)vh(Pi),
pi E
ch,
tf
i,
3 i=l where
i :f
m
-
6. =
measure union of
(nil, ni TE
0
=
a.
1'
6, such that Pi is a vertex of T;
the corresponding norm is denoted by
\*\,.
Then we define
It follows therefore from (2.45) that (2.43) can also be written
which is equivalent to
where in
cph
(2.48)'
(cp,)'
does not denote the positive part of
but the approximation of it defined by
LEAST SQUARE
In
(2.48)s
@Il
195
SOLUTION O F N O N LINEAR PROULEPIS
Eh
is a function of
t h-r~ ou _ g_h_ ( 2 . 2 3 )
and
.
r
i s a p o s i t i v e p a r a m e t er. Then w e c a n p r o v e t h e f o l l o w i n g :
- -~ P rop-osition
2 .1
-
If
t h e approximate c o n t r o l problem ~~
Min
(2.50) has a s o l u t i o n ,
Jh(Fh)
'h
I-1'
t h e n t h e s o l u t i o n o f - t h e p e n a l i z e d problem
(2.51) converges
Remark 2.17: ______
If
Ooh
= 0
notation) a variant of i n Sec.
2.4.3,
in
~~
the r e g u l a r i z e d f u n c t i o n a l i n t r o d u c e d
(2.51) i s a f i n i t e dimensional c o n t r o l
constraint.
We h a v e s o l v e d t h i s p r o b l e m
the n - on l i n e a r c o n j u g a t e
_g _r _ a d i e n t method
( s e e P o l a k [ 1, C h . 2 ,
more e f f e c t i v e ,
f o r our problem,
pp.
53-551)
w h i c h seems
than the F l e_ tche r.- R e e v e s _
The s c a l a r p r o d u c t u s e d i n t h i s a l g o r i t h m i s t h e
s c a l a r product important
+m.
(2.48) w e recover ( w i t h d i f f e r e~nt
u s i n g t h e P o l a k - R i b i & ~ ev e r s i o n o f
version.
+
P a r t A.
The p e n a l i z e d problem problem without
r
(2.50)
toasolutioLo-f
i n d u c e d by
H1(n)
over
s t e p i n t h e s o l u t i o n of
algorithm i s t h e computation of
(2.51)
Vh.
Therefore a very
by the ab o v e
the gradient
r _J h_
d5 h
~ (5,)
; this
196
R.
point i s discussed
GLOWINSKI and 0.
PIRONNEAU
i n t h e f o l l o w i n g s e c t i - o n 2.5.4.
A d e s c r i p t i o n o f t h e m o s t i m p o r t a n t non l i n e a r c o n j u g a t e g r a d i e n t a l g o r i t h m s i s g i v e n i n Sec.
2.5.4
Computation of
-- J h r
4.
( F ~ ) .
d5 h Owing t o t h e p r a c t i c a l i m p o r t a n c e of
.
Jhr
"h
we s h a l l d i s c u s s
i t s c o m p u t a t i o n w i t h some d e t a i l s : We have
I
We c a n e a s i l y compute t h e l a s t t e r m of
(2.53). of
(2.23)
About
t h e right-hand
s i d e of
t h e s e c o n d t e r m , w e o b t a i n by d i f f e r e n t i a t i o n
LEAST SQUARE S O L U T I O N O F N O N L I N E A R PROBLEMS
197
Since
(2.57) !
t h e second term o f with t h e l a s t
Jhr ( 5 h )
' 6!3
( 2 . 5 6 ) i s e a s i l y computed and by a d d i t i o n
t e r m of
t h e r i g h t hand s i d e o f
( 2 . 5 3 ) we o b t a i n
h'
Remark 2.18: I f -~ ..-.
i n s t e a d of t a k i n g
U.=l
i n (2.24),
one t a k e s
~
a
= 0
w i l l r e q u i r e the use o f
t h e n t h e c o m p u t a t i o n of
an adjoint s t a t e equ ation. -~
2.5.5
Computational c o n s i d e r a t i o n s . __
When s o l v i n g ( 2 . 5 1 )
by t h e non l i n e a r c o n j u g a t e g r a d i e n t
method d i s c u s s e d a b o v e , we h a v e t o s o l v e a t e a c h i t e r a t i o n the state equation
(2.23).
S i n c e t h e b i l i n e a r form o c c u r i n g
i n (2.23)
is positive de€inite
p o i n t of
6
( o n c e t h e v a l u e of'
h a s b e e n p r e s c r i b e d ) we c a n u s e t o s o l v e
e i t h e r i-t e r a t i v e m e t h o d s l i k e c o n j u g a t e g r a d_--.I ient, I_
S.S.O.R.
etc...
A x e l s s o n [ 13
qjh
,
(cf.
e.g.,
Young [ 11 )
P o l a k [ 1 1 , Concus-Golub
i n one
(2.23)
S.O.R., __
[l]
,
o r d i r e c t m e-. thods l i k e G a u s s ' s o r
Choleskyls.
2.6
Numerical e x p e r i me n t s .
The f o l l o w i n g n u m e r i c a l t e s t s by Mrs. B r i s t e a u a n d MM.
( i n two d i m e n s i o n s ) were done
P e r i a u x , P o i r i e r on I B M 3 7 0 - 1 6 8 ,
198
R.
G L O W I N S K I and 0.
PIRONNEAU
o t h e r t e s t s and r e s u l t s a r e r e p o r t e d
C 11 ,
21
,
i n Glowinski-Pironneau
~ l o w i n s ~ < i - ~ e r i a u x - ~ i r o n n[c11 a u,[: 21
Concerning 3-dimensional
problems,
.
some c a l c u l a t i o n s h a v e been
d o n e on t h e e f f e c t s o f o b l i q u e wind on t h e f l o w a t t h e i n t a k e (see Fig.
2.4)
of
a j e t e n g i n e a t l o w mach number ; t h e s e
-
c a l c u l a t i o n s w i l l be r e p o r t e d
elsewhere.
Aircraft
Velocity
\
J e t engine
F i g u r e 2.4 The f o l l o w i n g r e s u l t s c o n c e r n
-
A Naca 0012 p r o f i l e ,
A Kornls p r o f i l e , An O n e r a ' s p r o f i l e ,
A two-piece
2.6.1
airfoil.
Naca 0012 p r o f i l e .
The t e s t d e s c r i b e d b e l o w c o r r e s p o n d s i n f i n i t y of
.85
(M,
= .85)
t o a mach number a t
and i s w i t h o u t _ i n~c_i _ d -e n c e . ~
We
199
L E A S T SQUARE SOLUTION OF NON LINEAR PROCLEMS
have used __piecewise linear finite elements ing to a triangulation %h
correspond-
(k=l)
of 600 triangles and 341 nodes.
The entropy condition (2.12) was required via regularization. ~
~~
The convergence was obtained in 80 iterations for a CPU time of
4
mn.
O n Figure 2.5 we have shown the mach numbers at the midpoint
of the edges, on the skin of the airfoil, extrapolated from their values at the centroid o f the triangles. ed by Bauer-Garabedian-Korn's
Results obtain-
code are shown also.
We observe
the good agreement between the results, except for the amplitude of the shock jump,
This seems to be related to the
fact that in the B-G-K code we have used the discretization -. finite difference - scheme is achieved by a Eon conservative -
and that the small disturbance model is used.
(The schemes
we use are fully conservative). O n Figure 2 . 6 we show the results obtained for the same problem, using piecewise quadratic finite elements 2.6.2
(k=2).
Korn's profile.
The Korn's profile, which is not symmetric (see Fig. 2.7) has the property of giving a shock free flow if
for s o m e specific incidence angle).
Mm = 0.75
(and
On Figure 2.8 we have
shown the mach numbers on the skin of the airfoil;
we observe
that the computed flow is also shock free.
2.6.3
ONERA profile.
Wind tunnel measurements have been made by ONERA (Office Nationale d'Etudes et de Recherches AGrospatiales) for the
R.
200
GLOWINSKI and 0 . P I R O N N E A U
profile shown o n Fig. 2.9.
O n Fig.
2.9 is also indicated the
mach distribution on the skin o f the profile.
O n Fig. 2.10
the computed results are shown and the agreement with the measured results is very good, including the shock location and amplitude.
The agreement is less good after the shock
since the physical flow is n o longer potential.
We observe
that the computed shock is sharper than the measured shock. This fact is due, may b e , to the fact the actual fluid is slightly viscous, but very likely the main reason of this difference is the fact that the velocity fluctuations at the intrance o f the wind-tunnel induce fluctuations of the shock wave location; therefore the measurement taking same average value of the velocity has a smoothing effect,
2.6.4
Two-piece airfoil. ____~
The geometry shown on Fig. 2.11 corresponds to a &)ractical situation.
Kutta-Joukowsky conditions have to be imposed o n
both profiles.
T h e position o f the front piece makes the
funnel slightly convergent. incidence and (k=l)
Mm = . 5 5 .
The main airfoil is at 5 O of
Lagrangian elements of order 1
were used with 2936 triangles and 1 5 5 5 nodes.
Regularization i s used and convergence is obtained in iterations for a C P U time o f 1 6 mn. guessed by measuring
3.5
X
from
lo3
and at
IlV(@-C)Il
n=50
2;
L it is 2.
T h e precision can be at
n=O
its value is
O n each triangle it varies
in the subsonic region to
sonic region.
50
lom2
i n the super-
li
LEAST SQUARE S O L U T I O N OF N O N L I N E A R PROBLEMS
i
0
20 1
202
0
*a
R. G L O W I N S K I and 0 . P I R O N N E A U
W
m
N
F
LEAST SQUARE SOLUTION OF NON L I N E A R I’R0DLJ)MS
203
204
a U I
R.
h
x
G L O W I N S K I and 0 . P I R O N N E A U
h
h
4
LEAST SQUARE SOLUTION O F N O N L I N E A R PROBLEMS
'I;
205
R.
206
G L O W I N S K I and 0 .
PIRONNEAU
M
X 0.5
Figure 2.10
1.
L E A S T SQUARF: S O L U T I O N O F NON L I N E A R PROBLEMS
R. GLOWINSKI and 0. PIRONNEAU
208
3.
Numerical solutions of Navier-Stokes equations. ___ __
3.1
-
Generalities.
Or$entati&n.
In this Section 3 we should like to briefly introduce the reader to some aspects o f the numerical analysis o f the NavierStokes equations.
This is a very large subject in constant
evolution; for information prior to 1 9 7 6 see Temarn [l] and the bibliography therein.
In the following subsections we s h a l l see how the non linear least square methodology can be applied for solving the continuous Navier-Stokes equations in the -~ pressure-velocity ~~
formulation.
Similar methods can also be used for the stream
function-vorticity __ formulation; these methods will bedescribed in a forthcoming report. In practice the above problems have to-be approximated;
we
have used some mixed finite element methods which are described in the report mentioned above.
However we shall only
consider the continuous case whose formalism is simpler.
3.2
The - _N _a v i e r - S t o k e s - e - q u a t i o n s .
Let
0
be a domain o f
iRN
(N = 2 , 3
in practice) and
we consider the
stationary Navier-Stokes equations
(3.2)
uIT. = g
where
(
g'n dr = 0 ,
Jl-
r
= 30;
2 01'
LLAST S Q U A R E S U 1 , I J T I O N O F N O N L I K E A R I'ROB1,CMS
F o r e x i s t e n c e anti u r i i q u c i i e s s see, e.g.,
7.3
Lions
I31 ,
(3.I)-( 3 . 3 )
tlieorems concerning
Tctnam [ 11
A l e a s t s q u.a.r e f o r m u l a t i o n
,
L a d y z e n s k a y a [ 11
.
o f~-t h e Y a v i e r - S t o k e s
equatio _ n_s .
It can b c
Miri VEU
(3.4)
where i n
v 5 :6
i
-
I v(y-v) I
dx,
'R
(3.4)
and where
i s a function of
v
through
(3.5)
problcrn i n w h i c h t h e s t a t e e q u a t i o n
W e h a v e o b t a i n e d a c-ontrol
~
~
~~
i s a Stokcs problem.
If t h e o r i g i n a l p r o b l e m
(3.1)-(3.3) has a s o l u t i o n then the
l e a s t s q u a r e p r o b l e m has a s o l u t i o n f o r w h i c h t h e c o s t function takes
the value
0.
Remayk 3 .-1-.: O t h e r f o r m u l a t i o n s c a n b e u s e d ; f o r i n s t a n c e we can r e q u i r e f o r
etc..
.
v
the condition
v-v
= 0,
o r t a k e as c o s t
R. GLOWINSKI and 0. PIRONNEAU
2 10
3.4
Calculation o f t h e gradient of the cost function. ___ - ._ ~. ~
~
~~
In practical applications, an essential step, in view of using descent or conjugate gradient methods, is the calculation o f the d e r i v a t i v e , o f the cost function. Let
J
J(v) = ; 2
where
r,
the functional defined by
y
is a function of
v
lv(y-v)l
via ( 3 . 5 ) .
2
dx,
Then we f o r m a l l y
have =
V(y-v).V&(y-v)dx
(3.6) = -V
[
v(y-v)*06v
dx +
V
i
~ ( y - v *V6ydx )
W~V€(H’,(~))~
R
where
6y
is solution o f
I
+
-VA6y
v6p = - ( & v . v ) v
-
(v-V)bv,
At that stage it is convenient to introduce the following S t o k e s equation (which is the adjoint state equation i n o u r
problem)
-van + m (3.8)
y,v E
(H;(anN
on
a,
= 0,
nlr
van = If
= -v~(y-v)
0.
(H1(n))N,
then (3.8) has a unique solution in
x (L2(W/R).
Taking variational formulations of ( 3 . 7 ) , (3.8) we have ~
LEAST SQUARE
where
Wo
Taking
z
SOLUTTON
= { z
E
= 7
i n ( 3 . 9 ) and
I n t e g r a t i n g by- p a r t hand s i d e o f
7 - 2 = 03
(Hi(n))N,
(i.e.
z
= 6y
211
L I N E A R PROBLEMS
O F NON
*
i n (3.10)
we o b t a i n
using Green's formula) the right-
( 3 . 1 1 ) c a n a l s o be w r i t t e n
I t f o l l o w s t h e n from ( 3 . 6 ) , ( 3 . 1 1 ) , ( 3 . 1 2 ) t h a t w 6 v E (H:(R))~
which c a n also be w r i t t e i i
(3.14) (
where and
(J'(V),~V)=(-VA(Y-V)
- ,*)
(H:(n))N.
(3.15)
3.5
+
(V*V)n
+
(V.V)q
-
(V~)q,q)
d e n o t e s t h e d u a l i t y p a i r i n g between
(H-l(n))N
Therefore
J' (V) =
-vA(V-y)
+
(V*V)
n
-k
(v'V:Y
-
(vV)n
-
Numerical e x p e r i m e n t s .
I n p r a c t i c e we have a p p r o x i m a t e d t h e N a v i e r - S t o k e s by a mixed
f i n i t e e l e m e n t method.
equations
Then u s i n g a d i s c r e t e
212
R.
v a r i a n t of
GLOWINSKI and 0 . P I R O N N E A U
the l e a s t square formulation
( 3 . ) + ) , ( 3 . 5 ) we have
minimized t h e c o s t f u n c t i o n by a c o n j u g a t e g r a d i e n t a l g o r i t h m _ _ ~ . ~ . _ ( o f R i b i & r e - P o l a k t s t y p e ) . I n t h e conjugate g r a d i e n t algorithm 1 N
we used as a s c a l a r P r o d u c t a d i s c r e t e f o r m of t h e ( ~ ~ ( 0 ) ) s c a l a r product. The n u m e r i c a l t e s t p r e s e n t e d h e r e i s t h e t w o - d i m e n s i o n a l , time d e p e n d e n t f l o w p a s t a d i s k , o f a v i s c o u s i n c o m p r e s s i b l e f l u i d a t a Reynolds number o f 2 0 0 . f o r t h e time d i s c r e t i z a t i o n ,
U s i n g a n i_ m - p l i c i t scheme .__ ~
t h e numerical i n t e g r a t i o n i s
r e d u c e d t o a s e q u e n c e of s t a t i o n a r y a p p l y t h e t e c h n i q u e s . d e s c r i b e d above.
problems
t o which we
T h e r e f o r e a t each time
s t e p we h a v e t o s o l v e a l e a s t s q u a r e p r o b l e m ; however s i n c e
t o compute t h e f l o w
at
(n+l)At
we s t a r t o u r i t e r a t i v e
procedure u s i n g t h e informations a t of
nAt,
t h e convergence
the conjugate gradient algorithm i s f a i r l y f a s t .
S i n c e t h e domain o f t h e f l o w i s unbounded we have imbedded t h e d i s k i n a much l a r g e r d i s k on which a u n i f o r m boundary velocity i s prescribed. l y perturbated Fig.
3.1
S t a r t i n g from a u n i f o r m f l o w s l i g h t -
t o 3.4
show t h e e v o l u t i o n o f t h e f l o w
d u r i n g some t i m e i n t e r v a l , and p a r t i c u l a r l y t h e b e h a v i o u r of t h e f l o w behind t h e o b s t a c l e .
LEAST SQUARE
SOLUTION O F NOK L I N E A R PROBLEMS
F i g . 3.1
CYCLE DE TEMPS REYNOLDS
10
200.
214
R.
G L O W I N S K I and 0. P I R O N N E A U
Fig. 3 . 2
CYCLE DE TEMPS REY NClLDS
20
200.
LEAST S Q U A R E S O L U T I O N O F NON L I N E A R I’ROULEIIS
Fig. 3 . 3
CYCLE DE TEMPS REYNOLDS
40 200.
216
R.
G L O W I N S K I and 0 .
PIRONhTAU
Fig. 3 . 4
CYCLE DE TEMPS REYNOLDS
GO 200.
LEAST SQUARE S O L U T I O N OF NON L I N E A R PRODLEMS
4.
On c o n j u g a t e g r a d i e n t m e t h o d s .
An e s s e n t i a l t o o l i n t h e s o l u t i o n o f non l i n e a r p r o b l e m s by l e a s t s q u a r e methods i s t h e c_o_ n j-u.g_ a t. e_ - g r a d i e n t method t o b e d e s c r i b e d below:
N
be a s u f f i c i e n t l y smooth f u n c t i o n ,
Let
f: R
R
vf
af be {a fK ... , ,axN]
4
let
i t s g r a d i e n t , and l e t S b e a N x N 1 s y m m e t r i c , p o s i t i v e- d e_ f-i n i t e m a t r i x . Then t o s o l v e t h e m i n i =
~
~
r n i z a t i o n problem
I
x
Find
E RN
such t h a t
(4.1)
we c a n u s e t h e t w o f o l l o w i n g c o n j u g a t e g r a d i e n t a l g o r i t h m s t h e c o n v e r g e n c e o f which is s t u d i e d i n P o l a k [I]
(forS=I).
F i r s t method ( F l e t c h e r - R e e v e s ) :
(4.2)
xo
(4.3)
go =
s-l
(4.4)
w
-
-
0
E I R ~ given, Vf("),
0
= g .
Then a s s u m i n g t h a t
-x n
wn
-
and
a r e known we compute
x n+1
-
by
(4.5)
X n + l= I
where problem
p,
-n
n -p,y9
i s t h e s o l u t i o n f o r t h e one-dimensional minimization
R.
218
GLOWINSKI a n d 0 . PTRONNEAU
Then w e compute
-gn+'
(4.7)
gn+l =
(4.8)
wn+1 = -gn+1 + x n w n
wn+ 1
and
s-1
by
Vf(?n+l),
where
(4.9) i n (4.9),
denotes t h e u s u a l inner-product
( a , . )
Second method
of
N
R
.
(Polak-Ribisre).
T h i s method i s l i k e t h e p r e v i o u s o n e e x c e p t t h a t
(4.9) i s
r e p l a c e d by
-
( sgn+l, _g"+l-gn) (4.10)
Remark 4.1:
If
E RN
where
and where
A
is a
NxN
symmetric p o s i t i v e
d e f i n i t e m a t r i x , t h e n t h e two a b o v e a l g o r i t h m s a r e i n f a c t
similar,
s i n c e i t c a n be proved t h a t i n t h i s c a s e
-
( S g n + l , g") Remark 4 . 2 :
A t e a c h i t e r a t i o n of
= 0.
t h e two a b o v e a l o r i t h m s w e
have t o s o l v e a one d i m e n s i o n a l problem t o o b t a i n
pn.
For
t h e s o l u t i o n o f t h i s k i n d o f one d i m e n s i o n a l problem s e e , e.g.,
P o l a k [: 1 3
Remark 4.3: round-off
,
Brent [
13 .
The above a l g o r i t h m s a r e f a i r l y s e n s i t i v e t o
e r r o r s ; h e n c e d o u b l e p r e c i s io n may be r e q u i r e d o n
some c o m p u t e r s .
M o r e o v e r i t may b e c o n v e n i e n t t o r e s t a r t
LEAST SQUARE SOLUTION OF MON LINEAR PROBLEMS
periodically with
wn =
-g
n
.
219
Actually Powell [l] has recently
introduced a new type of restarting procedure which increases the efficiency of the algorithm, but which requires more storage.
The Powell's analysis shows also that the Polak-
RibiBre's algorithm may b e more convenient than the FletcherReeves' algorithm for some problems.
In fact f o r the type of
problems we are solving a similar observation has been done, from our computational experiments. Remark 4.4: F o r the problems described i n this paper a convenient choice f o r the matrix form of
S
is to take a discrete
-A.
5 . Conclusions, The methods presented in this paper may appear as very heuristical, however when convexity o r monotonicity are lacking they seem one o f the few possibilities for computing the solutions o f some difficult non linear problems, taking into account their special structure. Applications to non linear problems i n elasticity are to be done. Acknowledgments: Many thanks are due to Mrs. Bri,steau and MM. Perrier, Periaux, Poirier for their help at various stages of this report.
R. GLOWINSKI and 0. PIRONNEAU
220
Bibliography Axelsson, 0.
-
[l]
A class of iterative methods for finite
element equations. Ccmp.
Math. . Applied _ ~ _ .. Mech. Eng.,
Bauer, F., Garabedian, P., Korn, D.
~~
-
[l]
Supercritical _____ wing
sections, . -~ Lecture Notes in Economics and Math.
.
Systems, Vol. 66, Springer-Verlag, 1972. Bauer, F., Garabedian, P., Korn, D., Jameson, A. [l]
Supercritical wing sections (11), Lecture Notes
in Economics and Math. Systems, Vol. 108, SpringerVerlag, 1975. Begis, D., Glowinski, R.-[l] 616ments finis
Application de la m6thode des la r6solution d'un problbme de
domaine optimal. Vol. 2, NO 2, Brent, R.
-
Applied Math. and Optimization,
(1975), pp. 130-169.
[l] Algorithms for minimization without derivat~
~~~
~
ives, __ Prentice Hall, 1973. Cea, J., Geymonat, G.
-
[l] Une methode de lin6arisation via
l'optimisation.
Instituto d i Alta Mat. Symp. Mat.,
Vol. X, (1972), Bologna, pp. Ciarlet, P.G., Raviart, P.A.
-
431-451.
111 Interpolation theory over
curved element with applications to finite elements methods. Comp. Meth. Applied Mech. Eng., Vol. 1, (1972) , PP. 217-249.
[2] A mixed finite element method for the biharmonic equation, i n Mathematical aspects of the
LEAST SQUARE SOLUTION OF NON LINEAR PROBLEMS
finite elements in partial di-fferen&ial C. de B o o r Ed., Acad. Press, Concus, P., Golub, G.H.
-
221
equations,
1974, pp. 125-145.
[l] A generalized conjugate gradient
method for non symmetric systems of linear equations, in Computing Methods in Applied Sciences and Engineering, R. Glowinski and J.L. Lions Ed., -~ Lecturca Notes i n Ecsiiomics and Math. Systems,
VOI. 134, Springer-Verlag, 1976, pp. 56-65. Glowinski, R.
-
[ 13 Numerical methods f o r non linear
variatipnal problems, Lecture Notes at the Tata Institute, Bombay, to appear in 1977-1978. [2] Approximations externes par 616ments finis
d'ordre un et deux d u probl&me de Dirichlet pour
A 2 , in Topics i n Numerical Analysis , J. J.H. Miller ed., Acad. Press, (1973)) pp. 123-171. Glowinski, R., Marroco, A .
-
[l] Sur lfapproximation par
616ments finis dfordre un, et la r6solution par p6nalisation-dualit6 d'une classe de probl;?mes de Dirichlet non linGaires, Revue Franqaise d'dutoma___ ____ tique, d'Informatique et de Recherche Opgr-ationnelle, e
9
ann&e, aojt 1975, R-2, pp.
41-76.
Glowinski, R., Periaux, J., Pironneau, 0.
-
[l] Transonic
flow simulation by the finite element method via optimal control.
Proceedings of the Second Inter-
national Symposium on Finite Elements i n Flow Problems, Santa Margherita, Italy, June 1976,
PP. 249-259.
R. GLOWINSKI and 0 . PIRONNEAU
222
[2] Use o f optimal control theory for the numerical
simulation of transonic flow by the method of finite lements.
Proceedings o f the
5th Conference _____
on Numerical Methods i n Fluid Dynamics.
Enschede,
Netherlands, July 1976. Glowinski, R., Pironneau, 0.
-
[l] On the computation of
transonic flows, Proceed+-ngs o f the first FrancoJ_po&Ese
Co-lloquium on Functional and Numerical
__ Analysis, Tokyo, Kyoto, September 1976. [2] Calcul d'ecoulements transoniques par des
methodes dlel6ments finis et de controle optimal, in Computing Methods in Applied Sciences and E n- -g i n e e r i x ,
R. Glowinski and J.L.
Lions Ed.,
Lectures Notes i n Economics and Math. Systems, vol. 134, Springer-Verlag,
1976, pp. 276-296.
[ 3 ] Numerical methods f o r the first biharmonic equation and for the two-dimensional- Stokes problem, C.S. Report, University of Stanford, Computer Science Dpt., Hewitt, B.L.;
1977.
Illingworth, C.R.;
Lock, R.C.;
Mc Donnel, J.H.; Richards, C.; Walkden, F . ;
Mangler, K.W.; Ed.
[l] Computational method2 and problems i n aeronau-
tical fluid Acad. Press, 1976. ~ - dynamics, -___-Jameson, A.
- [l]
Transonic flow calculations, in Proceedings
of Conference o n Computational Fluid Dynamics, - Von .
Karman Institute, March 1976, Brussels (Belgium). [2] Three dimensional flows around airfoils with
LEAST SQUARE SOLUTION OF NON LINEAR PROBLEMS
223
shoks , in Computing Methods _in ApplA-e-d-Sci-e-Eces_and Engineering, Part 2 , R . -~
Glowinski and J.L. Lions
_ I
Ed., Lecture Notes in Comp. Sciences, V o l . Springer-Verlag,
131
11,
(1974), pp. 185-212.
Iterative solution of transonic flows over
airfoils and wings, including flows at Mach 1,
Comm. Pure Appl. Math., ~ _ _Vol. 27, (1974), pp.283-309.
[4] Numerical solution of non linear partial differential equations of mixed type, in Numerical Solution of Equations, 111, Synspade - Partial Differential - __. ~~~
~~
1975, B. ___
~
Hubbard Ed., Acad. Press, (1976),
PP. 275-320. Landau, L.; Lifchitz, F.
-
[l] M6canique des - Fluides, Editions ~
MIR, MOSCOU, 1953. Ladyzenskaka, O . A .
-
[l] The ___. mathematical theory of viscous
incompressibA-e fluids, Gordon Breach, New York, 1963. Lions, J.L.
-
[l] Gontrole optimal des _ systemes _ ~ gouvern6s par ~
des 6quations aux d6rivGes partielles.
Dunod, 1968.
[ 2 ] Some aspects of the optimal control of
distributed systems, Regional Conference Series in Applied Math., Siam Publication NO 6, 1972.
[ 3 ] Quelques m6thodes de rdsolution des problhmes aux limites non lineaires, Dunod, 1969. Morawetz, C.S.
-
[l] O n the non existence of continuous
transonic flows past profiles, C o m m . Pure Applied Math., V o l .
9, 1956, pp. 45-68.
224
R. GLOWINSKI and 0. PIRONNEAU
Murman, E.M.;
Cole, J . D .
-
[11 Calculation of plane steady
transonic flows, AIAA Journal, V o l .
9, (1971),
pp. 114-121. Polak, E.
-
[l]
Computational methods in optimization, Acad.
Press, 1971. Powell, M . J . D .
-
[l] Restart procedures f o r the conjugate
gradient method, Math. Programming 12,
(1977),
pp. 241-254. Strang, G.; Fix, G.
-
[l] &n--analysis of the finite element -
method, Prentice Hall, 1973. Temam, R.
-
[l] Navier-Stokes Equations, North-Holland,
1977 Young, D . M .
- I: 11
Iterative s-ol~~ix-of__large linear systems,
Acad. Press, 1971.
G.M. de La Penha, L.A. Medeiros ( e d s . ) Contemporary Developments i n Continuum Mechanics and P a r t i a l D i f f e r e n t i a l Equations @North-Holland P u b l i s h i n g Company (1978)
E L A S T I C BODIES I N A SIGNORTNI-TYPE ENVIRONMENT
P.
POD10 G U I D U G L I
%
Dipartimento d i Costruzioni Facol
ta
d i Ingegneria
U n i v e r s i t ' a d i Aiicona,
Noll's formulation of
Italy
t h e t r a c t i o n boundary-value
p r o b l e m o f e l a s t o s t a t i c s i s s p e c i a l i ~ e dt h r o u g h a s i m p l e
I t s dependence on a para-
c h o i c e of
environmental operator.
meter
l e a d s t o a r e p r e s e n t a t i o n o f s o l u t i o n s a s power
C
series i n G
E ;
t h e c o e f f i c i e n t s of
t h e s u c c e s s i v e powers
a r e s o l u t i o n s o f c e r t a i n Linear d i f f e r e n t i a l problems,
of
each
of w h i c h l e a v e s a c e r t a i n i n d e t e r m i n a c y i n t h e s o l u t i o n w h i c h
i s g e n e r a l l y removed by c o m p a t i b i l i t y c o n d i t i o n s d e r i v e d from
the Fredholm a l t e r n a t i v e .
When t h e body i s i n i t i a l l y a t e a s e ,
these conditions collapse into the l i n e a r conditions of Signorini,and generally determine the global r o t a t i o n .
Introduction.
In
1971, d u r i n g
Stability i n Elasticity",
a Symposium on " E x i s t e n c e and
W.
w h i c h he s k e t c h e d a method o f
N o l L delivered a lecture i n formulating t h e boundary-value
problem o f f i n i t e e l a s t o s t a t i c s . l e c t u r e remained u n p u b l i s h e d ,
Though t h e t e x t of t h i s
l a t e r P r o f e s s o r N o 1 1 was k i n d
enough t o p r o v i d e m e w i t h a copy of h i s m a n u s c r i p t .
The
q u e s t i o n s o f e l a s t i c i t y I i n t e n d t o d i s c u s s h e r e a r e most
*
Presently at Istituto di Scienza delle Costruzioni, Facolth di Ingegneria, Universita di Pisa, Italy.
226
POD10 GUIDUGLI
P.
conveniently dealt with within N o l l ’ s framework:
therefore,
I will adopt that framework, in a somewhat modified f o r m , to serve my purpose.
I identify a n elastic body
03
with the following
collection o f objects: (i)
B
A regular, bounded region
3B
has boundary
B
E ;
of Euclidean space
with outer unit n o r m a l n.
Whenever
convenient, I write vvmeas’v for the pair of the usual volume and surface measures on
B:
meas = (vol, surf). (ii)
b
A class
B,
o f deformations of
i.e., of inver-
B
tible and locally orientation-preserving mappings of open subsets of tries between
E ;
B
@
is assumed to contain all the isorne-
and open subsets of
b.
Given any
y E
g,
E ,
s o that, in
v x E B,
i(x) = x
particular, the identity mapping: to
onto
below
the deformation gradient
F(x)
def = VYI,
(A-1) x;
is the basic measure of deformation at
which are invertible second-order tensors. u(x) = y(x)
x
is the displacement vector at H(x)
has values
F o r each
x E B, (A-2)
x
in the deformation
y,
def = vulx
is the displacement gradient; F(x) =
-
F
clearly, for each
1 + H(x),
and
(-4-3) x
E B (A-4)
where 1 = Vi.
(A-5)
E L A S T I C B O D I E S I N A SIGNORINI-TYPE E N V I R O N M E N T
(iii) each
A n elastic response function
x E R
227
such that, f o r
5,
y E l ~ ,
and f o r each
( A- 6 ) x
i s the measure of the stress at the deformation
y;
when
R
has undergone
in particular,
is the ground stress at
F o r convenience, I think of S(x)
x.
as the first Piola-Kirchhoff stress tensor at
X.
The
prescription (6) satisfies the axioms of material frameindifference and balance of angular momentum if one assumes that T
5(F) = F & ( F F ) with
fJ
( A- 8 )
I
a mapping of the space of symmetric p o s i t i v e - d e f i n i t e
tensors into t h e space of symmetric t e n s o r s . A system of loads for
S
is any pair
4, = ( b , s )
such that:
b,
the body force, is a volume-integrable vector
field over
B;
s , the ___ surface _ traction, _ _ i s~ a surface-
integrable vector field over fB Ld(meas)
def =
aB;
1
Ld(meas) = 0, where
IB
( n b d(vol)
+
s d(surf).
The collection of
all such systems of loads is the load space, denoted b y Once a n elastic b o d y
fi
and a load space
6:
2.
are
given, I introduce the elastic response operator
.e:
l$
+
x,
Z(Y> =
(&),:(u)),
where
(A-9)
P.
228
POD10 GUIDUGLI
( A - 10) i(y)
i s t h e s y s t e m of
location
A
1.oads which hol-ds
i n t h e deformed
Y(B). F u r t h e r , i n t h e manner of N o l l , I i n t r o d u c e t h e
environmental r e a c t i o n o p e r a t o r
i(y)
63
i s t h e system o f
Z:
Q
+ C,
X(y) = (i;(y) , ~ ( y ) ) ;
l o a d s t h e e n v i r o n m e n t would e x e r t on
i f i t o c c u p i e d t h e deformed l o c a t i o n
y(B).
A typical
d i f f i c u l t y i n f i n i t e e l a s t i c i t y i s t o model p h y s i c a l l y m e a n i n g f u l l o a d i n g d e v i c e s by means of s e n s i b l e a s s i g n e m e n t s o f environmental r e a c t i o n o p e r a t o r s ( v i d . e . g . Batra (1972); Ericksen d e s c r i b e s a s y s t e m of
(1967);
Sewell
( 1 9 7 3 ) ) . The s i m p l e s t c a s e i s when dead l o a d s :
U s i n g t h e above m a c h i n e r y ,
t h e t r a c t i o n boundary-
v a l u e problem o f e l a s t o s t a t i c s i s f o r m u l a t e d a s f o l l o w s : A
Given a n e l a s t i c body
63,
with e l a s t i c response o p e r a t o r
f,,
which i s p l a c e d i n t o a n e n v i r o n m e n t d e s c r i b e d by t h e e n v i r o n -
-
mental r e a c t i o n o p e r a t o r
f,,
t(Y) I call (12).
(y,x(y))
find
y
E &
such t h a t :
= Z(Y).
(A-12)
a n e q u i l i b r i u m p a i r when
y
For an e q u i l i b r i u m p a i r , an easy consequence of
solves
(8)
and t h e d i v e r g e n c e theorem i s t h a t :
\
(y(x)-q) A t ( y ) l ,
d(me-1
= 0
(A-13)
B (here o f
q
i s any f i x e d p o i n t of
t h e m u t u a l c o n s i s t e n c y of
e);
( 1 3 ) i s an e x p r e s s i o n
a n e q u i l i b r i u m p a i r and i s
ELASTIC BODIES I N A SIGNORINI-TYPE E N V I R O N M E N T
i n t e r p r e t e d a s t h e requirement t h a t t h e system o f has vani shi ng t o t a l torque i n
229
loads
-
L
y(B).
S i g n o r i n i - t y p e environment.
The f o r m a l s e t t i n g d e v i s e d by Nol.1
( 1 9 7 1 ) a l l o w s one
t o d e a l i n a n a t u r a l way w i t h t h e f o l l o w i n g q u e s t i o n : When
63
i s a t r e s t i n t o a g i v e n e n v i r o n m e n t , what happens i f
t h e e n v i r o n m e n t i s changed? P e r s e , s u c h a q u e s t i o n l e a d s one t o c o n s i d e r p e r t u r b a t i o n techniques. non-trivial
I n t h i s s e c t i o n I o f f e r an answer f o r a classical c a s e f i r s t s t u d i e d by A .
Signorini i n the thirties;
more d e t a i l s w i l l b e found i n C a p r i z & Podio G u i d u g l i
(1977).
(i,% ( i ) )be a n e q u i l i b r i u m p a i r for a n e l a s t i c
Let
n
63
body
c h a r a c t e r i z e d by t h e e l a s t i c r e s p o n s e o p e r a t o r
.t(i) = l o ( i ) .
t,: (B-1)
A Signo r i n i - t y p e e n v i r o n m e n t i s d e s c r i b e d by a r a y -____ in
2 ,
with o r i g i n a t G - %
E (y)
zo,
o f s y s t e m s of dead l o a d s :
Eo(i) +
=
c
l,(i)
++
Y E 14.
(B-2)
One t h e n s e e k s a r e g u l a r F - f a m i l y of d e f o r m a t i o n s , i.e.,
a path i n
I$ E-Y
originating a t
yo = i
(B-3)
E
and s u c h t h a t i n a n e i g h b o r h o o d of
e q u i l i b r i u m p a i r for e a c h
t(Y& = tE(YE)' I n the s p i r i t of
0:
(B-4)
t h e s t a n d a r d methods o f a n a l y t i c n
p e r t u r b a t i o n , I assume t h a t
L
i s analytic a t
yo:
P.
230
with
v(n)i
Iyo
POD10 GUIDUGLI
a n-linear
c omple te ly symmetric transformati on.
Further, I seek a solution of type
(3) of equation (4) in the
form o f a formal power ser es in
:
E
-
-
Thus, I replace equation (4) by its Oth-order
“approximation” (l), plus a n infinite hierarchy of linear equations f o r the successive approximations solution
un
of
yE:
-
L u1 = C1,
with
L
def =
oil
and
t,
= t,(yo),
YO
L un =
the
xn,
n=2,3 ,..., with
-
cn
def n- 1 = -Qn(Yo;Cukll ).
(B-9)
If the system (9) i s solvable, the system of (1) and
(9) may be regarded as a reasonable model of equation (4)+ , o r rather of (A-12). -k
See next page.
The next Section will b e devoted to
231
ELASTIC BODIES I N A S I G N O R I N I - T Y P E E N V I R O N M E N T
s o l v a b i l i t y f o r t h e s y s t e m (9).
e s t a b l i s h c e r t a i n c o n d i t i o n s of
Fredholm c o n d i t i o n s .
L
Let
b e r e g a r d e d a s a l i n e a r t r a n s f o r m a t i o n from
a Banach s p a c e
X
i t s transpose.
I n t h e c a s e of
i n t o a Banach s p a c e
L
e l a s t o s t a t i c s both ker L
that
2[O]
t h e t r a c t i o n problem of are not injective.
L*
t h e n o n - l i n e a r problem ( A - 1 2 ) ; ker L
t h e s t r u c t u r e of
classification of
i s of
s o l u t i o n of n t h datum in
n-1
t h e nth
ln
thus, a
impnrtance f o r a
equations
(B-9)
when
Accord-
k e r L*$
[ O ] ,
have been s o l v e d , t h e
equation does e x i s t i f
and o n l y i f
the
b e l o n g s t o t h e o r t h o g o n a l complement o f k e r L*
Y*:
(tn,v*) Remark Fichera of
The f a c t
t h e d i v e r s e t y p e s of b i f u r c a t i o n s .
i n g t o a g e n e r a l lemma of a l t e r n a t i v e , a f t e r the first
be
t h e n i n d i c a t e s t h e e x i s t e n c e o f a bifurcation
i n t h e s o l u t i o n of knowledge of
a?d
L*
and l e t
Y,
-
= 0
I n h i s treatment
v*
of
k e r L* + +
.
(c-1)
a b s t r a c t l i n e a r transformations,
(1954) s t a t e s c l e a r l y the primitive algebraic nature
t h e lemma of a l t e r n a t i v e ,
‘Notice
E
that
and a d j u s t s i t s o a s
t o apply i n
( 9 ) 1 coincides formally with t h e equation o f t h e
t h e o r y o f i n f i n i t e s i m a l d e f o r m a t i o n s s u p e r i m p o s e d on l a r g e , but o f course the p o s i t i o n o f ( 9 ) a s t h e l S t - o r d e r approxima t i o n o f a n o n - l i n e a r problem w i t k i n a g i v e n p e r t u r b a t i o n scheme i s n o t a t a l l same.
++ I d e n o t e by
Y*
Y,
the dual of
p a i r i n g between e l e m e n t s of
Y
and by and
Y*.
(
,
)
the
P. PODIO GUIDUGLI
232
in various concrete contexts.
The role of the lemma in the
theory of branching o f solutions of non-linear equations is made clear in the book by Vainberg & Trenogin ( 1 9 7 4 ) through the notion of I1normally solvablef1operators.
L
of normally solvable operators
that
ker L
and
ker L*
A relevant class
is Fredholm's (those such
have finite, and equal, dimension);
it should be possible to find reasonable hypotheses under h
which the elastic response operator
L
approximated by a Fredholm operator
I,.
turns out to be
I call the solvability conditions(1) Fredholm I will provide below a n interpretation, i n terms
conditions.
of elasticity theory, of these conditions, o f their use to remove indeterminacies , and of the cases of "incompatibility", i.e.,
the cases when they fail to be fulfilled for some
n.
The Fredholm conditions will now be given a less general, but more explicit form. Let 0 < dim ker L = dim ker L"
and let ker L
r
{v,]
and
lcer L*.
and
(.:If
(c-2)
+m
be, respectively, a basis of
Then, if
equation ( B - 9 ) ,
(n-l)th
= r <
cn-
is any solution of the
the general solution of the same
equation has the representation
where
i (yn-l]y
numbers, (B-9)
is a list o f , i n principle, arbitrary real
However, the Fredholm condition for the nth equation
restricts the choice of the numbers
(U~];-~ and
n- 1
are given.
i yn,l
once
yo
)
233
ELASTIC BODIES IN A SIGNORINI-TYPE ENVIRONMENT
I introduce the following notation
(c-4) I then write (1) in the case
r
equations of second degree in the r
C,< +
i
Y1:
r
bki
c
unknowns
algebraic
r
as a system o f
n=2
Y; +
Gkij Y i Y p =
1
i= 1
'
0
k=l,2,...,r;
(C-5)
i,j = 1
i n the case
n
2
3
i unknowns Yn-1: r + C 'k,n-l i= 1
as a linear algebraic system in the
r
bki+ z C
Cikij Y j~ )i Y ~ =- 0~
... , r ,
k=1,2,
j=1
r
(C-6)
where the coefficients are defined in the following way:
st
Thus, the 1 response operator
i
-order approximation of the elastic
determines the possibility of solving
i
system (B-9), whereas the 2nd-order approximation of i determines the possibility of finding the y n for any
n.
A simple and yet non-trivial instance is when all the coefficients numbers
Yn-1
aki
vahish.
Then, if
det
are uniquely specified f o r each
as a consequence, series ( B - 6 )
bki n
0, the 2
2,
and,
is completely determined, the
indetermination in the solution of each equation (B-9) notwithstanding. from
(4)-(7)
Another example of the information deducible is obtained when
r=l.
In the instance, systems
P. POD10 GUIDUGLI
234
(5)
and (6) reduce to
c
+XY1
+
a(Y,)2 =
0,
bn-l P
+(J:+2aY1)Yn-1
(c-8)
= 0
and conditions of solvability, as well as incompatibility, for (8) can be read off at once.
Body initially at ease
In the case originally studied by Signorini (vid. e.g, Capriz & Podio Guidugli at ease in
B,
(1974)), (i) the body
@
is
i.e., the ground system of loads is null:
and the ground stress vanishes identically over
(ii)
@
B:
is hyperelastic, i.e., a strain-energy function
u = o(F TF)
exists such that
I put
with
L
a self-adjoint transformation whose kernel, as can
be seen at once, contains isometries
+,
+
3,
+. Further, I put
See next page.
the set of all inPinitesima1
ELASTIC BODIES IN A SIGNORINI-TYPE ENVIRONMENT
235
where
(v(2)sl-c, vVii ) c Ovji
1,r- vVii ) c v V j ~+
= d2)Q
+ (vvi)scvv.l J + (vvj)srvvil + sr (VVj)
T
,
vvil
(D-7
1
so that, i n particular,
Assume now, following Signorini, hypotheses sufficient to guarantee that
ker L = 3.
S
symmetry properties of
and
Then it follows from the that all the coefficients
Q
vanish; thus, although the (n-1) th equation (B-9) %i j determines u ~ to - within ~ an infinitesimal isometry, the situation occurs which I mentioned at the end of the previous section, and the Fredholm conditions of order
n
can be used,
in general, to specify the infinitesimal rotation of Moreover, when
ker L = 3 ,
ordern-1.
Capriz & Podio Guidugli ( 1 9 7 7 )
prove the following characterization of the Fredholm cofitions: The following three statements are equivalent: (i)
xn,
The systems of loads
n=1,2,...,
satisfy the
Fredholm conditionsfor a l l infinitesimal r o t a t i o n s . (ii) The systems of loads
Ln -
have vanishing total torque
in B: (x-q)
A
znlx d
(meas) = 0 ,
(iii) The system of loads
ic
n=1,2,...
.
(D-9)
has vanishing total torque in
Ye ( B ) : (D-10) X
'Here
sym is the symmetrizer o f second-order tensors.
An infinitesimal isometry is a displacement vector field which has the form w(x)
= w(q)
+ w(x-q),
with W a skew-symmetric second-order tensor, independent of X.
236
P.
POD10 GUIDUGLI
Aknow-1 e d gme n t Williams for h e l p i n g m e w i t h
I t h a n k P r o f e s s o r W.0.
t h e E n g l i s h w o r d i n g a n d p r o v i d i n g some u s e f u l comments.
References B a t r a , R.C.
(1972).
Capriz, G.,
and P o d i o G u i d u g l i , P.
Mech. C a p r i z , G.
F i c h e r a , G.
57,
Anal.
(1973).
(1974).
Anal.
(1954).
(1971).
4 8 , 163.
Arch.
Rational
1.
(1977). I n p r e p a r a t i o n .
I n N o n l i n e a r E l a s t i c i t y , R.W.
(Academic P r e s s , New York-London),
( L i b r e r i a V. N o l l , T$.
R a t i o n a l Mech.
and P o d i o G u i d u g l i , P .
E r i c k s e n , J.L. Ed.
Arch.
Dickey
161.
Lezioni s u l l e trasformazioni
lineari.
V e s c h i , Roma).
What i s a “ w e l l - p o s e d “ p r o b l e m i n f i n i t e
elasticity?
Italian
-
American Symposium on
“ E x i s t e n c e and S t a b i l i t y i n E l a s t i c i t y 1 ’ , U d i n e , I t a l y , J u n e 18-23. Sewell, M.J.
(1967).
V a i n b e r g , M.M.,
Arch.
R a t i o n a l Mech. A n a l .
a n d T r e n o g i n , V.A.
(1974).
T h e o r y of Branch-
i n g o f S o l u t i o n s of N o n - l i n e a r E q u a t i o n s . (Noordhoff, Leyden).
3 , 327.
G.M. de La Penha. L.A. Medeiros (eds.) Contenporary Developments i n Continuum Mechanics and Partial Differential Equations @North-Holland Publishing Conpany (1978)
ON THE NONLINEAR THEORY OF ELASTICITY
MORTON E. GURTIN Department of Mathematics Carnegie-Mellon University Pittsburgh, PA 15213
1. Introduction.
In these lecture notes we give a brief introduction to the nonlinear theory of elasticity.
We present a concise
derivation of the basic equations followed by a detailed discussion of the underlying boundary-value problems.
Our
discussion demonstrates why this theory is far more difficult than most nonlinear theories of mathematical physics.
It is
hoped that these notes will convince analysts that nonlinear elasticity is a fertile field i n which to work. The most important general references to the subject are :
[11
Truesdell, C. and W. Noll,
The non-linear field theories
Handbuch der Physik. III/3.
of mechanics.
Berlin:
Springer-Verlag (1965). [ 2 ] Wane;, C.-C. and C. Truesdell,
Elasticity.
[3]
Leyden: Noordhoff
Antman, S.S.,
Introduction to Rational
(1973).
Ordinary differential equations of non-linear
elasticity. Arch. Rational Mech. Anal.
[4] Ball, J . M . ,
a,307-393 (1976).
Convexity conditions and existence theorems
in non-linear elasticity.
337-403 (1977)
Arch. Rational Mech. Anal. 63,
MORTON E .
238
GURTIN
I have a l s o p r o f i t e d f r o m :
151
Antman, S . S . ,
Fundamental m a t h e m a t i c a l problems i n t h e
theory o f nonlinear e l a s t i c i t y . P a r t i a l D i f f e r e n t i a l Equations, Press
Numerical S o l u t i o n of 111.
New York:
Academic
(1976).
I n t h e t e x t when I r e f e r t o a t h e o r e m , c o u n t e r e x a m p l e , e t c . by name b u t o m i t t h e e x a c t r e f e r e n c e , t h e l a t t e r can b e found in [
2.
11
or [ 2 3 .
Notation.
W e u s e l'ower c a s e Greek l e t t e r s f o r s c a l a r s , l o w e r case Latin l e t t e r s f o r vectors
(elements o f
~ 3 1 ,u p p e r c a s e
Latin l e t t e r s f o r tensors ( l i n e a r transformations o f R3).
R'
into
Also:
C a I. b i '
a.b =
A * B = C A . .B IJ
ij'
AT
i s t h e transpose o f
A-T
= (AT)-17
A,
I = the identity tensor, a 8 b,
t h e t-. e n so r product o f
-
and
b,
is the tensor
a . b ., 1 3
w i t h components a A b = a 4 b
a
b 8 a,
Lin = t h e space o f a l l t e n s o r s , Lin'
= { F E L i n : d e t F > 01
Sym = [ F
t L i n : F = FT1
,
Orth+ = t h e space o f a l l proper orthogonal t e n s o r s . F i n a l l y , f o r a smooth ( i . e . , on a r e g i o n i n
R3,
div T
class
C')
tensor f i e l d
xcT(x)
i s t h e v e c t o r f i e l d w i t h canponents
239
ON THE NONLINEAR TIBORY O F E L A S T I C I T Y
aT..
(div T). = C
.
3 ax. J
3. ___Kinematics. 63
A body
p E 8
Points
is a compact, regular region in
are called material points. __ _____
R3.
A deformation of -.
~
n
fl
is a smooth one-to-one map
A motion is a
t E iR.
deformation for each x = r(p t) = rt(p)
at
= r
WJ
4
r: 9xIR + IR3
map
C2
r: 63
with
with
det
vr >
0.
rt = r(-,t) a
We use the following terminology:
is the place-occupied by
p
at __ time t;
t B) i
5 = {(x t):
x E R,,
t E W]
t)
is the trajectory;
is the veloc-iiy; is the spatial description of
p=ril (x) the velocitv: F(p,t) = vr(p,t)
is the deformation gradient.
We also use the notation summarized i n the following table:
63xR
domain gradient w.r.t. argument
first
divergence w.r.t. argument
first
derivative w.r.t. argumen t
second
V@ Div @
5 grad div
A
A
A‘
I t is also convenient to define the material time derivative
2 40
MORTON E .
12
GURTIN
A:
of a s p a t i a l f i e l d
Then
= v’
4.
+
(grad v)v.
Mass. A m a s s distribution f o r
t h e Bore1 s u b s e t s o f
8 )
B
i s a measure
m
(on
o f t h e form
i
po
with
po
a smooth, s t r i c t l y - p o s i t i v e
i s called the reference density;
C16 = n g ( p )
where
s c a l a r f i e l d on
R.
clearly,
i s the b a l l w i t h radius
6
centered at p,
n
V
and p
i s t h e Lebesgue volume measure on
i n a motion
where
r
x = r(p,t)
R’.
The d e n s i t y
i s t h e s p a t i a l f i e l d d e f i n e d by
R6
and
i s as a b o v e .
O f course,
conservation o f m a s s i s a t a c i t p a r t of t h i s d e f i n i t i o n . Proposition
-
(a)
p(x,t) det F(p,t) = po(p),
(b)
p’
+
x = rt(p);
d i v ( p v ) = 0.
5 . Stress. A s y s t e m of of:
f o r c e s for
8
i n a motion
r
consists
241
ON THE NONLINEAR THEORY OF ELASTICITY
(i) XI-
surface forces
s(n,x,t)
(ii)
x 3
s: {unit vectors)
with
-+ lR3
smooth;
9ody forces -
b: 3 + EL3
with
x-b(x,t)
continuous.
These fields are assumed to be consistent with balance of momentum which asserts that -___
f o r every part
with
nx
P
T,
63
and time
t.
the outward unit normal to
-
Cauchy's Theorem field
of
Here
apt
P t = rt(P) and
at
x.
There exists a symmetric spatial tensor
called the Cauchy stress, such that s(n,x,t) = ~ ( x , t ) n
and div T
+
b = p;
6. Constitutive _- Equation.
(equation of motion).
Change i n Observer.
Finite elasticity is based on the constitutive equation A
T = T(F) giving the stress at a material point gradient at
p
is known.
Here
p
when the deformation
MORTON E . GURTIN
242
is a smooth function; for convenience we have assumed that the material is homogeneous so that the response function . independent of the material point (r,T)
A pair
with
r
,.
T
is
p.
T = ?(vr)
a motion and
is
called an admissible process. A change in o b s e r v e r is a s m o o t h m a p
+ Q(t)x,t)
(x,t)l-(a(t) with
a(t) E R 3
and
Q(t) E Orth+
(R
4
-+
t E W.
for each
Axiom (invariance under a change in observer)
(r*,T*)
-
Given any
(r,T) and any change in frame (C), the
admissible process process
(c)
R4)
defined by r*(P,t) = a(t) + Q(t)r(P,t),
= Q(t)T(x,t)Q(t)T,
T*(x*,t)
x* = r*(p,t),
x = r(p,t)
is also admissible. Proposition
-
For every
F E Lin’
and
Q E Orth’,
Q+(F)Q~ = +(QF). Remark
-
(1)
It is not difficult to show that (I) holds if and
T = $(F)
only if the constitutive relation
T = F?(C)F T,
c
is of the form
= FTF.
This type of ”reduced constitutive relation”, while popular in the literature, seems to be of little use in studies of existence and uniqueness.
ON THE N O N L I m A H TIEORY O F E L A S T I C I T Y
7 . The P a r t i a l D i f f e r e n t i a l E q u a t i o n .
The c o n s t i t u t i v e r e l a t i o n and t h e e q u a t i o n o f motion combine t o g i v e t h e e q u a t i o n
+
d i v ?(Vr)
b = p;,
where
.
p det F = p
It i s important t o n o t e t h a t
div,
i s with r e s p e c t t o t h e p l a c e
x
t h e s p a t i a l divergence,
at.
in
Generally,
problems i n v o l v i n g s o l i d s 1 t h e deformed r e g i o n
Et
in i s not
known i n a d v a n c e , and for t h a t r e a s o n i t i s u s u a l l y more convenient
t o use t h e m a t e r i a l p o i n t
variable.
O f course,
one c a n c o n v e r t
p
a s independent div
t o an o p e r a t o r
i n v o l v i n g only d i f f e r e n t i a t i o n with r e s p e c t t o
p,
b u t then
t h e u n d e r l y i n g e q u a t i o n w i l l no l o n g e r b e i n d i v e r g e n c e form. F o r t h e above r e a s o n we i n t r o d u c e t h e Piola-Kirchhoff
stress S = ( d e t F)TFmT (considered a s a material f i e l d ) .
where
ap,
m
and
n
Then g i v e n any p a r t
a r e t h e outward u n i t normals t o
respectively, s o that
Sn
apt
63,
and
r e p r e s e n t s the surface f o r c e
measured p e r u n i t a r e a i n t h e u n d e f o r m e d
c o n f i g u r a t i o n . This
o b s e r v a t i o n , b a l a n c e of l i n e a r momentum, and t h e symmetry o f T
y i e l d the following A s opposed t o f l u f d s , where @ i s g e n e r a l l y known a p r i o r i , and where t h e s p a t i a l d e s c r i p t i o n i s u s u a l l y t h e m o s t appropriate.
244
MORTON E.
GURTIN
Proposition Div S
+
bo =
T
SFT = FS
where
bo = (det F ) b .
Remark
-
..
par,
,
Note that the above equations involve only material
fields and only material differentiation. I n view of the definition o f
relation for
S
T = +(F)
S,
the constitutive
is equivalent to a constitutive equation
of the f o r m S
= :(F)
with
5: and (I) and the relation
.S:
restrictions on
L i n + + Lin; SFT = FST
yield the following
Henceforth we neglect body forces and write
.
S;
S
for -
the underlying partial differential equation then takes
the form Div S(vr) = p 0 ? . Remark -
-
This is a partial differential equation i n divergence
form (recall that material field
Div
is with respect to
(p,t)t-r(p,t).
p)
for the
Note that, since
po
is the
reference density, we do not need balance o f mass; given a
245
ON THE NONLINEAR T H E O R Y OF ELASTICITY
r,
solution
p
if the density
computed using the expression
in
is needed it can be
r
p = po/det Vr.
The following table compares the salient features of the equation of elasticity with those o f
the Navier-Stokes
equations. ~~-
-
~
-
~
Navi er-S t okes
elasticity
~
Div S(Vr) = p
equations
-
v’+(grad v ) v = yAv grad T T , div v = 0
-
0
~-
restrictions on field
v a vector field, can live in a linear space
r one-to-one,
det vr > 0
__
S
nonlinearity
(grad v)v -
.
material properties _
__._
function
_
~
po
-
.
constitutive restrictions -
S,
__
?
- -~
W > O
-
8. The Elasticity Tensor. The derivative
A(F) (of
at
S
with respect to
F E Lin+;
to each tensor
A(F) H
= DS(F)
F) is called the elasticity tensor
is a linear transformation that assigns a tens or
A(F)H. When the reference is natural, that is when
s(1)
= 0,
(R) yield the following important restriction on
A(1):
246
MORTON E. GURTIN
Proposition
-
If the reference is natural, then A(I)W = 0, A(1)
W = 0
W.
for every skew tensor Remark .__
T
The above relations show that (when the reference is
natural)
A(I)
cannot -_ .
be positive definite;
in fact,
H.A(I)H = E*A(I)E, where E = z1 ( H + H T) is the symmetric part of
H.
It does, however, make sense to
talk about the positive definiteness of the restriction
The main open question of finite elasticity is: what physically acceptable restrictions should be placed on the response function
S
to insure successful mathematical
analysis of meaningful problems.
1
One possible restriction is the strong ellipticity condition: (a 4 b)*A(F)(a 1
8 b)
>
0
Generally, studies in partial differential equations center on investigating a given set of equations with a priori knowledge of the type of nonlinearity, etc. Note that here the thrust is different: one tries to deduce restrictions on S by investigating their consequences. Even negative results such as uniqueness everywhere (as we will see, we cannot expect to have unqualified uniqueness) are important, because they imply that the underlying restrictiomare not meaningful.
ON THE N O N L I N E A R THEORY O F E L A S T I C I T Y
f o r a l l nonzero t e n s o r p r o d u c t s
a 0 b.
2 47
T h i s assumption
renders the underlying p a r t i a l d i f f e r e n t i a l equations t o t a l l y h y p e r b o l i c and hence a p p r o p r i a t e f o r wave p r o p a g a t i o n s t u d i e s .
9. The Boundary-Value Problems o f E l a s t o s t a t i c s
.
W e now l i m i t our a t t e n t i o n t o t h e s t a t i c t h e o r y and consider the d i f f e r e n t i a l equation Div S ( V r )
= 0
supplemented by boundary c o n d i t i o n s of t h e form S(vr)n r where
a8 =
IJ
w i t h /J.
p r e s c r i b e d on d
t h e outward u n i t normal t o The f i e l d
p ,
p r e s c r i b e d on and
,
d i s j o i n t , and where
f
n
is
an.
S(Vr)n
i s c a l l e d the surface t r a c t i o n .
The s i m p l e s t example of a t r a c t i o n boundary c o n d i t i o n i s
dead
l o a d i n g i n which s(vr)n = s with
p k so(p)
a f u n c t i o n of
F o r many problems of
on
4
the m a t e r i a l p o i n t only. i n t e r e s t the prescribed
t r a c t i o n i s a f u n c t i o n of t h e d e f o r m a t i o n p r e s s u r e h a d i n g i n which each p l a c e
x
r.
An example i s
on t h e deformed
surface
r ( @ ) i s a c t e d on by a p r e s s u r e
terms o f
t h e Cauchy s t r e s s ) t h e t r a c t i o n boundary c o n d i t i o n
lT(x).
Thus ( i n
has t h e form
T(x)m = -n(x)m where
m
for
x E r(p),
i s t h e outward u n i t normal t o
ar(6).
This c o n d i m ,
248
MORTON E. GURTIN
in turn, is equivalent to (cf. the definition of the PiolaKirchhoff stress) S(Vr)n
= -(det Vr)Tf(r)Vr-Tn
(PI
on 0 .
It is not difficult t o verify that (de t Vr ) vr-Tn depends only on the tangential gradient
vTr
of
r
on p ;
thus the boundary condition (P) can be written more compactly as follows: S(vr)n
= so(r,v7r)
on
0 .
We now give some counterexamples which demonstrate, quite vividly, that uniqueness general is not to be expect-_ _ _ i n- -~
fi.
We assume i n examples ( A b ) ,
(Ba), and (Ca) that the
reference is natural. A.
The traction problem (a)
(D
= 38).
F o r the traction problem with dead loading a
translation of a solution yields another solution. (b)
(Armanni) Consider a thin hemispherical shell with
zero surface tractions.
Then
r = identity is a solution.
But there should be a second solution consisting o f the everted shell. (c)
(Ericksen)
Consider a rod subject to equal and
opposite tractions on its ends.
This type of loading should result in two types of solutions as shown below.
ON T H E NONLINEAR TIIEORY OF ELASTICITY
249
-I (d)
Also,
i n problem (c) w e would expect “buckled
solutions” of the form
provided the loads are sufficiently large.
B.
The displacement problem (a)
(John)
(jb =
as).
Consider a spherical shell with boundary
an.
condition
r(p) = p
identity.
But there are other deformations which leave the
on
Orie solution, o f course, is the
boundary unmoved, but deform the interior.
Indeed, consider
the deformation caused by a rotation of the inner boundary by a n integral multiple of
ZTI
about an axis through the center
of the sphere.
C.
The genuine mixed problem (a)
( 0
f 0,
p
f 0)
Consider a finite cylindrical rod with sides traction
free and ends rigidly fixed.
One solution is the identity.
Another corresponds to the deformation caused by a rotation ( i n its plane) of one of the ends b y a n integral multiple of ~ T T
about an axis through its center. (b)
We would also expect a situation similar to ( A d ) for
a rod which is loaded at one end, but which has the other end
MORTON E . GURTIN
2 50
fixed. Another difficulty intrinsic to finite elasticity A s noted before, to be meaning-
concerns the solution space. ful a solution
r: @ -+ R3
must b e
(a)
one-to-one, and have
(b)
det Vr > 0 . Condition (b) is severe and makes the theory quite
difficult.
Indeed, the collection of fields satisfying (b)
is not convex;
8
as a matter of fact, f o r
a torus this
collection can have an infinite number of connected components, none of which is convex (cf. Antman [ 3 3 ) .
Further, it is
usually not possible to extend the domain of to tensors
F
det F = 0,
with
becomes infinite as
since
S(F)
S
continuously
generally
det F + 0 .
Condition (a) is even more severe, since it is global.
O f course, one can drop this restriction provided
one is willing to accept solutions of the form:
R
An interesting question i n global analysis is: (b)
+
what
3
(a) 7
F o r the displacement problem an answer is furnished by the
fo 11owing
ON THE N O N L I N E A R T m O R Y OF ELASTICITY
-
Theorem (Meisters, Olech)'
r: 63
Let
4
R3
be smooth, and
suppos e that
det vr > 0, and
(i)
r
(ii) Then
r
I as
is one-to-one.
is one-to-one.
10. Variational Characterization o f the Problem.
The material is hyperelastic if there exists a
u: B +
stored energy function
such that
R
I n this case the mixed problem can be characterized by the principle o f minimum potential energy, at least f o r dead loadingc.
Thus consider the boundary conditions S(vr)n
= s
on P ,
r = r
On
0
with
so
E Ll(n ),
ro E
Ll(p);
7
and let
Def = the collection o f all deformations o f Kin = (r E Def: rIp = ro]
8,
.
Then the mixed problem is formally equivalent to the variational principle: 'Meisters, G . H . and C. a classical theorem on 6 3- 80 ( 1 9 6 3 ) . Here 363. separating set o f ~ 3
minimize 3 Olech, Locally one-to-one mappings and 30, Schlicht function, Duke Math. J. is assumed to b e an irreducible .
20r more generally when the loading is conservative.
3Ball [h]
has established the existence o f minimizers f o r a particular class o f stored energy functions.
MORTON E .
252
m(r) =
i
u(vr)
-
GURTIN
10
s *r
over
Two necessary c o n d i t i o n s f o r a deformation
4
Kin.
n
0
r E Kin
t o render
a l o c a l minimum a r e :
(i) (ii)
J
vu.A(Vr)Vu z
for all
0
B ( a 8 b).A(Vr)(a 8 b ) 2 0
u
E
Var,
f o r a l l tensor products
a 8 b. Here
and l o c a l means w i t h r e s p e c t t o t h e L,
norm of
t o p o l o g y g e n e r a t e d by t h e
t h e deformation g r a d i e n t .
C o n d i t i o n ( i i )i s
u s u a l l y r e f e r r e d t o as t h e Legendre-Hadamard c o n d i t i o n ; a theorem o f Hadamard t e l l s u s t h a t
11.
( i ) implies (ii).
U -n i q u e n e s s .
To a v o i d c o m p l i c a t i o n s w e assume t h a t
30.
( r e l a t i v e l y ) open s u b s e t o f t h e ( p u r e ) t r a c t i o n problem.
This,
pf
$I
is a
of course, rules out
We a l s o r e s t r i c t o u r a t t e n t i o n
t o dead l o a d i n g .
I t i s n o t d i f f i c u l t t o show t h a t a of
C2
solution
t h e mixed problem s a t i s f i e s t h e i d e n t i t y
Our r e s u l t s r e g a r d i n g u n i q u e n e s s p e r t a i n t o weak s o l u t i o n s ; t h a t i s , deformations A deformation
r E Kin r
E
Kin
that satisfy (W). is internally stable i f
r
ON THE N O N L I N E A R TIEORY OF ELASTICITY
253
B
a __ set
Pl c Kin
is internally stable i f every
f E
K
has this
property. The next result',
which is a simple consequence of
the above definition and the mean-value theorem, shows that stability implies uniqueness.
-
Theorem
X
C
Uniqueness holds i n any convex, internally stable
Kin.
Corollary fi)
(it)
-
Assume that the reference is natural and
A(I)ISym
P
= a63
elliptic.
for some
is positive definite,
or
(displacement problem) and
A(1)
is strongly
Then uniqueness holds i n
6 7 0.
The above theorem can be extended to situations
involving non-dead loads, but the extension i s nontrivial.
In particular, for loading of the form show that if
X c Kin
s0(r,vTr)
one can
is convex and uniformly internally
stable in the sense that vu-A(vr)Vu
2
xllull * H
for all
u E Var
and
r E
X,
(0)
and if the first two derivatives of s
0
are sufficiently si'iall
on K , t h e n uniqueness holds i n K.
'Gurtin, M.E. and S.J. Spector, On uniqueness i n finite elasticity. Arch. Rational Mech. Anal. Forthcoming. 3
'Spector, S.J., On uniqueness in finite elasticity with general loading. Forthcoming.
G.M. de La Penha. L.A. Medeiros (eds.) Contenporary Devel opmnts i n Continuum Mechanics and P a r t i a l D i f f e r e n t i a l Equations @North-Holland P u b l i s h i n g Company (1978)
CONSTITUTIVE EQUATIONS AND FREE SURFACES
DANIEL D. JOSEPH Department of Aerospace Engineering and Mechanics University of Minnesota, Minneapolis
Abs t ra c t The general theory of perturbations o f rigid body motions of simple fluids with applications to free surface problems is discussed.
The
general theory is utilized to explain phenomena exhibited i n the movie "Novel Weissenberg Effectst1by G . S .
Beavers and D.D. Joseph.
Introduction. The ideas behind the simultaneous perturbation of the domain and the constitutive equation may be explained without the extra, and largcly incidental, mathematical complications which follow from analysis of the nasty equations which govern rheologically complicated materials.
In a
simpler setting, we shall first consider a model problem in which we show that the distortion of a free surface due to motion may be expressed in terms of unknown rheological constants.
In the model problem we perturb the "state of
rest'?. A rich theory of perturbations of the 'Irest state" of real viscoelastic fluids following along lines laid out in this lecture has already been given by Joseph and Beavers
CONSTITUTIVE E Q U A T I O N S AND FREE SURFACES
(1977).
Many of
255
t h e s i m p l i f y i n g f e a t u r e s o f t h e t h e o r y of
P e r t u r b a t i o n s of t h e r e s t s t a t e o f a v i s c o e l a s t i c f l u i d a r e
a l s o p r e s e n t i n t h e y e t r i c h e r t h e o r y of p e r t u r b a t i o n s o f s t a t e s o f r i g i d motion of a v i s c o e l a s t i c f l u i d .
In fact,
s t a t e s o f r i g i d motions o f a v i s c o e l a s t i c f l u i d a r e g e n e r a l i z ed r e s t s t a t e s i n t h e s e n s e t h a t t h e Cauchy s t r a i n the extra s t r e s s
5
b o t h v a n i s h on r i g i d m o t i o n s .
G(s)
and
Moreover,
i n t h e t h e o r y g i v e n h e r e , and by J o s e p h ( 1 9 7 7 ) , t h e same m a t e r i a l f u n c t i o n s s u f f i c e t o d e s c r i b e t h e m o t i o n s which perturb a l l s t a t e s of steady r i g i d rotation including the rest s t a t e i n which t h e r e i s no r o t a t i o n .
I t i s t h e r e f o r e p o s s i b l e t o s e p a r a t e t h e problem o f f r e e s u r f a c e s on v i s c o e l a s t i c f l u i d s which a r e c l o s e t o steady r i g i d motions i n t o t w o p a r t s .
I n the f i r s t part
( $ 1 )we
c o n s i d e r t h e model problem i n which t h e i d e a s i n v o l v e d i n t h e s i m u l t a n e o u s p e r t u r b a t i o n o f t h e domain and c o n s t i t u t i v e e q u a t i o n a r e exposed i n a s i m p l e c o n t e x t .
I n t h e second p a r t ,
( $ 2 - 7 ) we p e r t u r b s t a t e s of s t e a d y r i g i d r o t a t i o n o f r e a l v i s c o e l a s t i c f l u i d s , b u t l e a v e t h e domain f i x e d . domain i s n o t f i x e d t h e c a n o n i c a l forms of
When t h e
the s t r e s s ( $ 6 )
and t h e e q u a t i o n s o f motion ( $ 7 ) a r e unchanged and t h e e q u a t i o n s which h o l d on t h e f r e e boundary may b e e a s i l y d e r i v e d by t h e methods u s e d i n t h e model problem ( $ 7 ) .
1. A model problem f o r domain p e r t u r b a t i o n s of
the r e s t state.
We s u p p o s e t h a t t h e r e i s a n a n a l y t i c f u n c t i o n
5(@) @ = 0
which i s n o t known e x c e p t t h a t i n a neighborhood o f
i t has a Taylor s e r i e s
256
DANIEL D.
1
a ( @ )= a = 5
where
(0)
a@
2
b = 5
and
1
+
a0
JOSEPH
bo3
00@
...
+
,
(1.1)
We may t h i n k t h a t
(0).
5 ( @ ) i s representative of the nonlinear p a r t of the stress Our i d e a i s t o f i n d t h e T a y l o r
i n some f i c t i t i o u s m a t e r i a l .
c o e f € i c i e n t s f o r (1.1) by m e a s u r i n g t h e f r e e s u r f a c e i n d u c e d by a dynamical p r o c e s s
v 20 + 5 ( @ )
F(E) =
= 0
in
b6,
(1.2)
subject t o the condition that
b6
where
i s a bounded r e g i o n o f s p a c e d e p e n d i n g on a
parameter
6
parameter
C .
and
g(x,E:)
a r e g i v e n d a t a d e p e n d i n g on a
I t i s i n s t r u c t i v e t o c a r r y o u t our a n a l y s i s i n e a s y stages.
F i r s t we p e r t u r b
leaving
6 ,
6
f i x e d and assume
that @(IE,E
1
1
=
@a(d
(1.4)
gn(d = 0,
(1.6)
c fG l 0
where,
in
6' 00 +
a n d , on
'
No0)
= 0,
ab6; Gn
where
gn(x)
we c o u l d s o l v e etc.
Y
i s nth
=
on(&)
-
d e r i v a t i v e of
( 1 . 5 ) and ( 1 . 6 ) for
g(&,s)
Go
at
c =
0.
we c o u l d f i n d
as t h e s o l u t i o n of l i n e a r b o u n d a r y v a l u e p r o b l e m s .
If
5,,@,, It
CONSTITUTIVE EQIJATIONS AND FREE
would b e h a r d t o s o l v e ( l . 5 ) l ,
i t i s nonlinear. we d o n ' t know n e i g h b o r h o o d of
3(0,),
even i f we knew
But we c a n n o t s o l v e ( 1 . 5 ) 1
3(o0)
2 57
SURFACES
bccausc.
a t all b e c a u s e
e x c e p t a s a power s e r i e s i n a s m a l l
zero.
The " r e s t s t a t e " i s now d e f i n e d as t h e s t a t e f o r
go(xlab
which
6
= 0.
assumption t h a t
(1.5)1
@,(x)(~,, =
Then
0.
W e couple t h i s d e f i n i t i o n with t h e has @
6
0
110
(x) -
E
solutions
b6
in
0
rijo
#
0
when @ o ( x _ ) - O
and, r e p l a c i n g
we g e t
etc.
These p r o b l e m s a r e l i n e a r and e a s y t o s o l v e even rihen
a
i s unknown ( i n f a c t ,
, @,
a r e independent o f
o2
= @21
+
a @ 2 2 , where
021
and
S o we c a n u s e t h e s o l u t i o n
a).
which p e r t u r b s t h e r e s t s t a t e t o f i n d
a
a n d , w i t h more work,
we c a n g e t e x p r e s s i o n s which i n v o l v e t h e o t h e r T a y l o r coefficients of
(1.1) a s w e l l .
Now suppose t h a t suppose f u r t h e r t h a t
of
convenience i s ,
symmetry. in
b6
6
i s f i x e d and
6
varies.
And
b o i s s o m e c o n v e n i e n t domain whose point
say, that
Ir0
h a s a h i g h degree o f
W e a r e g o i n g t o t r y t o s o l v e t h e dynamic problem
a s a s e r i e s whose c o e f f i c i e n t s c a n b e d e t e r m i n e d from
b o u n d a r y v a l u e problems posed on t h e symmetric domain F i r s t we map
Ir6
into
bo
t o one mapping, which i s a n a l y t i c i n p o i n t s on
ab6
into
all o :
b0.
w i t h a n i n v e r t i b l e one
6
and which t a k e s
D A N I E L D.
258
x_ = ~ ( ,0) x
(identity),
-1 = x_ (5,s)
(inverse),
-0
xo alr Let Ir6
H(x,6)
JOSEPH
alr o .
(1.8)
be any function defined i n the family of domains
and introduce the notation:
and
(1.9) etc.
It follows from the equations that
when
x
E b6,
x
-0
E lro,
and that (1.10)
W e may easily establish by mathematical induction, using the
the connection formulas, that
F(n)(so) = For example,
+'
F( 0 ) ( 5 , ) = 0
xC1] -VF(0)(~o)
in
= F(I)(x0) = 0.
0
b0
in
bo*
and
r (x,)
F 11
(1.11) = F(
' ) (&,)
+
It is a bit more complicat-
259
C O N S T I T U T I V E E Q U A T I O N S AND FREE SURFACES
ed a t t h e boundary. on
aho
Since
G(6) = 0
L mJ ( 5 , )
and t a n g e n t i a l d e r i v a t i v e s o f
G
C ml
must
It f o l l o w s
alro
t h a t on
n-v =
bo
on
v a n i s h b u t normal d e r i v a t i v e s need n o t v a n i s h .
n_
where
= 0
&his, G
on
i s t h e outward normal t o
[: 11 .n_ h g t vn = 5
and
a/an. W e may s e e k t h e s o l u t i o n o f
as a s e r i e s d e f i n e d i n
(1.2)
and
(1.3) i n
b5
b0
(1.13)
(1.14) etc.,
and on
The problems can't
ab,
(1.14),
( 1 . 1 5 ) a r e l i k e ( 1 . 5 ) and ( 1 . 6 ) .
s o l v e them b e c a u s e
W e
5 ( @ ) i s g i v e n only as a Taylor
s e r i e s (1.1) w i t h unknown c o e f f i c i e n t s and a n unknown c i r c l e of
convergence. The " r e s t s t a t e " f o r t h e domain p e r t u r b a t i o n ,
t h e " r e s t s t a t e " f o r t h e p e r t u r b a t i o n of may b e d e f i n e d b y t h e c o n d i t i o n condition implies t h a t Q,(O) P
0
in
b,
and
0 '0'
g[: 0 3
= @( o ) =
t h e boundary d a t a
(x,)] o
like
=
0.
This
0
on
aho,
hence,
2 60
D A N I E L D.
JOSEPH
(1.16)
The linear problems (1.16), (1.17) and higher order
etc.
problems are solvable and not too hard to actually solve, even when
a
is unknown.
In our rheological problems the boundary data
(e
in
our first example) pe’rturb the boundary (6 in our second example).
S o we may put
E
= 6
and construct a simple
example of a domain perturbation of the rest state with a free surface.
(c)
By a free surface we understand that there is a one
parameter family of domains
bc
which are unknown
Supposing now that o u r dynamical process (1.1) and (1.2 in
Uc
we might expect solutions i n each and every
responding to some possibly small
e
Ire
hold co r -
interval of the origin.
But no, this will not be possible because in addition to (1.2) we pose an additional boundary condition, which is analogous to, but much simpler than, the condition that the jump i n the normal component of stress is balanced by surface tension times mean curvature.
Because we have this extra condition
we can’t solve (1.1) and (1.2) i n every condition can be satisfied only when
kc
-
Ire ’
the extra
is properly chosen.
A s an example of the foregoing consider the two
dimensional problem specified in polar coordinates figure 1 where the boundary data correspond to a rest state ical process
@(r,B,c)
g(8,O)
g(8,c)
in
are given and
= g(O’(8)
and the function
(r,e)
= 0.
f(e,c)
The dynamwhich gives
261
C O N S T I T U T I V E E Q U A T I O N S AND FREE S U R F A C E S
r = 1 -+ f ( e , c )
t h e boundary
of
be
a r e unknown.
5(@);
t h e r e a d e r t h a t o u r aim i s t o show how t o f i n d
is,
W e remind that
t h e T a y l o r c o e f f i c i e n t s i n (1.1) b y ( f i c t i t i o u s ) e x p e r i -
m e n t a l measurements o f
r = 1
t h e (made u p ) b o u n d a r y
+
f(8,g).
(1.18)
Fig.
1.
Model of a f r e e s u r f a c e problem
F i r s t we s o l v e ( 1 . 1 8 ) when
v2@
z ( @ )=
+
0
in
bo.
bo
with
@ ( r y e y o=)
has
I# E 0
that
the r e f e r e n c e c o n f i g u r a t i o n
r
= 1.
where
in
Then,
since
W e seek t h e s o l u t i o n o€
f
[ 01 (8,)
= 0[
01 ( r o , e o )
r e l a t e d by a s h i f t i n g map
e = 0
= 0
Z(0)
bo
0
= 0,
and we f i n d t h a t on
r = 1 + f(8,o)
f(8,o) = 0
i s the unit circle
( 1 . 1 8 ) i n powers o f
and
9
and
bo
6
are
SO
2 62
D A N I E L D.
JOSEPH
h a v i n g a l l t h e p r o p e r t i e s r e q u i r e d of map t h e d e f o r m a t i o n o f
bo
Note t h a t f o r any f u n c t i o n f(n'
(8)
c
f E n 3( 8 ) .
is
(1.8).
F o r the shifting
a l o n g r a y s and
f(e)
of
8 = Bo
a l o n e , we have
Using t h e c o n n e c t i o n f o r m u l a s ( 1 . 9 )
we find
that
0[ 11
(ropeo) =
@
( 1) (ro,eo)
and
b o , r o = 1 we h a v e , from ( 1 . 1 8 ) 3 t h a t
On t h e boundary of
- f c 11 ( e o )
3m(0)@(1'(l,eo) that
f
c 11 ( 8 , )
o(')(ro,e0)
Fig.
= 0. and
2.
= 0.
Since
3@(0) = 0 ,
we f i n d
The boundary v a l u e problems s a t i s f i e d by
@( 2 )
(ropeo)
a r e g i v e n i n f i g u r e s 2 and 3.
The problem s a t i s f i e d by
0
( 1)
(ro,eo)
263
CONSTITUTIVE E Q U A T I O N S AND FREE SURFACES
c 21 ( e o )
f
3 that
We f i n d f r o m f i g u r e
These p r o b l e m s a r e e a s y t o s o l v e .
= a @( 1) 2( 1 9 6 o ) 9
where
@( 1)
(ro,eo)
f'igure
2.
I t f o l l o w s t h a t t h e f i r s t approximation t o the
b,
shape o f
i s t h e s o l u t i o n o f t h e problem shown i n
i s g i v e n by
2 r = I +
f(e,e)
= 1
+
a@(')
The n e x t a p p r o x i m a t i o n depends on
b
(i,e)s2+
qE3).
a s w e l l as
t h e r e f o r e deduce t h e v a l u e s o f d e r i v a t i v e s
of
by m o n i t o r i n g t h e c h a n g e s i n t h e s h a p e o f
a.
3(@)
We may at
@ = O
is
when
n ear t o zero.
2. C o n s t i t u t i v e e x p r e s s i o n s which p e r t u r b t h e r e s t s t a t e of -_
a simple f l u i d . ~
I n our r h e o l o g i c a l s t u d i e s t h e r o l e o f t h e nonlinear function
5 ( @ ) i s assumed by t h e n o n l i n e a r f u n c t i o n a l -
g i v i n g t h e c o n s t i t u t i v e l y determined p a r t of O S an i n c o m p r e s s i b l e f l u i d .
domain of
s = t
-
T
-3
the s t r e s s
The argument f u n c t i o n s i n t h e
a r e symmetric t e n s o r v a l u e d f u n c t i o n s of
f o r fixed
t
aiid
r,
called histories,
which a r e d e f i n e d o n t h e r e l a t i v e d e f o r m a t i o n t e n s o r
= V,X~(X,T)
of
T
where
the p a r t i c l e
&
Xt(x,T)
= {(&,T)
F (T)
-t
=
i s the position vector
which i s p r e s e n t l y a t
&.
Eq.
(2.1),
which assumes t h a t t h e s t r e s s d e p e n d s o n l y o n t h e f i r s t spatial
2 64
D A N I E L D.
JOSEPH
g r a d i e n t of t h e d e f o r m a t i o n , i s s t i l l t o o g e n e r a l t o be u s e d
t o s o l v e t h e problems which l e a d t o u n d e r s t a n d i n g how v i s c o e l a s t i c f l u i d s respond
t o applied forces.
I t i s w e l l known t h a t t h e e x t r a s t r e s s v a n i s h e s on n(0) = 0.
the zero h i s t o r y expansions o f
~(G(S))
I t i s t h e n n a t u r a l t o seek f u n c t i o n a l e x p a n s i o n s on
i n terms of
the r e s t h i s t o r y
gCo;G(s)l
3[G(s)l =
,G_(s)l
+ a,Co;G_(s)
3;2r:O;G(s),G;(s),G(s)l
+
+
9
. . a
where
i s an
n
linear form.
W e s h a l l assume ( G r e e n and R i v l i n
( 1 9 5 7 ) , Coleman-No11 ( 1 9 6 1 ) ~ P i p k i n ( 1 9 6 4 ) t h a t t h e s e n - l i n e a r forms
may b e r e p r e s e n t e d b y i t e r a t e d i n t e g r a l s o f t h e f o r m
a
'i j k k
. ..mv'
Gkk(sl)
...
1' 2 '
'
YSrl
1
G m v ( s n ) d s 1d s 2
. ..
d sn *
It i s v e r y e a s y t o f i n d t h e i s o t r o p i c forms o f t h e s e i n t e g r a l s f r o m invariance theory f o r a single tensor
94.7 i n Joseph ( 1 9 7 6 ) ) ,
a(")(C+(s))
G ( s ) ( s e e Exercise
If we i n t r o d u c e t h e n o t a t i o n
f o r t h e p a r t i a l sum o f
(2.2)
after
n
terms i t
(2.4)
CONSTITUTIVE E Q U A T I O N S AfJD FREE SURFACES
m
m
(n
So we may e x p e c t
i n which
G_(s)
stresses
a(").
l i n e a r forms i n G _ ( s i ) ) d sl...dsn
t h a t t h e s t r e s s i n any motion c l o s e t o one
= 0
can he e x p r e s s e d i n t e r m s of
But t h e s t r e s s e s
3(n)
t h e FrLchet
a r e n o t i n t h e form which
I c a l l canonical f o r perturbation o f the zero h i s t o r y .
The
c a n o n i c a l forms a r e t h e forms which t h e s t r e s s and e q u a t i o n s o f m o t i o n t a k e when d a t a g i v i n g r i s e t o s t e a d y r o t a t i o n a r e perturbed.
W - e d o n o t a l l o w k i n e m a t i c s which a r e i n c o n s i s t e n t ~~
~
F o r example,
w i t_ h __ t h e equ a t i o n s o f m.otion. _ -
though i t i s well-
known and e a s y t o d e m o n s t r a t e t h a t o n a l l r i g i d body m o t i o n s
G_(s) = 0,
only t h e s t e a d y r i g i d motions a r e compatible with
t h e equations.
R e l a t i v e t o a c o o r d i n a t e system t r a n s l a t i n g
w i t h t h e t r a n s l a t i o n a l v e l o c i t y o f t h e r i g i d body we may assume t h a t body a l s o r o t a t e s r i g i d l y w i t h v e l o c i t y Then
= 0,
3[0] = 0
A
5.
and t h e e q u a t i o n s o f motion
0
a r e s o l v a b l e i f and o n l y i f
a
A
x
is a gradient;
that is,
0
if
= 0.
S o we a r e r e s t r i c t e d t o p e r t u r b a t i o n s of r i g i d
r o t a t i o n s with constant
g.
T o f i n d t h e c a n o n i c a l f o r m s , t h e f o r m s which t h e s t r e s s and t h e e q u a t i o n s o f m o t i o n t a k e when d a t a g i v i n g r i s e
t o steady r i g i d rotation are perturbed, i t w i l l suffice t o i m a g i n e t h a t for ~ E a l j ( t ) , where by f l u i d ,
lj(t)
i s t h e r e g i o n occupied
t h e p r e s c r i b e d boundary v e l - o c i t y
266
D A N I E L D.
U(,li,t,€) =
61
A
X_
+
JOSEPH
Ff(x_,t),
VtER,
is a steady rigid rotation plus an arbitrary part proportional to
Now we suppose that the solutions o f all the govern-
11.
ing equations depend on
6
through the prescribed data and
that they may be differentiated a certain number of times at 6
= 0.
In the best case we would have analytic solutions and
convergent power series in
E .
In less good conditions we
suppose that some low order partial sums are asymptotic to true solutions.
In either event we must identify the boundary
value problems which govern the derivatives o f the solution at
E
= 0
and we call the stress and equations f o r these
derivatives canonical; canonical in the sense that the derivatives are independent of
E.
This natural method of
doing perturbations requires that we consider only those forms of the stress which are compatible with the solutions o f the equations, s o after all is done we get a good theory with which we can actually compute solutions to problems.
For
example, we have already noted the only rigid body rotations compatible with the equations of motion are steady. The canonical forms o f the stress and the equations o f motion are easiest to understand by actually deriving them.
We shall find that at each stage of the perturbation we shall need to ~ o l v efour linear partial differential equations f o r three components of velocity and a reaction pressure, as in the Navier-Stokes equations.
The strain history comes in as
an after-thought after the velocities are computed.
CONSTITUTIVE E Q U A T I O N S AMD FREE SURFACES
2 67
3. Kinematics f o r p e r t u r b a t i o n s .
Since
we h a v e , a s s u m i n g t h e p a r t i c l e
label
1
5 = Ijt(X,t,C i s independent o f
( n)
ax_t a T Moreover, u s i n g
n = 0,1,2,
that f o r
E ,
=
(g,T)
(3.4)
E(n)(~,~).
( 3 . 1 ) , (3.4) and ( 3 . 2 ) ,
xt(
0)
...,
(litt)
-p t( x _ , t )
=
(3.5)
we f i n d t h a t
x,
= 0
and
i-@ ( X , t ) The f u n c t i o n
,.,
E(&,T)
=
p ( x , t ) .
i s an a u x i l i a r y f u n c t i o n used t o
f a c i l i t a t e o u r c o m p u t a t i o n of p a r t i c l e p a t h s . To s i m p l i f y n o t a t i o n s , we d e f i n e
X(n) and
(TI
sin)
(_x,T),
n=1,2,.
..
2 68
D A N I E L D.
JOSEPH
Then, using (3.3) and the chain rule, we find that
where
The functions
z ( ~ ) ( T )
may be computed by integrating (3.10)
subject to the conditions (3.6) and (3.7). We turn next to the computation of derivatives of the strain tensor.
From the definition of the relative
deformation gradient given i n $2, we find, u s i n g (3.1) and
( 3 . 9 ) that
where
Then
where
and
2 69
CONSTITUTIVE EQUATIONS AND FREE SURFACES
4. Functional derivatives
the stress and the equations
of
governing the of ~ special _ ~ perturbation ~ _. __ _ motions _ ~with
-_
arbitrary motions.
Suppose that
G(s,E)
is the series given by ( 3 . 1 2 ) .
This series may be assumed to induce a functional expansion of the stress in powers of
Apart from a factorial,
6
(see Joseph, 1976; p.197):
3n is a functional derivative,
typically a Fr6chet derivative, evaluated on the history
g")(s)
of the special solution,
The linear arguments of
these derivatives, those following the vertical bar, are to be determined sequentially by solving the perturbation equations of motion which have yet to be specified.
The
functional derivatives given i n ( 4.1) are still too generally specified to be useful in the solution of problems. However, the first Frgchet derivative
z1[G(O)
assumed to b e in integral form when 2
Lh(O,m)
1 *]
may be
G _ ( s , c ) lies in a
Hilbert space whose scalar product is defined by a n
integral with a weight
h(s),
h(s)
-+
0
as
s
-+
-.
Such a
representation may be justified by appeal to the representation theorem of F. Riesz. Identifying independent powers of
6
i n the
270
D A N I E L D.
expansion o f
JOSEPH
t h e e q u a t i o n s o f m o t i o n , we may i d e n t i f y a n The z e r o t h
ordered sequence of p e r t u r b a t i o n problems. problem i s d e f i n e d by ( 5 . 1 ) , we f i n d t h a t i n
A t f i r s t order,
t > 0
b(t),
+
( 5 . 2 ) and ( 5 . 3 ) .
order
5 [ G ( O) -1 -
v and (7
-
) =
i'
The h i s t o r y of t h e v e l o c i t y
- (l) U
(5 ( 7 ) ,'r
Since
)
_F
i s prescribed i n
l~
(T
),
i s the g r a d i e n t of
T
S
(4.6)
0.
, (VI.2),
-
with
(4.2),
(4.3)
and
( 4 . 4 ) may b e viewed s e v e n l i n e a r e q u a t i o n s
i n t h e s e v e n unknown f u n c t i o n s
X(-t" ( ? , ' r ) ,
C(l)(z,t)
and
d l ) ( 5 ,t ) . A s i m i l a r l i n e a r problem f o r
(n 2
2)
a r i s e s at higher orders.
, -(n)
,
U(n) -
and p ( n)
I f t h e s e problems a r e
s o l v a b l e , t h e y a r e s e q u e n t i a l l y s o l v a b l e and t h e m o t i o n and
,
27 1
CONSTITUTIVE E Q U A T I O T J S A N D FREE SURFACES
s t r a i n h i s t o r y may b e g e n e r a t e d a s power s e r i e s . p e r t u r b a t i o n problems, integral,
1 61
S,[:(i(O)
with
These linear
r e p r e s e n t e d by a n
a r e not t o o g e n e r a l f o r mathematical s t u d i e s o f
e x i s t e n c e and u n i q u e n e s s .
5 . K i n e m a t-_ ics of a r b i t r a r y .m- o tio n s p e r t u r bi n g s- t e a d y r i g i d ~~
r o t a t i o n s of a simple f l u i d . ~~
N o w I am g o i n g t o d e r i v e a n a l g o r i t h m f o r computing
I want
m o t i o n s which p e r t u r b s t e a d y r i g i d r o t a t i o n s . solutions of
t h e b a s i c e q u a t i o n s for which
E‘O’ ( S
R i g i d body m o t i o n s have
)
I
but,
0
G ( O ) ( s ) z 0. i n general, 0
motions w i l l n o t s a t i s f y the e q u a t i o n s because conservative (see ( 2 . 6 ) ) .
I, therefore,
-m
<
< t
T
&=
5
0
i s not and p u t
= c-2 A S(l-1
U -( O ) ( 5 ( T ) , T )
for
set
A
such
a t a l l points i n
l j
(7
).
(5.1) Then
Z [ G-( O ’
(s)]
0
and p(O)(r,t) a t each The p a t h
x_
E
b(t)
xio’ -
+
2- 1~
xi2
r\
= const
and a t e a c h and e v e r y i n s t a n t
(X,T
)
z
S(T)
for
‘r
S
t
<
( t )= 5.
(5.2) t >
-m.
i s o b t a i n e d by
integrat i n e
5
-’T
= 61 A
C(T), -
Without l o s i n g g e n e r a l i t y , normal b a s i s
el,
g 2 , E~
such t h a t
Ri s a constant vector.
we c h o o s e a f i x e d o r t h o -
=
“7n
(5.3)
Then t h e p a r t i c l e p a t h i s g i v e n by
JOSEPH
DANIEL D.
g(ns)
where
is the unique orthogonal tensor rotating the
orthonormal basis is,
$,(ns),
e = Q(ns).&(Rs). -i
$,(ns),
Qij(ns)
ei
=
=
-
c o s 62s
sin 0s
ns
c o s 62s
sin
It follows from (5.4) that
3")
E
0.
z2, e3;
that
Qij(Qs)gj
0
and
g19
~ l ~ ~ - ( R s ) . ~that j ; is,
[Qij(ns)]
Hence,
into
Relative to the fixed basis,
Q(ns) where
g3
0
(5.5) 1
27 3
CONSTITUTIVE EQUATIONS AND FREE SURFACES
It follows that
The most natural measure of deformation for rotating simple fluids is the time derivative o f the Cauchy strain
We shall need the following expansion formula for
where the
3(n)
L(n'
=
are defined in terms
v[c(n)(x,7.)
Equations (5.8) show that & L n
and on
4 X,(Z,T)
-
&("'
for
A
Q(ns),
X( n ) ~ , n
depends o n
2
(')
and
1.
~ " ) ( $ ( T ) , T )
(5.12) for
k < n. In fact,
(5.13) and for
re2
To prove the expansion formula (5.11)*, that (5.10) reduces to (5.11)1. consequence o f the identity
we must first s h o w
This reduction is a direct
27 4
D A N I E L D.
''"''"'
JOSEPH
( 5 * 1 3 ) and ( 5 . 1 4 ) , we expand
F ( s , c ) = Q _ ( n s )+
c E'E"'
and E x , using ( 5 . 9 ) and (5.8), and collect the 1 coefficients o f independent powers o f E i n the induced expansion o f ( 5 . 1 1 ) . Some further transformations of the tensors are used in the analysis.
,(n)
(s)
These transformations are motivat-
ed by the fact that the perturbation problems to be derived lead to the sequential determination of the velocity coefficients
,(n)
( g ( )~,T
)
whose natural arguments are the
Components of the rotating vector now that w h e n b(t),
T = t,
we find easily that
Noting next that
we find that
and
where
5 = x
and
xio)
( 5 , ~ =)
div
g(n)(,,t)
S(T).
= 0
Noting in
CONSTITUTIVE EQUATIONS AND FREE SURFACES
275
and
Using ( 5 . 8 ) , we find that
and
where
and
L.0.t.
= lower order terms.
6. Canonical forms for the stress.
My constitutive hypothesis is that the FrGchet derivatives of ed by integrals.
Z[G_(s)] -
o n the zero history can be represent-
I also assume that kernels in these
integrals vanish at a rate sufficient to justify integrating by parts; for example,
DANIEL D.
27 6
JOSEPH
and
Explicit expressions for Beavers (1977). duce
B(s) -
g3
and
&
are given by Joseph 8s
This integration by parts allows us to intro-
= -dC_(s)/ds
as the fundamental measure of defor-
mation and leads ultimately to a theory in which perturbation velocities are sequentially computed from four equations governing three components of velocity and the pressure,as in an incompressible, Navier-Stokes fluid,
stress
?[G(s)]
Assuming that the
admits a FGechet expansion i n integrals with
good kernels, we get
T o obtain the canonical forms of the stress for the
theory o f rotating fluids, we identify independent coefficients
277
CONSTITUTIVE EQUATIONS AND FREE SURFACES
i n t h e s e r i e s expansion of
(5.12)
g(s,E).
of
5 [ 3 ( s , c ) ] i n d u c e d by t h e e x p a n s i o n
This l e a d s us t o
The c a n o n i c a l f o r m s o f t h e s t r e s s f o r p e r t u r b a t i o n s of
s t e a d y r i g i d r o t a t i o n a r e g i v e n t h r o u g h o r d e r two by m
Z [ ~ ( S , E ) ]=
E
G_(s)
G(s) Q T ( n s ) * [-* ( * ) ( s )
where
S -' " ( s )
+
sT(ns).A(l)(s)*y_T(,s)ds
+ B(s)]
.Q(ns)ds
i s d e f i n e d by (5.18) a n d
B(s)
+
by
The h i g h e r o r d e r s t r e s s e s a r e n o t h a r d t o d e r i v e . i n t h e form
(5.22). They a r e
278
D A N I E L D.
+
7. C a n o n i c a l f o r m s
JOSEPH
lower-order
of
terms.
t h e e q u a t i o n s of m o t i o n .
A f t e r e x p a n d i n g i n powers of
When P"'
n = 0,
y ( O ) ( ~ y ; t )=
+ 1/2 p l c A s 1 2
(6.7)
QA?,
3")
i s constant.
=
E ,
we f i n d t h a t
a!(')/at
= 0,
and
N o w we s h a l l d e m o n s t r a t e
that
[
m
div %(n) =
+ To e s t a b l i s h E(5,T)
=
g( s ) Q T ( n s )* V52 U- ( n )
lower-order
(7.2),
hl(S,T),
(C- ( t - s ) , t - s ) d s
terms.
(7.2)
w e f i r s t show t h a t f o r any t e n s o r
we have
T a k i n g components i n t h e f i x e d C a r t e s i a n b a s i s , we f i n d t h a t
where we h a v e u s e d
Q , ~=
we may a l s o v e r i f y t h a t
a c k / a x j'
Since
auj") ( 5 , T
)/ati
=
0,
CONSTITUTIVE E Q U A T I O N S AND FREE S U R F A C E S
Cornbilling
( 7 . 3 ) and ( 7 . 4 ) w i t h ( 6 . 7 ) , we p r o v e ( 7 . 2 ) .
(7.1)
Equations components
el,
1
,
g2,
e
TO
-3'
identically i n b(t)
U(n'(x,t)
( 7 . 1 ) and
b(t)
t
when
r e l a t i v e t o the f i x e d
e x p r e s s ( 7 . 2 ) i n terms of
S
5 E ab(t)
for
fern
The i n i t i a l - h i s t o r y y("(x,t)
and
t 2 0,
and
Q
U -")
n = 2,
When
are also we may n = 1,
(7.6)
( 7 . 6 ) determine
The s o l u t i o n i s i n d e p e n d e n t of
5 , ~ )a t f i r s t o r d e r .
p a t h may b e computed a s a n a f t e r t h o u g h t once known:
i s prescrib-
( 5- ( t - s ) , t - s ) d s = O .
problem a s s o c i a t e d w i t h
3 ( 1) (
!(n),
A
i n sequence.
I . ( S ) Q T ( W *V!
p ( 1) ( 5 , t ) .
the p a r t i c l e path
(x,t)
y(n'
(~(n).o)&o)=
w r i t e t h e p e r t u r b a t i o n problems
-
and
hold
Boundary c o n d i t i o n s , s a y
0.
Noting t h a t
div v ( n ) ( x , t ) = 0
t > 0
for
= g(I1)(T,t)
prescribed.
When
the
we n o t e t h a t
Equations
ed i n
may b e e x p r e s s e d i n t e r m s o f
g(n) - ci ui( n )
~ 1( n ) of
c a r t e s t a n basis U(n)
27 9
The p a r t i c l e
!(')
( 5 ,t )
is
280
D A N I E L D.
JOSEPH
G( s ) Q T ( n s )adiv
B(s)ds
5 -
[
+ div
where
4
is given by (6.19) and
terms on the right of are known.
(7.8)
Y(s1,s2)L ( l ) ( ~ l ) * J ( l(s2)dslds ) 2
B(s)
by ( 6 . 2 2 ) .
dl) -
(7.8) are known when
The
and
Hence, we may solve the initial-history problem
associated with (7.8) for
y(2)(x(,t)
and
p(2)(x,t).
Then
we can compute the path at second order:
It follows that at each order we may compute three velocity components
c(n'(s,t)
and a pressure
~ ( ~ ) ( x , t ) from a n
inhomogeneous, linear, initial-history problem associated with (7.1) and
div
c(n) =
0.
&( n, appears
T h e particle path
(.L)
as an auxiliary quantity which may be computed when .L < n
is known.
,
In other words, at each stage o f the
sequence, we solve four equations i n four unknowns. The mathematical problem defined by this perturbation sequence may b e stated as follows: b(t) on
for ab(t)
t > 0, h(x,t) for
div a(x,t) = 0
in
t > 0, find and
b(t) g(5,t)
~ ( 5 , t )= &(lf,t)
for
and in
Given
f(s,t)
t
and
S
0
@(&,t)
b(t)
in
q(5,t)
such that
for
t < 0,
28 1
CONSTITUTIVE E Q U A T I O N S AND F R E E SURFACES
g(r;,t)
.a(x,t) =
ba p { s
+
(c
+
?)-VS
A
-[
t > 0
x E ab(t),
for
+
51
T
and
V@
2
G(s)Q (Qs)*V? ? ( < - ( t - s ) , t - s ) d s = f ( x , t ) .
I t i s known ( S l e m r o d , 1 9 7 6 ) t h a t u n d e r v e r y mild c o n d i t i o n s G(s),
on t h e s h e a r - r e l a x a t i o n modulus
R
= 0
n f
The s t a h i l i t y r e s u l t p r o v e d by
has a unique s o l u t i o n .
Joseph (1977) s u g g e s t s t h a t i f a 0,
t h e problem w i t h
s o l u t i o n o f t h i s p r o b l e m e x i s t s when
t h e n i t i s unique. Finally, I note that i n l i m i t
+
0
the theory o f
r o t a t i n g s i m p l e f l u i d s c o l l a p s e s i n t o my p r e v i o u s t h e o r y o f p e r t u r b a t i o n s of
the s t a t e of r e s t
The c a n o n i c a l forms o f
(Joseph,
1976).
t h e s t r e s s and t h e e q u a t i o n s
of m o t i o n t a k e a p a r t i c u l a r l y s i m p l e f o r m i n a r o t a t i o n a l l y s y m m e t r i c c o o r d i n a t e s y s t e m d e f i n e d by particles
a t zeroth
order.
t h e c i r c u l a r p a t h s of
These e q u a t i o n s a r e d e r i v e d and
t h e i r p r o p e r t i e s a r e d i s c u s s e d and some r h e o l o g i c a l problems a r e s o l v e d by J o s e p h
(1977).
I n the derivation of
t h e c a n o n i c a l f o r m s of
e q u a t i o n s o f m o t i o n we assumed t h a t
5 =
Xt(&,t,~)
of
t h e domain
problem,
labels
u s e d i n t h e backward i n t e g r a t i o n o f p a t h l i n e s
a r e independent
xt(,,T,e)
the particle
the
b,(t)
of
If
E.
d e p e n d s on
t h e boundary
a b C( t )
as i n a f r e e s u r f a c e
C ,
we p r o c e e d a s i n t h e model problem e x c e p t
that
the
mapping f u n c t i o n w i l l depend p a r a m e t r i c a l l y o n t h e t i m e ;
i s , we map
b,(t)
into
bo
where
t i m e w i t h a n i n v e r t i b l e mapping
bo
i s independent
that
of
5 = ~ ( ~ o , tp o, s ~ s e s)s i n g
D A N I E L D.
282
the properties and i f
that
E abo,
x
-0
xo then
JOSEPH
z ~ ( & ~ , t , O ) i s i n d e p e n d e n t of t i m e , 5
E abc(t).
A f t e r c a r r y i n g out
a n a l y s i s l i k e t h a t g i v e n i n $ 2 , we c a n d e r i v e e x a c t l y t h e same c a n o n i c a l e q u a t i o n s w i t h
b0.
a r e posed o n t h e domain s e t on
b0
x
-0
replacing
5.
A l l equations
The f r e e s u r f a c e e q u a t i o n s a r e
a n d , s i n c e t h e r e a l f r e e s u r f a c e depends on t i m e ,
t h e s e e q u a t i o n s c a n c o n t a i n d e r i v a t i v e s of t h e boundary v a l u e s of t h e mapping f u n c t i o n . The main a p p l i c a t i o n s o f
the theory j u s t described
t o f r e e s u r f a c e problems have s o f a r been c o n f i n e d t o
1977
p e r t u r b a t i o n s o f t h e r e s t s t a t e ( s e e J o s e p h 6% B e a v e r s ,
The m o s t s u c c e s s f u l
for a r e v i e w of t h e s e a p p l i c a t i o n s ) .
a p p l i c a t i o n s o f a r h a s been t o rod climbing ( t h e Weissenberg effect).
O n t h e b a s i s of
t h e s e c o n d o r d e r t h e o r y a l o n e we
h a v e b e e n a b l e t o e x p l a i n and e v e n t o p r e d i c t many of n o v e l e f f e c t s which a p p e a r i n t h e movie by G . S . myself.
B e a v e r s and
These e f f e c t s a r e p e c u l i a r t o non-Newtonian
l i k e STP.
They i n c l u d e c l i m b i n g on r o t a t i n g r o d s ,
bifurcation
( t h e b r e a t h i n g i n s t a b i l i t y ) of
symmetric c l i m b ,
the c r i t i c a l radius,
t e m p e r a t u r e and achesion remove g r a v i t y on e a r t h
b u c k l i n g of
fluids
a Hopf
the steady axi-
t h e b i g e f f e c t of
a normal s t r e s s a m p l i f i e r
,
the
f u i d towers,
(how t o t h e mean
c l i m b on a n o s c i l l a t i n g r o d and t h e symmetry b r e a k i n g b i f u r c a t i o n of a x i s y m m e t r i c t m e - p e r i o d i c
li t y )
f l o w ( t h e flower instabi-
. T h i s w o r k was s u p p o r t e d by t h e U . S .
O f f i c e and u n d e r N S F G r a n t
19047.
A r m y Research
283
C O N S T I T U T I V E E Q U A T I O N S AND F R E E S U R F A C E S
References Coleman, R., and Noll. W. ( 1 9 6 1 ) . Viscoelasticity. Green, A.E.,
Foundations of Linear
33, 239. Rev. Modern Phys. -_
and Rivlin, R.S. ( 1 9 5 7 / 5 8 ) .
The Mechanics of