ON GENERALIZED REYNOLDS TRANSPORT THEOREM

J. P. Jaric1 , R.D. Vignjevic2 and S. Mesarovic3

1
Matematicki faklultet, Univerzitet u Beogradu, Beograd, Jagiceva 5
2
Structural Integrity Theme, Brunel University London, Kingston Lane, Uxbridge UB8 3PH
3
School of Mechanical & Materials Engineering, Washington State University, PO Box 642920
Pullman, Washington 99164-2920, USA

Abstract. Here we present a general form of Reynolds transport theorem in n-dimensional Riemannian space
VN . We assume that the Riemannian subspace V t V changes its position continuously with time t
M N

in VN . For a field x, t defined on R t V t we calculate

M M

d
x, t dv
dt R t
M
M
making use of the notion of convected coordinates. In this way we derived transport theorem for subspaces of
VN . The derived transport theorem is presented in different forms: independent and dependent of
parameterization and in a unified form. It was applied to some cases typically used in differential geometry and
continuum mechanics in order to validate new developments.
1.Introduction

Reynolds transport theorem is a fundamental theorem used in formulating the basic

conservation laws in continuum mechanics. These conservation laws (law of conservation of
mass, law of conservation of linear momentum, and law of conservation of energy) are
adopted from classical mechanics and thermodynamics where the system approach is
normally followed. Analogous to the classical Reynolds transport theorem in continuum
mechanics, the surface transport theorem is essential in the study of thin films undergoing
large deformations, epitaxial growth and phase boundary evolution. It is also important in the
modeling of a moving singular surface which carries a certain structure of its own as it
migrates through and interacts.
There is a vast amount of literature on transport theorem. A large number of references
can be found in books written by J.C. Slattery (2007), (2005). However they do not include more
recent contributions and approaches. We'll take a brief look at some that are relevant to our
approach, i.e. those directly related to the generalization of a transport theorem known in the
literature under the name Reynolds theorem.

We start with D.E. Betounes (1986) who presented a general transport theorem
connected with the motion of an arbitrary p -dimensional submanifold (the continuum) in
a given n -dimensional semi Riemannian manifold (the ambient space). The study provides a
natural generalization and unification of the classical kinematical treatments of moving
particles, curves, surfaces and solids as submanifolds of R 3 ( p 0,1, 2,3 and n 3 ). The
paper emphasizes the significance of the mean curvature normal in the transport formulas for
all dimensions which hereto has been noted for two dimensional submanifolds (surfaces), and
relates this to submanifold divergence operator.

1
While reach in geometrical insights, this approach seems to lie beyond the grasp of the large
number of researches and users of continuum mechanics.

In a paper published in the same year H. Petryk and Z. Mroz (1986) derived the
expressions for the first and second time derivates of integrals and functionals defined on
volume and surface domains which vary in time. The derivatives of integral functionals
defined on varying domains are of importance in studying stability conditions for damaged
structures, in the analysis of phase-transition processes or propagation of discontinuity
surfaces. The result presented in their paper is somewhat more general as it pertains to
piecewise regular surfaces and contains the edge terms. Also they stated that the second
derivatives of volume or surface integrals of functionals defined on varying domains have not
been studied.
P. Cermelli, E. Fried and M.E. Gurtin (2005) proved a transport theorem for smooth
surface S t which evolves with time t in Euclidean space, expressed in terms of the
parameter independent derivatives.
In general, the challenge of deriving transport theorem can be accomplished in many
ways, by the use of a number of integral transport theorems, each of which corresponds to a
particular reference frame in which the time rate of an integral quantity can be measured. For
instance M. E. Gurtin, A. Struthers & W. O. Williams, (1989) develop a space-time approach.
Following this approach a general transport theorem for moving surface St R3 within a non-
fixed region R3 was derived by J.P. Jaric (1992).
With help of the theory of generalized derivatives R.E. Estrada and R.P. Kanwal (1991)
presented a general transport theorem for moving surface S (t ) R3 .
G. Scovazzi, T.J.R. Hughes (2007), R. Fosdick and H. Tang (2009) made an attempt
to present a unified approach to continuum mechanics on arbitrarily moving domains. They
developed a space-time approach which incorporates the more traditional consideration of
time and space as separate entities. They point out that it is desirable to formulate the
transport theorem in a single unified way, using classical approach, in terms of standard
quantities from differential geometry and explicitly displaying the features that are common
to all submanifolds regardless of their finite dimensions.
P. Lidström (2010) stated " that a generalization of the present discussion on moving regions
in three-dimensional Euclidean space to the n-dimensional Riemannian manifold setting could
be an interesting undertaking".
The work presented in the above publications was the main motivation and the starting
point for our contribution described in this paper. Our objective was to drive transport
theorem in unified way in terms of standards quantities from differential geometry in
V V dimensional Riemannian spaces.
M N

In order to achieve this objective, in Sections 2 and 3, we review and derive relevant
geometric and kinematic formulas of Riemannian spaces in an intrinsic coordinate system, as
well as in a convected system. The main forms of the Generalized Reynolds transport theorem
in VN are presented in Section 4. The theorem was derived in invariant forms for the used
coordinate systems. Section 5 provides explicit forms for several known cases used in
continuum mechanics. We present transport theorem depending on parameterization In
Section 6.
So derived transport relation enables us to consider in a unified way several particular cases of
importance in continuum mechanics. We remind the reader that in the process of deriving
these relations we used in abstract way the notion of “material body”, “material boundary”
and “ flow” in the space of higher dimension than three. In Section 7 and 8 we consider a

2
moving subspace VM t and normal time derivative in VM t for “fluids”, in Section 9
"Material
VM t in fluids”. A derivation of transport relation for "surfaces of discontinuity" is given
in Section10. This allowed us to state General balance laws as given in Section 11. A short
conclusion given in Section 12. Appendix provides some elements of geometry and
kinematics of the boundary RM t of RM t .

2. Geometry of V t V
M N

Here and throughout i, j,k assume values 1 to N , the Greek letters , , , , ¸

, assume values from 1 to M ; , range from 1 to M 1 and , range from 1
to N M . The range and summation convention apply unless an explicit statement to the
contrary is made. When subscript or superscript is in brackets, e.g. it is just a number and
not an index.

Let us consider an N - dimensional Riemannian space VN with positive definite

metric tensor gmn given in an allowable coordinate system x k , k 1,2,.., N . We shall
denote by x  x k  a typical point in VN . By definition, vectors
x
gk  (2.1)
x k
represent basis of VN . The same holds for its reciprocal basis g k defined by g j gk k
j ,

j, k  1,2,.., N , gmn gk gl is positive definite metric tensor of VN ..

Let V be a subspace of VN , i.e. V V .

M M N

(Our considerations are confined to Euclidean space E N as the space of physical events.)
Further, we assume that the Riemannian subspace V changes it position continuously with
M

time t in VN . We emphasize this by writing VM t and u for its intrinsic coordinate

system. Then

V t : x x u,t , (2.2)
M

where u u denotes typical point of VM t , or in componential form

xk
xk x k u ,t , rank M, (2.3)
u
By elimination of u in (2.3), we obtain equivalent representation of VM t , i.e.
VM t : f x k ,t 0, (2.4)
where
f
rank N M. (2.5)
xk

3
Simple example M=1, N=3.
The family of curves given in parametric form
x1 a t cos u , x 2 a t sinu , x 3 b t sin 2u
in E 3 , may be represented as the intersection of circular cylinders
1 2 2 2
f x 1,x 2 ; t x1 x2 a t 0
and hyperbolic paraboloids
2 b t
f x 1,x 2 ,x 3 ; t x 1x 2 x3 0
2
a t

Each of these representations for VM t has some advantage and further we shall use both of
them. It is important to note that (2.3) is very useful in considering kinematics of VM t
although it depends on the way it is parameterized. For instance, x k x k c , t represents
the path of a point x of VM t in V , with respect to fixed parameters u , i.e. when
N

u c . On the other hand, the representation (2.4) is independent of parameterization with

respect to intrinsic coordinate system u . It means that any transformation of intrinsic
coordinate systems
u
u u U ,t , J U ,t 0
U
does not change the representation (2.4). As a consequence the set of vectors gradf are
independent of parameterization with respect to intrinsic coordinate system u . Moreover,
they are linearly independent, because of (2.5) and may be taken as the basis vectors of the
space VN M , the complement of space VM t to the space VN .
Now
VM t : x  x  U, t   x  u  U, t  , t  , (2.6)
and from this we have
x x u u
A = a , (2.7)
U u U U
x xi x
where a u , t gi and A U , t are the basis vectors of VM t
u u U
with respect to coordinate systems u and U , respectively. Thus these basis vectors depend
on parameterization.

Moreover, assuming that (2.3) satisfies (2.4) it follows that

f xk
gradf a 0 (2.8)
xk u
from which we conclude that gradf are orthogonal to a and hence to any vector in
VM t . They are also orthogonal to covariant basic vectors a in VM t defined through the

4
relations a a . Since the set of vectors a and gradf are linearly independent
they may be taken as the basis vectors of the space VN . It is important to notice that unit
vectors
n gradf / gradf , (2.9)
are also independent of parameterization.

Moreover, because of (2.8),

n a 0. (2.10)

Further we shall make use of n , which are the reciprocal vectors of the vectors n in
VM t , defined by n n , , 1,....,N M as n ,n are orthogonal to
VM t . Generally, the vectors n , as well as n , are not orthonormal. In VN M we may
consider n (or n ) as a nonholonomic basis.
Let
n n n , n n , det n 0 (2.11)
Also we write n n n and then
n n (2.12)
and
n n n , n n n . (2.13)
More generally, coefficients n and n may be used as a tool for rising and lowering
indices for the quantities defined in the nonholonomic basis n or n . Note that n ,n
and vectors n do not depend of any parameterization of VM t .
Then vectors g k at the points x of VM t may be decomposed as
gk x,k a α nk n , (2.14)

xk
where x,k gk aα and n k gk n
. This relation is of crucial importance for
u
the decomposition of any tensor quantity defined on VM t to its components in VM t and
VN M .
Often we shall make use of the relation between metric tensors of VM t defined in
the two coordinate systems u and U , i.e. the relation
u u
A U ,t a . (2.15)
U U

xk xl
where, by definition, A t A A and a t a a gkl .
u u
Then from (2.15) follows

5
 A = A U ,t J 2a(t ) , (2.16)
where A U , t det A 0, a t det a 0.

And finally, we are in position to define the volume element dv t of VM t which is

M

used in the derivations to follow.

1
dv t a 2dτ t , (2.17)
M M

where
dτ t da u (2.18)
M

is the extension (see J.L. Synge, A. Schild (1978)).

3. Kinematics of V t V
M N

Making use of (2.4) we obtain

f
gradf v = 0, (3.1)
t
where gradf v is directional derivative of f v in the direction of velocity v
dx
v (3.2)
dt
d
For definiteness here and further by we denote time derivative of any quantity along
dt
u const . Notice that
f
V v n t . (3.3)
gradf
represents the scalar-normal velocity in n direction which is independent of
parameterization. Having this in mind we shall write
v v a V n v tan V n , (3.4)
Here
v tan v a , (3.5)
denotes the tangential component of v in VM t (This notation for vectors in VM t is used
throughout the paper). v tan obviously, depends on parameterization since a do.
Note that V n is the same for any parameterization and that v differs for different
parameterizations only through vtan . This property of vtan is very important for various
forms of the transport theorem.

Hereafter we shall consider the coordinate system U A as convected system in VM t .

In continuum mechanics terms material or Lagrangian coordinates are used since the material

6
particles can be labeled by these coordinates and, as such, they do not change their values
during the motion of the material body. Also, at any particular time t , the relation
u
u u U ,t , J U ,t 0 (3.6)
U
defines the coordinate transformation between the convected coordinate system U and
intrinsic coordinate system u . It is very important to mention here that, although fixed, the
choice of convective coordinates U A is arbitrary. Further we consider
x
V (3.7)
t
as the velocity of x of VM t in V along the path of convected coordinates U , i.e.
N

U C , where C are arbitrary constants. Generally, we shall denote by the time

t
T U ,t T U ,t
derivative of any quantity along U C . Of cause for any
t t
x x
quantity T which is explicit function of U . Particularly in this case .
t t
From (2.6) we obtain the relation between velocities v and V of the point x of VM t , i.e.
x x dx x u u
V v a . (3.8)
t t dt u t t
Obviously, from (3.8), (3.3) and (2.10) we conclude that
f
V V n v n t , (3.9)
gradf
proving that the scalar-normal velocity in n direction is independent of parameterization.
Instead of (3.8) more compact representation is given by
V V n Vtan , (3.10)
where
u
Vtan v a . (3.11)
t
It may be also written as
Vtan V Α , (3.12)
where
u U
V v . (3.13)
t u
We may also write
V V n Vtan , (3.14)
where
V n V . (3.15)
In particular, for the hypersurface , i.e. when M N 1,

7
 11 1
n n 11 1 and V 1 V .

4. Generalized Reynolds transport theorem in VN

Let x, t be defined on RM t . We want to calculate

x, t dv (4.1)
t M
RM t

along the path of convected coordinates U .

u
Since d a u d a U , then
U
1 1 1
dv =a 2 d =a 2 d a u =A2 d a U (4.2)
M M

Next, making use of (2.6) we write

x, t x u U, t , t , t U, t . (4.3)
Thus in view of (4.2) and (4.3) we have

1 1 1
x, t dv = A2 d U A2 A2 d U , (4.4)
t M t a t t a
RM t M M

where M is defined with respect to convected coordinates U , and therefore the integration
with respect to U is independent of t .
Now,
1
1 21 A
A2 A A , (4.5)
t 2 t
where A A .
It is convenient to write (2.15) in equivalent form
A A A .
Then
A A A V V
A A A A A A , (4.6)
t t t t U U

where , according to (3.8), we have that

A x x V
.
t t U U t U
Thus
V V A A
A V . (4.7)
U U U

Next

8
 A
A n (4.8)
U
and
V A V . (4.9)
Formally we may write (4.8) as
A , n , (4.10)
where
, (4.11)
Here are the components of symmetric tensor of the second order for each value of
(see David Lovelock, Hanno Rund, 1989, p.270).
or
A , n , (4.12)

where
n . (4.13)
Next we calculate (4.7) and obtain

V V
A V V V , V ,
U U
where we made use of (4.8) and (4.9).
Therefore (4.6) is given by
A
V , V , 2V (4.14)
t
Now
1
1 A 1 21 A 1
A2 1
AA A2 divVM Vtan V K , (4.15)
t 2A 2
t 2 t
where
divVM Vtan A V, a V, , (4.16)

K( ) A (4.17)
( See R. Aris, 1962 Vectors, Tensors, and the Basic Equations of Fluid Dynamics; Prentice–
Hall).

Remark 4.1. In order to get the geometric meaning of K( ) we use (4.8) and obtain
A n
n A .
U U
Hence
n n
K( ) A A A A DivV t n
U U M

9
 n
a divV t n , (4.18)
u M

and it does not depends of any parameterization.

Further,

U, t x, t grad V
t t t
grad V A V n
t
V , V grad n
t
Vtan Grad V V ,
t M
n

where Grad V , A and grad n . The term

M
n

V (4.19)
t n

represents the normal time derivative of following convected VM t . Obviously is

independent of parameterization.
Remark 4. 2. To justify the name it is useful to write V n in a unique compact invariant
form as
Vn n V n ,
where n is a unit vector orthogonal to RM t . Then
1/2 1/2
Vn V V n n n V V ,
and
n V / Vn n ,

Accordingly we have simple expression for V , i.e.

V Vtan Vnn .
Since Vtan n 0 we have that
Vn V n
which is of the form of classical expression for the scalar normal-velocity of the surface
evolving in three dimensional Euclidean space.
Then

V V grad n Vn grad n
t n t t
(See P.Cermelli, E. Fried and M.E. Gurtin (2005), Roger Fosdick and Huang Tang (2009)) .
u
Remark 4. 3. Making use of a = A it easy to see that
U
Grad V , A = , a
M
grad V t . M

10
Also, DivV Vtan M
divV M
t Vtan . We used them when we wont to emphasize what coordinate
are used. Therefore

U,t x,t Vtan Grad V Vtan grad V (4.20)

t t M M

Substituting (4.18) and (4.20) into (4.4) we obtain

1
x, t dv Vtan Grad V DivVM Vtan V K A2 d U
tR t
M
RM
M

Vtan grad V t divV t Vtan V K dv .

M M M
RM t
But
Vtan grad V M
t divVM
t Vtan divV M
t Vtan ,
and hence

x, t dv V K divVM t Vtan dv .(4.21)

tR t
M
RM t
M
M

Further, using the divergence theorem we obtain (when the boundary consists of "material
points", Appendix A.6)

x,t dv V K dv Vtan ν dv , (4.22)

t M M M -1 M 1
RM t RM t RM t

where ν is normal to RM t RM t but tangent to RM t . Notice that the term

M -1

V Vtan ν (4.23)
M -1

on right side, under the integral RM t is intrinsic to the motion of RM t . According to

P. Cermelli at all. [3] we refer it as the scalar normal-velocity of RM t RM t .

Thus our transport theorem with respect to convective coordinate reads as

x,t dv V K dv V dv (4.24)
t M M M 1
RM t RM t RM t

We underline again that the first integral is intrinsic to RM t while the second integral is
intrinsic with respect to RM t .
Since the first integral on the right-hand side of (4.24) is invariant to any
parameterization we may equally applied it for parameters u . In this case the scalar normal-
velocity of RM t with respect to parameters u , i.e. v R t M 1
, needs to be used
M

(Apendix A14). Then we write basic generalized transport theorem with respect to
nonconvective coordinate u in the form
d
x,t dv V K dv v R t M 1
dv .
dt R t M
R t
M
R t
M M 1
M M M

(4.25)
11
 5. Examples
We proceed by writing transport theorem for several cases using either (4.24) or (4.25).
We start with (4.24) , mostly used in continuum mechanics, and then with (4.25) by remark
of this section.

I. First we take VN as tree dimensional Euclidean space E 3 , i.e. V3 E3 and N 3.

i. “Body” u E 3 .
M 3.
0
Then N M 0, K 0 and V 0.
Let R3 t v t , R3 t v t , dv dv and dv da . Then from (4.24)
3 2

we have well known transport theorem

x, t dv dv V da , (5.1)
tv t v t
t v t

where V Vtan ν ; ν is unit vector in v t orthogonal to v t . Classical case.

ii. Surface in E 3 .
M 2 . Then N M 1.
1
Let V2 t S t , R2 t A t ,K 1 K, V V , dv da , and
2

dv ds .
1

Then form (4.24) we have

x, t da VK da V dl , (5.2)
tAt At At

Here V Vtan ν ; ν is unit vector in A t orthogonal to A t , K divS t n the

total (i.e. twice the mean) curvature of S t , and n is unit-normal vector of S t .
(see . Slattery, J.C (2007), M.E. Gurtin, A. Struthers and W. O. Williams 1989, Jovo P.
Jaric 1992).

iii. Curve in L t E3 .
M 1.
a) General case.

Then N M 2 . Let l t L t and

R1 t l t : xi x i s, t ,

where s is arch length. Equivalently, the intersection of S t : f x k ,t 0

, 1,2 , define l t .

12
 We shall denote by n and K divR t n their unit vectors of S t
1

and total curvature, respectively. V is defined through (3.9) and dv ds .

1

Then form (4.24) we have

1 2 t
x, t ds V K 1 V K 2 ds V |t2 (5.3)
tl t l t
1

where

V grad V n , 1,2
t n t
V Vtan t .
Note that here ν t is unit tangent vector of the curve l t .

b) Curve in plane.
In particular, if the curve l t is plane then one of K is zero. For definiteness,
1
let K 1 K ,V V and K 2 0 . Then

t
x, t ds VK ds V |t2 , (5.4)
tl t l t
1

V Vgrad n,
t n t
where n is a unit normal vector of the plane curve l t ( See Gurtin, M.E., 1993,
1995, N.K. Simha, K. Bhattacharya , 2000).
Without loss of generality we may define l t with respect to the coordinates x i ,
x t
i 1,2 , i.e. l t : x i x i s, t , a1 s, t
t , a1,1 k n and K k
s s
is the curvature of l t (See P.Cermelli, E. Fried and M.E. Gurtin 2005)
c) If the curve l t is straight line, i.e. the intersection of two planes then

K1 K 2 0, and
t
t
s, t ds ds V |t2 . (5.5)
tl t l t
t 1

This is classical Leibnitz integral formula.

II. Part of surface moving on the surface N M 2,

Let R2 t S t , R2 t S t , dv da , dv ds . Then
2 1

and
t

13
 x, t da da V ds , (5.6)
tS t S t
t S t

where V Vtan ν ; ν is unit vector in S t orthogonal to S t ( P.Cermelli,

E. Fried and M.E. Gurtin, 2005, Roger Fosdick and Huang Tang 2009).

III. Curve on the surface : N 2 ,M 1

1
Let R1 t l t ,K 1 K, V V , dv ds .
1

Then
t
x, t ds VK ds V |t2 , (5.7)
tl t l t
1

where V Vgrad n and V Vtan ν ;

t n t
n is a unit normal vector of the curve l t in S t R2 t , ν  t , where t is
unit tangent vector of the curve l t and hence V Vtan t ( See Gurtin, M.E., 1993,
Gurtin, M.E, and Podio-Guidugli, P. 1998, N.K. Simha, K. Bhattacharya , 2000). Notice
that (5.5) and (5.6) are of the same form. In both cases the curve is a subspace of the
surface although in the first case the surface is the plane.

Remark 5.1 In all these cases transport theorems holds also for the coordinates
d
u , i.e. (4.24) , when we substitute with and V V V tan with
t dt
v R t M 1
as it is defined in Appendix. The results stated in Appendix
M

allow us, particularly, to calculate v RM t M 1

explicitly for one dimensional spaces.

6. Transport theorem depending on parameterization

In some cases it is useful to use transport theorem in terms depending on parameterization.

This can be done simply if we substitute (4.20), now written as,
x,t Vtan grad V (6.1)
t M

into (4.21) or (4.24). Here we confine ourself to the case given by (4.21), since it is more used
in practice . It is convenient to to put it in more familiar form.
Then

x,t dv x,t Vtan grad V div VM t Vtan V K dv

tR t
M
RM t
t M M
M

But
Vtan grad V M
t divVM t Vtan Vtan grad V M
t grad V M
t Vtan divVM t Vtan divVM t Vtan
so that
14
 x,t dv divVM t V V K dv .(6.2)
tR t
M
t M
M M

Further,
V K V divV t n divV t V n gradV t V n
M M M

divV t V n
M

since gradV t V n 0.
M

Hence
divVM t Vtan V K divVM t Vtan V n divVM t V
Accordingly

x,t dv x,t divVM t V dv (6.3)

tR t
M
RM t
t M
M

(See Slaterry , J.C. 2007, pp.61, (1.3.3-6)).

7. Migrating VM t in “fluids”

The transport relation given by (4.25) permit us to consider several particular cases of
importance in continuum mechanics in a unified way. We remind the reader that in the
process of deriving transport theorem we used in abstract way the notion of “material body”,
“material boundary” , “ flow” in the space of higher dimension than three. Derived a transport
theorem in that form exhibits no influence on the flow field. We shall proceed by deriving
alternative versions of (4.25) that account for that influence. First we denote by w x,t the
velocity of “fluid”. We may write it as
w W n w tan , (7.1)

where
w tan w a . (7.2)
Then vr v w represent is the relative velocities of motion of VM t with respect to
“fluid”. We may decompose it as
vr v w= V W n v tan w tan .
Now looking at the transport theorem the only term which may be influenced by motion of
“fluid “
is v R t M 1
. For definiteness we write
M

vr v RM t w (7.3)
and
V mig
R t
vr M 1
(7.4)
M

15
for the relative velocities of motion and normal migrational velocity of RM t with respect
to “fluid”, respectively.
Then , since n M 1
0,

V mig
R t
vr M 1
v RM t M 1
w tan M 1
,
M

and hence
v RM t M 1
V mig
R t
w tan M 1
. (7.5)
M

Substituting this into (4.25) we obtain

d
x,t dv V K dv V mig
R t
w tan M 1
dv
dt R t M
R t
M
RM t
M M 1
M M

.
Next we using the divergence theorem,
w tan M 1
dv divV t w tan dv ,
M 1 M M
RM t RM t

we obtain transport theorem for migrating VM t in “fluid”, in the form

d
x,t dv V K divV w tan dv V mig
R t
dv
dt R t
M
RM t
M M
RM t
M M 1
M

(7.6).

8. Migrating normal time derivative VM t in “fluids”

The normal time derivative of following convected VM t

V
t n
has to be modified for the migration VM t through the “fluids”. We shall start from the
expression (7.6) written as
d
x,t dv V K grad w tan div VM t w tan dv
dt R t M
R t
VM t M
M M

V mig
R t
dv
M M 1
RM t
Taking into account that
k
grad = ,k g ,k x,k a α nk n , aα n grad VM t n
n n
then
grad w tan grad w tan
VM t

since n w tan 0 . Hence

16
 grad w tan V grad w tan
VM t
t n

grad V n w tan grad V,

t t
where
V = V n w tan . (8.1)
represents migrational normal velocity for VM t computed relative to "fluid". Indeed, it is
normal when computed relative to "fluid" since
V w V -w n .
Let

grad V grad w tan . (8.2)

t VM t

We call it normal time derivative of x, t following VM t corresponding to

migrationally normal velocity field V = V n w tan .
Therefore corresponding transport theorem reads as
d
x,t dv V K div VM t w tan dv V mig dv
dt R t M
R t
M
RM t
R t
M M 1
M M

(8.2).

9. "Material VM t in fluids”

Let us assume that VM t "is material", i.e convected to the "fluid". Then "fluid" velocity w
is the "material" velocity for VM t , i.e. V w , the migrational normal velocity field for

VM t coincides with "fluid" velocity, w = V n w tan , x,t ,

t
V mig
R t
0.
M

Then transport theorem reads as

x,t dv x,t w n K divVM t wtan dv .(9.1)

tR t
M
RM t
t M
M

Note that K divV t n . Then

M

w n K w n divV t n
M

= divV t w n n n grad w n divV t w n n

M VM t M

so that
w n K divV t w tan divV t w n n w tan divV t w.
M M M

Therefore the above transport theorem becomes

17
 x,t dv x,t divVM t w dv . (9.1)
tR t
M
RM t
t M
M

As a very special case we consider VM t as "material" but RM  t  "non-material". Then

V mig
R t
0 and
M

d d
x,t dv x,t w n K divVM t wtan dv V mig dv
dt R t
M
RM t
dt M
RM t
R t
M M 1
M

,
or
d d
x,t dv x,t divVM t w dv V mig dv
dt R t
M
RM t
dt M
RM t
R t
M M 1
M

(9.2)

10. Surfaces of discontinuity

Usually in continuum mechanics the fields representing properties of the material composing
a body vary smoothly throughout space. However, there may exist surfaces across which the
variation occurs so rapidly that they can be regarded as being surfaces of discontinuity.
Examples of such singular surfaces are the interfaces within a heterogeneous media, shock
waves in gases, vortex sheets in fluids separating a main stream from the dead water, and
cracks in solids. Moving surfaces of discontinuity are often referred to as a waves. In these
cases Reynolds transport theorem has to be modified in order to include the influence of
surfaces of discontinuity t on

x,t dv .
tv t
We shall generalized transport theorem for the integral

x,t dv
tR t
M
M

to a region RM t containing a hypersurface M 1 t across which x,t may suffer a

jump discontinuity.

Definition. An oriented smooth surface M 1 t in a region RM t is called a singular

hypersurface relative to a field x,t defined on RM t , if x,t is smooth in

RM t M 1 t and suffers a jump discontinuity across M 1 t . The jump of x,t is
defined as
(10.1)

18
where and are the one side limits from the two regions of RM t separated by

M 1 t and designated as RM t and RM t respectively.

Let
M 1 t : u ,t 0. (10.2)
Denote by M 1 its outward unit normal and by ω velocity of its points. Then

ω M 1 is the normal speed of M 1 t with the directions pointing into RM t .

Applying (4.22) to the two regions RM t and RM t bounded by RM t and M 1 t

and RM t and M 1 t , respectively, we write the transport theorem as

x,t dv = V K dv
tR t
M
RM t
M
M

Vtan M 1
dv ω M 1 dv . (10.3)
M 1 M 1
RM t M 1 t
or, using Green- Gauss theorem, as
x,t dv V K divV t V dv V ω M 1 dv
tR t
M
RM t
M M
t
M 1
M M 1

.(10.4)

11. General balance laws

Reviewing physical principles of continuum physics it is evident that the physical

balance principles all take the form of integral equations. It has been shown that from these
balance principles can be derived corresponding field equations by way of suitable conditions
of continuity.

The basic laws of mechanics can all be expressed, in general, in the following form,
d
dv nda dv (12.1)
dt P Pt Pt
t

for any bounded regular subregion P of the body B and the vector field n , the outward unit
normal to the boundary of the region Pt in the current configuration. The quantities and
are tensor fields of certain order m , and is a tensor field of order m 1.
The relation (12.1) , called the general balance of in integral form,is intrpreted as asserting
that the rate of increase of the quantity in a part P of a body ia affected by the inflow of
through the boundary of P and the growth of within P . We call the flux of and
the supply of . In general, the supply may contain contributions of supplies from
external sources and internal productions due to the motion of the body (C. Truesdell and R. A.
Toupin 1960).

19
 We now proceed to state general balance laws for V t V containing a
M N

hypersurface M 1 t across which x,t may suffer a jump discontinuity. Inspired

by the above stated balance law in this case we write

x,t dv n RM t dv dv . (12.2)
tR t
M
RM t
M 1
RM t
M
M

Here n RM t is now the outward unit normal to the boundary RM t of RM t (in VM t );

and are the flux and the supply of , respectively.
From this equation we can obtain the local balance equation at a regular point as well
at a singular point. First we apply the Green-Gauss theorem of the first integral on the right
side of (10.1) and obtain that
ndv divVM t dv M 1 dv .(12.3)
M 1 M M 1
RM t RM t M 1 t
( See C. A. Eringen,1980, pp 524, A2.3)

Then substituting this and (11.4) into (12.2) we obtain

V K divV t V dv V ω M 1 dv
M M M 1
RM t M 1 t

divVM t dv M 1 dv dv ,
M M 1 M
RM t M 1 t RM t
and from this

V K divV t V dv V ω M 1 dv 0
M M M 1
RM t M 1 t

.
Finaly , in standard procedure, C. Truesdell and R. A. Toupin 1960, we are able to write local
balance law as

V K divV t V 0 in RM t , (12.4)
M

and jump conditions

V ω M 1 on M 1 t . (12.5)

12. Conclusion
Thus the objective of our paper is to drive transport theorem in unified way in terms of
standards quantities from differential geometry inV V dimensional Riemannian spaces.
M N

We present transport theorem in a single unified way by using classical approach expressed in
terms of standards quantities from differential geometry and explicitly displaying the features
that are common to all submanifolds regardless of their finite dimensions.
In order to applied it in continuum mechanics we applied both coordinate systems: convected
U coordinate as well as any arbitrary u system on V V . This is illustrated in Section
M N

5. We present transport theorem depending on parameterization. We consider migrating

VM t in “fluids”, migrating normal time derivative VM t in “fluids”, "Material VM t in
fluids”. We generalized transport relation for "surfaces of discontinuity" and then stated General
20
balance laws. We underline the importance of Appendix where some elements of geometry
and kinematics of the boundary RM t of RM t are derived in order to clarify the
influence of the motion of the boundary surface in transport relation.
Every Section can be used to write explicit form of there derived expressions of any order. In
fact we can write tables of these expressions.

References
1. R. Aris, Vectors, Tensors, and the Basic Equations of Fluid Dynamics. Prentice–Hall,
1962
2. D.E. Betounes, Kinematics of submanifolds and the mean curvature normal, Archive for
Rational Mechanics and Analysis, Volume 96, Issue 1, pp.1-27, 1986
3. P. Cermelli, E. Fried and M.E. Gurtin., Transport relations for surface integrals arising in
the formulation of balance laws for evolving fluid interfaces. Journal of Fluid Mechanics,
544, 339-351 (2005)
4. Roger Fosdick and Huang Tang, Surface Transport in Continuum Mechanics,
Mathematics and Mechanics of Solids, August 2009, Vol. 14, 6, pp. 587-5987.
5. C. A. Eringen, Mechanics of Continua, Robert E. Krieger Publishing Company,
Huntington, New York, pp 524, A2.3, 1980
6. R. Estrada, & R. P. Kanwal, Non-classical derivation of the transport theorems for wave
7. Fronts, J. Math. Anal. Applic. 159, 290–297, 1991
8. M.E. Gurtin, Thermomechanics of evolving phase phase boundaries in the plane, In
Oxford, 1993
9. M.E., Gurtin, The nature of configurational forces, Arch. Rat. Mech. Anal., 131, 67-100,
1995
10. M.E Gurtin, & P. Podio-Guidugli, Confgurational forces and a constitutive theory for
crack propagation that allows for kinking and curving. J. Mech. Phys. Solids, Vol. 46,
No. 8, pp. 1343-1378, 1998
11. M.E. Gurtin, A. Struthers and W. O. Williams, A transport theorem for moving interfaces,
Q. Appl. Math. 47, 773-777, 1989
12. J.P. Jaric, On a transport theorem for moving interface, Int. J. Engng Sci Vol. 30, No 10,
pp.1535-1542, 1992
13. J.P. Jaric; Z. Golubovic, The Balance Laws of the interline and the Bulk Material, ZAMM
71 12, 518-521, 1991
14. P. Lidstrom, Moving region in Euclidean Space and Reynolds' transport theorem,
Mathematics and Mechanics of Solids 16(4):366-380, 2005
15. D, Lovelock, H. Rund , Differential forms, and variational principles, Dover edition,
1989
16. H. Petryk & Z. Mroz, Time derivatives of integrals and functionals defined on varying
volume and surface domains. Arch. Mech., 38, 697–724, 1986
17. G. Scovazzi, T.J.R. Hughes, Lecture Notes on Continuum Mechanics on Arbitrary Moving
Domains, Sandia National Laboratories, Albuquerque, New Mexico 87185 and
Livermore, California 94550, Printed November 2007
18. N.K. Simha, K. Bhattacharya, Kinetics of phase boundaries with edges and junctions in a
three-dimensional multi-phase body, Jornal of mechanics and Physics of Solids, 48, 2619-
2641, 2000 518-521
19. J.C. Slattery, Advanced transport phenomena, Cambridge University Press, 2005
20. J.C. Slattery, Interfacial Transport Phenomena, Springer, New York, 2007
21. J.L., Synge & A. Schild, Tensor Calculus, University of Toronto Press, Dover publication
1978

21
22. C. Truesdell & R. A. Toupin, The classical field theories, In S. Flugge, editor, Handbuch
der Physik, vol. 3/1, Berlin, Springer-Verlag, 1960

Appendix. Elements of geometry and kinematics of RM t .

So far we have considered some elements of geometry and kinematics of VM t .
E3 the motion of boundary of body contribute to transport theorem of
But we know that even in
some quantity defined on the body. In order to clarify if for general case of motion of RM t
with its boundary RM t we want to study id more closely. Thus we assume as we do up to
now that RM t be a finite region in VM t , bounded by the closed M 1 -space
RM t . We shall first consider some elements of geometry and kinematics of RM t
consisting of the same surface ("material ") particle. Its parametric equations in convected
coordinates of are
U U , (A.1)

or
U 0. (A.2)
Then Grad as well as its unit vector
Grad
M 1
(A.3)
Grad
are independent of any parametrization of in RM t . Note that vector M 1
is in
RM t and ortogonal to RM t . In view of (2.6) the same boundary RM t is given in
V t by
N

x x U ,t . (A.4)

Then
x x
V
t t
so that
x
M 1
V M 1
(A.5)
t

at the points of RM t defined by (A.1). From (3.10), having in mind that n M 1

0,
this can be written, as
x
M 1
Vtan M 1
(A.6)
t
which is independent of any parametrization of . It represents the speed of propagation of
RM t while its vector of displacement is given by

22
 x x
V RM t M 1 M 1 M 1 M 1
. (A.7)
t t
In the case of the coordinate system u that is not convective determination of the
speed of movement of the border is more complex. Then

u
RM t : u u ,t , rank M 1, 1,...,M 1, (A.8)
or
RM t : u ,t 0. (A.9)

Furthermore, grad RM t is vector in RM t normal to RM t which is independent of

parameterization with respect to intrinsic coordinates of RM t . The same holds for its
unit vector
M 1
grad RM t / grad RM t . (A.10)

We shall denote by u u c , t the path of point u u in RM t when c .

u
Then a represents the velocity of points of the path. From (A.8) and (A.9) we
t c

obtain
u
gradR t 0 ,
t M
t
and from this , in view of (4.10),

u t
M 1
(A.11)
t gradR t
M

u
We shall call M 1
the speed of propagation of in RM t in RM t . Obviously
t
it is independent of any parameterization with respect to intrinsic coordinates . Its
displacement vector v RM t is given by

u
v RM t v RM t M 1 M 1 M 1
t
(A.12)
( For this terminology we refer to Truesdeel and Toupin, CFT, pp. 499).
So far we consider kinematics of RM t as subspace of RM t .
Now substituting (4.8) into (2.2) we obtain representation of RM t in VN , i.e
x = x u ,t x u ,t ,t . (A.13)
Then
23
 x x x u u
v RM t v+
t c t u t c t

represents the velocity of the points of the paths c in V . From here we may conclude
N

u
that v RM t
v is relative velocity of RM t with respect to motion of RM t in
t
V . Further,
N

v RM t M 1
v M 1
v RM t .
From (3.4) we have v M 1
v tan M 1
. Then , since n M 1
0 we have
v RM t M 1
v tan M 1
v RM t . (A.14)
Therefore v RM t M 1
, i.e. the normal component of v RM t
is independent of

parameterization of RM t and thus intrinsic to its motion. This relation can be also used
to calculate its values when the motion of RM t and RM t are known.

24