0% found this document useful (0 votes)

41 views

Introduction To Tensors

The document discusses constitutive equations in continuum mechanics. It states that five equations are obtained from conservation of mass and energy, while three more are obtained from linear momentum balance, for a total of eight equations. Constitutive relations provide the remaining eleven equations needed to solve for the sixteen unknowns in the system, including stress, displacement, temperature, and other fields. The constitutive relations must satisfy the principle of material frame-indifference, meaning they have the same form regardless of the observer's frame of reference.

Uploaded by

tewedaj

Available Formats

Download as PDF, TXT or read online on Scribd

0% found this document useful (0 votes)

41 views

Introduction To Tensors

Uploaded by

tewedaj

Available Formats

Download as PDF, TXT or read online on Scribd

You are on page 1/ 37

i i

“CM˙Final” — 2015/3/13 — 11:09 — page 203 — #225

i i

4 Constitutive Equations

The equation of mass conservation and the energy equation yield two equations, while
the linear momentum balance equation yields three equations (one for each displacement
component), thus giving a total of five equations. The balance of angular momentum is
used up implicitly by assuming the stress tensor to be symmetric, and thus having only
six independent components. The unknowns to be solved for are the density, displace-
ment field, stress tensor components, temperature, internal energy, entropy and conduc-
tivity vector, viz, ρ, u (3 unknowns), τ (6 unknowns), θ, e, η and q (3 unknowns), giving a
total of sixteen unknowns. So we need eleven more equations to solve for all the remain-
ing unknowns. These equations are provided by the constitutive relations, an example of
which is

τ = τ̃ ( F, θ ), (6 equations)
q = q̃(ρ, θ ), (3 equations)
e = ẽ(ρ, θ ), (1 equation)
η = η̃ (ρ, θ ). (1 equation)

The fact that the mathematical model developed so far is incomplete is also evident
on physical grounds. While the equations developed so far are valid for any material,
irrespective of whether it is solid, liquid or gas, it is clear that the nature of the underlying
material should be taken into account. For example, for a given set of forces, the resulting
deformations are different if the body is made of steel or of rubber.
In this chapter, we discuss the restrictions imposed on constitutive relations by the prin-
ciple of material frame-indifference, also known as the principle of material objectivity. This prin-
ciple states that constitutive equations must be invariant (in the sense that they have the
same form) under changes of frame. As an example, consider an object attached to the end
of a spring, and spinning at steady speed. It is tacitly assumed that the spring constant,
which has been found in a non-rotating condition, is unaffected by the rigid rotation. The
principle of material frame-indifference extends this assumption to general constitutive re-
lations, so that, for example, we postulate that the constitutive relation for the stress in a
given body is unchanged even if a rigid-body motion is superimposed on it.

4.1 Frame of Reference

A frame of reference is the background against which observations are made. A change of
frame of reference may be interpreted as meaning that two observers who are in arbitrary
motion with respect to each other, and have their clocks set differently, but who use the
same units of length and time, are equally qualified to describe a given phenomenon. A

Downloaded from https://www.cambridge.org/core. Indian Institute of Technolgy Bombay, on 26 Aug 2018 at 09:30:21, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/CBO9781316134054.005

i i

i i
i i

“CM˙Final” — 2015/3/13 — 11:09 — page 204 — #226

i i

204 Continuum Mechanics

z
y
Z x
x O frame
c(t)
Y
x∗
X
O∗ frame
Fig. 4.1 Frames of reference.

particular event to which a place and time ( x, t) are assigned by the framing O may be
assigned another place and timing ( x∗ , t∗ ) in another framing O∗ , as shown in Fig. 4.1. Let
c(t) be the vector connecting the origins of these two coordinate systems as shown. If x,
x∗ and c are all expressed in terms of the basis vectors of either the O or the O∗ frame,
then we have the relation x∗ = x + c; if ei∗ are the (Cartesian) basis vectors associated
with the O∗ frame, it follows that x∗ · ei∗ = x · ei∗ + c · ei∗ . However, in practice, it is more
convenient to express x in terms of the basis vectors of the O frame. Let ei denote the
basis vectors associated with the O frame, and let Q(t) = ek ⊗ e∗k , with components with
respect to the starred frame given by Qij = ei∗ · e j , i.e., Q = [e1 | e2 | e3 ]. Then we have
ei∗ = (ei∗ · e j )e j = Qij e j , so that x · ei∗ = x · ( Qij e j ) = Qij x · e j . If we use the symbols
x∗ , x and c to represent the components of these vectors with respect to the O∗ , O and
O∗ frames, respectively, i.e., if x∗ ≡ ( x∗ · e1∗ , x∗ · e2∗ , x∗ · e3∗ ), x ≡ ( x · e1 , x · e2 , x · e3 ), and
c ≡ (c · e1∗ , c · e2∗ , c · e3∗ ), then x∗ = Q(t) x + c(t). Assuming that the clock in O is set
differently from that in O∗ by an amount a, we thus have the following relations:

x ∗ = Q ( t ) x + c ( t ),
(4.1)
t∗ = t − a.

A change of frame of reference should not be confused with a change of coordinate axes.
In any given frame, we have an infinite number of coordinate systems, and the components
of a vector and a tensor with respect to any two different coordinate axes are related as
per Eqns. (1.99) and (1.100). However, the Q matrix in these relations is time-independent
(although, of course, v or T themselves could be time-dependent), in contrast to the Q
matrix in Eqn. (4.1), which is time-dependent. The invariance of the governing equations
under change of coordinate axes in a given frame of reference is ensured merely by writing
the equations in tensor form. Following Einstein’s ‘covariance principle’, however, we
postulate the more extensive principle that there are no preferred reference frames (such
as the ‘inertial frame’ in the classical setup), and that the axioms that we have stated have
the same form in reference frames that are in arbitrary motion with respect to each other.
We will investigate the implications of this principle in Section 4.4. We now consider a few
examples illustrating the first relation in Eqn. (4.1).
Consider the case shown in Fig. 4.2a. The frame O is fixed at the point under obser-
vation, and consequently, the position vector of this point is the zero vector with respect

i i

i i
i i

“CM˙Final” — 2015/3/13 — 11:09 — page 205 — #227

i i

Constitutive Equations 205

Fig. 4.2 Two observers in (a) translatory; (b) translatory and rotatory relative motion.

to O. On the other hand, since the observer in frame O∗ is translating to the left, the posi-
tion vector of the observed point is c(t), as shown in the figure. Thus, the position vectors
observed by O and O∗ are related as
x ∗ = c ( t ) = I ( 0 ) + c ( t ),
which is a special case of Eqn. (4.1) with Q(t) = I.
Now consider the situation shown in Fig. 4.2b. The observer in frame O∗ is stationary.
Assume that a Cartesian coordinate system in this frame is fixed at the center of the rotating
disc as shown. The observer in frame O has the coordinate system fixed at a distance
R, which is the distance of the mass from the center before the disc starts spinning. The
observer O spins with the disc. The position of the center of the mass with respect to the
observer in frame O∗ is
x ∗ = ( R + δ) cos θ = δ cos θ + R cos θ,
y∗ = ( R + δ) sin θ = δ sin θ + R sin θ,
z∗ = 0,
where δ is the elongation of the spring, and θ (t) is the angle made by the groove with the
horizontal. The coordinates of the center of the mass with respect to the observer in frame
O are x = δ, y = 0 and z = 0. Thus, once again, we see that the position vectors in the two
frames are related via Eqn. (4.1), with c(t) = ( R cos θ, R sin θ, 0), and
 
h i cos θ − sin θ 0
Q(t) = e1 | e2 | e3 =  sin θ cos θ 0 .
 

0 0 1
Note that c(0) = [ R, 0, 0] is not the zero vector. Some other examples are presented by
Malvern [207].
If scalar, vector and tensor functions, φ, v and T, transform according to the ‘expected’
transformations,
φ∗ ( x∗ , t∗ ) = φ( x, t),

i i

i i
i i

“CM˙Final” — 2015/3/13 — 11:09 — page 206 — #228

i i

206 Continuum Mechanics

v∗ ( x∗ , t∗ ) = Q(t)v( x, t),

T ∗ ( x∗ , t∗ ) = Q(t) T ( x, t) Q(t) T , (4.2)

they are said to be objective. We shall soon see that some kinematical quantities are objec-
tive, while some others are not. Note that two different observers observe the same value,
magnitude and eigenvalues, respectively, of objective scalar, vector and tensor quantities.

4.2 Transformation of Kinematical Quantities

Suppose the motion of a continuum is described by x = χ( X, t) in frame O, where X
are the referential coordinates. We assume that the framings O and O∗ coincide at some
instant of time, and take the configuration of the body at this instant of time to be the
reference configuration. Thus, X as observed from O or O∗ is identical1 . The motion of the
continuum as seen from frame O∗ is related to that seen from frame O by the relation

χ∗ ( X, t∗ ) = Q(t)χ( X, t) + c(t), (4.3)

where Q ∈ Orth+ , t∗ = t − a, Q(0) = I and c(0) = 0. From Eqn. (4.3), we see that the
configurations χ∗ ( X, t∗ ) and χ( X, t) are related by a rigid deformation (see Fig. 4.3). Two
motions related by Eqn. (4.3) will be referred to as equivalent.

x∗ = χ∗(X, t∗ )

Q(t)x + c(t)
X

x = χ(X, t)

Fig. 4.3 Change of observer can be given the alternative interpretation of a superposed rigid body
motion.

We now determine how the various kinematical quantities transform under a change of
observer. When a quantity is a function of time alone, we use a superposed dot to indicate
the time derivative, e.g., Q̇ ≡ dQ/dt. Note that since t∗ = t − a, the time derivative of any
quantity with respect to either t∗ or t is the same.
Since F = ∇χ, we have

F ∗ = ∇χ∗ = Q(t)∇χ = QF. (4.4)

1 This assumption can be relaxed by relaxing the constraints on Q (0) and c (0) stated below Eqn. (4.3), i.e., by

letting Q(0) = Q0 , which is not necessarily the identity tensor, and c(0) = c0 , which is not necessarily zero (as
in the rotating disc example in the previous section), so that X ∗ = Q0 X + c0 . In this case, some of the relations
derived below get modified, e.g., Eqn. (4.4) gets modified to F ∗ = QFQ0T ; however, whether a quantity is objective
or not is not affected by this change. For a more detailed discussion, see [198].

i i

i i
i i

“CM˙Final” — 2015/3/13 — 11:09 — page 207 — #229

i i

Constitutive Equations 207

Thus, the deformation gradient transforms like a vector under a change of frame. Since
det Q = 1, we have det F ∗ = det F. The transformations for the right and left Cauchy–
Green strain tensors are
C ∗ = ( F ∗ ) T F ∗ = F T Q T QF = F T F = C,
B∗ = F ∗ ( F ∗ ) T = QFF T Q T = QBQ T .
Hence, the transformations for the Lagrangian and Eulerian measures of strain are
U ∗ = U,
V ∗ = QV Q T ,
E∗ = E, (4.5)
∗ T
Ē = Q ĒQ . (4.6)

Since F = RU, it follows that R∗ = QR. Since B ∈ Sym, it can be expressed as ∑ik=1 λi Pi ,
where λi are the distinct eigenvalues and Pi the associated eigenprojections. Thus, B∗ =
∑ik=1 λi Gi , where Gi := QPi Q T are also projections. By using Eqn. (1.312), it follows that
!
k k
log B∗ = ∑ (log λi )Gi = Q ∑ λi Pi Q T = Q(log B) Q T . (4.7)
i =1 i =1

The velocity transformation relation follows from material time differentiation of

Eqn. (4.3):
Dχ∗
v∗ =
Dt
Dχ
=Q + Q̇χ + ċ
Dt
= Qv + Q̇x + ċ. (4.8)
Consider two points fixed in the unstarred frame, having coordinates x1∗ and x2∗ . Since
v1 = v2 = 0, we get

v1∗ − v2∗ = Q̇Q T ( x1∗ − x2∗ ) = w × ( x1∗ − x2∗ ),

where w is the axial vector of the skew tensor Q̇Q T . From the above relation, we see that
w represents the angular velocity of the unstarred frame with respect to the starred frame.
To find the relation for the acceleration, we take the material derivative of Eqn. (4.8) to
get
a∗ = Qa + 2Q̇v + Q̈x + c̈. (4.9)
From Eqns. (4.8) and (4.9), it is apparent that the velocity and acceleration vectors are not
objective. Physically, to see that the velocity vector is not indifferent consider a small ball
moving outward along a straight radial groove on a spinning disc; with respect to an ob-
server on the spinning disc, the ball moves radially outward with nonzero velocity, while
with respect to an observer on the ball, the velocity of the ball is zero. Thus, the two ob-
servers observe different velocity vectors, and one cannot be obtained from the other by an
orthogonal transformation.

i i

i i
i i

“CM˙Final” — 2015/3/13 — 11:09 — page 208 — #230

i i

208 Continuum Mechanics

One observation that we make at this point is that although physical scalar field vari-
ables (such as temperature, density and so on) are objective, artificially constructed scalar
field variables (e.g., v · v) need not be objective.
Using Eqn. (2.49) and the invertibility of F (and hence, also of F ∗ ), we have
DF ∗ ∗ −1 DF
L∗ = ( F ) = (Q + Q̇F ) F −1 Q T ,
Dt Dt
which yields

L∗ = QLQ T + Q̇Q T . (4.10)

Using the fact that Q̇Q T is skew (Problem 38, Chapter 1), we get
1h ∗ i
D∗ = L + ( L∗ ) T = QDQ T (4.11)
2
∗ 1h ∗ i
Ŵ = L − ( L∗ ) T = QŴ Q T + Q̇Q T (4.12)
2
Thus, D is objective, whereas Ŵ is not. Taking the trace of both sides of Eqn. (4.10), and
again using the fact that Q̇Q T ∈ Skw, we get
∇ x∗ · v∗ = tr L = ∇ x · v. (4.13)
From Eqn. (2.61) and (2.62) (or from Eqns. (4.5) and (4.6)), we see that the material rates
of the Lagrangian and Eulerian strain measures are not objective. Instead, we have
DE∗
= ( F ∗ )T D∗ F ∗
Dt
= F T Q T QDQ T QF
= F T DF
DE
=
Dt
∗
D Ē
= D ∗ − ( L∗ )T Ē∗ − Ē∗ L∗
Dt
D Ē T h T
i
=Q Q + Q̇ ĒQ T + Q Ē Q̇ .
Dt

4.3 Principle of Frame-Indifference

The principle of frame-indifference states that the governing laws must have the same
form in reference frames moving arbitrarily with each other, i.e., there is no ‘preferred’
frame such as the ‘inertial’ frame in the classical setup. In this section, we shall examine
the implications of this principle by applying it to all the balance laws.
Thus, first consider the law of mass conservation. Since ρ0 ( X ) = ρ∗ J ∗ = ρJ, and since
J = J, we have ρ∗ = ρ. By virtue of Eqn. (4.13), we have ∇ x∗ · v∗ = ∇ x · v, and thus, the
∗
equation of mass conservation given by
Dρ Dρ∗
+ ρ∇ · v = + ρ∗ ∇ x∗ · v∗ = 0
Dt Dt∗

i i

i i
i i

“CM˙Final” — 2015/3/13 — 11:09 — page 209 — #231

i i

Constitutive Equations 209

has the same form in all frames.

In the previous section, we have derived the transformation relations for various kine-
matical quantities under a change of frame. We saw that certain kinematical quantities
were objective, whereas some others were not. We now postulate that the traction vector,
and hence the stress tensor, are objective, and analyze the consequences of this postulate.
We emphasize that the requirement of objectivity for the stress tensor is an axiom, and not
a derived relation, as in the case of the kinematical quantities. We shall see that this axiom
when combined with the requirement that the linear momentum balance equation has the
same form in all frames yields the transformation relation for the body force b, which is,
in general, not of the form b∗ = Q(t)b, i.e., b is not objective. Note that in contrast to
the rather controversial principle of material frame-indifference, which is discussed in the
following section, the axiom that the stress tensor is objective is universally agreed upon.
To see the physical motivation behind treating contact forces and body forces differ-
ently, we consider again the example of a spring mounted on a rotating table. From the
viewpoint of an external observer, the body force is merely that of gravity, while from the
viewpoint of an observer whose reference frame is fixed to the rotating table, there is an
outward force, seemingly exerted by an external field, in addition to the force of gravity.
Thus, the body forces observed by the two observers are different. In contrast, both ob-
servers observe the same spring constant (which could be a function of the angular velocity
of the table), and the same extension, and hence observe the same contact force between
the table and spring (or the mass and the spring).
The condition that the traction vector be objective is written as

t ∗ ( x∗ , t∗ , n∗ ) = Q(t)t ( x, t, n) ∀ x, t and n, (4.14)

and x∗ = Q(t) x + c(t). Since by Cauchy’s relation, t = τn, Eqn. (4.14) can be written in
terms of the Cauchy stress tensor as

τ ∗ = Q(t)τQ T (t), (4.15)

i.e., the stress tensor is objective.

To see how the term ∇ x · τ in Eqn. (3.19) transforms under a change of frame, we note
that
∂τij∗
∗
∇ x∗ · τ = ei∗
∂x ∗j
∂τrs ∗
= Qir Q js e
∂x ∗j i
∂τrs
= Qir Q js Q e∗
∂xm jm i
∂τrs ∗
= Qir δms e
∂xm i

i i

i i
i i

“CM˙Final” — 2015/3/13 — 11:09 — page 210 — #232

i i

210 Continuum Mechanics

∂τrs ∗
= Qir e
∂xs i
= Q∇ x · τ,

where we have assumed that the Cauchy stress tensor τ is objective. From the above equa-
tion, we see that the term ∇ · τ transforms as a vector. Thus, if Eqn. (3.19) is to have the
same form in all frames, then the term ρ(b − a) must also transform as a vector, i.e., we
must have

ρ∗ (b∗ − a∗ ) = ρQ(b − a). (4.16)

Hence, using Eqn. (4.9), the transformation law for the body force vector is

b∗ = Qb + 2Q̇v + Q̈x + c̈,

or, alternatively,
h i
b = Q T b∗ − (Ẇ + W 2 )( Qx) − 2W ( Qv) − c̈ , (4.17)

where W = Q̇Q T ∈ Skw. Since W ∈ Skw, Q T W Q and Q T Ẇ Q are also skew-symmetric

tensors. Note that Q T W 2 Q = ( Q T W Q)( Q T W Q). Using Q̇ = W Q, one can easily verify
˙
that Q T Ẇ Q = Q T W Q. Hence, the axial vector of Q T Ẇ Q is simply the time derivative of
the axial vector of Q T W Q. Thus, if we denote the axial vector of Q T W Q (= Q T Q̇) by Ω,
Eqn. (4.17) can be written as

b = Q T [b∗ − c̈] − Ω̇ × x − Ω × (Ω × x) − 2Ω × v. (4.18)

If w denotes the axial vector of W, it can be easily shown (see Problem 17, Chapter 1)
that Ω = Q T w is the axial vector of Q T W Q, or, in other words, the components of Ω
are simply the components of the vector w with respect to the unstarred coordinate frame.
Since Qij = ei∗ · e j , we have (noting that the starred frame is stationary with respect to itself)

[ Q T W Q]ij = [ Q T Q̇]ij = (e∗k · ei )(e∗k · ė j ) = [(ei · e∗k )e∗k ] · ė j = ei · ė j ,

so that
 
ė2 · e3
Ω = ė3 · e1  ,
 
(4.19)
ė1 · e2
w = QΩ.

In terms of the Euler angle representation of Q given by Eqn. (1.108), the axial vector
corresponding to W is
 
θ̇ cos ψ + φ̇ sin θ sin ψ
w = θ̇ sin ψ − φ̇ sin θ cos ψ . (4.20)
 

ψ̇ + φ̇ cos θ

i i

i i
i i

“CM˙Final” — 2015/3/13 — 11:09 — page 211 — #233

i i

Constitutive Equations 211

The axial vectors of Q T W Q and Q T Ẇ Q T are given by

 
θ̇ cos φ + ψ̇ sin θ sin φ
Ω = wQT W Q = −θ̇ sin φ + ψ̇ sin θ cos φ , (4.21a)
 

φ̇ + ψ̇ cos θ
 
d
dt ( θ̇ cos φ + ψ̇ sin θ sin φ )
Ω̇ = wQT Ẇ Q =  dtd (−θ̇ sin φ + ψ̇ sin θ cos φ) . (4.21b)
 
d
dt ( φ̇ + ψ̇ cos θ )

If we assume the ‘∗’ frame to be ‘stationary’ with a body force b∗ = − gez , and the
unstarred frame to be rotating and translating with respect to the ‘∗’ frame, then we see
from Eqn. (4.18) that b now includes among other forces, the centrifugal and Coriolis forces.
This is consistent with experience, since we do feel an outward force when sitting in a
vehicle going around a curve, or a backward push when a vehicle suddenly accelerates
forward. When Q and ċ are constant, these ‘extra’ forces vanish, i.e., b∗ = Qb, and such a
change of frame is called a Galilean transformation.
As a simple example, consider the disc shown in Fig. 4.2b rotating with a constant angu-
lar velocity Ω = Ωez , and with the origin of the rotating frame of reference coinciding with
the center of the disc. Assume that the mass on the disc has reached a steady-state position
with respect to the rotating frame of reference, i.e., v = 0. Since c is now zero at all times,
and since Ω̇ = 0, the only term that contributes to the in-plane body forces in Eqn. (4.18)
is the centrifugal body force term. Since e1 = (cos θ, sin θ, 0), e2 = (− sin θ, cos θ, 0),
e3 = (0, 0, 1), from Eqn. (4.19), we get Ω = θ̇ez , and the term −Ω × (Ω × x) evaluates
to ( R + δ)Ω2 e1 , which is consistent with the outward force that the body experiences on
the spinning disc.
We now consider more complicated examples. The spinning disc of radius R shown
in Fig. 4.4 spins about its own axis with an angular velocity γ, and the rod of circular
cross-section to which it is attached spins about the Y-axis with an angular velocity β. The
xyz coordinate frame (assumed to be the unstarred frame) is fixed to the spinning disk,
and is given to be parallel to the XYZ coordinate frame (the ‘∗’-frame) at t = 0. Both γ
and β are given to be constant. We are interested in computing the body forces at a point
x = ( R, 0, 0) with respect to the xyz system, assuming that the body force with respect to
XYZ is b∗ = (0, 0, − g). The matrix Q has components Qij = ei∗ · e j , which in the case of
this example, can be written as Qij = (ei∗ · ê∗m )ê∗m · e j = Qim Qmj , i.e., as a product of the Q
matrices corresponding to the rotations β and γ, respectively. The vector c is the position
vector of the origin of xyz with respect to XYZ. Thus, we get

  
cos βt 0 sin βt cos γt − sin γt 0
Q= 0 1 0   sin γt cos γt 0
  

− sin βt 0 cos βt 0 0 1
 
cos βt cos γt − cos βt sin γt sin βt
= sin γt cos γt 0 , (4.22)
 

− sin βt cos γt sin βt sin γt cos βt

i i

i i
i i

“CM˙Final” — 2015/3/13 — 11:09 — page 212 — #234

i i

212 Continuum Mechanics

z
y γ
Z

γt x
R
L
βt Y

X
Fig. 4.4 Body forces experienced by a point on a spinning disc.

 
L sin βt
c =  0 .
 

L cos βt

By means of Eqn. (4.19) (or by comparing the entries of the Q matrix with the Euler angle
representation given by Eqn. (1.108), observing that θ = βt, ψ = π/2, and φ = γt − π/2,
and substituting these values into Eqns. (4.20)–(4.21b)), we get,
     
γ sin βt β sin γt βγ cos γt
w =  β , Ω =  β cos γt , Ω̇ = − βγ sin γt . (4.23)
     

γ cos βt γ 0

Note that w and Ω are the angular velocities of the xyz frame expressed with respect to
the XYZ and xyz frames, respectively. Since the point x = ( R, 0, 0) is fixed with respect to
xyz, we have v = 0. Substituting these quantities into Eqn. (4.18), we get the required body
force as
 
g sin βt cos γt + Rβ2 cos2 γt + Rγ2
b =  − g sin βt sin γt − Rβ2 sin γt cos γt  .
 

− g cos βt + Lβ2 − 2Rγβ sin γt

Note that the body forces experienced by the point are oscillatory!
As another example, consider the flow through a pipe of radius R that is spinning about
its own axis with a constant angular speed Ω a , and about the Z-axis with constant angular
speed Ω p , as shown in Fig. 4.5.

i i

i i
i i

“CM˙Final” — 2015/3/13 — 11:09 — page 213 — #235

i i

Constitutive Equations 213

Z Ωp

Y
er
eθ θ + Ωa t

ez
Ωp t
Ωa
X

Fig. 4.5 Flow in a pipe spinning about two axes; the origins of the er –eθ –ez and XYZ frames are
shown separated simply for clarity, although actually they coincide.

The vector c is zero, and the matrix Q = [e1 | e2 | e3 ] is given by

− sin(Ω p t) cos(Ω a t + θ ) sin(Ω p t) sin(Ω a t + θ ) cos(Ω p t)

 

Q =  cos(Ω p t) cos(Ω a t + θ ) − cos(Ω p t) sin(Ω a t + θ ) sin(Ω p t)  ,

 

sin(Ω a t + θ ) cos(Ω a t + θ ) 0

so that by means of Eqn. (4.19), we get

Ω p sin(Ω a t + θ ) Ω p Ω a cos(Ω a t + θ )
   

Ω = Ω p cos(Ω a t + θ ) , Ω̇ = −Ω p Ω a sin(Ω a t + θ ) .

   

Ωa 0

Note again that Ω is simply the angular velocity of the pipe expressed with respect to the
er –eθ –ez frame. The centrifugal term in the body force, Ω × (Ω × x) = −∇ |Ω × x|2 /2,
can be absorbed into the ∇ p term in Eqn. (7.104). Using Eqn. (4.18), the body force com-
ponents (excluding the centrifugal force) at a point x = rer + zez in the rotating frame
are

br = − sin(Ω a t + θ ) g + Ω p Ω a z sin(Ω a t + θ ) − 2Ω p vz cos(Ω a t + θ ) + 2Ω a vθ ,

bθ = − cos(Ω a t + θ ) g + Ω p Ω a z cos(Ω a t + θ ) + 2Ω p vz sin(Ω a t + θ ) − 2Ω a vr ,
bz = −Ω p Ω a r sin(Ω a t + θ ) − 2Ω p vθ sin(Ω a t + θ ) + 2Ω p vr cos(Ω a t + θ ),

where (vr , vθ , vz ) are the velocity components of the fluid particle at x as measured with
respect to the rotating frame. Substituting the above body force components into Eqns.
(A.2) of Appendix A, Vol. II, and solving these equations in conjunction with the other
governing equations (such as the continuity equation), subject to the boundary condition
v|r= R = 0, and the given initial condition, we can solve for the velocity field of the fluid in
the pipe. Note that the Coriolis terms can have a significant influence on the solution.
We have already seen that provided the balance of linear momentum holds, the balance
of angular momentum is equivalent to the symmetry of the stress tensor. Thus, the balance

i i

i i
i i

“CM˙Final” — 2015/3/13 — 11:09 — page 214 — #236

i i

214 Continuum Mechanics

of angular momentum is frame-indifferent (in the sense that it has the same form in any
arbitrary frame) since the balance of linear momentum is frame-indifferent and since

(τ ∗ )T = Q(t)τ T Q(t)T = Q(t)τQ(t)T = τ ∗ .

We now show that balance of energy coupled with the axiom of frame-indifference
is equivalent to all the balance laws that have been derived so far! As usual, let ρ( x, t),
t ( x, t, n), b( x, t), e( x, t), Qh ( x, t) be the density, surface tractions, body force per unit mass,
the internal energy per unit mass, and the heat supply per unit mass, respectively. In
addition, let γ( x, t, n) represent the heat flux across a surface with unit normal n. Note that
γ( x, t, n) depends on the normal to the surface across which the heat conduction is taking
place, and is, thus, analogous to the traction term t ( x, t, n).
As we have seen in Eqn. (3.75), the first law of thermodynamics states that the rate of
increase of total energy (kinetic and internal) of any portion of the body equals the rate of
work done on that portion by the body forces and surface tractions, plus the rate of increase
of heat energy from the heat source, and the heat flow across the boundary, i.e.,

d 1
Z Z Z
ρ e+ v·v dV = (t · v + γ) dS + ρ(b · v + Qh ) dV. (4.24)
dt V (t) 2 S(t) V (t)

We assume that under a change of observer, the following hold:

ρ∗ ( x∗ , t) = ρ( x, t), e∗ ( x∗ , t) = e( x, t), γ∗ ( x∗ , t, n∗ ) = γ( x, t),

Q∗h ( x∗ , t) = Qh ( x, t), t ∗ ( x∗ , t, n∗ ) = Qt ( x, t, n), ( b ∗ − a ∗ ) = Q ( b − a ),

with the transformations relations for v∗ and a∗ given by Eqns. (4.8) and (4.9). Then, we
have the following fascinating result:

Theorem 4.3.1. Let the balance of energy given by Eqn. (4.24) hold. Assume that
under a change of observer, the form of Eqn. (4.24) still holds. Then there exist functions
τ and q such that t = τn and γ = −q · n, and conservation of mass, balance of linear
momentum and balance of angular momentum all hold.
Conversely, if there exists a function τ such that t = τn, and the balance principles
all hold, then the balance of energy has the same form under a change of observer.

Proof. The proof that we present is similar to that in [209]. Let

1
f := ρ e + v · v .
2

Then Eqn. (4.24) can be written using transport theorem I (note that transport
theorem II cannot be used since conservation of mass is not assumed to hold) as

Df
Z Z Z
+ f ∇ · v dV = (t · v + γ) dS + ρ(b · v + Qh ) dV. (4.25)
V (t) Dt S(t) V (t)

i i

i i
i i

“CM˙Final” — 2015/3/13 — 11:09 — page 215 — #237

i i

Constitutive Equations 215

If Eqn. (4.25) holds under a change of observer, then, since ρ∗ = ρ, Q∗h = Qh and
γ∗ = γ, J ∗ = J (so that dV ∗ = dV),

Df∗
Z Z Z
+ f ∗ ∇ x∗ · v∗ dV = (t ∗ · v∗ + γ) dS + ρ(b∗ · v∗ + Qh ) dV,
V (t) Dt S(t) V (t)
(4.26)

where f ∗ = ρ∗ (e∗ + v∗ · v∗ /2). Subtracting Eqn. (4.25) from Eqn. (4.26), and
using Eqn. (4.13), we have

D ∗
Z Z
( f − f ) + ( f ∗ − f )∇ · v dV = (t ∗ · v∗ − t · v) dS
V (t) Dt S(t)
Z
+ ρ(b∗ · v∗ − b · v) dV. (4.27)
V (t)

Using the expressions for f ∗ and f , the left-hand side of the above equation is
given by

1 ∗ ∗ Dρ ρ D ∗ ∗
Z
LHS = (v · v − v · v) + ρ∇ · v + (v · v − v · v) dV
V (t) 2 Dt 2 Dt

1 Dρ
Z
∗ ∗
= Qv + ( Q̇x + ċ) · ( Q̇x + ċ) + ρ∇ · v + ρ(v · a − v · a) dV.
V (t) 2 Dt

Substituting t ∗ = Qt, (b∗ − a∗ ) = Q(b − a), and the above formula into
Eqn. (4.27), we get

1 Dρ
Z
Qv + ( Q̇x + ċ) · ( Q̇x + ċ) + ρ∇ · v + ρ[ Q( a − b) · ( Q̇x + ċ)] dV
V (t) 2 Dt
Z
= Qt · ( Q̇x + ċ) dS. (4.28)
S(t)

Eqn. (4.28) holds for all Q(t) ∈ Orth+ , and ċ ∈ V. Choose Q ≡ I (Q̇ = 0),
and ċ = constant. Then, Eqn. (4.28) reduces to

1 Dρ
Z Z
v · ċ + ċ · ċ + ρ∇ · v + ρ( a − b) · ċ dV = t · ċ dS.
V (t) 2 Dt S(t)

Now, using an argument analogous to that in Theorem 3.5.1 yields the existence
of a tensor τ such that t = τn. Substituting this result into Eqn. (4.25), and again
using a similar argument yields the existence of a vector function q such that
γ = −q · n. Since n∗ = Q(t)n, and since γ is objective, we get q∗ = Q(t)q.
Using t = τn and the divergence theorem, the right-hand side of Eqn. (4.28)
can be simplified as
Z Z
Qt · ( Q̇x + ċ) dS = Qτn · ( Q̇x + ċ) dS
S(t) S(t)

i i

i i
i i

“CM˙Final” — 2015/3/13 — 11:09 — page 216 — #238

i i

216 Continuum Mechanics

Z
= τ T Q T ( Q̇x + ċ) · n dS
S(t)
Z
= ∇ · [τ T Q T ( Q̇x + ċ)] dV
V (t)
Z Z
= (∇ · τ ) · Q T ( Q̇x + ċ) dV + τ : Q T Q̇ dV
V (t) V (t)
Z Z
= Q(∇ · τ ) · ( Q̇x + ċ) dV + τ : Q T Q̇ dV.
V (t) V (t)

Substituting this result into Eqn. (4.28), we get

1 Dρ
Z
Qv + ( Q̇x + ċ) · ( Q̇x + ċ) + ρ∇ · v dV =
V (t) 2 Dt
Z Z
Q[∇ · τ + ρ(b − a)] · ( Q̇x + ċ) dV + τ : Q T Q̇ dV. (4.29)
V (t) V (t)

Now choosing Q ≡ I (Q̇ = 0), and ċ = λe, where λ ∈ <, and e is a unit vector,
we get

λ2

Dρ
Z Z
λ(v · e) + + ρ∇ · v dV = λ [∇ · τ + ρ(b − a)] · e dV.
V (t) 2 Dt V (t)
(4.30)
Differentiating Eqn. (4.30) twice with respect to λ, we get

Dρ
Z
+ ρ∇ · v dV = 0,
V (t) Dt

which is valid for an arbitrary volume V (t), and hence implies the conservation
of mass, i.e.,
Dρ
+ ρ∇ · v = 0. (4.31)
Dt
Substituting this result into Eqn. (4.30), and using the arbitrariness of V (t) and
e, we get
∇ · τ + ρ(b − a) = 0, (4.32)
which is nothing but the balance of linear momentum.
For W ∈ Skw, let Q(t) = eW t . Then Q T Q̇ = W (since Q̇ = W Q = QW).
Using Eqns. (4.31) and (4.32), Eqn. (4.29) reduces to
Z
τ : W dV = 0,
V (t)

which being valid for any arbitrary volume V (t), and any arbitrary skew tensor
W, yields the fact that τ is symmetric (see Problem 8, Chapter 1). Thus, the
balance of angular momentum holds.

i i

i i
i i

“CM˙Final” — 2015/3/13 — 11:09 — page 217 — #239

i i

Constitutive Equations 217

To prove the converse, note that since now we are given that t = τn,
Eqn. (4.29) represents the difference between Eqn. (4.25) and (4.26). Given that
Eqn. (4.25) holds, we want to prove that Eqn. (4.26) holds. By the above com-
ment, it suffices to show that Eqn. (4.29) holds. Since by assumption, mass con-
servation, balance of linear momentum and balance of angular momentum hold,
and since Q T Q̇ ∈ Skw, Eqn. (4.29) holds, and thus the converse is proved.

A methodology similar to the above is used to derive the governing equations for liquid
crystals. See [73], [185]–[187] and [313] for more details.
From Eqn. (3.64), we see that since τ : D (which is also equal to T : Ḟ/J and S : Ė/J) is
frame-indifferent, the quantity
Z Z
t · v dS + ρ(b − a) · v dV,
S(t) V (t)

which is nothing but the mechanical power minus the rate of change of kinetic energy of
the material volume (see Eqn. (3.65)), is also frame-indifferent. Thus, the mechanical power
itself is not frame-indifferent. This is also evident from Eqn. (4.24), since e, γ and Qh are
frame-indifferent.
Though the stress tensor is postulated to be objective, the stress rate is not objective in
general, as can be seen by taking the material derivative of Eqn. (4.15):

Dτ ∗ T
= Q̇τQ T + Qτ̇Q T + Qτ Q̇ ,
Dt
which shows that the stress rate is objective if and only if

T
Q̇τQ T + Qτ Q̇ = 0 ∀ Q ∈ Orth+ , (4.33)

which in turn is valid if and only if τ ( x, t) = λ( x, t) I, for all x and t. To see this, let
W ∈ Skw, and let Q = eW t . From the results of Section 1.10, we know that Q̇ = W Q.
Substituting this expression into Eqn. (4.33) and noting that W is arbitrary, we get

W QτQ T = QτQ T W ∀W ∈ Skw,

By Problem 18, Chapter 1, this proves that QτQ T is of the form λI, and hence that τ = λI.
Many rate-dependent materials need to be described in terms of stress rates, and the
resulting constitutive models must be frame-indifferent. Hence, it is essential to derive
stress rates that are objective. The simplest way of generating an objective stress rate uses
the fact that the second Piola–Kirchhoff stress tensor remains unchanged under a change
of observer, which is proved quite easily as follows:

S ∗ = J ∗ ( F ∗ ) −1 τ ∗ ( F ∗ ) − T
= JF −1 Q−1 QτQ T Q−T F −T
= JF −1 τF −T
= S.

i i

i i
i i

“CM˙Final” — 2015/3/13 — 11:09 — page 218 — #240

i i

218 Continuum Mechanics

Thus, we also have

DS∗ DS
= . (4.34)
Dt Dt
Taking the material derivative of the relation S = JF −1 τF −T , and using Eqns. (2.50) and
(2.51), we get
DS −1 −T
= J̇F −1 τF −T + J Ḟ τF −T + JF −1 τ̇F −T + JF −1 τ Ḟ
Dt
= (tr L) JF −1 τF −T − JF −1 LτF −T + JF −1 τ̇F −T − JF −1 τL T F −T
h i
= JF −1 τ̇ − Lτ − τL T + (tr L)τ F −T
= JF −1 τ ◦ F −T , (4.35)
where
τ ◦ := τ̇ − Lτ − τL T + (tr L)τ (4.36)
is known as the Truesdell stress rate. To see that it is objective, we note from Eqn. (4.35) that
1 DS T
τ◦ = F F .
J Dt
Now using Eqn. (4.34), we have
1 ∗ DS∗ ∗ T
[τ ◦ ]∗ = F (F )
J∗ Dt
1 DS T T
= QF F Q
J Dt
= Qτ ◦ Q T .
The Jaumann stress rate (also known as the corotational stress rate) is another objective
stress rate, which is defined by

τ ∇ := τ̇ − Ŵ τ + τ Ŵ, (4.37)
where Ŵ is a skew-symmetric part of L. This stress rate is obtained by substituting L =
D + Ŵ into Eqn. (4.36), and shifting all terms with D to the left, i.e.,

τ ∇ = τ ◦ + Dτ + τD − (tr D )τ.
That τ ∇ is objective is easily seen from the above equation, since τ ◦ , D and τ are all objec-
tive.
Writing Eqn. (4.37) as
T
τ ∇ = τ̇ + Ŵ τ + τ Ŵ,
substituting Ŵ = L − D, and shifting all terms with D to the left, we get the convective
stress rate given by

τ := τ̇ + L T τ + τL. (4.38)

i i

i i
i i

“CM˙Final” — 2015/3/13 — 11:09 — page 219 — #241

i i

Constitutive Equations 219

Objectivity of the convective stress rate is a result of the relation

τ = τ ∇ + Dτ + τD.

By replacing Ŵ with ṘR T (since ṘR T is skew-symmetric) into Eqn. (4.37), we obtain
the Green–Naghdi stress rate:

τ 4 := τ̇ − ṘR T τ + τ ṘR T .

If Ē is defined in a manner analogous to τ , then from Eqns. (2.62) and (4.38), we see
that

Ē = D. (4.39)

Similar frame-indifferent rates can also be formulated for the heat flux vector. Using
the result of Problem 3, Chapter 2, we see that the fluxes in the reference and deformed
configurations q0 and q, respectively, are related by q0 = JF −1 q. Since F ∗ = QF and
q∗ = Qq, we have q0∗ = q0 , which in turn implies that q̇0∗ = q̇0 . Taking the material time
derivative of the relation q0 = JF −1 q, and using Eqns. (2.50) and (2.51), we get

q̇0 = JF −1 [q̇ + (∇ · v)q − Lq] .

Since ( F −1 )∗ = F −1 Q T , this implies that the bracketed term given by

1 ∂q
q := F q̇0 = q̇ + (∇ · v)q − Lq = + (∇q)v + (∇ · v)q − Lq (4.40)
J ∂t x

is objective. As in the stress case, there are an infinite number of such objective flux rates.
For example, other objective rates are

q4 := q̇ − Lq. (4.41)
∇
q := q̇ − Ŵ q, (4.42)

where Ŵ = ( L − L T )/2. Objective rates such as those given by Eqns. (4.40)–(4.42) can be
used in a non-Fourier heat conduction law (see Eqns. (7.14) and (7.64)).

4.4 Principle of Material Frame-Indifference

Ideally, one would like to extend the principle of frame indifference to constitutive relations
as well, i.e., we would like to postulate that constitutive relations have the same form in
all frames of reference. However, such a postulate has not been universally accepted, and
indeed, starting with the works of Muller [228], there have been a number of works that
have claimed that the constitutive relations need not have the same form in all frames.
This, in turn, has resulted in several works that have tried to clarify the issues involved
(see, e.g., [19], [199], [316]).
To see what is at the heart of the controversy, consider again the example of a spring
mounted on a rotating table with one end attached to the table, and the other to a mass (see
Section 4.1). As noted by Bertram and Svendsen ([19], [316]), the first aspect of material

i i

i i
i i

“CM˙Final” — 2015/3/13 — 11:09 — page 220 — #242

i i

220 Continuum Mechanics

frame-indifference, namely, observer invariance, merely requires that the material response
function be objective, which, in the context of our example, means that two observers ob-
serve the same spring constant. Suppose that the spring constant has been found to be k0
under stationary conditions. If now the turntable is rotated with an angular speed ω, then
although observer invariance states that two observers would observe the same spring
constant, that by itself does not rule out the possibility that the spring constant is a function
of ω, i.e., one could have k = k̂(ω ) with k̂(0) = k0 . Muller’s example of a constitutive rela-
tion obtained from the kinetic theory of gases [228] also depends on the spin of the frame
of reference, and is thus, analogous to the dependence of k on ω. From a practical view-
point, observer invariance by itself is clearly incapable of relating k̂ (ω ) to k0 . However,
material frame-indifference as traditionally formulated (see, e.g, [108] or [337]) implicitly
assumes that in addition to observer invariance, form-invariance also holds. The hypothesis
of form-invariance states that the constitutive relations have the same form in all frames of
reference, and hence depend only on variables intrinsic to the material; thus, for example,
dependence on the rotation rate of the frame of reference is excluded. In the context of our
spring example, form invariance implies that k̂(ω ) = k0 . This hypothesis, if valid, is clearly
useful from a practical viewpoint, since a designer using the value of the spring constant
found under stationary conditions can analytically determine the extension of the spring
for any ω, which otherwise would have to be determined experimentally for every ω.
If τ̂ ∗ and τ̂ represent the constitutive relations for τ ∗ and τ, respectively, i.e., if τ ∗ =
τ̂ (l ∗ ), and τ = τ̂ (l ), where l is a list of kinematical parameters on which τ is assumed to
∗

depend, and l ∗ is the same list of parameters as observed by O∗ , then Eqn. (4.15) implies
that

τ̂ ∗ (l ∗ ) = Q(t)τ̂ (l ) Q(t) T , (4.43)

which is the condition of observer invariance mentioned above. As discussed, τ̂ ∗ and τ̂

could be different functions, so that the above relation by itself does not lead us to any
conclusion about the nature of the constitutive relations (indeed, as observed by Liu [199],
Eqn. (4.43) can be regarded as the definition of τ̂ ∗ , once τ̂ is given). However, if, in addition,
we assume form-invariance, i.e.,

τ̂ ∗ (l ∗ ) = τ̂ (l ∗ ), (4.44)

then, from Eqn. (4.43), it follows that

τ̂ (l ∗ ) = Q(t)τ̂ (l ) Q(t) T . (4.45)

The above condition, often alluded to as the invariance of the constitutive relation under su-
perposed rigid-body motion, or, as the principle of material frame-indifference is a restriction that
leads us in many cases to the so-called reduced form of the constitutive relations. For ex-
ample, in the case of our spring example, Eqn. (4.43) does not preclude a dependence of the
spring constant on the angular speed ω, but Eqn. (4.45) does. We have shown that granted
observer invariance, form-invariance is equivalent to the hypothesis of invariance under
superposed rigid body motion, i.e., Eqns. (4.43) and (4.44) together imply Eqn. (4.45). It is
also easily seen that Eqns. (4.44) and (4.45) together imply Eqn. (4.43), and that Eqns. (4.43)
and (4.45) together imply Eqn. (4.44). Thus, any two hypothesis out of observer invariance,
form-invariance and invariance under superposed rigid body motions implies the third.

i i

i i
i i

“CM˙Final” — 2015/3/13 — 11:09 — page 221 — #243

i i

Constitutive Equations 221

Note that while observer invariance involves the same motion as observed by two dif-
ferent observers, invariance under superposed rigid body motions involves two differ-
ent motions as seen by one observer. Thus, clearly, these two notions are conceptually
quite different. Muller’s examples of constitutive relations are observer-invariant, but
not form-invariant, or, equivalently, not invariant under superposed rigid body motion.
Notwithstanding these examples (which, to the best of our knowledge, have not been ver-
ified experimentally), form-invariance, or, equivalently, the principle of material frame-
indifference, is satisfied by a large class of materials.
Murdoch ([230]–[232]) has claimed that the restrictions on constitutive relations can be
deduced by purely objective considerations without the requirement of form-invariance
of constitutive relations. Liu [199] remarks that such a claim if true would be a truly re-
markable result, since one can easily come up with a counterexample of a constitutive
relation that satisfies the observer invariance condition given by Eqn. (4.43), but is not
form-invariant, and hence does not satisfy Eqn. (4.45) such as, for example,
τ̂ ( E) = αE, (4.46)
where α is a material constant. By defining τ̂ ∗ ( E∗ ) := αQ(t) E∗ Q(t) T and recalling that
E∗ = E, we see that
τ̂ ∗ ( E∗ ) = Q(t)τ̂ ( E) Q(t) T .
Thus, the observer invariance condition given by Eqn. (4.43) is satisfied. On the other hand,

τ̂ ( E∗ ) = αE∗ = αE = τ̂ ( E) 6= Q(t)τ̂ ( E) Q(t) T ,

so that the the restriction of material frame-indifference given by Eqn. (4.45), or, equiva-
lently, that of form-invariance given by Eqn. (4.44) is not satisfied. The above controversy
can be resolved if one notes that Murdoch has a condition similar to that of form invari-
ance (condition O.4. on page 311 of [231]; since this condition is not stated in mathematical
form, it is not clear whether this statement is weaker than or equivalent to form-invariance)
in his definition of objectivity, while objectivity as generally understood and as interpreted
by Liu [199, 200] is simply Eqn. (4.43).
Summarizing, notwithstanding claims that in principle material frame-indifference could
be violated, in the absence of any experimental evidence, we shall henceforth in this book
adopt the principle of material frame-indifference as an axiom. This condition is only a
necessary condition in the sense that a constitutive relation that violates material frame-
indifference such as the one in Eqn. (4.46) is inadmissible, but, conversely, a constitutive
relation that satisfies material frame-indifference may or may not model the actual physi-
cal behavior of a material accurately.
Using Eqns. (3.38), (3.49) and (4.4), the material frame-indifference condition given by
Eqn. (4.45) can be written in terms of the first and second Piola–Kirchhoff stress as
T̂ (l ∗ ) = Q(t) T̂ (l ), (4.47a)
∗
Ŝ(l ) = Ŝ(l ). (4.47b)
We now present some examples of constitutive relations that satisfy material frame-
indifference. For a Newtonian fluid, the constitutive relation for the Cauchy stress is
τ̂ = − p(ρ, θ ) I + [λ(ρ, θ )tr D ] I + 2µ(ρ, θ ) D.

i i

i i
i i

“CM˙Final” — 2015/3/13 — 11:09 — page 222 — #244

i i

222 Continuum Mechanics

Since ρ∗ = ρ, θ ∗ = θ and D ∗ = QDQ T , we have p∗ = p, λ∗ = λ, µ∗ = µ, so that

τ̂ (ρ∗ , θ ∗ , D ∗ ) = Qτ̂ (ρ, θ, D ) Q T , and material frame-indifference is satisfied. On the other
hand, if we were to replace D by L in the above constitutive relation, then since L∗ =
QLQ T + Q̇Q T , material frame-indifference is violated, i.e., τ̂ (ρ∗ , θ ∗ , L∗ ) 6= Qτ̂ (ρ, θ, L) Q T ,
while, by defining τ̂ ∗ (ρ∗ , θ ∗ , L∗ ) := Qτ̂ (ρ, θ, L) Q T , observer invariance is satisfied. As
another example, since E∗ = E, and since the Lamé parameters λ and µ are constants, the
constitutive relation Ŝ = λ(tr E) I + 2µE satisfies the constraint given by Eqn. (4.47b), and
hence is frame-indifferent.

4.5 Constitutive Relations for Simple Materials

We expect the stress in a body at any given instant of time to depend on what has happened
to the body in the past. Also, considering only mechanical influences, we expect the stress
at a time t to be determined by the history of the motion of the body upto time t. Since
stress represents contact force, we further expect the stress at a point to be determined by
the history of the motion in the immediate neighborhood of the point. Since a deformation
is characterized locally by its deformation gradient, we are lead to define the following
class of materials:
Definition: A material is said to be simple if for each particle X and each time t, there exists
a functional such that the Cauchy stress at a point x = χ( X, t) is given by
∞
τ ( x, t) = τ̂ ( X, F ( X, t − s)), (4.48)
s =0

where F is the deformation gradient, and the variable s, which varies from 0 to ∞, has been
introduced to represent the history dependence that the constitutive relation of a material
might have. The value s = 0 corresponds to the present time t, while large s gives the
distant past. This function, which is a material property is called the constitutive equation of
the simple material. Thus, the stress at a point is determined by the history of the defor-
mation gradient relative to some configuration at that point. The adjective ‘simple’ is used
to indicate that the deformation gradient is the only characteristic of the motion involved
in the constitutive equation.
Note that we have assumed that the constitutive relation for a simple material is not
dependent explicitly on time. Thus, we are assuming that aging does not occur. Also
note that the constitutive relation does not depend on the motion χ, since, otherwise, the
Cauchy stress would vary if the body is rigidly translated. In addition, the stress at a point
x is assumed not to depend on the deformation gradient F (Y, t) evaluated at all other
points Y of the domain, as well as on higher-order gradients, e.g., ∂2 χk /∂Xi ∂X j .
In spite of the simplicity of this model, we shall find that most constitutive models
currently in use fall in this category. For example, a material is said to be elastic if the
history dependence in Eqn. (4.48) is excluded. In this case the Cauchy stress depends only
on the current value of the deformation gradient, i.e.,
τ ( x, t) = τ̂ ( X, F ( X, t)), x = χ( X, t). (4.49)
We note, however, that the constitutive relation given by Eqn. (4.49) may be taken to be a
good approximation only for ‘small’ values of F; for large values of F, plasticity or fracture
may occur invalidating this relation.

i i

i i
i i

“CM˙Final” — 2015/3/13 — 11:09 — page 223 — #245

i i

Constitutive Equations 223

As another example, we shall see that the constitutive relation for a Newtonian fluid
depends on the velocity gradient L, which can be expressed as ḞF −1 . Thus, we see that
due to the presence of the term Ḟ, the stress at a given instant of time is influenced by the
history of the deformation, even though not a significant amount of such history might
need to be considered.
By virtue of the relations of the first and second Piola–Kirchhoff stresses with the Cauchy
stress, we also have the following equations corresponding to Eqn. (4.48):
∞
T ( X, t) = T̂ ( X, F ( X, t − s)),
s =0
∞
S( X, t) = Ŝ ( X, F ( X, t − s)).
s =0

In what follows, we will often make the assumption that the body is homogeneous
or isotropic. A homogeneous body is one that has identical properties at all points. A
body is isotropic with respect to certain properties if these properties are identical in all
directions. These two properties of homogeneity and isotropy are completely independent
of each other, i.e., one could have a non-homogeneous isotropic body or a non-isotropic
homogeneous body.
For a homogeneous body, the explicit dependence of the constitutive relations on X can
be dropped. Thus, for a homogeneous body, we have
∞
τ ( x, t) = τ̂ ( F ( X, t − s)).
s =0

Homogeneity is a property satisfied in a given reference configuration, which may no

longer hold true if a ‘deformed configuration’ is chosen as the reference configuration.
This same statement holds even for isotropy, as we shall see later.
We now consider the restrictions imposed by material frame-indifference on the consti-
tutive relations of a simple material. By Eqn. (4.3) we know that
χ∗ ( X, t∗ ) = Q(t)χ( X, t) + c(t),
where t∗ = t − a. We need to find F ∗ ( X, t∗ − s) associated with this motion. We have
F ∗ ( X, t∗ ) = Q(t)∇χ( X, t)
= Q(t) F ( X, t)
= Q(t∗ + a) F ( X, t∗ + a).
Consequently,
F ∗ ( X, t∗ − s) = Q(t∗ + a − s) F ( X, t∗ + a − s)
= Q(t − s) F ( X, t − s). (4.50)
The list of kinematical variables in Eqn. (4.45) comprises only of the deformation gradient,
while the response function for the Cauchy stress is as given by Eqn. (4.48). Thus, by
virtue of Eqn (4.50), the response function for the Cauchy stress satisfies material frame-
indifference if and only if
∞ ∞
τ̂ ( X, F ∗ ( X, t∗ − s)) = τ̂ ( X, Q(t − s) F ( X, t − s))
s =0 s =0

i i

i i
i i

“CM˙Final” — 2015/3/13 — 11:09 — page 224 — #246

i i

224 Continuum Mechanics

∞
= Q(t) τ̂ ( X, F ( X, t − s)) Q(t)T .
s =0

Since for any F ∈ Lin+ , we can find a motion χ such that ∇χ = F (the motion χ =
FX + c(t) does the job), we can write the above equation as
∞ ∞
τ̂ ( X, Q(t − s) F ( X, t − s)) = Q(t) τ̂ ( X, F ( X, t − s)) Q(t) T ∀ F ∈ Lin+ , Q ∈ Orth+ , (4.51)
s =0 s =0

We now have the following theorem:

Theorem 4.5.1. Let τ̂ : V0 × Lin+ → Sym be the response function for the Cauchy
stress. Then the following are equivalent:
(i) τ̂ satisfies the axiom of material frame-indifference, i.e., Eqn. (4.51) holds for all
X ∈ V0 , and for all t ∈ <.
(ii) For all X ∈ V0 , and for all t ∈ <,

∞ ∞
τ̂ ( X, F ( X, t − s)) = R( X, t) τ̂ ( X, U ( X, t − s)) R( X, t) T ∀ F = RU ∈ Lin+ .
s =0 s =0
(4.52)

(iii) There exists a mapping S̃ : V0 × Psym → Sym, such that, for all X ∈ V0 , and for
all t ∈ <,
∞ ∞
Ŝ ( X, F ( X, t − s)) = S̃ ( X, F T F ( X, t − s)) ∀ F ∈ Lin+ , (4.53)
s =0 s =0

∞
where Ŝ is the response function for the second Piola–Kirchhoff stress.
s =0

Proof. We shall show that (i) =⇒ (ii) =⇒ (iii) =⇒ (i). Let

F ( X, t) = R( X, t)U ( X, t) be the polar decomposition of F. We then have
F ( X, t − s) = R( X, t − s)U ( X, t − s). Since R( X, t − s) ∈ Orth+ , we choose
Q(t − s) ≡ R( X, t − s) (we can make this choice since X is fixed), and since we
also have U ∈ Lin+ , Eqn. (4.51) implies that
∞ ∞
τ̂ ( X, F ( X, t − s)) = τ̂ ( X, R( X, t − s)U ( X, t − s))
s =0 s =0
∞
= R( X, t) τ̂ ( X, U ( X, t − s)) R( X, t)T .
s =0

To show that (ii) implies (iii), note that the above equation can be written as
∞ ∞
τ̂ ( X, F ( X, t − s)) = FU −1 ( X, t) τ̂ ( X, U ( X, t − s))U −1 F ( X, t) T ,
s =0 s =0

so that
∞
∞

−1
Ŝ ( X, F ( X, t − s)) = (det F ) F τ̂ ( X, F ( X, t − s)) F −T
s =0 s =0

i i

i i
i i

“CM˙Final” — 2015/3/13 — 11:09 — page 225 — #247

i i

Constitutive Equations 225

∞
= S̃ ( X, F T F ( X, t − s)),
s =0

where
∞ ∞
S̃ ( X, C ( X, t − s)) := (det U )U −1 ( X, t) τ̂ ( X, U ( X, t − s))U −1 ( X, t),
s =0 s =0
√
U= C, ∀C ∈ Psym.
Note that we have written the above relation for all C ∈ Psym, since for any
C ∈ Psym, we have C = UU = U T U, where U ∈ Psym ⊂ Lin+ .
Finally, (iii) implies (i) since
∞ ∞
1
τ̂ ( X, Q(t − s) F ( X, t − s)) = Q(t) F ( X, t) Ŝ ( X, Q(t − s) F ( X, t − s)) F ( X, t) T Q(t) T
s =0 det( QF ) s =0
∞
= (det F )−1 Q(t) F ( X, t) S̃ ( X, F T F ( X, t − s)) F ( X, t)T Q(t)T
s =0
h ∞ i
= Q(t) (det F )−1 F ( X, t) Ŝ ( X, F ( X, t − s)) F ( X, t)T Q(t)T
s =0
∞
= Q(t) τ̂ ( X, F ( X, t − s)) Q(t)T .
s =0

Using Eqns. (4.47), we get the following restrictions on the constitutive relations for the
first and second Piola–Kirchhoff stress:
∞ ∞
T̂ ( X, Q(t − s) F ( X, t − s)) = Q(t) T̂ ( X, F ( X, t − s)) ∀ F ∈ Lin+ , Q ∈ Orth+ , (4.54)
s =0 s =0
∞ ∞
Ŝ ( X, Q(t − s) F ( X, t − s)) = Ŝ ( X, F ( X, t − s)) ∀ F ∈ Lin+ , Q ∈ Orth+ . (4.55)
s =0 s =0

∞ ∞
It is clear that the response functions T̂ and Ŝ satisfy the axiom of material frame-
s =0 s =0
indifference if and only if the above relations are satisfied. Equation (4.52), known as
Richter’s theorem, says that the response function τ̂ is completely determined by its re-
striction to Psym. Another form of Richter’s theorem that is useful is obtained by defining
∞ ∞
τ̃ ( X, C ( X, t − s)) := U −1 ( X, t) τ̂ ( X, U ( X, t − s))U −1 ( X, t), C = U 2.
s =0 s =0

∞
Then, we have the result that the response function τ̂ ( X, C ( X, t − s)) is frame-indifferent
s =0
if and only if the stress is given by the constitutive relation
∞
τ ( x, t) = F ( X, t) τ̃ ( X, C ( X, t − s)) F ( X, t) T . (4.56)
s =0

Eqn. (4.53), which relates the right Cauchy–Green tensor (and hence the Lagrangian
measure of strain) to the stress tensor, explains why constitutive relations are known as
stress–strain relations.

i i

i i
i i

“CM˙Final” — 2015/3/13 — 11:09 — page 226 — #248

i i

226 Continuum Mechanics

4.6 Material Symmetry

The treatment in this and the following section is based on [35].
In general, a transformation of an object that leaves some property of the object invari-
ant is called a symmetry transformation. For example, if a square is rotated by an integer
multiple of 90o in its plane, it appears the same as before the rotation. Thus, such a rota-
tion, which leaves the appearance of the square invariant, is a symmetry transformation.
We now show by verifying the properties discussed in Section 1.12 that the set of symmetry
transformations is a group.
The composition of any of these symmetry transformations is itself a symmetry trans-
formation, e.g., a 180o clockwise rotation followed by a 90o anticlockwise rotation is just a
clockwise 90o rotation. Thus, the operation of composition of symmetry transformations
of the square is closed.
The composition of these symmetry transformations is also associative, e.g.,

[(90o CW) ◦ (90o CW)] ◦ (270o CCW) = 90o CW ◦ [(90o CW) ◦ (270o CCW)].

The inverses of these symmetry transformations are also symmetry transformations,

e.g., the inverse of the 90o CW rotation is just the 90o CCW rotation.
Finally, the identity transformation, i.e., do nothing to the square, is a symmetry trans-
formation.
Thus, the set of all symmetry transformations of the square is a group with respect to
composition.
Now we consider symmetry transformations of material bodies2 . In the context of con-
tinuum mechanics, the material symmetry of a body is characterized by those changes of
reference configuration that do not alter the response function. Referring to Fig. 4.6, the
stress response with respect to the configuration κ is related to that with respect to the
configuration µ by
∞ ∞
τ̂ κ ( X, ∇( f ◦ g )( X )) = τ̂ µ (Y, ∇ f (Y )),
s =0 s =0

where Y = g ( X ). Let F = ∇ f (Y ) and G = ∇ g ( X ). Then, by the chain rule

∇( f ◦ g )( X ) = ∇ f (Y ) ◦ ∇ g ( X ) = FG.

Hence,
∞ ∞
τ̂ κ ( X, FG ) = τ̂ µ (Y, F ) ∀ F, ∈ Lin+ . (4.57)
s =0 s =0

Since X and Y are the positions occupied by the same material particle in different config-
urations, the stress response with respect to the configurations is the same if
∞ ∞
τ̂ κ ( X, A) = τ̂ µ (Y, A) ∀ A ∈ Lin+ , (4.58)
s =0 s =0

2 We shall write F ( X, t − s) simply as F.

i i

i i
i i

“CM˙Final” — 2015/3/13 — 11:09 — page 227 — #249

i i

Constitutive Equations 227

µ
Y = g(X)
f

f ◦g

X κ

Fig. 4.6 Stress response with respect to two different reference configurations.

Thus, from Eqns. (4.57) and (4.58), the condition for the stress response at a point X to be
the same is
∞ ∞
τ̂ κ ( X, FG ) = τ̂ κ ( X, F ) ∀ F, G ∈ Lin+ . (4.59)
s =0 s =0

We need to impose certain restrictions on G in the above equation in order that

Eqn. (4.59) makes sense in a physical setting; more specifically, we need G ∈ Unim+ . The
reason is as follows. By applying Eqn (4.59) repeatedly, we see that
∞ ∞
τ̂ κ ( X, FG n ) = τ̂ κ ( X, F ) ∀n ≥ 1.
s =0 s =0

Since F, G ∈ Lin+ , we have F n ≡ FG n ∈ Lin+ . If det G > 1, then we have det G n → ∞,

∞ ∞
but the stress τ̂ ( X, F n ) = τ̂ ( X, F ) remains the same. Since ρκ = ρµ det G, we can find a
s =0 s =0
configuration, say µ0 , with deformation gradient G n with respect to κ, whose density given
by ρκ / det G n can be made arbitrarily close to zero by choosing n large enough, but whose
stress is still the same as that in the configuration κ. Clearly, this does not make sense
physically. Similarly, we can rule out the case det G < 1, since in that case we can have a
configuration whose density approaches infinity, but whose stress remains the same as in
the original configuration. Based on the above discussion, we restrict G to lie in the proper
unimodular group. Physically, FG with G ∈ Unim+ corresponds to a density-preserving
deformation (det G = 1), followed by a deformation with gradient F.
We define the material symmetry group of a particle X that occupies the position X in the
reference configuration κ as the set
∞ ∞
∂κX := { H ∈ Unim+ : τ̂ κ ( X, F H ) = τ̂ κ ( X, F )}. (4.60)
s =0 s =0

Note that the above definition is for a single particle X. Thus, H ∈ ∂κX is not a field, but
rather, it is just the gradient at X of a deformation of κ into a configuration with the same
material properties at the particle X.

i i

i i
i i

“CM˙Final” — 2015/3/13 — 11:09 — page 228 — #250

i i

228 Continuum Mechanics

Two mathematical structures are said to be isomorphic if there is a one-to-one corre-

spondence between them that preserves their structure. Thus, two reference configurations
κ and µ are said to be materially isomorphic at a particle X if they have the same properties
at X, i.e., if

ρκ ( X ) = ρµ (Y ) and τ κ ( X ) = τ µ (Y ).

In view of the above considerations, we can say that the two reference configurations κ and
µ are materially isomorphic at X if and only if G = ∇ g ∈ ∂κX .
As with the plane rotations of the square, the set of symmetry transformations is a
group with respect to composition of the symmetry transformations.

Theorem 4.6.1. The material symmetry group ∂κX is a group with respect to tensor
multiplication.
Proof. We shall drop the explicit dependence of τ̂ on X for notational conve-
nience. Let H 1 , H 2 ∈ ∂κX . Then necessarily H 1 , H 2 ∈ Unim+ , and since Unim+
is a group with respect to tensor multiplication, H 1 H 2 ∈ Unim+ . Also, for any
F ∈ Lin+ , we have
∞ ∞
τ̂ κ ( F ( H 1 H 2 )) = τ̂ κ (( F H 1 ) H 2 )
s =0 s =0
∞
= τ̂ κ ( F H 1 ) (since H 2 ∈ ∂κX )
s =0
∞
= τ̂ κ ( F ) (since H 1 ∈ ∂κX ).
s =0

Therefore, H 1 H 2 ∈ ∂κX , i.e., closure is satisfied.

The remaining axioms are easy to check.
1. If H 1 , H 2 , H 3 ∈ ∂κX , then H 1 , H 2 , H 3 ∈ Unim+ , and by the associativity of
tensor multiplication
( H 1 H 2 ) H 3 = H 1 ( H 2 H 3 ).

2. The neutral element is I ∈ ∂κX .

3. We need to show that if H ∈ ∂κX , then H −1 ∈ ∂κX . H ∈ ∂κX implies that
H ∈ Unim+ , which in turn implies that there exists H −1 ∈ Unim+ , since
Unim+ is a group. It remains to be shown that H −1 ∈ ∂κX . We have
∞ ∞
τ̂ κ ( F H −1 ) = τ̂ κ (( F H −1 ) H ) (since H ∈ ∂κX )
s =0 s =0
∞
= τ̂ κ ( F ( H H −1 )) (associativity)
s =0
∞
= τ̂ κ ( F ),
s =0

which proves that H −1 ∈ ∂κX .

i i

i i
i i

“CM˙Final” — 2015/3/13 — 11:09 — page 229 — #251

i i

Constitutive Equations 229

Since the formulation of material symmetry is in terms of an underlying configuration

κ, we now study the effect of using a different reference configuration µ, not necessarily
materially isomorphic with κ.

Theorem 4.6.2. Let the reference configurations κ and µ be connected through a de-
formation g as shown in Fig. 4.6 with G = ∇ g, where G is not necessarily a proper
unimodular tensor. Then for any material particle X,

H ∈ ∂κX ⇐⇒ G ( X ) HG −1 ( X ) ∈ ∂µX , (4.61)

where X = κ ( X ). This rule is referred to as Noll’s rule, and written suggestively as

∂µX = G ( X )∂κX G −1 ( X ).

Proof. We drop the explicit dependence of τ̂ on X for notational convenience.

Suppose H ∈ ∂κX . Then, H ∈ Unim+ , and

det[ G ( X ) HG −1 ( X )] = (det G ( X ))(det H )(det G −1 ( X ))

= det H
= 1.

Thus, G ( X ) HG −1 ( X ) ∈ Unim+ . We also have

∞ ∞
τ̂ µ ( FG ( X ) HG −1 ( X )) = τ̂ κ ( FG ( X ) HG −1 ( X ) G ( X )) (by Eqn. (4.57))
s =0 s =0
∞
= τ̂ κ ( FG ( X ) H )
s =0
∞
= τ̂ κ ( FG ( X )) (H ∈ ∂κX )
s =0
∞
= τ̂ µ ( F ) (by Eqn. (4.57)).
s =0

Thus, G ( X ) HG −1 ( X ) ∈ ∂µX .
Conversely, if G ( X ) HG −1 ( X ) ∈ ∂µX , then the following argument using
Eqn. (4.57) shows that H ∈ ∂κX :
∞ ∞
τ̂ κ ( F H ) = τ̂ κ ( F HG −1 ( X ) G ( X ))
s =0 s =0
∞
= τ̂ µ ( F HG −1 ( X ))
s =0
∞
= τ̂ µ ( FG −1 ( X ) G ( X ) HG −1 ( X ))
s =0
∞
= τ̂ µ ( FG −1 ( X ))
s =0

i i

i i
i i

“CM˙Final” — 2015/3/13 — 11:09 — page 230 — #252

i i

230 Continuum Mechanics

∞
= τ̂ κ ( FG −1 ( X ) G ( X ))
s =0
∞
= τ̂ κ ( F ).
s =0

Observe that if the reference configuration µ is obtained from the reference configura-
tion κ by a dilatation, i.e., G = λI where λ > 0, then ∂κX = ∂µX (see Problem 1). Thus, it is
possible to have the same symmetry group with respect to different configurations κ and µ
at a given particle X with unequal densities at the particle in the two configurations.
For orthogonal members of the symmetry group, we can get special results by using
the principle of material frame-indifference.

Theorem 4.6.3. Let Q ∈ Orth+ . Then Q ∈ ∂κX if and only if

∞ ∞
τ̂ κ ( X, QFQ T ) = Q τ̂ κ ( X, F ) Q T . (4.62)
s =0 s =0

Proof. As usual, we drop the dependence on X for notational convenience. Sup-

pose that Q ∈ ∂κX . Then, we have
∞ ∞
τ̂ κ ( F ) = τ̂ κ ( FQ) ∀ F ∈ Lin+ .
s =0 s =0

Since QFQ ∈ Lin+ , we have

∞ ∞
τ̂ κ ( QFQ T ) = τ̂ κ ( QFQ T Q)
s =0 s =0
∞
= τ̂ κ ( QF )
s =0
∞
= Q τ̂ κ ( F ) Q T (by material frame-indifference).
s =0

Conversely, suppose that

∞ ∞
τ̂ κ ( QFQ T ) = Q τ̂ κ ( F ) Q T ∀ F ∈ Lin+ .
s =0 s =0

By using the principle of material frame-indifference, the above equation be-

comes
∞ ∞
τ̂ κ ( QFQ T ) = τ̂ κ ( QF ) ∀ F ∈ Lin+ .
s =0 s =0

With Q T FQ playing the role of F, this becomes

∞ ∞
τ̂ κ ( F ) = τ̂ κ ( FQ),
s =0 s =0

which implies that Q ∈ ∂κX , since Q ∈ Unim+ .

i i

i i
i i

“CM˙Final” — 2015/3/13 — 11:09 — page 231 — #253

i i

Constitutive Equations 231

The two previous results were derived without using a constitutive equation that au-
tomatically satisfies material frame-indifference. If we start with the reduced form of the
constitutive equation given by Eqn. (4.56), we get the following result:

∞
Theorem 4.6.4. In terms of the reduced response function τ̃ , the material group of the
s =0
particle X relative to the configuration κ is given by
∞ ∞

∂κX = H ∈ Unim+ : H τ̃ κ ( H T CH ) H T = τ̃ κ (C ) ∀C (t) ∈ Psym .
s =0 s =0

Proof. By definition, a unimodular tensor H belongs to ∂κX if and only if

∞ ∞
τ̂ κ ( F H ) = τ̂ κ ( F ) ∀ F ∈ Lin+ .
s =0 s =0

By using Eqn. (4.56), the above requirement for a given particle X can be written
as
∞ ∞
F (t) H τ̃ κ (( F H ) T ( F H ))( F (t) H ) T = F (t) τ̃ κ (C ) F T (t) ∀ F ∈ Lin+ .
s =0 s =0

Using the facts that F is invertible and F T F = C, we get

∞ ∞
H τ̃ κ ( H T CH ) H T = τ̃ κ (C ) ∀C (t) ∈ Psym.
s =0 s =0

The requirement that the above condition hold for all C (t) ∈ Psym follows from
the fact that every C ∈ Psym can be written as UU = U T U, where U belongs to
Psym, and hence to Lin+ . Thus, every F ∈ Lin+ generates C = F T F ∈ Psym,
and every C ∈ Psym can be written as F T F, where F ∈ Lin+ .

As a corollary of the above theorem, we have the following companion to Theorem 4.6.3:

Theorem 4.6.5. Let Q ∈ Orth+ . Then Q ∈ ∂κX if and only if

∞ ∞
τ̃ κ ( QCQ T ) = Q τ̃ κ (C ) Q T ∀C (t) ∈ Psym.
s =0 s =0

Since C ∈ Psym is a measure of strain, Theorem. 4.6.5 can be expressed as saying:

‘Rotating the strain by an element of the symmetry group rotates the stress in the same
manner’.

4.7 Classification of Materials

In this section, we give precise definitions of various terms such as isotropy, solid,
fluid etc.

i i

i i
i i

“CM˙Final” — 2015/3/13 — 11:09 — page 232 — #254

i i

232 Continuum Mechanics

If the material properties at a particle X of a body are unaltered by rotations of the

reference configuration, we say that the material is isotropic. Since the rotation of the ref-
erence configuration corresponds to the examination of properties in different directions,
this definition corresponds to the intuitive idea of the material properties of an isotropic
body being independent of direction. More precisely, we have
Definition: A body is isotropic at a point if there exists a reference configuration κ, called
undistorted, such that

Orth+ ⊂ ∂κX ,

or, in other words,

∞ ∞
τ̂ κ ( X, FQ) = τ̂ κ ( X, F ) ∀ F ∈ Lin+ , Q ∈ Orth+ . (4.63)
s =0 s =0

Physically, the above equation corresponds to the idea that the stress remains the same
even if we carry out an arbitrary rotation of the body about the point X before carrying out
the deformation F. Since χ = FQX + c(t) in this deformed configuration, an alternative
interpretation is that the Cauchy stress remains unaltered when the reference configura-
tion is rotated by Q around X. Note that while characterizing frame-indifferent materials,
we rotate the deformed configuration, so that the product QF appears in the characteriza-
tion, while in the case of characterizing isotropy, we rotate the material and then carry out
the deformation, which results in the appearance of the product FQ. If Eqn. (4.63) is not
satisfied for some Q ∈ Orth+ , then the material is said to be anisotropic at X.
It is important to note that isotropy is a property that holds only with respect to spe-
cial configurations, namely, the undistorted configurations κ in Eqn. (4.63). The material
symmetry group relative to any other reference configuration µ is given by Eqn. (4.61) as

∂µX = G ( X )∂κX G −1 ( X ),

where G = ∇ g is the deformation gradient of the deformation taking κ to µ. It is possible

that ∂µX may not contain any orthogonal tensors, let alone all of them. However, if the ref-
erence configuration µ is obtained by a rigid deformation of an undistorted configuration
κ, then µ is also an undistorted configuration. This follows since G = Q ∈ Orth+ , and
thus, for every R ∈ Orth+ , we have Q T RQ ∈ Orth+ ⊂ ∂κX , which on applying Noll’s
rule yields the fact that R ∈ ∂µX , or, in other words, Orth+ ⊂ ∂µX . Thus, while undistorted
configurations are special, they are not unique.
By combining the notions of isotropy and material frame-indifference, we have the fol-
lowing result:

∞
Theorem 4.7.1. The response functional τ̂ is isotropic and material frame-indifferent
s =0
∞
at a particle X with respect to a configuration κ if and only if there exists a mapping τ̄ ,
s =0
such that
∞ ∞
τ̂ κ ( X, F ( X, t − s)) = τ̄ κ ( X, FF T ( X, t − s)) ∀ F ∈ Lin+ , (4.64)
s =0 s =0

i i

i i
i i

“CM˙Final” — 2015/3/13 — 11:09 — page 233 — #255

i i

Constitutive Equations 233

and that satisfies

∞ ∞
τ̄ κ ( X, QBQ T ) = Q τ̄ κ ( X, B) Q T ∀ B(t) ∈ Psym, Q(t) ∈ Orth+ . (4.65)
s =0 s =0

Proof. We drop the dependence on X in the proof. Note that in the polar de-
composition F = V R, V ∈ Lin+ and R ∈ Orth+ . Assuming that the particle is
isotropic at X, by Eqn. (4.63) we have
∞ ∞ ∞ ∞
τ̂ κ ( F ) = τ̂ κ (V R) = τ̂ κ (V ) = τ̄ κ ( B),
s =0 s =0 s =0 s =0

where V 2 = B = FF T . Thus,
∞ ∞
Q τ̄ κ ( B) Q T = Q τ̂ κ ( F ) Q T
s =0 s =0
∞
= τ̂ κ ( QFQ T ) (by Theorem 4.6.3)
s =0
∞
= τ̄ κ ( QFQ T QF T Q T )
s =0
∞
= τ̄ κ ( QFF T Q T )
s =0
∞
= τ̄ κ ( QBQ T ).
s =0

Since every B(t) ∈ Psym can be expressed as V V = V V T , where V ∈ Lin+ , the

above equation holds for all B(t) ∈ Psym.
Conversely, assuming Eqns. (4.64) and (4.65) hold
∞ ∞
τ̂ κ ( FQ) = τ̄ κ ( FQQ T F T )
s =0 s =0
∞
= τ̄ κ ( FF T )
s =0
∞
= τ̂ ( F ) ∀ Q ∈ Orth+ ,
s =0

which proves that the particle is isotropic. Substituting B = FF T into Eqn. (4.65),
we have
∞ ∞
τ̄ κ ( QF ( QF ) T ) = Q τ̄ κ ( FF T ) Q T ,
s =0 s =0

which in turn implies that

∞ ∞
τ̂ κ ( QF ) = Q τ̂ κ ( F ) Q T ,
s =0 s =0

thus proving material frame-indifference.

i i

i i
i i

“CM˙Final” — 2015/3/13 — 11:09 — page 234 — #256

i i

234 Continuum Mechanics

If we pour a fluid from a container into a test-tube, the fluid undergoes large deforma-
tions. Yet, intuitively, it is clear that the density at a fluid particle is preserved, and that
the stress response at that particle is the same. Thus, if the stress response at a particle X
is unaltered by any density preserving deformation, we say that the material at X is fluid,
i.e., a fluid particle has the maximum amount of material symmetry.
Definition: A body is fluid at a particle X if there exists a reference configuration κ such
that

∂κX = Unim+ .

In contrast to the undistorted configuration in the definition of an isotropic particle, the

reference configuration in the definition of a fluid particle is of no special significance as
the following result shows.

Theorem 4.7.2. Let a simple body be fluid at a particle X. Then for any configuration
µ, ∂µX = Unim+ .

Proof. Let H ∈ Unim+ . Then Ĥ := G −1 ( X ) HG ( X ) ∈ Unim+ = ∂κX . By

Eqn. (4.61), the corresponding element of ∂µX is

G ( X ) ĤG −1 ( X ) = G ( X ) G −1 ( X ) HG ( X ) G −1 ( X ) = H,

i.e., H ∈ ∂µX . Hence, Unim+ ⊂ ∂µX . By definition, ∂µX ⊂ Unim+ , and therefore
∂µX = Unim+ .

Thus, a fluid particle has no preferred reference configurations; this is in contrast to the
‘undistorted configuration’ in the definition of an isotropic particle.
Suppose that a body is a fluid at a particle X. Then ∂κX = Unim+ for any reference
configuration κ, and since Orth+ ⊂ Unim+ , Orth+ ⊂ ∂κX . Thus, we have

Theorem 4.7.3. Fluid particles of bodies are isotropic with all reference configurations
being undistorted.

If we apply a load to a bar so that it gets deformed, and now choose the deformed bar
as the new reference configuration, and apply the same load, then one would expect the
response of the bar in the starting and new reference configurations to be different. Thus,
we would expect the stress response of a ‘solid’ to be altered by nonrigid deformation of
its reference configuration, although rigid deformations may or may not alter its response.
More precisely, we define a solid as follows:
Definition: A body is solid at a particle X if there exists a reference configuration κ, called
undistorted, such that

∂κX ⊂ Orth+ .

If ∂κX = I, then the solid is called triclinic.

i i

i i
i i

“CM˙Final” — 2015/3/13 — 11:09 — page 235 — #257

i i

Constitutive Equations 235

As in the isotropic particle case, the undistorted reference configuration κ in the

definition of a solid particle is of special significance, except when the solid is triclinic.
If µ is some other reference configuration, then by Noll’s rule

∂µX = G ( X )∂κX G −1 ( X ).

For Q ∈ Orth+ , generally G ( X ) QG −1 ( X ) does not belong to Orth+ . So, generally, ∂µX is
not a subgroup of Orth+ . However, we have

Theorem 4.7.4. Let a body be solid at a particle X with κ an undistorted reference

configuration. Then any reference configuration µ obtained by a rigid deformation of κ
is also undistorted.

The proof follows from the fact that Orth+ being a group, if R, Q ∈ Orth+ , then
RQR T ∈ Orth+ , and hence ∂µX ⊂ Orth+ .
Suppose that a body is solid at a particle X. Then there exists a reference configuration
κ such that ∂κX ⊂ Orth+ . For a particle to be a fluid, we need ∂κX = Unim+ . Since by
Theorem 1.12.10, Orth+ is a proper subgroup of Unim+ , we conclude that

Theorem 4.7.5. A particle of a body cannot be both solid and fluid.

However, a particle can be both solid and isotropic. To see this, suppose that a body is
solid at a particle X. Then there exists a reference configuration κ such that ∂κX ⊂ Orth+ . If
the particle X is also isotropic, then there exists a reference configuration κ̂ such that

Orth+ ⊂ ∂κ̂X ⊂ Unim+ .

Consequently, by the maximality of Orth+ in Unim+ (Theorem 1.12.11), either

∂κ̂X = Orth+ or ∂κ̂X = Unim+ .

The latter possibility implies that X is a fluid particle. However, by hypothesis, it is a solid
particle, and it cannot be both solid and fluid (Theorem 4.7.5). Hence,

∂κ̂X = Orth+ . (4.66)

By Eqn. (4.61),

Q̂ ∈ ∂κ̂X = Orth+ ⇐⇒ Q = G Q̂G −1 ∈ ∂κX ⊂ Orth+ ,

T
where G is the gradient of the deformation that takes κ̂ into κ. Since G = QG Q̂ , we have
T T
G T G = Q̂G T Q T QG Q̂ = Q̂G T G Q̂ ∀ Q̂ ∈ Orth+ ,
which, by Theorem 1.8.1, implies that G T G = λI for all possible deformation gradients G.
Since G T G ∈ Psym, we have λ > 0. Therefore,
T
1 1
√ G √ G = I.
λ λ

i i

i i
i i

“CM˙Final” — 2015/3/13 — 11:09 — page 236 — #258

i i

236 Continuum Mechanics

In terms of R defined by

1
R := √ G ∈ Orth+ ,
λ

we can write Q ∈ ∂κX as

Q = R Q̂R T . (4.67)

Now let Q̃ ∈ Orth+ . Then since Orth+ is a group, R T Q̃R ∈ Orth+ = ∂κ̂X . Letting R T Q̃R
play the role of Q̂ in Eqn. (4.67), we get

R( R T Q̃R) R T = Q̃ ∈ ∂κX .

Thus, Orth+ ⊂ ∂κX . But, by hypothesis, ∂κX ⊂ Orth+ . Hence,

∂κX = Orth+ .

Summarizing, we have

Theorem 4.7.6. Suppose that a particle X of a body is both solid, meaning that there
exists a reference configuration κ such that ∂κX ⊂ Orth+ , and isotropic, meaning that
there exists a reference configuration κ̂ such that Orth+ ⊂ ∂κ̂X . Then

∂κX = ∂κ̂X = Orth+ .

Thus, for an isotropic solid particle, the material symmetry groups relative to the special
reference configurations employed in the definition of solid and isotropic (both termed
undistorted) each√ coincide with the orthogonal group. Note that the deformation gradient
G is given by λR, i.e., in general, it is a scalar multiple of a rotation matrix. Hence,
the undistorted reference configuration κ and κ̂ need not be the same, and, in general,
need not even have the same density. However, by choosing λ = 1 and R = I, we can
make the undistorted reference configurations coincide, so that the undistorted reference
configuration associated with the isotropy is also an undistorted reference configuration
for the solidity, and vice versa.
We have seen that an isotropic particle of a body can be fluid (Theorem 4.7.3) or solid
(Theorem 4.7.6). A spectacular consequence of the maximality of the orthogonal group in
the unimodular group (Theorem 1.12.11) is that these are the only two possibilities for an
isotropic material.

Theorem 4.7.7. An isotropic particle of a body is either solid or fluid.

Proof. Suppose that a body is isotropic at a particle X. Then there exists a refer-
ence configuration κ such that

Orth+ ⊂ ∂κX ⊂ Unim+ .

i i

i i
i i

“CM˙Final” — 2015/3/13 — 11:09 — page 237 — #259

i i

Constitutive Equations 237

By the maximality of Orth+ in Unim+ , either

∂κX = Orth+ ,

in which case ∂κX ⊂ Orth+ , and X is a solid particle, or

∂κX = Unim+ ,

in which case X is a fluid particle.

In addition to fluid and solid particles, we can have yet another category of particles,
which we refer to as fluid crystals.
Definition: A particle X of a body is a fluid crystal if it is not solid, i.e., X is a fluid crystal
if there does not exist a reference configuration κ such that ∂κX ⊂ Orth+ , or, equivalently, if
∂κX 6⊂ Orth+ for all configurations κ.
Thus, for a fluid crystal, ∂κX always contains some nonorthogonal elements. While a
fluid crystal cannot be solid, it can be fluid or isotropic, as seen from the following result:

Theorem 4.7.8. A fluid crystal particle of a body is fluid if and only if it is isotropic.
Proof. Suppose if a fluid crystal particle is isotropic. Then by Theorem 4.7.7, it is
either a solid or fluid particle. But a fluid crystal, by definition, cannot be solid.
Hence, it is fluid.
Conversely, if a fluid crystal particle is fluid then it has to be isotropic, since
all fluid particles are isotropic by Theorem 4.7.3.

Based on the above discussion, a particle can either be solid or fluid or an anisotropic
fluid crystal. We need to show that anisotropic fluid crystals exist (at least mathematically).
First, we construct a proper subgroup of ∂κX that contains some nonorthogonal elements.
Let e ∈ V be a unit vector, and consider

G(e) := { H ∈ Unim+ : He = ±e}.

If H 1 , H 2 ∈ G(e), then H 1 H 2 ∈ G(e), since H 1 H 2 ∈ Unim+ , and H 1 H 2 e = ± H 1 e = ±e.

The other properties of associativity, and existence of identity and reverse elements can
also be proved easily. Thus, it follows that G(e) is a group.
As an example, let e = e1 . Then we see that
 
1 0 0
1 1 0
 

0 0 1

defines a proper unimodular tensor that is not in G(e), so G(e) 6= Unim+ . Similarly,
 
1 0 1
H = 0 1 0
 

0 0 1

i i

i i
i i

“CM˙Final” — 2015/3/13 — 11:09 — page 238 — #260

i i

238 Continuum Mechanics

defines an element of G(e), but it is nonorthogonal since H T H 6= I. Thus, G(e) is a group

such that
G(e) 6= Unim+ and G(e) 6⊂ Orth+ .
Next, we try and find a body, some particle of which has G(e) for its material symmetry
group. Consider the body defined by the constitutive relation
τ ( X, t) = α[ F ( X, t)e] ⊗ [ F ( X, t)e].
Since the constitutive relation involves only the deformation gradient, it is an example of
an elastic body (though not a solid elastic body, as we shall soon show). A direct calcu-
lation shows that the above constitutive relation satisfies the principle of material frame-
indifference.
Since α is a constant, this material is homogeneous, and its material symmetry group at
any particle X is given by
∂κX = { H ∈ Unim+ : ( F He) ⊗ ( F He) = ( Fe) ⊗ ( Fe) ∀ F ∈ Lin+ }.
We now show that ∂κX = G(e).
First, suppose that H ∈ G(e). Then H ∈ Unim+ , and He = ±e. Hence, we get
( F He) ⊗ ( F He) = ( Fe) ⊗ ( Fe) ∀ F ∈ Lin+ ,
i.e., H ∈ ∂κX , or G(e) ⊂ ∂κX .
Next, suppose that H ∈ ∂κX . Then H ∈ Unim+ , and
( F He) ⊗ ( F He) = ( Fe) ⊗ ( Fe) ∀ F ∈ Lin+ ,
This can be written as
F H (e ⊗ e) H T F T = F (e ⊗ e) F T ,
which, by the invertibility of F, implies
H (e ⊗ e) H T = e ⊗ e,
or,
( He) ⊗ ( He) = e ⊗ e.
Operating on both sides by e, we get
(e · He) He = e.
Clearly, e · He 6= 0 since, otherwise, the right-hand side of the above equation would be
zero. Taking the cross product of both sides of the above equation with e, and using Theo-
rem 1.2.1, we conclude that He = βe, where β is a constant to be determined. Substituting
He = βe into the above equation, we get
β2 = 1,
or, in other words, β = ±1. Therefore, He = ±e, i.e., H ∈ G(e), or ∂κX ⊂ G(e).
Thus, ∂κX = G(e), and we have a theoretical example of an anisotropic fluid crystal.
Summarizing, we have the following classification of materials:

i i

i i
i i

“CM˙Final” — 2015/3/13 — 11:09 — page 239 — #261

i i

Constitutive Equations 239

A particle X of a body is
• isotropic if Orth+ ⊂ ∂κX , κ undistorted;
• solid if ∂κX ⊂ Orth+ , κ undistorted;
• a fluid crystal if ∂κX 6⊂ Orth+ ∀ configurations κ.

• fluid if ∂κX = Unim+ ∀ configurations κ.

EXERCISES

1. Show that changes of configuration (κ → µ) defined by deformations of the form

g ( X ) = Y 0 + α ( X − X 0 ), α ∈ <,

do not affect the material symmetry group of a body, i.e., ∂κX = ∂µX . Deformations of
this type are called dilatations if α > 0, and central inversions if α < 0.

i i

EN - Complete Sinumerik Operate Milling - FULL PDF
100% (1)
EN - Complete Sinumerik Operate Milling - FULL PDF
714 pages
boolean circuits
No ratings yet
boolean circuits
54 pages
Applications of The Discrete Element Method in Mechanical Engineering
No ratings yet
Applications of The Discrete Element Method in Mechanical Engineering
14 pages
Closed-Form Time Derivatives of The Equations of Motion of Rigid Body Systems
No ratings yet
Closed-Form Time Derivatives of The Equations of Motion of Rigid Body Systems
17 pages
Dado 2005
No ratings yet
Dado 2005
12 pages
Mathematical and Computer Modelling: A. Labuschagne, N.F.J. Van Rensburg, A.J. Van Der Merwe
No ratings yet
Mathematical and Computer Modelling: A. Labuschagne, N.F.J. Van Rensburg, A.J. Van Der Merwe
11 pages
Dynamic Analysis of Geometrically Nonlinear Truss Structures S
No ratings yet
Dynamic Analysis of Geometrically Nonlinear Truss Structures S
7 pages
Denavit Hietnberg Parameters Research
No ratings yet
Denavit Hietnberg Parameters Research
17 pages
Marching Procedure For Tensegrity 2007
No ratings yet
Marching Procedure For Tensegrity 2007
26 pages
The Pedal Curves Generated by Alternative Frame Vectors and Their Smarandache Curves
No ratings yet
The Pedal Curves Generated by Alternative Frame Vectors and Their Smarandache Curves
28 pages
On The Numerical Implementation of Elasto-Plastic Constitutive Equations For Metal Forming
No ratings yet
On The Numerical Implementation of Elasto-Plastic Constitutive Equations For Metal Forming
16 pages
Design Optimization of Planar Mechanisms: Abstract
No ratings yet
Design Optimization of Planar Mechanisms: Abstract
12 pages
Warping+Bernoulli Beam Model
No ratings yet
Warping+Bernoulli Beam Model
6 pages
5th Order Polynomial Hilton Express
No ratings yet
5th Order Polynomial Hilton Express
10 pages
Mixed T-M Shell Elements
No ratings yet
Mixed T-M Shell Elements
26 pages
Prototype Design of A Translating Parall
No ratings yet
Prototype Design of A Translating Parall
29 pages
Lecture 3 (Updated)
No ratings yet
Lecture 3 (Updated)
21 pages
Briottet 1997
No ratings yet
Briottet 1997
19 pages
Stability and Dynamics of Offshore Singl PDF
No ratings yet
Stability and Dynamics of Offshore Singl PDF
25 pages
Plastic Behaviour at Cross Section Level
No ratings yet
Plastic Behaviour at Cross Section Level
27 pages
Alain Goriely and Michael Tabor - Nonlinear Dynamics of Filaments III: Instabilities of Helical Rods
No ratings yet
Alain Goriely and Michael Tabor - Nonlinear Dynamics of Filaments III: Instabilities of Helical Rods
20 pages
1973 - Journal of Structural Mechanics - 1 - 497-518 - WITTRICKW-An Algorithm For Computing Critical Buckling Loads of Elastic Structures
No ratings yet
1973 - Journal of Structural Mechanics - 1 - 497-518 - WITTRICKW-An Algorithm For Computing Critical Buckling Loads of Elastic Structures
23 pages
Bianchi Type-I, type-III and Kantowski-Sachs Solutions in F (T) Gravity
No ratings yet
Bianchi Type-I, type-III and Kantowski-Sachs Solutions in F (T) Gravity
8 pages
A Review of The Extended Finite Element Method On Macrocrack and Microcrack Growth Simulations
100% (1)
A Review of The Extended Finite Element Method On Macrocrack and Microcrack Growth Simulations
39 pages
8 Vertical Stresses Below Applied Loads
No ratings yet
8 Vertical Stresses Below Applied Loads
13 pages
Project Description V1
No ratings yet
Project Description V1
9 pages
Damping Sandwich Honeycomb
No ratings yet
Damping Sandwich Honeycomb
19 pages
amm 638-640 1716
No ratings yet
amm 638-640 1716
5 pages
Symmetry & Antisymmetry-1
No ratings yet
Symmetry & Antisymmetry-1
6 pages
Fatehi Et Al 2014 Modelling and Control of An Overhead Crane System With A Flexible Cable and Large Swing Angle
No ratings yet
Fatehi Et Al 2014 Modelling and Control of An Overhead Crane System With A Flexible Cable and Large Swing Angle
15 pages
PretwistedBeamReport
No ratings yet
PretwistedBeamReport
44 pages
ME 212 - Dynamics - Final Study Guide
No ratings yet
ME 212 - Dynamics - Final Study Guide
4 pages
Mathematical and Dynamical Modelling of A Quadcopter
No ratings yet
Mathematical and Dynamical Modelling of A Quadcopter
17 pages
Flexural Vibration of Propeller Shafts Using Distr
No ratings yet
Flexural Vibration of Propeller Shafts Using Distr
10 pages
Mit - 1998-01 (1) Shear Stress in Beams
No ratings yet
Mit - 1998-01 (1) Shear Stress in Beams
31 pages
master
No ratings yet
master
12 pages
10-Parametric Equations and Polar Coordinates
No ratings yet
10-Parametric Equations and Polar Coordinates
11 pages
Kintzel Fourth Order Tensors Part2
No ratings yet
Kintzel Fourth Order Tensors Part2
23 pages
Finite Element Method in Structure Ass-ALE
No ratings yet
Finite Element Method in Structure Ass-ALE
19 pages
Three Moment Equation Paper
No ratings yet
Three Moment Equation Paper
9 pages
Chen-Li2013 Article Size-dependentFreeVibrationAna
No ratings yet
Chen-Li2013 Article Size-dependentFreeVibrationAna
15 pages
Stiffness Matrix
50% (2)
Stiffness Matrix
25 pages
Constitutive
No ratings yet
Constitutive
3 pages
Mathematical Ship Modeling For Control Applications Perez Blanke
No ratings yet
Mathematical Ship Modeling For Control Applications Perez Blanke
23 pages
Geometric Spin Foams, Yang-Mills Theory and Background-Independent Models
No ratings yet
Geometric Spin Foams, Yang-Mills Theory and Background-Independent Models
28 pages
Project DescriptionV2
No ratings yet
Project DescriptionV2
9 pages
the thing to be uploaded
No ratings yet
the thing to be uploaded
23 pages
Axisymmetric Bending of Circular and Annular Sandwich Plates With Nonlinear Elastic Core Material
No ratings yet
Axisymmetric Bending of Circular and Annular Sandwich Plates With Nonlinear Elastic Core Material
12 pages
175-3-1309
No ratings yet
175-3-1309
12 pages
Volterra Integral Equations
No ratings yet
Volterra Integral Equations
14 pages
Article 2
No ratings yet
Article 2
16 pages
Micro-Mechanical Model of A Hyper-Elastic Open-Cell Foam: P. Hård Af Segerstad & S. Toll
No ratings yet
Micro-Mechanical Model of A Hyper-Elastic Open-Cell Foam: P. Hård Af Segerstad & S. Toll
5 pages
Lattice Construction of Mixed 'T Hooft Anomaly With Higher-Form Symmetry
No ratings yet
Lattice Construction of Mixed 'T Hooft Anomaly With Higher-Form Symmetry
8 pages
pp_232_CVC01
No ratings yet
pp_232_CVC01
22 pages
Slip Line Field Theory Project
No ratings yet
Slip Line Field Theory Project
14 pages
1952 G Gerard & S Wildhorn A Study of Poisson's Ratio in The Yield Region NASA TN 2561
No ratings yet
1952 G Gerard & S Wildhorn A Study of Poisson's Ratio in The Yield Region NASA TN 2561
31 pages
Flow behaviour
No ratings yet
Flow behaviour
8 pages
GE For FM
No ratings yet
GE For FM
9 pages
Preliminary Analysis of The Scaling Exponents in Channel Ow Turbulence
No ratings yet
Preliminary Analysis of The Scaling Exponents in Channel Ow Turbulence
9 pages
Mathematical Analysis 1: theory and solved exercises
From Everand
Mathematical Analysis 1: theory and solved exercises
Alessio Mangoni
5/5 (1)
Theory of Elastisity, Stability and Dynamics of Structures Common Problems
From Everand
Theory of Elastisity, Stability and Dynamics of Structures Common Problems
Konstantin Kazakov
No ratings yet
Physics For 9 Class (Unit # 3)
No ratings yet
Physics For 9 Class (Unit # 3)
6 pages
STEM 12 Module 8
100% (1)
STEM 12 Module 8
18 pages
CI 4 Velocity
No ratings yet
CI 4 Velocity
44 pages
Waves Gizmo Student Exploration Sheet
100% (1)
Waves Gizmo Student Exploration Sheet
8 pages
TASI Lectures On The Holographic Principle PDF
No ratings yet
TASI Lectures On The Holographic Principle PDF
57 pages
GRADE 11 AP PHYSICS 1-1
No ratings yet
GRADE 11 AP PHYSICS 1-1
5 pages
JEE - Physics - Kinematics
No ratings yet
JEE - Physics - Kinematics
28 pages
Toilet Paper Tie
100% (1)
Toilet Paper Tie
12 pages
CFD Simulation of Core Annular Flow Through Sudden Contraction and Expansion
No ratings yet
CFD Simulation of Core Annular Flow Through Sudden Contraction and Expansion
12 pages
MECDYNAMICS
No ratings yet
MECDYNAMICS
15 pages
Class-7: DPP # 01 Topic: Motion & Time Physics
No ratings yet
Class-7: DPP # 01 Topic: Motion & Time Physics
2 pages
Fundamentals of Physics Textbook Pages 283 294
No ratings yet
Fundamentals of Physics Textbook Pages 283 294
12 pages
2022-23 Class 8-Geog - Chpt.1 Local Time and Standard Time NOTES
No ratings yet
2022-23 Class 8-Geog - Chpt.1 Local Time and Standard Time NOTES
3 pages
G 8 Unit 2 Lesson 1
No ratings yet
G 8 Unit 2 Lesson 1
42 pages
Phys 121 Exam Eqn Sheet
No ratings yet
Phys 121 Exam Eqn Sheet
3 pages
9 Class Phy CH 1
No ratings yet
9 Class Phy CH 1
42 pages
Measurements (1)
No ratings yet
Measurements (1)
18 pages
Kinematics Part 2
No ratings yet
Kinematics Part 2
25 pages
General Physics 1 q1 Week 3
No ratings yet
General Physics 1 q1 Week 3
24 pages
Forces Answers
No ratings yet
Forces Answers
27 pages
Dynamics of Rigid Bodies Notes
No ratings yet
Dynamics of Rigid Bodies Notes
2 pages
On Gravity and The Forces Acting On Heavenly Bodies
No ratings yet
On Gravity and The Forces Acting On Heavenly Bodies
5 pages
Chapter 3: Spatial Descriptions & Transformations: Object Location
No ratings yet
Chapter 3: Spatial Descriptions & Transformations: Object Location
44 pages
Free-Body Diagrams
100% (1)
Free-Body Diagrams
16 pages
Physics Paper-1: Marks:60 Time 3 Hrs
No ratings yet
Physics Paper-1: Marks:60 Time 3 Hrs
1 page
Topic 6 (Telling Time)
100% (1)
Topic 6 (Telling Time)
9 pages
P1 - Midterm Review
No ratings yet
P1 - Midterm Review
7 pages
Time in A Block Universe
No ratings yet
Time in A Block Universe
25 pages
Density: Q) Which Weighs More:-A Kilogram of Feathers or A Kilogram of Iron?
No ratings yet
Density: Q) Which Weighs More:-A Kilogram of Feathers or A Kilogram of Iron?
24 pages