Download as docx, pdf, or txt
Download as docx, pdf, or txt
You are on page 1of 17

Equations of motion for a viscous fluid

Real and Ideal Fluids


A fluid is a material which flows. Its capacity to flow, which distinguishes it from
a solid, results from the fact that its particles can be readily displaced under the
action of shearing stresses. Like solids, the fluids known as the liquids offer
considerable resistance to compression and tensile stresses but all fluids deform
readily and
continuously under the action of a shearing stress, so long as that stress operates. If
the rate of shear deformation is small, the fluid offers negligible resistance. With
increased rates of deformation, it offers an increased resistance, which disappears,
however, once the deforming motion ceases. The resistance arises from the
existence of viscosity in the fluid.
Theoretical investigations into fluid mechanics in the 19 th century were mainly
based on the ideal fluid, i.e. a fluid which is inviscid and incompressible. It is
only since 20th. century that the effects of viscosity and compressibility have
been taken into account in any great way. In the flow of inviscid fluids, no
tangential forces (shear stresses) exist between adjacent layers; only normal
forces (pressures) do. The theory of flows of ideal fluids is mathematically very
highly developed and indeed in many cases gives a satisfactory description of real
flows, as for example, in the cases of wave motion and the formation of liquid jets.
On the other hand, the theory of ideal fluids is useless when faced with the
problem of calculating the drag of a body. It predicts that a body moving
subsonically
and uniformly through an infinitely extended fluid will experience no drag at all.
This unacceptable result of the theory of ideal fluids is due to the fact that in a real
fluid, in addition to the normal forces, tangential forces also occur both between
layers in the fluid and between the fluid and the solid walls. These tangential or
friction forces of real fluids are connected to a physical property called the
viscosity of the fluid.
Viscosity
Consider the flow between two long parallel plane plates, one of which is at rest
while the other moves with constant velocity U in its own plane. The distance
between the plates is h (Fig. 1). We assume that the pressure in the entire fluid is
constant. From experiment it is seen that the fluid adheres to both plates so that the
velocity at the lower plate is zero, while at the upper plate the fluid moves with
velocity U . In addition, we assume the simplest case (Newtonian fluid, constant
temperature), i.e. there is a linear velocity distribution between the plates.

1
Therefore the velocity is proportional to the distance z from the lower plate, and we
have
Z
u ( z )= U … … … … … … (1)
h

In order to maintain the state of the motion in the fluid, a tangential force in the
direction of motion must act on the upper plate. This keeps the friction forces of
the
fluid in equilibrium. According to experimental results, this force (force per
unit surface area of the plate = shear stress τ) is proportional to U /h. In the
general case, this can be replaced by du /dz . The constant of proportionality
between τ and du /dz , which we will denote by μ, depends on the nature
of the fluid, i.e. it is a physical property of the fluid.
Therefore, we have the elementary law of fluid friction in the form
du
τ =μ … … … … … … … … .(2)
dz
The value μ is a physical property of the fluid which is strongly dependent
on the temperature and is called the viscosity of the fluid. The friction law
given by Eq. (2) is called Newton’s law of friction. The shear stress
kg N
has the units or2 and the velocity gradient du /dz the unit sec−1.
sec 2 . m m
Therefore μ has the units of Pa . sec. Fluids where there is a nonlinear relation
between the shear stress τ and the velocity gradient du /dz are called non–
Newtonian fluids. Refer fig.2. Since all gases and many technically important
2
liquids, e.g. water, demonstrate Newtonian behaviour (i.e. Eq. (2) holds), we will
consider only Newtonian fluids in this course. As already stated, the viscosity is a
physical property. Since it establishes the momentum transport perpendicular to the
main flow direction, viscosity is also called a transport property of the fluid.

Isotropic Newtonian fluids


All gases and many liquids, in particular water, are members of this category. A
fluid is called isotropic if the relation between the stress tensor and the rate
of deformation tensor is the same in all directions. If this relation is linear, we
are dealing with a Newtonian fluid.
Description of Flow Fields
The equations of motion for a general (Newtonian) fluid will now be established.
In doing so, the fluid will be considered to be a continuum. In a continuum the
smallest volume element considered is still homogeneous, i.e. the dimensions of
the volume element are still very large compared to the average distance between
the molecules in the fluid.
In three–dimensional motion, the flow field is given by the velocity vector
⃗ ^ ^j v+ k^ w
V = iu+

3
Where i^ , ^j∧k^ the unit are vectors and u , v∧w are the magnitudes of the component
velocities along these directions in a Cartesian co-ordinate system. These velocity
components along with the pressure p and the temperature T describe a non-
isothermal field. To determine these five quantities, there are five equations
available: Continuity equation which ensures that mass in the flow field is
conserved; the three momentum equations that ensure the conservation of
momentum and finally the energy equation dealing with the First law of
thermodynamics.
Continuity Equation
Consider a control volume V bounded by a surface S that is fixed in space. The
assumption that mass is conserved means that for any fixed control
volume enclosed by a control surface—the rate of change of the mass contained in
that volume is equal to the rate at which mass is passing through the surface
from outside to inside, minus the rate at which mass is passing
from inside to outside. This can be expressed as an equation in integral form over
the control volume.

The mass of an element of fluid in the control volume is ρdv . So, the total mass
inside the control volume is given by∭ ρdv .
Now, the rate of decrease of mass in the control volume is expressed as
−d ∂ρ
dt ∭ ρdv=−∭
∂t
dv

Now, if mass is conserved, this expression must equal the total rate of mass flux
out of the control volume. The rate of outward mass flux across the total surface
area of the control volume is ∬ ⃗ṁ. d ⃗S , where⃗
ṁ is the local mass flux rate (= ρ ⃗
V¿

4
across the surface of the control volume; it is a vector. The quantity ⃗
dS represents
the local area vector.
Equating these two,
∂ρ
−∭ dv=¿ ∬ ⃗
ṁ. d ⃗S ¿
∂t
Using Stokes theorem, the above equation is rewritten as follows:
∂ρ
−∭ dv=¿ ∭ ( ⃗ ⃗
∇ ¿ . ṁ)dv ¿¿
∂t
Now, after removing the volume integrals, the continuity equation is finally
written as
∂ρ ⃗ ⃗
+ ∇ . ṁ=0
∂t
∂ρ ⃗ ⃗ ) =0 … … …(1)
¿, +∇ . ( ρ V
∂t
Generally, ρ=ρ ( x , y , z ,t ) . For steady flow, ρ=ρ ( x , y , z ).i.e. all fluid properties are,
by definition, independent of time. For this case,
∂ρ
=0
∂t
Accordingly, equation (1) reduces to the following form:
⃗ ⃗ ) =0
∇ .( ρ V
A further assumption that affords great simplification is incompressible fluid,
where the density changes are negligible i.e. ρ=constant . So, in this case, the
density term can be slipped out of the divergence in equation (1) which now
assumes the following form
⃗ ⃗ =0 … … … … .(1 a)
∇ .V
In this vector form, the equation is still quite general and can readily be applied to
other coordinate systems.
In three- dimensional Cartesian co-ordinate system,
⃗ ∂ ^ ∂ ^ ∂
∇=i^ +j +k
∂x ∂ y ∂z
⃗ ^ ^ ^
¿ ṁ=i ρu+ j ρv+ k ρw
On substitution into equation (1a), one obtains
∂u ∂ v ∂w
+ + =0 … … .(1 b)
∂x ∂ y ∂z
Momentum Equation
The momentum equation is the basic law of mechanics which states that
mass times acceleration in a direction is equal to the sum of the forces in that
direction. For the purpose of analysis, let us Consider a volume V bounded by a

5
material surface S that moves with the flow, always containing the same material
elements.
Inertia term: Let us consider an element of mass of fluid ( δm ) . In terms of the
local density, this element of mass is expressed asδm = ρdv
If ui be the velocity of this element of mass along the i−th direction, its momentum
in this direction is ( ui δm ) .
So, the rate of change of momentum of the element of mass along the i−th
direction is
d (ui δm) d ( ρui dv ) du
= = ρdv i
dt dt dt
The above expression represents the inertia force on an element of mass. On
integration over the whole mass, the rate of change of momentum along the i
direction is equal to
d ui
∭ρ dt
dv

External forces: The external forces acting on the element are of two types: body
forces and surface forces.
Body force may result from such effects as gravitational, electric and magnetic
fields. They act on the entire mass within the element. Refer fig. below where only
the body force due to gravity is shown. Now, in general, if F i represents the body
force per unit volume (associated with the massδm ¿ along the idirection, then the
total body force in the idirection, is equal to ∭ F i . dv

By surface forces, it is implied that such forces are active on the surfaces of the
fluid element. Refer fig. 7.5. In general, there are two types of surface forces - the
forces acting normal to a plane are called normal forces and those acing parallel to
a plane are called tangential forces. When we divide these forces by the respective
areas involved, we get stresses. For example, σ YY represents a normal stress acting

6
on the Y-plane while the τ YX and τ YZ are tangential stresses acting along the Y-plane
but acting along x∧z directions respectively.

In indicial notation, such stresses are denoted by a single symbol i.e. τ ji where the
second subscript i indicates the direction in which the τ ji acts. The first subscript j
indicates the direction which is perpendicular to area dA j along/on which τ ji acts.
In general, the total surface force on the mass of the fluid in thei−th direction
∂ τ ji
¿ ∬ τ ji d A j=∭ dv
∂xj
Taking into account all these surface and body forces, the conservation equation is
as follows:
d (ui ) ∂ τ ji
∭ρ dt
dv=∭ F i dv +∭
∂xj
dv

Removing the volume integrals, one obtains,


d ui ∂τ
ρ =F i+ ji … … ..(11)
dt ∂ xj
Unlike velocity⃗
V , which is a three-component vector, stress τ ji is a nine-component
tensor and require two subscripts to define each component. Refer appendix at the
∂ τ ji
end. It is not these stresses but their gradients i.e. ∂ x that cause a net force. In this
j

equation, τ ji is a double index quantity: the first position( j) indicates to which axis
the surface element is perpendicular, and the second position (i) states in which
direction the stress τ ji is pointing. For i=1,2,3 and j =1,2,3, it is seen that τ ji
comprises of nine components as shown below:
τ 11 τ 12 τ 13
τ 21 τ 22 τ 23
τ 31 τ 32 τ 33

7
The state of stress is, therefore, determined by nine scalar quantities and these
form the stress tensor. The nine components of the stress tensor, in together, are
also called the stress matrix.
The surface forces are due to the stresses on the sides of the control surface. These
stresses are the sum of hydrostatic pressure plus viscous stresses σ ijthat arise from
motion with velocity gradients. Thus,
τ ji =−P δ ij +σ ij

On substitution, we get the following expansion of the stress tensor.

Here the stresses have been decomposed additively into a part with the normal
stress (−P) that is the same in all directions, and a part which deviates
from this (deviator stresses).
The viscous stress tensor i.e. σ ijis symmetric: thus two tangential forces whose
indices only differ in their order are equal i.e. σ 12=σ 21. This will be shown by
considering the equation of motion of a fluid element. In general, this motion
can be decomposed into a translation and a rotation. For our aims we only
need to consider the latter. We restrict the rotation about the y-axis. The law of
rotation is
dI . θ̈ y =Torque about the y −axis

Here dI is the moment of inertia of the element considered. The element volume
(dV ) is dxdydz and its density is ρ . θ̈ y is the rotational acceleration about the y-axis.
dI . θ̈ y =Torque about the y −axis
¿ , dI . θ̈ y =( σ xz −σ zx ) dV
Now the moment of inertia dI is proportional to the fifth power of the linear
dimension of the parallelepiped, while the volume element is proportional to the

8
third power. Making the transition to a very small volume element, the left hand
side of the above equation vanishes faster than the right hand side. Therefore
we have
σ xz −σ zx =0∨σ xz=σ zx
On substitution into the momentum equation, we obtain
d ui ∂ σ ij ∂ P
ρ =F i+ −
dt ∂ x j ∂ xi
This is the differential momentum equation in its full glory, and it is valid for any
fluid in any general motion, particular fluids being characterized by particular
viscous stress terms.
General State of Deformation of Flowing Fluids

Refer fig. above. A fluid element of initially square shape changes its position in
the flow field. During the movement its shape may change. As can be seen, it may
undergo a dilatation (characterized by expension or contraction of the element); a
solid body rotation may occur and finally, it may undergo a deformation in shape.
Our present objective is to calculate the strain rate associated with these processes
so that we can formally connect it to the stress σ ij .
Translation of the fluid element
The fluid element merely changes its position in the flow field without being
deformed at all. If there are no velocity gradients at all in a flow, then there are no
accompanying stresses in the flow field. So, ϵ̇ ij =0 i . e . this phenomenon does not
contribute to the strain rate.

9
Volumetric dilatation
The volumetric dilatation rate is a measure of how much a fluid element is
shrinking or expanding per unit time. Consider the following two-dimensional
fluid element.

Let us consider a two dimensional fluid element with unit thickness in the z-
direction. At time t =0 , the volume of the element ( V t )=dx . dy . After an infinitesimal
time period dt ,there is a change in the length of the sides of the element . The new
∂u ∂v
side lengths are dx + ∂ x dx . dt and dy + ∂ y . dy .dt

10
So, the new volume is given by

(
V t +dt = dx+
∂u
∂x )(
dx .dt dy +
∂v
∂y
. dy . dt )
So the volumetric strain rate is given by

1 dV V t+ dt
=
−V
=
t
( dx +
∂u
∂x
dx . dt )( dy+
∂v
∂y
. dy . dt )−dx . dy

V dt V t . dt dx .dy . dt

Neglecting terms containing (dt )2 ,it can be shown that


∂u ∂v
ϵ̇ ij = +
∂x ∂y
For a three dimensional element,
∂u ∂v ∂w ⃗ ⃗
ϵ̇ ij = + + = ∇ .V
∂ x ∂ y ∂z
Throughout this transformation, the shape of the element, described by the angles
between the sides, remains unchanged since all right angles are conserved. Thus ϵ̇ ij
describes the local instantaneous volume dilatation of the fluid element. If the fluid
is incompressible, we have ⃗ ⃗ =0 ;consequently, ϵ̇ ij =0.So, for an incompressible
∇ .V
fluid, the strain rate due to volume dilatation is zero. However, for compressible
fluid, we have, from continuity equation
ϵ̇ ij = ⃗ ⃗ = −1 dρ
∇ .V
ρ dt
Note that volume dilatation occurs when is set equal to j ; consequently, there is
i
no mixed derivative.
Pure shear deformation
The relative velocity field has a completely different form if one of the off–
diagonal terms of the matrixϵ̇ ij , e.g. ∂u/∂y, does not vanish but is, for example,
positive. The situation is depicted below:

The original right angle at point A changes by

11
∂u
. δy . δt
∂y ∂u
tanδθ= = . δt
δy ∂y
For small values of the deformation angle,we set tanδθ ≈ δθ so that
∂u
δθ= . δt
∂y
The shear angular velocity is given by
∂u
. δy
dθ ∂ y ∂u
= =
dt δy ∂y
∂u
∧∂ v
Now, let us consider the case of deformation of the element when both ∂ y are
∂x
non-zeros and positive quantities. The shape of the deformed element is as shown
in the following figure:

According to the diagram,


∂u
. δy . δt
∂y ∂u
∆ θ 1= = . δt
δy ∂y
∆ θ1 ∂ u
¿ , d θ̇1= lim =
δt → 0 δt ∂y
∂v
Similarly,d θ̇2= ∂ x . So the average rate of shearing strain or rate of angular
deformation is expressed as follows:
ϵ̇ xy=
1
2
( d θ̇1 +d θ̇2 )= (
1 ∂u ∂ v
+
2 ∂ y ∂x )
In tensor notation, it is expressed as follows:
ϵ̇ ij = (
1 ∂ ui ∂u j
+
2 ∂ x j ∂ xi )
for i ≠ j

12
∂u i −∂u j
When ∂ x = ∂ x , the fluid element will undergo rotation about point A. In this
j i

case, we can conclude that the right angle at point A is not altered i.e. a solid body
rotation without any deformation is obtained. Refer fig. below

Thus we have considered translation, rigid body rotation, pure shear and volume
dilatation. Only the final two motions lead to a deformation of the fluid element
about the reference point; the first two merely lead to a change in position.

Constitutive relation
A Constitutive relation is one that expresses the relationship between the stress and
the rate of strain tensor for incompressible flow of a Newtonian fluid is expressed
as follows:
τ ji =−P δ ij +σ ji

Here , σ ji =2 μ ϵ̇ ij =μ
∂u i ∂ u j
+
∂ x j ∂ xi( )
The reason behind assuming σ ji =2 μ ϵ̇ ij instead of σ ji =μ ϵ̇ ij has been discussed in
detail by Cohen and Kundu (1) and Bachelor (2). They will not be discussed here.
Thus the general form is
τ ji =−P δ ij + μ
( ∂ ui ∂ u j
+
∂ x j ∂ xi )
Modifications for compressible flow
So far we did not consider the volume dilatation because of the incompressibility
condition. For compressible flows, the strain rate due to volume dilatation is
ϵ̇ ij = ⃗ ⃗ = −1 dρ
∇ .V
ρ dt
To include it into the stress equation, we multiply this strain rate by a quantity λ
where λ represents the second co-efficient of viscosity. The expression for stress is
consequently modified as follows:
τ ji =−P δ ij + μ
( ∂ ui ∂ u j
+
∂ x j ∂ xi )
+δ ij λ ¿

13
Note that Pis the thermodynamic pressure related to ρ and T by an equation of state
(e.g., the thermodynamic pressure for a perfect gas is p = ρRT ).
Setting i= j, we obtain
∂u ⃗ ⃗
τ xx =−P+ 2 μ + λ∇.V
∂x
∂v
τ yy =−P+2 μ +λ ⃗
∇ .⃗
V
∂y
∂w ⃗ ⃗
τ zz =−P+2 μ + λ∇.V
∂z
Adding these three equations, we obtain
τ xx +τ yy+ τ zz=−3 P+2 μ ( ∂∂ ux + ∂∂ vy + ∂∂wz )+3 λ ⃗∇ .V⃗
¿ , τ xx +τ yy + τ zz =−3 P+ 2 μ ⃗ ⃗ +3 λ ⃗
∇ .V ∇.⃗
V

¿ , τ xx +τ yy + τ zz =−3 P+(2 μ+3 λ) ⃗


∇.⃗
V

If we define the mean pressure P as the negative one-third of the sum of the
three normal stresses, then
−3 P=−3 P+(2 μ+ 3 λ) ⃗ ⃗
∇ .V

¿ , P=P− ( 2
3
μ+ λ ⃗ )⃗
∇ .V

Here, P∧P represent the mechanical pressure and thermodynamic pressure


( 2 ) is often termed the coefficient of bulk viscosity
respectively. The factor 3 μ+ λ

This equation implies that, unless either ( 3 μ+ λ) or .Uis equal to zero, the
2

mean (mechanical) pressure in a deforming viscous fluid is not equal to the


thermodynamic pressure.
Physically, this factor is connected with the dissipation mechanism during a
change of volume at a finite rate. From the macroscopic viewpoint, it characterizes
the resistance to dilatation of an infinitesimal bulk element at constant shape. In
this context, let us reconsider the constitutive relation i.e.

τ ji =−P δ ij + μ
( ∂ ui ∂ u j
+
∂ x j ∂ xi )
+δ ij λ ¿

Setting i= j=1, we obtain


∂u
τ 11=−P+ 2 μ +λ¿
∂x

14
( 2
)
In 1845, Stokes simply assumed that 3 μ+ λ = 0. The resulting negative value of λ
implies that the tension i.e.(τ 11) required to produce a specified proportional rate of
stretching along one principal axis is reduced if the fluid is locally expanding.
It is strictly zero only for dilute monoatomic gases. The second law of
(2 )
thermodynamics requires that both μ and 3 μ+ λ to be non-negative.
Consider propagation of disturbances, in particular sound waves in hypothetical
equilibrium fluids. These waves propagate with certain velocity through the
medium and can be damped by the following irreversible processes: viscosities,
thermal conduction, and chemical reactions. The bulk viscosity accounts for the
dissipation of energy due to compression or expansion through transferring kinetic
energy into internal degrees of freedom (such as chemical reactions, excitation of
atomic/molecular levels, etc.).
Direct measurement of the bulk viscosity is difficult to carry out, and to this end
there are as yet no reliable results. It is probable that the bulk viscosity vanishes
for low density gases, i.e. under conditions where only binary collisions occur. In
dense gases, the numerical value of the bulk viscosity seems to be very small.
In all following discussions, we refrain from considering the bulk viscosity.
Navier Stokes equation
We have seen earlier that the differential momentum equation, as derived earlier, is
expressed as follows:
d ui ∂ τ ij
ρ =F i+ … … …( )
dt ∂ xj
The constitutive relation is
τ ji =−P δ ij + μ
( ∂ ui ∂ u j
+
∂ x j ∂ xi)… …..( )

Substituting the constitutive relation into the linear momentum equation, we obtain

ρ
d ui
dt
=F i−
∂P ∂
+
∂ xi ∂ x j
μ +
[(
∂ ui ∂ u j
∂ x j ∂ xi )]
The simplest assumption is frictionless flow. Accordingly, we set μ=0. So the
linear momentum equation reduces to
d ui ∂P
ρ =F i−
dt ∂ xi
This is Euler’s equation for inviscid flow.
Now we consider the expanded version for a two-dimensional Cartesian
coordinate system.
15
ρ
du
dt
=F x −
∂P ∂
+
∂x ∂x
μ 2
∂u
∂x [ ( )] [ (
+

∂y
μ
∂u ∂ v
+
∂ y ∂x )]
ρ
dv
dt
=F y −
∂P ∂
+
∂y ∂y
μ 2
∂v
∂y [ ( )] [ (
+

∂x
μ
∂u ∂v
+
∂y ∂x )]
For incompressible flow, use of the continuity equation leads to the following
simplified version of the Navier Stokes equation.

( )
2 2
du ∂P ∂u ∂u
ρ =F x − +μ 2
+ 2
dt ∂x ∂x ∂y

ρ
dv
dt
=F y −
∂P
∂y
+μ 2 (
∂2 v ∂ 2 v
+ 2
∂x ∂ y )
For steady flow, the above equations reduce to the following:

(
ρ u.
∂u
∂x
+v
∂u
∂y )
=F x −
∂P
∂x

∂2 u ∂2 u
+
∂ x2 ∂ y2 ( )
( ) ( )
2 2
∂v ∂v ∂P ∂ v ∂ v
ρ u. +v =F y − +μ +
∂x ∂y ∂y ∂ x2 ∂ y2
For non-isothermal cases, we have to consider the energy equation as well which
is given below without proof:

(
ρ Cp u .
∂T
∂x
+v
∂T
∂y
=K) (
∂2 T ∂ 2 T
+
∂ x2 ∂ y2
+μ ∅
)
Where the viscous energy dissipation function is expressed as follows:

[( ) ( ) ] ( )
2 2 2
∂u ∂v ∂ v ∂u
∅=2 + + +
∂x ∂y ∂x ∂y

Appendix
A scalar is a physical quantity that it represented by a dimensional number at a
particular point in space and time. Examples are hydrostatic pressure and
temperature.
A vector keeps track of two pieces of information (typically magnitude and
direction) for a physical quantity. Examples are position, force and velocity. The
vector has three components.
What happens when we need to keep track of three pieces of information for a
given physical quantity? We need a tensor. Examples are stress and strain. The
tensor has nine components. For stress, we keep track of a magnitude, direction and
which plane the component acts on. See the figure below

16
References:
1. ‘Fluid Mechanics’, Fourth Edition, Pijush K. Kundu and Ira M. Cohen,
Academic Press, 2008.
2. ‘An introduction to fluid mechanics’, G. K. Batchelor, Cambridge
University Press, 1967.

17

You might also like