COURSE NUMBER: ME 323

Fluid Mechanics II
3 credit hour

Internal flows

Course teacher
Dr. M. Mahbubur Razzaque
Professor
Department of Mechanical Engineering
BUET 1
 Flow in a circular pipe
Viscous effects in a flow result in the introduction of Reynolds
number,

The Reynolds number was observed to be a ratio of the inertial force

the viscous force.

Hence, when this ratio becomes large, it is expected that the inertial
forces may dominate the viscous forces.

For sufficiently low Reynolds number (Re < 2000 in a pipe) a laminar
flow results, and at sufficiently high Re number a turbulent flow
occurs. We consider laminar flow first and then turbulent flow.

2
 Entrance Flow And Developed Flow

In the entrance region the velocity profile changes in the flow

direction as shown in Fig.
direction, Fig 7.1.
71

A developed flow results when the velocity profile ceases to change in

the flow direction.
direction The idealized flow from a reservoir begins at the
inlet as a uniform flow (there is a thin boundary layer on the wall, as
shown); a wall viscous layer then grows over the inviscid core length,
L until the viscous stresses dominate the entire cross section; the
L,
profile then continues to change in the profile development region
due to viscous effects until a developed now is achieved. 3
The inviscid
Th i i id core length
l h is
i approximately
i l one-fourth
f h off the
h entrance
length, LE, depending on the conduit geometry, shape of the inlet, and
the Reynolds number.

For a laminar flow in a circular pipe with a uniform profile at the inlet,
the entrance length is given by

where the Reynolds number is based on the average velocity and the
diameter. For engineering applications a value of 2000 is the highest
Reynolds number for which laminar flow is assured. 4
 Laminar Flow In A Pipe

Consider incompressible,
incompressible steady,
steady developed laminar flow in a pipe,
pipe as
sketched in Fig. 7.4. There are two methods of analysis: an elemental
approach and a direct solution of the x-component Navier-Stokes
equation We shall use the elemental approach.
equation. approach

An elemental volume of the fluid is shown in Fig. 7.4. This volume

can be considered an infinitesimal control volume into which and from
which fluid flows and to which the momentum equation can be
applied.
5
Since the velocity profile does not change in the x-direction, the
momentum flux in equals the momentum flux out and the resultant
force is zero. Consequently,

which can be simplified to

where we have used sinθ = - dh/dx, the vertical direction being

denoted by h. The shear stress in this flow is related to the velocity
gradient and the viscosity, giving

Which can be integrated to give the velocity distribution,

6
where A is a constant of integration. Using u = 0 at r = ro, we can
evaluate A and find the velocity distribution to be

a parabolic profile. It is often referred to as Poiseuille flow, which

means steady,
d laminar
l i andd developed
d l d flow fl ini a circular
i l pipe.
i

Now the average velocity V is found to be

(7.3.13)
(7 3 13)
Or expressing the pressure drop Δp in terms of the average velocity we
have, for a horizontal pipe 7
where we have used Δp/L = - dp/dx since dp/dx is constant for a
developed flow. Note that the pressure drop is a positive quantity,
whereas the ppressure ggradient is negative.
g

The maximum velocity is obtained from Eq. 7.3.5 by putting r = 0

Comparing Eq. (7.3.13) and Eq. (7.3.15), we get,

The shearing stress is determined to be

Letting τ = τ0 at r = r0, we see that the pressure drop Δp over a length L

of a horizontal section of pipe is 8
 (7.3.18)
where we have again used dp/dx = - Δp/L.
Δp/L If we introduce the friction
factor f, a dimensionless wall shear, defined by
(7.3.19)

we see that (7.3.20)

where hL is the head loss with dimension of length.
g This equation
q is
often referred to as the Darcy-Weisbach equation and is valid for both
laminar and turbulent flows. Combining Eqns. 7.3.14, 7.3.18, and
7.3.19,
7.3. 9, we findd that

for laminar flow in a pipe. Substituting this back into Eq. 7.3.20, we see
that

i.e. the head loss is directly proportional to the velocity to the first
power which is applied to developed,
power, developed laminar flows in conduits,
conduits
including conduits of shape other than circular.
9
10
At the exit of the tube the pressure is zero; hence

The average velocity is found to be

0 036 m compared with p/γ = 2 m) , so the

This is quite small (V2/2g = 0.036
assumption of negligible velocity head is valid. Our pressure
calculation is acceptable. Using Eq.7.3.13, we can find the viscosity to
be

11
Consider the incompressible, steady, developed flow of a fluid between parallel 
plates, with the upper plate moving with velocity U, as shown in Fig. 7.5.

Let us take an elemental volume of unit depth in the z‐direction, as sketched in 
Fig. 7.5. If we sum forces in the x‐direction,
Fig. 7.5. If we sum forces in the x direction, we can write
we can write

12
13
14
15
16
hL= 12μLV/ρga2 or
17
18
Fully developed, steady flow between concentric, rotating cylinders, as shown in
Fully developed steady flow between concentric rotating cylinders as shown in
Fig. 7.6 , has particular application in the field of lubrication, where the fluid may 
be oil and the inner cylinder a rotating shaft. In such a case, 

The laminar flow solution we will find will be valid up to a Reynolds number of 
Th l i fl l i ill fi d ill b lid R ld b f
1700, providing the angular velocity of the outer cylinder ϖ2 = 0, as is often the 
case. Above Re  = 1700 a secondary laminar flow (a flow with a different velocity 
distribution) may develop, and eventually a turbulent flow develops. In fact, 
) y p, y p ,
numerous laminar flows (all different) have been observed for Re = 1700.

19
Body forces will be neglected in this derivation, or the cylinders will be ssumed
to be vertical. Since pressure does not vary with u, an element in the form of a
thin cylindrical shell will be used, as shown below. The resultant torque acting
on this element is zero because it has no angular acceleration; this is expressed
as

20
21
22
 Turbulent Flow In A Pipe

Most flows encountered in practical applications are turbulent flows in

pipes. Even though for carefully controlled laboratory conditions,
laminar flows have been observed up to Reynolds numbers of 40000 in
a pipe flow, turbulent flow is assumed to occur in a pipe under
standard operating conditions whenever the Reynolds number
> 2000

Water at room temperature flowing in a small 5-mm-diameter pipe; the

average velocity
l i needsd only
l beb 0.8
0 8 m/s
/ for
f turbulent
b l fl
flow to occur.
This is the situation whenever we drink water from a drinking
fountain.

For larger-diameter pipes the average velocity is sufficiently large so

that a turbulent flow is produced in most situations.

23
In a turbulent flow all three velocity components are nonzero. If we
measure theh components as a function
f i off time,
i graphs
h similar
i il to those
h
shown below would result for a flow in a pipe where u, v, w are in the
x-, r-, and θ-directions, respectively.

There is seldom any interest (to the engineer) in the details of the
randomlyy fluctuatingg velocityy components;
p ; hence we introduce the
notion of a time-average quantity.
24
The velocity components (u, v, w) are written as

where a bar over a quantity denotes a time average and a prime

denotes the fluctuating part. Using the component u as an example, the
time average is defined as,

where T is a time increment large enough to eliminate all time

dependence
p from . In a developed
p ppipe
p flow would be nonzero
and as is observed above.
25
 Turbulent shear stress

In turbulent flow, fluid particles move randomly throughout the flow.

Consider, at an instant in time a fluid particle moves through the
incremental area dA due to the velocity fluctuation v'; it enters a
neighboring layer of fluid which is moving with a higher x-component
velocity, thereby providing a retarding effect on the neighboring layer.

A fluid particle that moves to a neighboring layer that is traveling with a

lower x-component velocity would tend to accelerate the slower moving
fluid. The x-component
p force that results due to the random motion of a
fluid particle passing through the incremental area dA would be

where u' is the negative change in x-component velocity due to the

momentum exchange and ρvdA is the mass flux through the area; the
negative sign provides a positive dF.
26
If we divide both sides by the area dA, we obtain a “stress” that we call
the
h turbulent
b l shear
h stress. It is
i

where we know that (u'v') is, on the average, a negative quantity since a
positive v produces a negative u. This “shear stress” is actually a
momentum exchange but since it has the same effect as a stress, we call it
a shear stress.

The time-average turbulent shear stress, often called the apparent shear
stress, which is of primary interest, is

However, would be zero since a w' -component (in the θ -direction)

would not move a fluid particle into a layer of higher or lower x-
p
component velocity.
y Also,, usingg the same logic.
g

27
The total shear stress at a particular location would be due to both the
viscosity
i i andd the
h momentum exchangeh d
described
ib d above;
b that
h is,
i

The shear stress can be related to the pressure gradient by summing

forces on the horizontal cylindrical element shown in Fig. 7.8. This
results

The turbulent shear ggoes to zero at the wall since the velocityy
perturbations are zero at the wall, and the total shear is zero at the
centerline where r = 0 or y = r0, as shown in Fig. 7.9. 28
The viscous shear is nonzero only in a very thin viscous wall layer, of
thickness
hi k δv, near the
h wall,
ll as shown
h i part (b).
in (b) Theh turbulent
b l shear
h
reaches a maximum near the wall in the viscous wall layer.

The differential equation

q that must be solved if the time-average
g velocityy
distribution is to be determined is found by combining the two preceding
equations; it is

Where, r + y = r0 so that dy = - dr.

29
For developed flow = const.; hence if we know how varies with
r, the differential equation could be solved. The quantity cannot be
d
determined
i d analytically,
l i ll however,
h so the
h solution
l i to Eq. 7.6.7
6 cannot be
b
attempted until an empirical expression is found for .

Instead of using the quantity as the unknown in Eq. 7.6.7, we often

introduce the eddy viscosity η, defined by the relationship

Note that it has the same dimensions as the kinematic viscosity. In terms
of the eddy viscosity the differential equation becomes

If we view the turbulent pprocess as the random and chaotic mixingg of

particles of fluid, we may introduce the mixing length lm, the distance a
particle travels before interacting with another particle. Based on
reasoningg related to momentum interchange
g we relate the eddyy viscosityy
to the mixing length with
30
Hydraulically smooth wall

The time-average velocity profile in a pipe is quite sensitive to the

magnitude of the average wall roughness height e, as sketched nelow.

Most materials are “rough”g when viewed with sufficient magnification,

g ,
with glass and plastic assumed to be smooth with e = 0.

The laminar shear is significant only near the wall in the viscous wall
layer with thickness δv. If the thickness δv is sufficiently large, it
submerges the wall roughness elements, so that they have negligible
effect on the flow; it is as if the wall were smooth.
smooth Such a condition is
often referred to as being hydraulically smooth.
31
If the viscous wall layer is relatively thin, the roughness elements
protrude out of this layer and the wall is rough.

The relative roughness e/D and the Reynolds number can be used to
determine if a pipe is smooth or rough.

Velocity distribution
There are two common expressions for the empirical velocity distribution
in a developed turbulent flow. One method considers flows with smooth
walls and flows with roughg walls separately.
p y If the flow has a smooth
wall we identify two regions of the flow, the wall region and the outer
region. In the wall region the characteristic velocity and length are the
shear velocityy defined by
y and the viscous length
g

32
The dimensionless velocity distribution in the wall region, for a smooth
pipe, is

In the interval the

b ff zone, the
buffer th experimental
i t l data
d t
do not fit either of the curves. The
viscous wall layer has thickness
δν andd it
i is
i in
i the
h viscous
i l
layer
that turbulence is thought to be
initiated. This layer possesses a
time-average, linear velocity
distribution, but instantaneously
the layer is very time dependent.
The outer edge of the wall region is quite dependent on the Reynolds
number, as shown; for low Re it may be located near 33
For rough
g p p the viscous wall layer
pipes y does not pplay
y an important
p role
since turbulence initiates from the protruding wall elements. The
characteristic length is the average roughness height e; the dimensionless
velocity profile for the rough pipe is

In the outer region,

region the velocity defect is normalized with ;
the characteristic length is r0, and the empirical relation, for both smooth
and rough pipes, is

34
An additional empirical equation is needed to complete the profile for
0.15 <y/r
y/r0 ≤ 1. In the overlap region of the wall region and the outer
region we combine the equations above to obtain an expression for the
maximum velocity; for a smooth pipe it is

and for a rough pipe we find that

Before umax can be found we must know uτ; before uτ can be found we
must know τ0. To find τ0 we use the pressure gradient, or the friction
factor.
factor

Power-law profile
A alternative,
An lt ti andd simpler
i l formf th t adequately
that d t l describes
d ib the
th turbulent
t b l t
flow velocity distribution in a pipe is the power-law profile.
35
where y is measured from the pipe wall and n is an integer between 5 and
10 Using this distribution the average velocity is found to be
10.

This distribution is compared with a laminar profile in the above Figure.

Figure
The value of n in the exponent is related to the friction factor f by the
empirical expression

36
The constant n varies from 5 to 10 depending on the Reynolds number
and
d the
th pipe
i wallll roughness
h e/D.
/D

The power-law profile cannot be used to obtain the slope at the wall since
i will
it ill always
l yield
i ld (du/dy)
(d /d )wall = ∞ for
f all ll n. Thus
Th iti cannot be
b usedd to
predict the wall shear stress. In addition, it gives a positive slope du/dy at
the centerline of the pipe, where the slope should be zero, so it is not
valid near the centerline.

Head loss in a pipe

The most calculated quantity in pipe flow is the head loss. If the head loss
is known in a developed
p flow,, the ppressure changeg can be calculated;; for
a pipe the energy equation provides us with

37
The head loss that results from the wall shear in a developed flow is
related to the friction factor (see Eq.
Eq 7.3.20)
7 3 20) by the Darcy-Weisbach
Darcy Weisbach
equation, namely,

Consequently, if the friction factor were known, we could find the

head loss and then the pressure drop.

The friction factor/depends on the various quantities that affect the

flow, written as

where the average wall roughness height e accounts for the influence
of the wall roughness elements. A dimensional analysis provides us
with

where e/D is the relative roughness. 38

39
Experimental data that relate the friction factor to the Renolds
number
b have
h b
been obtained
bt i d for
f fully
f ll developed
d l d pipe i flow
fl over a
range of wall roughnesses. The results of these data are presented
in Fig. 7.13, which is commonly referred to as the Moody
diagram.

There are several features of the Moody diagram that should be

noted.

- For a given
gi en wall
all roughness,
ro ghness measured
meas red by
b the relative
relati e roughness
ro ghness
e/D, there is a sufficiently large value of Re above which the
friction factor is constant, thereby defining the completely
turbulent regime, The average roughness element size is
substantially greater than the viscous wall layer thickness, so that
viscous effects are not significant; the resistance to the flow is
produced by the drag of roughness elements that protrude into
the flow. 40
For the smaller relative roughness e/D values it is observed that,
as Re
R decreases,
d the
h friction
f i i factor
f i
increases i the
in h transition
ii
zone and eventually becomes the same as that of a smooth pipe.
The roughness
g elements become submerged g in the viscous wall
layer so that they produce little effect on the main flow.

For Reynolds numbers less than 2000,

2000 the friction factor of
laminar flow is shown. The critical zone couples the turbulent
flow to the laminar flow and may represent an oscillating flow
that
h alternately
l l exists
i between
b turbulent
b l and d laminar
l i fl
flow.

g
The e values in this diagram are for new ppipes.
p With age
g a ppipe
p
will corrode and become fouled, changing both the roughness
and the pipe diameter, with a resulting increase in the friction
factor Such factors should be included in design considerations;
factor.
they will not be reviewed here.
41
Three categories of problems can be identified for developed
turbulent flow in a pipe length L:

A category 1 problem is straightforward and requires no iteration

procedure when using the Moody diagram.
diagram

Category 2 and 3 problems are more like problems encountered

in engineering design situations and require an iterative trial-and-
error process when using the Moody diagram. 42
EXAMPLE (Category 1)
Water at 20oC is transported
p for 500 m in a 4-cm-diameter
wrought iron horizontal pipe with a flow rate of 0.003 m3/s.
Calculate the pressure drop over the 500-m length of pipe.

43
This answer is given to two significant numbers since the friction
factor is known to at most two significant numbers. The pressure
drop is found to be

44
EXAMPLE (Category 2)

45
46
47
EXAMPLE (Category 3)

48
49