Prof. Dr.

Atl BULU
1



Chapter 2



Flows under Pressure in Pipes



If the fluid is flowing full in a pipe under pressure with no openings to the atmosphere, it
is called pressured flow. The typical example of pressured pipe flows is the water
distribution system of a city.


2.1. Equation of Motion

Lets take the steady flow (du/dt=0) in a pipe with diameter D. (Fig. 2.1). Taking a
cylindrical body of liquid with diameter r and with the length x in the pipe with the
same center, equation of motion can be applied on the flow direction.


Figure 2.1.



The forces acting on the cylindrical body on the flow direction are,

a) Pressure force acting to the bottom surface of the body that causes the motion of
the fluid upward is,

F
1
= Pressure force = ( )
2
r p p +
D
r
F
1
F2
x r
2

x
y
x
Flow

Prof. Dr. Atl BULU
2

b) Pressure force to the top surface of the cylindrical body is,

F
2
= Pressure force =
2
r p


c) The body weight component on the flow direction is,


sin
2
x r X =

d) The resultant frictional (shearing) force that acts on the side of the cylindrical
surface due to the viscosity of the fluid is,


Shearing force = x r 2


The equation of motion on the flow direction can be written as,

( ) = + x r x r r p r p p 2 sin
2 2 2
Mass Acceleration (2.1)

The velocity will not change on the flow direction since the pipe diameter is kept
constant and also the flow is a steady flow. The acceleration of the flow body will be
zero, Equ. (2.1) will take the form of,

0 2 sin
2 2
= x r x r pr


r
x
p

= sin
2
1
(2.2)


The frictional stress on the wall of the pipe
0
with r = D/2,

2
sin
2
1
0
D
x
p

= (2.3)


We get the variation of shearing stress perpendicular the flow direction from Equs.
(2.2) and (2.3) as,


Prof. Dr. Atl BULU
3
2
0
D
r
= (2.4)



Fig. 2.2


Since r = D/2 y,

=
2
1
0
D
y
(2.5)


The variation of shearing stress from the wall to the center of the pipe is linear as can
be seen from Equ. (2.5).


2.2. Laminar Flow (Hagen-Poiseuille Equation)

Shearing stress in a laminar flow is defined by Newtons Law of Viscosity as,


dy
du
= (2.6)

Where = (Dynamic) Viscosity and du/dy is velocity gradient in the normal direction
to the flow. Using Equs. (2.5) and (2.6) together,



r
y

0
y
D/2
Prof. Dr. Atl BULU
4
dy
D
y
du
dy
du
D
y

=
=

2
1
2
1
0
0

By taking integral to find the velocity with respect to y,

= dy
D
y
u
2
1
0

cons
D
y
y u +

=
2
0

(2.7)


Since at the wall of the pipe (y=0) there will no velocity (u=0), cons=0. If the specific
mass (density) of the fluid is , Friction Velocity is defined as,

0
=

u (2.8)


Kinematic viscosity is defined by,

2
0 0 2

= =
= =
u
u


The velocity equation for laminar flows is obtained from Equ. (2.7) as,

=

D
y
y
u
u
2 2

(2.9)

Using the geometric relation of the pipe diameter (D) with the distance from the pipe
wall (y) perpendicular to the flow,
Prof. Dr. Atl BULU
5

=
= =

2
2
2
1
2
2 2
r
D
D
y
D u
u
r
D
y y
D
r

=
2
2 2
4
r
D
D
u
u

(2.10)


Equ. (2.10) shows that velocity distribution in a laminar flow is to be a parabolic
curve.

The mean velocity of the flow is,


A
udA
A
Q
V
A

= =

Placing velocity equation (Equ. 2.10) gives us the mean velocity for laminar flows as,


8
2

=
Du
V (2.11)


Since

0 2
=

u

And according to the Equ. (2.3),


2
sin
2
1
0
D
x
p

=


Prof. Dr. Atl BULU
6
2
sin
2
1
2
D
x
p
u

Placing this to the mean velocity Equation (2.11),

sin
32
2
x
p D
V (2.11)


We find the mean velocity equation for laminar flows. This equation shows that
velocity increases as the pressure drop along the flow increases. The discharge of the
flow is,


V
D
AV Q
4
2

= =

sin
128
4
x
p D
Q (2.12)

If the pipe is horizontal,


x
p D
Q

128
4
(2.13)


This is known as Hagen-Poiseuille Equation.


2.3. Turbulent Flow

The flow in a pipe is Laminar in low velocities and Turbulent in high velocities.
Since the velocity on the wall of the pipe flow should be zero, there is a thin layer
with laminar flow on the wall of the pipe. This layer is called Viscous Sub Layer and
the rest part in that cross-section is known as Center Zone. (Fig. 2.3)

Prof. Dr. Atl BULU
7



Fig. 2.3.



2.3.1. Viscous Sub Layer

Since this layer is thin enough to take the shearing stress as,
0
and since the flow
is laminar,


dy u du
u
dy
du
2
2
0

=
= = =

By taking the integral,


cons y u u
dy u u
+ =
=

2
2

Since for y =0 u = 0, the integration constant will be equal to zero. Substituting
=/ gives,


y
u
u

= (2.14)
Viscos Sublayer
Center Zone
y

0
Prof. Dr. Atl BULU
8


The variation of velocity with y is linear in the viscous sub layer. The thickness of the
sub layer () has been obtained by laboratory experiments and this empirical equation
has been given,

=
u

6 . 11 (2.15)


Example 2.1. The friction velocity u
*
= 1 cm/sec has been found in a pipe flow with
diameter D = 10 cm and discharge Q = 2 lt/sec. If the kinematic viscosity of the liquid
is = 10
-2
cm
2
/sec, calculate the viscous sub layer thickness.



mm cm
u
2 . 1 12 . 0
1
10
6 . 11
6 . 11
2
= =
=
=

2.3.2 Smooth Pipes

The flow will be turbulent in the center zone and the shearing stress is,


( ) v u
dy
du
+ = (2.16)

The first term of Equ. (2.16) is the result of viscous effect and the second term is the
result of turbulence effect. In turbulent flow the numerical value of Reynolds Stress
) ( v u is generally several times greater than that of ( ) dy du . Therefore, the
viscosity term ( ) dy du may be neglected in case of turbulent flow.

Shearing stress caused by turbulence effect in Equ. (2.16) can be written in the similar
form as the viscous affect shearing stress as,


dy
du
v u
T
= = (2.17)
Prof. Dr. Atl BULU
9


Here
T
is known as turbulence viscosity and defined by,



dy
du
l
T
2
= (2.18)


Here l is the mixing length. It has found by laboratory experiments that l = 0.4y for

0
zone and this 0.4 coefficient is known as Von Karman Coefficient.
Substituting this value to the Equ. (2.18),


y
dy
u du
u
du
y
dy
dy
du
y u
dy
du
y u
dy
du
y
dy
du
dy
du
y
T
T

=
=
=

= =

= =
=
5 . 2
4 . 0
4 . 0
16 . 0
16 . 0
16 . 0
2
2 2 0
2
2
0
2

Taking the integral of the last equation,


cons Lny u u
y
dy
u u
+ =
=

5 . 2
5 . 2
(2.19)
The velocity on the surface of the viscous sub layer is calculated by using Equs.
(2.14) and (2.15),


Prof. Dr. Atl BULU
10

=
= =
=
u u
u
y
y
u
u
6 . 11
6 . 11
2

Substituting this to the Equ. (2.19) will give us the integration constant as,

+ =

=
=
u
Ln u u cons
u
Ln Ln u u cons
u
Ln u u cons
Lny u u cons

5 . 2 5 . 5
6 . 11 5 . 2 6 . 11
6 . 11 5 . 2 6 . 11
5 . 2



Substituting the constant to the Equ. (2.19),

+ =
u
Ln u u Lny u u

5 . 2 5 . 5 5 . 2


5 . 5 5 . 2
5 . 5 5 . 2
+ =
+ =

yu
Ln
u
u
u
yu
Ln u u
(2.20)


Equ. (2.20) is the velocity equation in turbulent flow in a cross section with respect to
y from the wall of the pipe and valid for the pipes with smooth wall.


The mean velocity at a cross-section is found by the integration of Equ. (2.20) for are
A,


Prof. Dr. Atl BULU
11

A
udA
A
Q
V
A

= =



75 . 1 5 . 2
75 . 1
2
5 . 2
+ =

+ =

Du
Ln
u
V
u
D u
Ln V
(2.21)


2.3.3. Definition of Smoothness and Roughness

The uniform roughness size on the wall of the pipe can be e as roughness depth. Most
of the commercial pipes have roughness. The above derived equations are for smooth
pipes. The definition of smoothness and roughness basically depends upon the size of
the roughness relative to the thickness of the viscous sub layer. If the roughnesess are
submerged in the viscous sub layer so the pipe is a smooth one, and resistance and
head loss are entirely unaffected by roughness up to this size.




Fig. 2.4



Since the viscous sub layer thickness () is given by,

=
u

6 . 11 pipe roughness size

e is compared with to define if the pipe will be examined as smooth or rough pipe.


Flow
e
Pipe Center
Pipe wall
Prof. Dr. Atl BULU
12
a)

= <
u
e

6 . 11


The roughness of the pipe e will be submerged in viscous sub layer. The flow in the
center zone of the pipe can be treated as smooth flow which is given Chap. 2.3.2.
b)

>
u
e

70

The height of the roughness e is higher than viscous sub layer. The flow in the center
zone will be affected by the roughness of the pipe. This flow is named as Wholly
Rough Flow.


c)

< <
u
e
u

70 6 . 11


This flow is named as Transition Flow.


2.3.4. Wholly Rough Pipes

Pipe friction in rough pipes will be governed primarily by the size and pattern of the
roughness. The velocity equation in a cross section will be the same as Equ. (2.19).


cons Lny u u + =

5 . 2 (2.19)


Since there will be no sub layer left because of the roughness of the pipe, the
integration constant needs to found out. It has been found by laboratory experiments
that,


30
e
y o u = =


The integration is calculated as,


Prof. Dr. Atl BULU
13
30
5 . 2 5 . 2
30
5 . 2
30
5 . 2 0
e
Ln u Lny u u
e
Ln u const
const
e
Ln u

=
=
+ =


The velocity distribution at a cross section for wholly rough pipes is,

e
y
Ln
u
u
e
y
Ln u u
30
5 . 2
30
5 . 2
=
=

(2.22)

The mean velocity at that cross section is,

73 . 4
2
5 . 2
73 . 4
2
5 . 2
+ =

+ =

e
D
Ln
u
V
e
D
Ln u V
(2.23)


2.4. Head (Energy) Loss in Pipe Flows

The Bernoulli equation for the fluid motion along the flow direction between points
(1) and (2) is,


L
h
g
V p
z
g
V p p
z + + + = +
+
+
2 2
2
2
2
2
1
1

(2.24)


If the pipe is constant along the flow, V
1
= V
2
,


( )
1 2
z z
p
h
L

(2.25)




Prof. Dr. Atl BULU
14



Figure 2.5




If we define energy line (hydraulic) slope J as energy loss for unit weight of fluid
for unit length,



x
h
J
L

= (2.26)



Where x is the length of the pipe between points (1) and (2), and using Equs. (2.25)
and (2.26) gives,

sin
1 2

=
x
p
J
x
z z
x
p
J
(2.27)


g
V
2
2

p p +

h
L

p

Z
1

Z
2
Horizontal Datum
Energy Line
H.G.L
Flow

1
2
x
Prof. Dr. Atl BULU
15
Using Equ. (2.3),


J
D
x
p D
D
x
p
4
sin
4
2
sin
2
1
0
0
0

=
(2.28)


Using the friction velocity Equ. (2.8),



J
gD
J
D
u
u
u
4 4
2
2
0
0

= =
=
=

gD
u
J
2
4

= (2.29)



Energy line slope equation has been derived for pipe flows with respect to friction
velocity u
*
. Mean velocity of the cross section is used in practical applications instead
of frictional velocity. The overall summary of equational relations was given in
Table. (2.1) between frictional velocity u
*
and the mean velocity V of the cross
section.












Prof. Dr. Atl BULU
16
Table 2.1. Mathematical Relations between u
*
and V



Laminar Flow
(Re<2000)




8
2

=
Du
V



Turbulent
Flow
(Re>2000)






Smooth Flow

<
u
e

6 . 11

+ =

75 . 1
2
5 . 2

Du
Ln u V


Wholly Rough Flow

>
u
e

70

+ =

73 . 4
2
5 . 2
e
D
Ln u V





After calculating the mean velocity V of the cross-section and finding the type of low,
frictional velocity u
*
is found out from the equations given in Table (2.1). The energy
line (hydraulic) slope J of the flow is calculated by Equ. (2.29). Darcy-Weisbach
equation is used in practical applications which is based on the mean velocity V to
calculate the hydraulic slope J.




g
V
D
f
J
2
2
= (2.30)


Where f is named as the friction coefficient or Darcy-Weisbach coefficient. Friction
coefficient f is calculated from table (2.2) depending upon the type of flow where

VD VD
= = Re .


Prof. Dr. Atl BULU
17

Table 2.2. Friction Coefficient Equations



Laminar
Flow
(Re<200
0)




Re
64
= f






Turbulent
Flow
(Re>2000)






Smooth Flow

<

u
e

6 . 11





75 . 1
32
Re 5 . 2
8
+

=
f
Ln
f


Wholly Rough
Flow

>

u
e

70





73 . 4
2
5 . 2
8
+

=
e
D
Ln
f


Transition Flow

< <

u
e
u

70 6 . 11

+ +

=
e D
Ln
e
D
Ln
f
Re
2 . 9
1 5 . 2 73 . 4
2
5 . 2
8






The physical explanation of the equations in Table (2.2) gives us the following results.

a) For laminar flows (Re<2000), friction factor f depends only to the Reynolds
number of the flow. ( ) Re f f =
b) For turbulent flows (Re>2000),

1. For smooth flows, friction factor f is a function of Reynolds number of the flow.
( ) Re f f =
Fric
turb
diag
show



Sum


2. For tran
roughn
3. For wh
(e f f =

ction coeffic
bulent flow,
gram has be
ws the func
mmary
a) Energy

For a pipe w
nsition flow
ness of the p
holly rough
) D e
cient f is c
, the calcul
een prepare
ctional relati
loss for uni
with length
ws, f depend
pipe (e/D), f
h flows, f
calculated fr
lation of f
ed to overco
ions betwee
it length of
L, the ener
18
ds on Reyno
(Re, e f f =
f is a fun
from the eq
will alway
ome this dif
en f and Re,
Figure 2

pipe is calc

D
f
J =
gy loss will
olds number
) D e
nction of t
quations giv
s be done
fficulty. It i
, e/D as curv
.6
culated by D
g
V
2
2

l be,
r (Re) of the
the relative
ven in Tabl
by trial an
is prepared
ves. (Figure
Darcy-Weisb
Prof. Dr. A
e flow and r
e roughnes
le (2.2). In
d error me
by Nikura
e 2.6)
bach equati
Atl BULU
relative
s (e/D)
case of
thod. A
adse and

ion,
Prof. Dr. Atl BULU
19

JL h
L
= (2.31)


b) The friction coefficient f will either be calculated from the equations given in
Table (2.2) or from the Nikuradse diagram. (Figure 2.6)


2.5. Head Loss for Non-Circular Pipes

Pipes are generally circular. But a general equation can be derived if the cross-section
of the pipe is not circular. Lets write equation of motion for a non-circular prismatic
pipe with an angle of to the horizontal datum in a steady flow. Fig. (2.7).



Figure 2.7


( ) on accelerati Mass x A x P pA A p p = + sin
0



Where P is the wetted perimeter and since the flow is steady, the acceleration of the
flow will be zero. The above equation is then,

=
=


sin
0 sin
0
0
x
p
P
A
x A x P pA
(2.32)



A
x A W =

0
p+p
p
Flow
P=Wetted
Perimeter


x
Prof. Dr. Atl BULU
20


Where,


P
A
R = = Hydraulic Radius (2.33)


Hydraulic radius is the ratio of wetted area to the wetted perimeter. Substituting this
to the Equ. (2.32),

sin
sin
0
0
x
p
R
x
p
R


Since by Equ. (2.27),

sin

=
x
p
J


Shearing stress on the wall of the non-circular pipe,

RJ =
0
(2.34)

For circular pipes,


R D
D
D
D
P
A
R
4
4
4
2
=
= = =

(2.35)


This result is substituted (D=4R) to the all equations derived for the circular pipes to
obtain the equations for non-circular pipes. Table (2.3) is prepared for the equations
as,




Prof. Dr. Atl BULU
21


Table 2.3.


Circular Pipes

Non-Circular pipes
J
D
4
0
=

g
V
D
f
J
2
2
=

( ) e D f f Re, =

VD
= Re


RJ =
0



g
V
R
f
J
2 4
2
=

( ) e R f f 4 Re, =

R V4
Re =


2.6. Hydraulic and Energy Grade Lines

The terms of energy equation have a dimension of length[ ] L ; thus we can attach a
useful relationship to them.


L
h z
g
V p
z
g
V p
+ + + = + +
2
2
2 2
1
2
1 1
2 2
(2.36)


If we were to tap a piezometer tube into the pipe, the liquid in the pipe would rise in
the tube to a height p/ (pressure head), hence that is the reason for the name
hydraulic grade line (HGL). The total head

+ + z
g
V p
2
2

in the system is greater

than

+ z
p

by an amount
g
V
2
2
(velocity head), thus the energy (grade) line (EGL)
is above the HGL with a distance
g
V
2
2
.

Some hints for drawing hydraulic grade lines and energy lines are as follows.


1.
2.
3.
4.
5.
6.
7.
By definiti
the velocity
HGL and E
Head loss f
downward
when a pum
rise in the
the pump.
If energy is
EGL and H
In a pipe or
the water in
used to loc
the outlet e
the upstream
For steady
(diameter, r
unit of leng
HGL will b
If a flow pa
size, the ve
EGL and H
because the
larger veloc
If the HGL
atmospheri
on, the EG
y head. Thu
EGL will co
for flow in
in the direc
mp supplies
EGL occur
s abruptly t
HGL will dro
r channel w
n the system
ate the HGL
end of a pipe
m end, whe
y flow in
roughness,
gth will be
be constant
assage chan
elocity there
HGL will ch
e head per
city.
L falls below
ic pressure (
22
Figure

L is positio
us if the vel
oincide with
a pipe or c
ction of flow
s energy (an
rs from the
taken out of
op abruptly
where the pre
m because
L at certain
e, where the
ere the press
a pipe tha
shape, and
constant; th
and parallel
nges diamete
e in will also
hange. Mor
unit length
w the pipe,
(Fig.2.8).
e 2.8
oned above
locity is zer
h the liquid s
channel alw
w. The only
nd pressure
upstream si
f the flow b
y as in Fig
essure is zer
0 = p at
n points in th
e liquid cha
sure is zero
at has unif
so on) alon
hus the slop
l along the l
er, such as i
o change; h
reover, the s
h will be la
p is neg
the HGL a
ro, as in lak
surface. (Fig
ways means
y exception
e) to the flo
ide to the d
by, for exam
.
ro, the HGL
these point
he physical
arges into th
in the reser
form physi
ng its length
pe ( L h
L

length of pi
in a nozzle
hence the di
slope on the
arger in the
gative, there
Prof. Dr. A
an amount e
ke or reserv
gure 2.8)
the EGL w
n to this rule
ow. Then an
downstream
mple, a turb
L is coincide
ts. This fact
l system, su
he atmosphe
rvoir. (Fig.2
ical charac
h, the head
) L of the E
ipe.
or a change
istance betw
e EGL will
e conduit w
eby indicati
Atl BULU

equal to
voir, the
will lope
e occurs
n abrupt
m side of
bine, the
ent with
t can be
uch as at
ere, or at
2.8)
cteristics
loss per
EGL and
e in pipe
ween the
l change
with the
ing sub-
Prof. Dr. Atl BULU
23



Figure 2.9


If the pressure head of water is less than the vapor pressure head of the
water ( -97 kPa or -950 cm water head at standard atmospheric pressure),
cavitation will occur. Generally, cavitation in conduits is undesirable. It
increases the head loss and cause structural damage to the pipe from
excessive vibration and pitting of pipe walls. If the pressure at a section in
the pipe decreases to the vapor pressure and stays that low, a large vapor
cavity can form leaving a gap of water vapor with columns of water on
either side of cavity. As the cavity grows in size, the columns of water
move away from each other. Often these of columns of water rejoin later,
and when they do, a very high dynamic pressure (water hammer) can be
generated, possibly rupturing the pipe. Furthermore, if the pipe is thin
walled, such as thin-walled steel pipe, sub-atmospheric pressure can cause
the pipe wall to collapse. Therefore, the design engineer should be
extremely cautious about negative pressure heads in the pipe.


















Prof. Dr. Atl BULU
24