Chapter 7

Flow in Closed Conduits

 Review Example

h
L

D
Given Data: h, L, D

Find : velocity at exit assuming no energy loss

 Review Example Cont’d.

V 22
Consider the head loss as: h t1 − 2 =C×
2g
P1 V1 2 P2 V 22
+ Z1 + = + Z2 + + h t1 − 2 ⇒
γ 2g γ 2g
V 22 V 22 2 gh
0+h+0= 0+0+ +C ⇒ V2 =
2g 2g 1+ C
1

2
 Review Example Cont’d.
Parameters in Fluid Mechanics

• Fluid Parameters/Conditions: density, viscosity and so on

• Flow Parameters/Conditions: velocity, pressure and so on

• Boundary Parameters/Conditions: pipe length, diameter,

roughness and so on

The coefficient C depends on all three groups of parameters

2 gh
V2 =
1+ C
 Introduction
Pipe flow characteristics

Reynolds demonstrated that the type of flow that occurred

depended on the interrelationship between four flow
parameters:
1. Average flow velocity (V)
2. Fluid density (ρ)
V ⋅ D ⋅ ρ V ⋅ D inertia force
3. Pipe diameter (D) R = = =
µ ν
e
viscous force
4. Fluid viscosity (µ )
In fact he showed that the following non-dimensional
number was very crucial; it was thereafter known as the
Reynolds Number
 Introduction
Flow Equations
For steady-incompressible flow:

• Continuity Equation Q = Vin Ain = Vout Aout

• Momentum Equation ∑F S = ρQ (VSout − VSin )

• Bernoulli/Energy Equation
2 2
Pin Vin Pout Vout
zin + + + H pump = zout + + + H turbine + htin −out
γ 2g γ 2g
In compact form: Hin + H pump = Hout + Hturbine + htin−out s s
L o
r gy
n e
E
 Energy Loss ( ht )

The total energy loss term, ht , basically comes from

two sources:
• Distributed (Major) Losses (h f ): Losses due to the
friction or roughness of the pipes. This is usually
the largest energy loss in a pipeline system.
• Local (Minor) Losses ( hm ): Losses at local points
on the pipe system such as bends, valves,
connections, elbows, etc.
Hence: ht = hf + ∑hm
 Head Loss due to Pipe Friction

• Consider in the following only fully developed flow

• From previous chapter:
hf
1 L τo ⋅ P
h f = ⋅ ⋅τ o = ⋅L Q=V⋅A
γ Rh A⋅γ
P = Wetted Perimeter D
L
A = X-sectional Area
γ = Specific Weight A D
τo = Wall Shear Stress Rh = =
P 4
Rh= Hydraulic Radius
 Head Loss due to Pipe Friction Cont’d.
Wall Shear Stress
Dimensional analysis

 τo VDρ e  τo e
τ o = τ o (V , D, e, ρ , µ ) ⇒ φ 
 , 
, =0⇒ = f ( Re , )
 ρ ⋅V 8 µ D  ρ ⋅V 8
2 2
D
Also we know:
1 L A D e
h f = ⋅ ⋅τ o Rh = = f = f ( Re , )
γ Rh P 4 D
L V2
hf = f ⋅ ⋅ Darcy-Weisbach Equation
D 2g (1845)
2
Note: Head loss is directly proportional to V
 Rough and Smooth Pipe
In fact when flow occurs in a pipe, even if it is turbulent flow, there
is a small region close to the pipe wall where turbulent fluctuations
are damped out and laminar flow prevails. This is called the
viscous sub-layer. If the pipe roughness elements are contained
within this layer, their effect is not felt by the bulk flow field and
hence the pipe flow behaves as if it is a smooth pipe. If the
roughness elements protrude through the sub-layer into the flow
field then they affect the gross flow as a roughness; rough pipe.

Smooth Pipe ( e < δ ) Rough Pipe ( e > δ )

 Energy Dissipation in Laminar Flow

• Molecular Viscosity ( µ ): A temperature-dependent fluid

property, which is always known
• Newton’s Viscosity Law:
dV
τl = µ
dy
• V is the instantaneous velocity, which in general is a
function of time and space
 Energy Dissipation in Turbulent Flow

• Cascade of Energy: The transfer of energy from large to

small scale eddies
• Similar to Newton’s viscosity law in laminar flow:
dV dV
τ t = µt , τ m = µm ⇒ τtotal = τ t + τ m
dy dy

• Eddy Viscosity: A flow parameter, refers to the internal

friction generated as flow becomes irregular and
turbulent as it passes over irregularities on the surface
• Eddy viscosity is much greater than molecular viscosity
 Friction Factor for Laminar Flow
Laminar flow is amenable to mathematical analysis the
Newton’s law of viscosity:
dV
τ =µ
dy

By employing this relationship it can be shown that:

64
f = (A)
Re

Notes: 1) This is a purely theoretical equation

2) Independent of pipe roughness
 Friction Factor for Laminar Flow Cont’d.

Combine (A) with DW equation:

L V2 64 V ⋅ D⋅ ρ V ⋅D
hf = f ⋅ ⋅ , f = , Re = =
D 2g Re µ ν

32 L
hf = ⋅ν ⋅ 2 ⋅V Hazen-Poiseuille Equation
g D (1820)

Note: The head loss in laminar flow is proportional to V , but

for the case of turbulent flow it is proportional to V 2
 Friction Factor for Turbulent Flow

• Turbulent flow cannot be analyzed theoretically and

so we must take recourse to experimentally derived
correlation equations.

• Blasius Equation (1913): For smooth pipes only

0 . 316 5
f = valid for R e ≤ 10
R e0 .25
 Friction Factor for Turbulent Flow Cont’d.

Nikuradse (1932) took smooth pipes and artificially

roughed them by attaching small sand grains of size ( e)
onto the pipe wall. He performed a series of tests with
pipes roughened by the addition of different sized
particles. He found that if e was “very small” then the
following friction factor equation worked well. He called
these pipes smooth pipes.
1
f
( )
= 2.0 log Re f − 0.8 (B)

This is an implicit equation and must be solved by

iteration or trial and error.
 Friction Factor for Turbulent Flow Cont’d.

Nikuradse also found that if e was “very large” then the

following friction factor equation worked. He called
these rough pipes.

1 e
= −2.0 log  + 1.14 (C)
f D

Note: There is no dependence on Reynolds number

and this is an explicit equation for friction factor
 Friction Factor for Turbulent Flow Cont’d.

Colebrook and White (1938) conducted experiments

on commercially available steel pipes. They found that
the following equation described the friction factor and
covered all types of pipes (rough, smooth and
intermediate).

1 e 9.35 
= 1.14 − 2.0 log +  (D)
f D R f 
 e 

Note: The equations B and C are special case of this

equation
 Moody Diagram

Moody (1944) took the Colebrook-White equation and

plotted friction factor versus Reynolds number for
various values of relative roughness ( e D ) to produce
something which is known as the Moody diagram. We
can use Moody diagram as an alterative to solving this
equation. Moody also proposed an approximate
equation for the Colebrook-White relation, which is
explicit and predicts friction factor to within +/- 5%.
Figure 10.7 Resistance coefficient f versus Re for
sand-roughened pipe. [After Nikuradse (32)].
 Figure 10.8 Resistance coefficient f versus Re. Reprinted with minor
variations. [After Moody (31).Reprinted with permission from the A.S.M.E.].
Moody Diagram

D C

A
B

6 × 10 4
 Approximations to Colebrooke-White Eq.
• Barr Equation (1994): with accuracy +/- 1%.
0.25
f = for Re ≥ 1.0 × 105
e 5.1286 2
[log( + 0.89
)]
3.7 D Re

• Swamee-Jain Equation (1976): with accuracy +/- 1%

0 . 25 e
f = for 10 -6
≤ ≤ 10 − 2
e 5 . 74 2 D
[log( + 0 .9 )]
3 .7 D Re
and 5000 ≤ R e ≤ 10 8
 Approximations to Cloebrooke-White Eq.

• Haaland Equation (1983): with accuracy +/- 2%

1  6.9  e  
1.11

≈ −1.8 log  +   
f  Re  3.7 D  

Notes: 1) All approximations are explicit in friction factor

2) Easy to use with computers and calculators
3) Reasonably accurate compared with the results
of Moody diagram
 Absolute Roughness Coefficients (e)

pipe material pipe roughness e (mm)

glass, drawn brass, copper 0.0015
commercial steel or wrought iron 0.045
asphalted cast iron 0.12
galvanized iron 0.15
cast iron 0.26
concrete 0.18-0.6
rivet steel 0.9
PVC 0.12
Flow Chart for Friction Factor

Find Reynolds number

• fluid properties (ρ, µ) V ⋅ D⋅ ρ
Re =
• pipe geometry (D) µ
• flow speed (V)

Turbulent Flow
Re ≥ 4000
Laminar Flow Re
Re ≤ 2000
Find relative roughness
64
f =
Re Look up Moody diagram
for friction factor
 Wall Shear Stress

• We often need to know the shear stress on the pipe wall

caused by the moving fluid (to know if the fluid will scour
the pipe). This is given by:
V2
τo = f ⋅ ρ ⋅
8

• The following quantity called the shear velocity or

friction velocity is also used in hydraulic computations:
τo
u =
*

ρ
 Local (minor) Losses

• These are losses in pressure associated with the fluid

encountering:
™restrictions in the system (valves)
™changes in flow direction (elbows, bends, tees, etc.)
™changes in pipe size (reducers, expanders)
™losses associated with fluid entering or leaving a pipe
• Minor losses occur locally
 Local Losses Cont’d.
Computational Approach

• Most minor losses can not be obtained analytically, so

they must be measured
• A local/minor head loss coefficient, Km, is associated
with each component
• Total local losses:
2
V
hm = ∑ K m ⋅
2g
• Km ≈ constant, depends on fitting
 Local Losses Cont’d.
Head Loss due to Gradual Expansion (Diffuser)
1 2

h EE = K EE
(V 11 − V 22 )
22
θ

2g 1
2
22
V 22  A 22
22
 0.8
h EE = K EE  − 1  0.7
2 g  A11  0.6
0.5
KE 0.4
0.3
0.2
0.1
0
0 20 40 60 80
diffuser angle ( θ )
 Local Losses Cont’d.
Sudden Contraction
22
 1  V 22 A cc
h cc =  − 1 22 C cc =
C  2g A 22
 cc 

V1 V2

vena contracta
Note: Losses are reduced with a gradual contraction
 Local Losses Cont’d.
Sudden Contraction
22
 1  V 22
h cc =  − 1 22
V1 V2
C  2g
 cc  1
0.95
0.9
0.85
Cc 0.8
0.75
0.7
0.65
0.6
0 0.2 0.4 0.6 0.8 1
A2/A1
 Local Losses Cont’d.
Entrance Losses
22
Losses can be V
h ee = K ee
reduced by accelerating 2g
the flow gradually K ee ≈ 1 . 0
Re-entrant Edge
and eliminating the
vena contracta
K ee ≈ 0 . 5
Sharp Edge

K ee ≈ 0 . 04
Rounded Edge
 Local Losses Cont’d.
Head Loss in Bends
• Head loss is a function of High pressure
the ratio of the bend radius
to the pipe diameter( R D )
Possible
• Velocity distribution returns to separation
normal several pipe diameters R from wall
downstream
V 22 D
h bb = K bb
2g Low pressure
Kb varies from 0.6 - 0.9
 Minor Losses Cont’d.
Head Loss in Valves

• Function of valve type and valve position

• The complex flow path through valves can result in high
head loss (of course, one of the purposes of a valve is
to create head loss when it is not fully open)

22
V
h vv = K vv
2g
 Problems for Single Pipes

• Type I Problems: Re-design of an existing system

Given: diameter, length, pipe material, flow rate
Find : required head
• Type II Problems: Existing behaviour analysis
Given: diameter, length, pipe material, available head
Find : the flow discharge or velocity
• Type III Problems: Design problem
Given: length, head, pipe material and flow discharge
Find : the minimum pipe size to carry this discharge.
 Numerical Example
Illustration

Discharge line
+180 m

+153 m
pump
+150 m

Suction line
Datum
 Numerical Example Cont’d.

• Given Data: te r
a wa
se
Fluid Flowing : Brine with S pg = 1.1 , µ = 8.63×10−4 Pa.S
Pipe Material : Cast iron with e = 0.18 mm
Suction line : D=20 cm and L = 12 m
Discharge line: D=15 cm and L = 30 m
Minor loss coefficients:
Kentry = 0.5
Kelbow = 0.9
Kvalve = 0.2
Kexit = 1.0
Efficiencies: for pump = 85% and for motor = 80%
 Numerical Example Cont’d.

• Find:

(i) pump head and power for 140 lit/s flow discharge
(ii) flow for -6.0 m pressure at point 2 - pump inlet
(iii) the diameter of the discharge line for 140 lit/s
flow discharge and head pump 45.0 m
 4

Numerical Example Cont’d.

3

Solution 1
2

(i) Write energy equation between 1 and 4

2 2
P1 V1 P4 V4
z1 + + + H p = z4 + + + (ht )1− 4
γ 2g γ 2g
H p = (z 4 − z1 ) + (ht )1− 4

H p = ( z 4 − z1 ) + (ht )S . L. + (ht )D. L.

2
 LS . L.  VS . L.
(ht )S .L. =  K entry + 2 K elbow + f 
 D  2g
 LD. L.  VD. L.
2
(ht )D.L. =  K valve + K elbow + K exit + f 
 DD. L.  2 g
 Numerical Example Cont’d.

Suction line Discharge line

D = 20 cm D = 15 cm
A = 0.031 m2 A = 0.0177 m2
V = 4.456 m/s V = 7.922 m/s
V2 V2
2g
= 1.012 m 2g
= 3.199 m

Re = 1.136 × 106 Re = 1.515 × 106

e e
D = 0.0009 D = 0.0012
f = 0.0195 f = 0.0208
 Numerical Example Cont’d.

 12 
Thus: (ht )SL = 0.5 + 2 × 0.9 + 0.0195 1.012 = 3.51 m
 0.2 
 30 
(ht )DL = 0.2 + 0.9 + 1.0 + 0.0208  3.199 = 20.03 m
 0.15 

Finally: H p = (180 − 150) + 3.51 + 20.03 = 53.54 m

γ ⋅Q ⋅ H p
Power requirement: P =
η P ⋅η m

P=
(1.1×1000 × 9.81)× 0.140 × 53.54  1 
 
0.85 × 0.80  1000 
P = 118.95 kW ≈ 160 hp
 Numerical Example Cont’d.

(ii) Find: Flow discharge for -6.0 m pressure at point 2

Discharge line
+180 m

2
+153 m
pump
+150 m

Suction line
Datum
 Numerical Example Cont’d. 4

(ii) Write energy equation between 1 and 2

3
2 2 1
P1 V1 P V
z1 + + = z 2 + 2 + 2 + (ht )1− 2 2

γ 2g γ 2g
Q2  LSL  P2
1 + K + 2 K + f  = ( z − z ) −
γ
entry elbow 1 2
A2 2 g  DSL 
 P2 
 z − z − 
1 2
γ
Q = ASL 2g  
L
1 + K entry + 2 K elbow + f 
 D 
1
   2

0.2 2  (150 − 153) − (− 6)   0.058

Q= π  2 g  ⇒ Q =
4 1 + 0.5 + 2 × 0.9 + f 12   3.3 + 60 f
 0.2  
Iterative Method for Problems Type II
take an initial
guess f i

i = i +1 calculate Q

no
calculate V , R e

if using MD,
f i ≈ f i +1 calculate f i + 1

yes

stop
 Numerical Example Cont’d.

Now, using iterative method with f o = 0.02 as an initial

guess for friction factor ( e D = 0.18 / 200 = 0.0009 ):

Iteration No. Q (cms) V (mps) Re = V ⋅ D ⋅ ρ µ f (Moody Diagram)

i =1 0.1135 3.613 9 ×105 0.0195
i=2 0.1139 3.626 9 ×105 0.0195

Clearly, after two iteration the friction factor remains

unchanged. Hence, the flow discharge would be:

Q = 0.1139 m3 / sec ≈ 114 lit/sec

 Numerical Example Cont’d.

(iii) Find: The diameter of the discharge line for 140 lit/sec
flow discharge and head pump 45.0 m

Discharge line
+180 m

+153 m
pump
+150 m

Suction line
Datum
 Numerical Example Cont’d.

(iii) Write energy equation between 1 and 4 4

H p = ( z 4 − z1 ) + (ht )SL + (ht )DL 3

1

(ht )DL = H p − (z4 − z1 ) − (ht )SL 2

= 45 − 30 − 3.51 = 11.49 m
 L Q2
(ht )DL =  K valve + K elbow + K exit + f  2
⇒
 D   D2 
 π  2 g
 4 
 30  0.140 2 × 16
11.49 = 0.2 + 0.9 + 1.0 + f  4 2 ⇒
 D  D π × 2 × 9.81

7094.842 D 5 − 2.1 D − 30 f = 0
Iterative Method for Problems Type III
take an initial
guess f i

i = i +1 calculate D

no
calculate Q , V , R e , e D

if using MD,
f i ≈ f i +1 calculate f i + 1

yes

stop
 Numerical Example Cont’d.

Now, using iterative method with f o = 0.02 as an initial

guess for friction factor ( e = 0.18 mm ):
Iteration No. D (mm) e D V (mps) Re = V ⋅ D ⋅ ρ µ f
i =1 168 0.001 6.316 1×106 0.021
i=2 169 0.001 6.241 1× 106 0.021

Clearly, after two iteration the friction factor remains

unchanged. Hence, the designed diameter would be:

D = 169 mm⇒ use D = 170 mm

 Numerical Example Cont’d.
EGL and HGL

EG
L
HG
L +180 m

+153 m
pump
+150 m
EGL
HGL

Datum
 Other Resistance Equations

h
General Form: V = C⋅R ⋅S
x
h
y
f
where S f =
f

Sf = the slope of an open channel or

Sf = energy gradient = friction loss pre unit length of pipe
Rh = hydraulic radius
C = must be determined for practical application
V = velocity
Q = VA = flow discharge
 Other Resistance Equations

h
General Form: V = C⋅R ⋅Sx
h
y
f
where S f =
f

L
• Hazen-William Equation: V = 0.844 ⋅ C HW ⋅ Rh
0.63
⋅Sf
0.54

• Manning Equation (1889): 1 2/3 1/ 2

V = ⋅ Rh ⋅ S f
n
• Chezy Equation (1769): 12 12
V = Cchezy ⋅ Rh ⋅ S f
• Darcy-Weisbach Equation (1845): 12 12
V = 8 g f ⋅ Rh ⋅ S f
 Aging of Pipes

• The Moody diagram is for new and clean pipes.

• Pipes become rougher during their life-time because of:
corrosion, deposition of materials
carried by water and incrustation (crust
or hard-coating of anything such as lime
or sediment from water on the inner
surface of the pipe wall)
• Cloebrook and Withe (1937) found that:
• pipe roughness et = enew +α ⋅ t
 Non-circular Conduits

• Empirically, it is found that f − Re − e D correlations for

circular conduits matches the data for non-circular
conduits if the diameter is replaced by the so-called
“equivalent hydraulic diameter, De ” defined as:
A
De = 4 Rh = 4 ⋅
P

Rectangular Circular Annuli Triangular

 Non-circular Conduits Cont’d.

4ab 2ab
Rectangular: De = = b
2( a + b ) a + b
a
V ⋅ De ⋅ ρ V ⋅ De
Reynolds No.: Re = =
µ ν
e
Relative roughness:
De
L V2
Darcy-Weisbach: hf = f
De 2 g
 Non-circular Conduits Cont’d.
• Laminar-to-turbulent transition always R e ≈ 2000
cr
if D is evaluated as equivalent diameter.
V2
• We treat “minor” losses by hm = K m ⋅
2g
• How “good” is the equivalent-diameter concept?
- for turbulent flow: reasonably ok.
- for laminar flow: not applicable.

Remember: f =
64 …laminar flow in circular pipe
Re
K …K ≡ shape factor
In general: f =
Re
 Non-circular Conduits Cont’d.

In particular: for circular cross-section: K = 64

for wide channel : K = 96
for square cross-section : K = 24

Plotted in Moody diagram:

W
f id
e
ch
a
cir nn
cu el
la
r
sq
u ar
e

V ⋅ De
Re =
2000 ν