CHAPTER TWO [BTME2223 APPLIED FLUID MECHANICS]

Chapter Two: Flows in Bounded Systems

Motivation
Piping system is an important aspect of system design in any energy consuming system. An
economical operation is always the alternative to minimize the work and cost saving in choosing the
right tools for a system. For example, the minimum horsepower pump to match the requirement of
the work. In order to maximize the work efficiency, there are few factors can be considered in a
piping system such as the elevation head, velocity head and friction head. These are the resistances
where it determines the amount of work required to transport liquid from one to another place.

Chapter Objectives:
 Calculate the major and minor losses associated with pipe flow in piping networks and
determine the pumping power requirements.

Chapter Outline:
2.1 Introduction
2.2 Head Loss in Series Piping System
2.3 Minor Losses in Series Piping System
2.4 Total Energy and Hydraulic Gradient Line
2.5 Incompressible Flow through Pipe Network

2.1 Introduction
A pipe is a closed conduit through which the fluid flows under pressure. Hence, it is known as
pressure flow. Mostly, the pipes are circular in shape. But in some special cases, non-circular pipes
may be used. Essentially, the pipe should run full, that is, there exists no free surface in the pipes. In
case the pipe runs partially full as in the sewerage lines, the air will enter into the pipe and there will
exist a free surface. Such fluid flow problems are treated with laws of open channel flow
(not cover in the syllabus). The fluid flow is always resisted by viscous forces, at the boundary as
well as between adjacent fluid layers. The viscous resistance is overcome by the drop in the total
energy line in the direction of flow. The fluid resistance depends on the type of flows. Separate laws
govern the frictional resistance to laminar and turbulent flows. These are discussed in this chapter.
Besides frictional resistance, the energy loss also occurs due to non-uniformity of flow caused due to
various pipe fittings are described in Section 2.3. A few problems of turbulent flow through pipe are
also dealt with herein under.

2.2 Head Loss in Series Piping System

2.2.1 Laminar Fully Developed Pipe Flow
As the water flows from a faucet at a very low velocity,
the flow appears to be smooth and steady. The stream has
a fairly uniform diameter and there is little or no evidence
of mixing of the various parts of the stream. This is called
laminar flow, a term derived from the word layer,
because the fluid appears to be flowing in continuous
Fig. 2.1: Illustration of laminar flow in a circular pipe
layers with little or no mixing from one layer to the
adjacent layers.

June2012[Rev2.10:May2017]-pkq\dmee\afm\c2 2-1
For NON-COMMERCIAL use only.
 CHAPTER TWO [BTME2223 APPLIED FLUID MECHANICS]

In other words, fluid friction laws for laminar flows are:

a) The fluid resistance is directly proportional to the velocity of flow.

b) It is proportional to the area of contact surface.

c) It is independent of the nature of the contact surface.

d) It is independent of the pressure.

e) It greatly varies with the temperature.

Laminar flow is defined as the fluid particles moves in orderly

and in straight lines parallel to the pipe walls. For a laminar
flow in pipe, the Reynolds number is always below 2000.
From previous studies, the relationship between shear stress
and velocity gradient (see Fig. 2.3):

𝑑𝑢
𝜏=𝜇 (2.1)
𝑑𝑦

Fig. 2.2: Laminar Flow in pipe

Fig. 2.3: Velocity distribution in laminar flow in a circular pipe

Let’s consider a straight and cylindrical pipe with length of 𝑙 as shown in Fig. below, the pressure
loss during the laminar flow can be obtained in the equilibrium of fluid.

Fig. 2.4: Element of fluid in a pipe

June2012[Rev2.10:May2017]-pkq\dmee\afm\c2 2-2
For NON-COMMERCIAL use only.
 CHAPTER TWO [BTME2223 APPLIED FLUID MECHANICS]

The pressure at upstream, 1 is denoted as 𝑃1 and as the flow further away to downstream, 2, the
pressure is drop along the pipe by ∆𝑃. Thus, the pressure at downstream is denoted as 𝑃1 − ∆𝑃. Since
the fluid is equilibrium, the driving force due to the pressure (surface force) can be written as below:
𝐹 = 𝑃1 𝐴 − (𝑃1 − ∆𝑃)𝐴 = ∆𝑃 ∙ 𝜋𝑟 2 (2.2)
However, as the fluid flows along the pipe, the shear force is created along the wall surface.
d𝑢
𝐹 = 𝜏𝐴 = −𝜇 ∙ 2𝜋𝑟𝑙 (2.3)
d𝑟
Therefore, it leads the equation to become the following form and then integrate with respect to radius.
d𝑢 ∆𝑃 𝑟
=− ∙
d𝑟 𝑙 2𝜇
∆𝑃 𝑟 2
𝑢=− ∙ +𝐶
𝑙 4𝜇
At the centre of the pipe (𝑟 = 0), 𝐶 = 𝑢 = 𝑢max . However, as the fluid flow is approaches near to
the wall (𝑟 = 𝑅), the velocity, 𝑢 will become zero. Thus,
∆𝑃 𝑅 2
𝐶= ∙
𝑙 4𝜇
∆𝑃 1
𝑢= ∙ (𝑅 2 − 𝑟 2 ) (2.4)
𝑙 4𝜇
If the flow velocity is constant, the flowrate will change as the area of the pipe alters.
𝛿𝑄 = 𝑢𝛿𝐴 = 𝑢 ∙ 2𝜋𝑟𝛿𝑟
Substitute Eq. (1.4) into it and take the integration at both sides,
∆𝑃 1 𝑅 2
𝑄= ∙ ∫ (𝑅 − 𝑟 2 ) ∙ 2𝜋𝑟𝛿𝑟
𝑙 4𝜇 0
∆𝑃 𝜋𝑅 4 𝜋𝑑 4 ∆𝑃
𝑄= ∙ = (2.5)
𝑙 8𝜇 128𝜇𝑙
Rearrange the above equation where the pressure loss along a pipe system is given by
128𝜇𝑙𝑄
∆𝑃 = (2.6)
𝜋𝑑 4
This equation is known as Hagen-Poiseuille Equation. To get the pressure loss, or also known as head
loss, the formula is given by
∆𝑃 = 𝜌𝑔ℎ𝑓
therefore, we can summarise Eq. (2.6) in term of mean velocity and head loss.
32𝜇𝑙𝑢
ℎ𝑓 = (2.7)
𝜌𝑔𝑑 2
where mean velocity
𝜋
𝑄 = 𝑢 ∙ 𝑑2 (2.8)
4

June2012[Rev2.10:May2017]-pkq\dmee\afm\c2 2-3
For NON-COMMERCIAL use only.
 CHAPTER TWO [BTME2223 APPLIED FLUID MECHANICS]

Example 2.1
Glycerine of viscosity 0.9 Nsm-2 and density 1260 kg/m-3 is pumped along a horizontal pipe 6.5 m
long of diameter d= 0.01 m at a flow rate of Q = 1.8 litres/min-1. Determine the flow Reynolds
number and verify whether the flow is laminar or turbulent. Calculate the pressure loss in the pipe
due to frictional effects and calculate the maximum flow rate for laminar flow conditions to prevail.

Mean velocity,
−3
𝑄 1.8 × 10 ⁄60
𝑢= = 𝜋 = 0.382 ms−1
𝐴 2
4 × 0.01
The Reynolds number,

𝜌𝑢𝑑 1260 × 0.382 × 0.01

𝑅𝑒 = = = 5.35
𝜇 0.9

Therefore, flow is laminar as 𝑅𝑒 < 2000. Frictional losses may be

calculated from the Hagen-Poiseuille equation [Eq. (2.6)],
−3
128𝜇𝑙𝑄 128 × 0.9 × 6.5 × 1.8 × 10 ⁄60
∆𝑃 = =
𝜋𝑑4 𝜋 × 0.014

∆𝑃 ≈ 715 × 103 Nm-2 #

Upper limit of laminar flow conditions is reached when

𝑅𝑒 𝑄
=
𝑅𝑒CR 𝑄CR

𝑅𝑒CR
𝑄CR = 𝑄
𝑅𝑒
2000
𝑄CR = 1.8 ×
5.35

𝑄CR ≈ 672.9 litres min−1

*verify the answer with Darcy-Weisbach equation.

June2012[Rev2.10:May2017]-pkq\dmee\afm\c2 2-4
For NON-COMMERCIAL use only.
 CHAPTER TWO [BTME2223 APPLIED FLUID MECHANICS]

2.2.2 Head Loss with the Friction Factor

When the faucet is nearly fully open, the water has a rather high velocity. The elements of fluid
appear to be mixing chaotically within the stream. This is a general description of turbulent flow.
Turbulent flow is defined as the flow motion is locally and completely random but the overall
direction of flow is one way. For a turbulent flow in pipe, the Reynolds number is always above 4000.
Fluid friction laws for turbulent flows:

a) The fluid resistance is directly proportional to the (velocity)n of flow.

b) It is proportional to the area of contact surface.

c) It is independent of the nature of the contact surface.

d) It is independent of the pressure.

e) It is proportional to the density of fluid.

f) It fluid resistance slightly varies with the temperature.

Fig. 2.5: Turbulent Flow in pipe

When applying pipe flow formulas to practical problems, it is
customary to use a control volume analysis. Consider
incompressible steady flow between sections 1 and 2 of the
inclined constant area pipe in Fig. 2.6. The one-dimensional
continuity relation reduces to
𝑄2 = 𝑄2 = constant

Since the pipe is of constant area. The steady flow energy

equation (refer to previous module) becomes
𝑃1 𝑈12 𝑃2 𝑈22
+ 𝛼1 + 𝑍1 = + 𝛼2 + 𝑍2 + ℎ𝑓 (2.9)
𝜌𝑔 2𝑔 𝜌𝑔 2𝑔

since there is no pump or turbine between 1 and 2. For fully

developed flow, the velocity profile shape is the same at
Fig. 2.6: Control volume of steady, fully developed
sections 1 and 2. flow between two sections in an inclined pipe

June2012[Rev2.10:May2017]-pkq\dmee\afm\c2 2-5
For NON-COMMERCIAL use only.
 CHAPTER TWO [BTME2223 APPLIED FLUID MECHANICS]

Thus, 𝛼1 = 𝛼2 (kinetic energy correction factor, commonly assumed to be 1 in problems that

involving turbulent pipe flow) and, since 𝑈1 = 𝑈2 , Eq. (2.9) reduces to head loss versus pressure drop
and elevation change:
𝑃1 𝑃2 ∆𝑃
ℎ𝑓 = (𝑍1 − 𝑍2 ) + ( − ) = ∆𝑍 + (2.10)
𝜌𝑔 𝜌𝑔 𝜌𝑔

The pipe head loss equals the change in sum of pressure and gravity head – that is, the change in
height of the hydraulic grade line (HGL). Finally, apply the momentum relation to the control volume
in Fig. (2.6), accounting for applied x-directed forces due to pressure, gravity, and shear:
∑ 𝐹𝑥 = ∆𝑃(𝜋𝑅 2 ) + 𝜌𝑔(𝜋𝑅 2 )𝐿 𝑠𝑖𝑛 𝜙 − 𝜏𝑤 (2𝜋𝑅)𝐿 = 𝑚̇(𝑈2 − 𝑈1 ) (2.11)

Rearrange this and we find that the head loss is also related to wall shear stress:
∆𝑃 4𝜏𝑤 𝐿
∆𝑍 + = ℎ𝑓 = (2.12)
𝜌𝑔 𝜌𝑔 𝑑

where we have substituted ∆Z = L sin ϕ from the geometry of Fig. 2.6. Note that, regardless of
whether the pipe is horizontal or tilted, the head loss is proportional to the wall shear stress. The
dimensionless parameter 𝑓 is called Darcy friction factor, after Henry Darcy (1803-1858), a French
engineer whose pipe flow experiments in 1857 first established the effect of roughness on the pipe
resistance. The quantity 𝑘 is the wall roughness height, which is important in turbulent (but not
laminar) pipe flow. Or the alternative form for friction factor is given by:
8𝜏𝑤
𝑓=
𝜌𝑈 2
4𝑓𝑙 𝑈 2
ℎ𝑓 = (2.13)
𝑑 2𝑔

Systematic results tested by Nikuradse on the turbulent flow in smooth and rough pipes demonstrated
a perfect relationship between 𝑓, Reynoldss number (𝑅𝑒) and 𝑘⁄𝐷 (relative roughness, ratio of
roughness height to pipe diameter). It is to be noted that for 𝑅𝑒 under 2000, a simple relationship
exists between 𝑓 and 𝑅𝑒 which is completely independent of roughness of the wall. The relationship
is described by Eq. (2.14). For laminar region (𝑅𝑒 < 2000) through the pipe, the relationship is
16
𝑓= (2.14)
𝑅𝑒
Beyond the critical range of 𝑅𝑒, a transitional region exits in which the value of 𝑓 are uncertain and
the flow might be either laminar or turbulent. In turbulent flow, there are various formulae suggested
by researchers, some of which are discussed below:

a) Blasius suggested the following equation for smooth (pipe) turbulent flow. This equation is
valid for 𝑅𝑒 values up to 105 .
0.079
𝑓= 1 (2.15a)
𝑅𝑒 4
b) Nikuradse simplified the Karman-Prandtl’s universal resistance equation in explicit form for
turbulent flow in smooth pipe. The equation is valid for 𝑅𝑒 values up to 3.4 × 106 .
0.221
𝑓 = 0.00332 + 0.237 (2.15b)
𝑅𝑒

June2012[Rev2.10:May2017]-pkq\dmee\afm\c2 2-6
For NON-COMMERCIAL use only.
 CHAPTER TWO [BTME2223 APPLIED FLUID MECHANICS]

Moody diagram (as attached in Appendix 2A) is an alternative method and used to help determine
the value of the friction factor 𝑓 for turbulent flow. The value of 𝑅𝑒 and the relative roughness must
be known. Therefore, the basic data required are the pipe inside diameter, the pipe material, from
which the wall roughness (𝑘) to be found, the flow velocity, and the kind of fluid and its temperature,
from which the viscosity can be found. Several important observations can be made from these curves:
a) For a given 𝑅𝑒 of flow, as the 𝑘/𝐷 is increased, the 𝑓 increases.

b) For a given 𝑘/𝐷, the 𝑓 decreases with increasing 𝑅𝑒 until the zone of complete turbulence is
reached.

c) Within the zone of complete turbulence, the 𝑅𝑒 has no effect on the 𝑓.

d) As the 𝑘/𝐷 decreases, the value of the 𝑅𝑒 at which the zone of complete turbulence begins
also increases.

June2012[Rev2.10:May2017]-pkq\dmee\afm\c2 2-7
For NON-COMMERCIAL use only.
 CHAPTER TWO [BTME2223 APPLIED FLUID MECHANICS]

Example 2.2
Calculate the loss of head due to friction and the power required to maintain flow in a horizontal
circular pipe of 40 mm diameter and 750 m long when water flows at a rate:
(a) 4.0 litres min-1;
(b) 30 litres min-1.
Assume that for the pipe the absolute roughness is 0.00008 m and coefficient of dynamic viscosity
1.14 × 10-3 Nsm-2.

Given that b) If the flowrate is 30 litres per min,

𝑑 = 0.04 m 𝑙 = 750 m 𝑘 = 0.00008 m 10−3 m3

𝑄 = 30 × = 5 × 10−4 m3 /s
60 s
𝜇 = 1.14 × 10−3 Nsm−2
The velocity in the circular pipe,
a) If the flowrate is 4 litres per min,
𝑄 5 × 10−4
−3
10 m 3 𝑢= = = 0.3978 m/s
𝑄 =4× = 6.6667 × 10−5 m3 /s 𝐴 𝜋 × 0.042
60 s 4

The velocity in the circular pipe, Check the Reynoldss number,

𝑄 6.6667 × 10−5 𝜌𝑢𝑑 1000 × 0.3978 × 0.04

𝑢= = 𝜋 = 0.053 m/s 𝑅𝑒 = =
𝐴 2
𝜇 1.14 × 10−3
4 × 0.04
𝑅𝑒 ≈ 13959
Check the Reynoldss number,
Given the wall roughness of the pipe is 0.00008 m. Thus,
𝜌𝑢𝑑 1000 × 0.053 × 0.04 we cannot treat this pipe as a smooth pipe. Then the
𝑅𝑒 = =
𝜇 1.14 × 10−3 relative roughness would be,
𝑅𝑒 = 1861.23 𝑘
𝜀= = 0.002
𝑑
The flow in circular pipe is laminar. By using Eq. (2.13)
and Eq. (2.14), From the Moody diagram, at line of relative roughness,
𝜀 = 0.002 at the right axis and Reynoldss number of
16
𝑓= ≈ 8.596 × 10−3 14000 at bottom axis, the friction factor,
𝑅𝑒
𝑓 is 0.008. Thus, apply and substitute 𝑓 into Eq. (2.13).
4𝑓𝑙 𝑢2 4 × 8.596 × 10−3 × 750 × 0.0532
ℎ𝑓 = ∙ = 4𝑓𝑙 𝑢2 4 × 0.008 × 750 × 0.39782
𝑑 2𝑔 2 × 9.81 × 0.04 ℎ𝑓 = ∙ =
𝑑 2𝑔 2 × 9.81 × 0.04
ℎ𝑓 = 0.0923 m#
ℎ𝑓 ≈ 4.84 m#
The power to maintain the flow,
The power to maintain the flow,
ℙ = 𝜌𝑔ℎ𝑓 𝑄 = 9810 × 0.0923 × 6.6667 × 10−5
ℙ = 𝜌𝑔ℎ𝑓 𝑄 = 9810 × 4.4 × 5 × 10−4
ℙ = 0.0604 W#
ℙ ≈ 24 W#

Please verify the answer using Hagen-Poiseuille

equation.

June2012[Rev2.10:May2017]-pkq\dmee\afm\c2 2-8
For NON-COMMERCIAL use only.
 CHAPTER TWO [BTME2223 APPLIED FLUID MECHANICS]

2.3 Minor/Geometrical Losses in Series Piping System

The dissipation of energy of flow in a pipe line consists of loss due to frictional resistance and loss
due to local disturbance. Conventionally, the former is termed as major loss (Section 2.2) and this
section discusses the minor losses.

In general the minor losses result from changes in velocity of flow either a change in its magnitude
or its direction or a change in both. Whenever, a weak disturbance occurs in a pipeline, eddy currents
are set up in the flow (because of separation of flow) which causes more turbulence in the flow as
compared to normally present in the flow. The eddying turbulence persists for considerable length
downstream and is eventually damped out by the fluid viscosity. The eddy motion derives the energy
from the main flow which is finally dissipated as viscous resistance and thus a portion of useful
energy is lost. The magnitude of this localised loss is proportional to the abruptness of the velocity
change. Though the disturbing effect is usually confined to a very short length of path, the effects
may not disappear for a considerable distance downstream. Moreover, if the transitions are carefully
streamlined, the flow separation may be avoided and thus the resulting loss of energy may be
minimized.

The minor losses are found to very approximately linearly with the velocity head, that is,
𝑈2
ℎ𝑓 = 𝐾 (2.16)
2𝑔
the term 𝐾 is known as energy loss coefficient. Except the loss of head caused due to sudden
enlargement, most of the minor losses are determined experimentally. The minor losses caused due
to the following transitions are discussed hereafter:

a) Loss due to sudden enlargement.

b) Loss due to sudden contraction.

c) Loss at entrance.

d) Loss at exit.

e) Loss at bend.

f) Loss due to pipe fittings (valves).

June2012[Rev2.10:May2017]-pkq\dmee\afm\c2 2-9
For NON-COMMERCIAL use only.
 CHAPTER TWO [BTME2223 APPLIED FLUID MECHANICS]

2.3.1 Loss due to Sudden Enlargement

Generally, the losses due to pipe or duct fittings are determined experimentally. However, the case
of a sudden expansion in a pipe or duct may be determined analytically. The head loss is given by
Borda-Carnot relationship.
2
(𝑈1 − 𝑈2 )2 𝑈22 𝐴2 𝑈12 𝐴1 2
ℎ𝑓 = = ( − 1) = (1 − ) (2.17)
2𝑔 2𝑔 𝐴1 2𝑔 𝐴2

Fig. 2.7: Sudden enlargement

To avoid the head losses incurred by the

sudden enlargement, diffusers are
commonly employed. Fig. below illustrate
a typical conical diffuser and the variation
in pressure loss across with diffuser
included angle. The total loss across the
diffuser is made up of 2 components. The
first is due to fluid friction along the length
of the diffuser. The second is the separation
loss. The angle of the diffuser is
sufficiently small to limit separation, or
flow eddy, losses, but not too small as to
increase the length of the diffuser to the
point where the frictional losses become
Fig. 2.8: Losses of pressure in a conical diffuser
predominant.

June2012[Rev2.10:May2017]-pkq\dmee\afm\c2 2-10
For NON-COMMERCIAL use only.
 CHAPTER TWO [BTME2223 APPLIED FLUID MECHANICS]

2.3.2 Loss due to Sudden Contraction

There is little or no loss between the upstream large section
𝐴𝐶
conduit and the vena contracta formed within the smaller
conduit just downstream of the junction. It has been shown
experimentally that the majority of the pressure loss occurs
as a result of the eddies formed as the flow area expands
from the vena contracta area up to the full cross-section of
the downstream pipe. The head loss through sudden
contraction is given by
2
𝑈22 1 Fig. 2.9: Vena Contracta
ℎ𝑓 = ( − 1) (2.18)
2𝑔 𝐶𝐶
where 𝐶𝐶 = 𝐴𝐶 ⁄𝐴1, the coefficient of contraction and 𝐴𝐶 is the area at vena contracta.

2.3.3 Losses due to Piping Fittings, Bends, Pipe Entry and Exit
Recall back the general equation for minor losses, the head loss can be determined in Eq. (2.16) and
the 𝐾 can be obtained from the standard value.

Table 2.1: Various coefficients of loss for fittings, bends, pipe entries and exit
Fitting Coefficient of Loss, K
Gate valve 0.25-25
Globe valve 10
90 deg. elbow 0.9
45 deg. elbow 0.4
Large-radius 90° bend 0.6
Tee junction 1.8
Sharp pipe entry 0.5
Rounded pipe entry 0.2
Sharp pipe exit 1.0

The losses can be represented in terms of the equivalent length 𝑙𝑒 , of the un-obstructed pipe:
2 2
4𝑓𝑙𝑒 𝑢2 𝑢2
ℎ𝑓 = =𝐾
2𝑑𝑔 2𝑔
4𝑓𝑙𝑒
𝐾= (2.19)
𝑑

2.3.4 Pipe Connected in Series

When pipes of different diameters are connected end to end to form a pipe line and fluid flows through
each of them in turn, the pipes are said to be connected in series. Thus, the head loss in the pipe line
is the summation of all local losses. The Bernoulli equation when written to include the effects of
friction and of minor losses becomes
𝑃1 𝑉12 𝑃2 𝑉22
+ + 𝑍1 = + + 𝑍2 + ∑ ℎ𝑓 (2.20)
𝜌𝑔 2𝑔 𝜌𝑔 2𝑔

June2012[Rev2.10:May2017]-pkq\dmee\afm\c2 2-11
For NON-COMMERCIAL use only.
 CHAPTER TWO [BTME2223 APPLIED FLUID MECHANICS]

We refer to Eq. (2.20) as the modified Bernoulli equation. As an

example, consider the system shown in Fig. 2.10. A pipe is 𝐿1
connected to two tanks. We will examine five different control 𝐿2 𝐿
3
volumes applied to this same pipe, and write the modified
Bernoulli equation for each case. In Fig. 2.10a, the control
volume includes all the fluid in the pipe and the fluid in both Fig. 2.10a
tanks. Section 1 is the free surface of the tank on the left and
Section 2 is the free surface of the tank on the right. We now evaluate each property at both sections:
𝑃1 = 𝑃2 = 𝑃atm = 0
𝑉1 = Surface velocity ≈ 0 (compared to velocity in pipe); 𝑉2 ≈ 0
𝑍1 = height at Section 1; 𝑍2 = height at Section 2
∑ ℎ𝑓 = losses encountered by a fluid particle in travelling from Section 1 to Section 2

The modified Bernoulli equation from Eq. (2.20) becomes

4𝑓𝐿1 4𝑓𝐿2 4𝑓𝐿3 𝑉2
𝑍1 = 𝑍2 + (𝐾inlet + + 𝐾90°elbow + + 𝐾90°elbow + + 𝐾valve + 𝐾exit )
𝑑 𝑑 𝑑 2𝑔

In Fig. 2.10b, the control volume includes the fluid in the tank at the
left and all the fluid in the pipe. Section 1 is the free surface of the
liquid in the tank and Section 2 is just at the end of the pipe. After
Section 2, the liquid could be discharged to the atmosphere, to
another, or to another pump. Its destination after Section 2 is of no
concern with regard to the analysis we formulate. The exit loss is not Fig. 2.10b
accounted for in Fig. 2.10b because the pressure loss in a fitting is
realized by the fluid only after it passes through it. Thus,
𝑃2 𝑉22 4𝑓𝐿1 4𝑓𝐿2 4𝑓𝐿3 𝑉2
𝑍1 = + + 𝑍2 + (𝐾inlet + + 2𝐾90°elbow + + + 𝐾valve )
𝜌𝑔 2𝑔 𝑑 𝑑 𝑑 2𝑔

where the exit velocity 𝑉2 equals the velocity in the pipe. Fig. 2.10c
shows the right tank removed as does Fig. 2.10b. In Fig. 2.10c,
however, our control volume does not end abruptly with the end of
the pipe. Instead, we use a large surface area as Section 2. The
pressure at the exit of the pipe is usually not equal to atmospheric
pressure, so we allow the fluid expand until its pressure does equal Fig. 2.10c
𝑃atm . Thus Section 2 is assumed to be the location where the liquid
pressure has become equal to atmospheric pressure. Moreover, because the area at Section 2 is so
large, the velocity of the liquid (or its kinetic energy) is reduced to a negligible value (compared to
the velocity in the pipe). In other words at Section 2, the pressure equals atmospheric pressure and
the kinetic energy of the liquid has dissipated. Thus,
4𝑓𝐿1 4𝑓𝐿2 4𝑓𝐿3 𝑉2
𝑍1 = 𝑍2 + (𝐾inlet + + 2𝐾90°elbow + + + 𝐾valve + 𝐾exit )
𝑑 𝑑 𝑑 2𝑔

June2012[Rev2.10:May2017]-pkq\dmee\afm\c2 2-12
For NON-COMMERCIAL use only.
 CHAPTER TWO [BTME2223 APPLIED FLUID MECHANICS]

In Fig. 2.10d, we have the same pipe leading to a tank. The inlet
of the pipe could be fed from a reservoir, by a pump, or by another
pipeline. It makes no difference in our analysis. Section 1 is at the
pipe inlet and Section 2 is the free surface of the liquid in the tank.
Fig. 2.10d
Thus,
𝑃1 𝑉12 4𝑓𝐿1 4𝑓𝐿2 4𝑓𝐿3 𝑉2
+ + 𝑍1 = 𝑍2 + ( + 2𝐾90°elbow + + + 𝐾valve + 𝐾exit )
𝜌𝑔 2𝑔 𝑑 𝑑 𝑑 2𝑔

Fig. 2.10e shows the pipe without tanks attached. The liquid source or its
ultimate destination do not affect out analysis. The locations of Section 1
and Section 2 are shown. For Fig. 2.10e, the modified Bernoulli equation
reduces to
𝑃1 𝑃2 4𝑓𝐿1 4𝑓𝐿2 4𝑓𝐿3 𝑉2 Fig. 2.10e
+ 𝑍1 = + 𝑍2 + ( + 2𝐾90°elbow + + + 𝐾valve )
𝜌𝑔 𝜌𝑔 𝑑 𝑑 𝑑 2𝑔

As indicated, it is extremely important to clearly define the boundary of the control volume. We are
now equipped to model piping problems. Again, we examine problems in which pressure drop ∆𝑃
(or ∆ℎ), volume flowrate 𝑄, or diameter 𝑑 is unknown which will be discussed in the examples later.

Example 2.3
In a water pipeline there is an abrupt change in diameter from 140 mm to 250 mm. If the head lost
due to separation when the flow is from the smaller to the larger pipe is 0.6m greater than the head
lost when the same flow is reversed, determine the flow rate.

Given then upstream pipe diameter, 𝑑1 is 140 mm, the From the table, the coefficient can be obtained by
downstream pipe diameter, 𝑑2 is 250 mm, and the head loss interpolating the value between it. Thus, 𝐾 ≈ 0.33.
difference is 0.6 m. Thus, the head loss due to sudden contraction
would be,

𝑢22 𝑢22
ℎ𝑐 = 𝐾 = 0.33
2𝑔 2𝑔

Observe that, the velocity at downstream of sudden

When the flow is from smaller to larger pipe the loss is due contraction is equal to the velocity at upstream of
to sudden enlargement and is given by Eq. (2.17), sudden enlargement. So,
2
𝑢12 𝑑12 𝑢12 𝑢12 𝑢12
ℎ𝑒 = (1 − 2 ) = (1 − 0.314)2 = 0.47 ℎ𝑐 = 0.33
2𝑔 𝑑2 2𝑔 2𝑔 2𝑔
When the flow is reversed, the loss is due to sudden
Also, the head loss difference for sudden
contraction. Area ratio,
enlargement and contraction is given by 0.6 m.
𝐴2 𝑑22 0.142
= = = 0.314
𝐴1 𝑑12 0.252 ℎ𝑒 − ℎ𝑐 = 0.6
The coefficient of contraction, 𝐶𝐶 , or the coefficient of
friction, 𝐾 can be obtained from the table provided. 𝑢12 𝑢12
0.47 − 0.33 = 0.6
2𝑔 2𝑔

𝑢1 = 9.17 m/s#

June2012[Rev2.10:May2017]-pkq\dmee\afm\c2 2-13
For NON-COMMERCIAL use only.
 CHAPTER TWO [BTME2223 APPLIED FLUID MECHANICS]

2.4 Total Energy (or Head) and Hydraulic Gradient (or Pressure Line)
The nature of the solution process for pipe flow problems can depend strongly on which the various
parameters are independent parameters (the “given) and which is the dependent parameter (the
“determine”). There are three common types of problems. Type I problem specify the desired
flowrate or average velocity and determine the necessary pressure difference or head loss. For
example, if a flowrate of 2gal/min is required for a dishwasher that is connected to the water heater
by a given pipe system, what pressure is needed in the water heater?

Type II problem specify the applied driving pressure (or, alternatively, the head loss) and determine
the flowrate.

In Type III problem, we specify the pressure drop and the flowrate and determine the diameter of the
pipe needed.

In order to carry on the analysis of pipe, we will use the Bernoulli’s equation, the total energy (head,
𝐻) is given as follow
2
𝑃 𝑢
𝐻= + +𝑧 (2.21)
𝜌𝑔 2𝑔
which can be represented by a vertical ordinate above the pipe centreline. The line joining the top of
these ordinate is called the total energy line (or total head line), where,
𝑃
= Pressure Head
𝜌𝑔
2
𝑢
= Velocity Head
2𝑔
𝑧 = Elevated Head

If a piezometer is inserted at any cross section in a

pipe carrying a liquid, the liquid rises in the
piezometer to a height corresponding to the sum of
Pressure head and Elevated head above the pipe
centreline. A line passing through the surfaces of the
liquid in piezo-meters is called the Hydraulic
Gradient or pressure line. However, the hydraulic
gradient is observed below the total energy line.

Fig. 2.11: Energy Change in a Fluid System

June2012[Rev2.10:May2017]-pkq\dmee\afm\c2 2-14
For NON-COMMERCIAL use only.
 CHAPTER TWO [BTME2223 APPLIED FLUID MECHANICS]

Example 2.4 (Self-exercise)

Consider a horizontal pipe line connecting two opened reservoirs as shown in the following figure.
The difference of water level in two tanks is 50 m. Considering the friction and non-uniformities
listed below, determine the discharge in m3/s and draw the total energy line and hydraulic grade line.
Take 𝑓 = 0.015 , 𝐾 = 0.2 (rounded entrance), 𝐶𝐶 = 0.78 (sudden contraction), 𝐾 = 10 (Globe
valve).

Solution:

Further calculation is shown in the following page.

June2012[Rev2.10:May2017]-pkq\dmee\afm\c2 2-15
For NON-COMMERCIAL use only.
 CHAPTER TWO [BTME2223 APPLIED FLUID MECHANICS]

From the steady state equation (Bernoulli equation), Thus, substitute the 𝑈2 and 𝑈3 into the simplified equation,
𝑃𝑜1 𝑈12 𝑃𝑜2 𝑈22 𝑈12
+ + 𝑍1 = + + 𝑍2 + ∑ ℎ𝑓 50 = (0.2 + 60 + 0.5625 + 1.875 + 0.0796 + 30 + 10 + 30
𝜌𝑔 2𝑔 𝜌𝑔 2𝑔 2𝑔
The reservoir is open to atmosphere, thus, 𝑃𝑜1 = 𝑃𝑜2 = 0 + 1)
and 𝑈1 = 𝑈2 = 0. Apply these conditions and the equation 𝑈12
can be further simplified into 50 = 133.7171
2𝑔
𝑍1 = 𝑍2 + ∑ ℎ𝑓 𝑈12 = 7.336
Given that the difference of water level in two tanks is 50 𝑈1 = 2.708 m/s
m. Therefore,
𝑍1 − 𝑍2 = ∑ ℎ𝑓 Therefore, the flowrate in the pipe is
𝑄 = 𝐴1 𝑈1
50 = ∑ ℎ𝑓 𝜋
𝑄 = (0.1)2 (2.708) = 0.02127 m3 /s#
4
So, from the diagram given, the losses in the piping system
then,
are:
𝑈2 = 0.667 m/s 𝑈3 = 2.708 m/s
Entry loss : 𝑈12
ℎ𝑓1 = 0.2 For the total energy at
2𝑔
At Point 0 : 𝐸0 = 55 m
Loss in the first : 4𝑓𝑙1 𝑈12
straight pipe ℎ𝑓2 = At Point 1, Entry : 𝐸1 = 𝐸0 − ℎ𝑓1 = 54.9253 m
𝑑1 2𝑔
loss
4 × 0.015 × 100 𝑈12 At before sudden : 𝐸2𝑎 = 𝐸1 − ℎ𝑓2 = 32.4994 m
=
0.1 2𝑔 enlargement
𝑈12 At after sudden : 𝐸2𝑏 = 𝐸2𝑎 − ℎ𝑓3 = 32.2892 m
ℎ𝑓2 = 60
2𝑔 enlargement
Loss in sudden : 𝑈12 𝑑12
2
𝑈12 At before sudden : 𝐸3𝑎 = 𝐸2𝑏 − ℎ𝑓4 = 31.6089 m
enlargement ℎ𝑓3 = (1 − 2 ) = 0.5625 contraction
2𝑔 𝑑2 2𝑔
At after sudden : 𝐸3𝑏 = 𝐸3𝑎 − ℎ𝑓5 = 31.5792 m
Loss in the : 4𝑓𝑙2 𝑈22
second straight ℎ𝑓4 = contraction
𝑑2 2𝑔
pipe At before the globe : 𝐸4𝑎 = 𝐸3𝑏 − ℎ𝑓6 = 20.3663 m
4 × 0.015 × 100 𝑈22 valve
=
0.2 2𝑔 At after the globe : 𝐸4𝑏 = 𝐸4𝑎 − ℎ𝑓7 = 16.6287 m
𝑈22 valve
ℎ𝑓4 = 30
2𝑔 At before exit : 𝐸5𝑎 = 𝐸4𝑏 − ℎ𝑓8 = 5.4158 m
2
Loss in sudden : 𝑈32 1 𝑈32 At after exit : 𝐸5𝑏 = 𝐸5𝑎 − ℎ𝑓9 = 5.0420 m
contraction ℎ𝑓5 = ( − 1) = 0.0796
2𝑔 𝐶𝐶 2𝑔
For the hydraulic energy at
Loss in the third : 4𝑓𝑙3 𝑈32 4 × 0.015 × 50 𝑈32
ℎ𝑓6 = = At Point 0 : 𝐻0 = 55 m
straight pipe 𝑑3 2𝑔 0.1 2𝑔
At Point 1, Entry : 𝑈21
𝑈32 𝐻1 = 𝐸1 − = 54.5515 m
ℎ𝑓6 = 30
loss 2𝑔
2𝑔 At before sudden : 𝑈21
Loss in fitting : 𝑈32 𝑈32 𝐻2𝑎 = 𝐸2𝑎 − = 32.1256 m
enlargement 2𝑔
ℎ𝑓7 = 𝐾 = 10
2𝑔 2𝑔 At after sudden : 𝑈22
𝐻2𝑏 = 𝐸2𝑏 − = 32.2665 m
Loss in the fourth : 4𝑓𝑙4 𝑈32 4 × 0.015 × 50 𝑈32 enlargement 2𝑔
straight pipe ℎ𝑓8 = =
𝑑4 2𝑔 0.1 2𝑔 At before sudden : 𝑈22
𝐻3𝑎 = 𝐸3𝑎 − = 31.5862 m
𝑈32 contraction 2𝑔
ℎ𝑓8 = 30
2𝑔 At after sudden : 𝑈23
𝐻3𝑏 = 𝐸3𝑏 − = 31.2054 m
Exit loss : 𝑈32 contraction 2𝑔
ℎ𝑓9 =
2𝑔 At before the globe : 𝑈23
𝐻4𝑎 = 𝐸4𝑎 − = 19.9925 m
valve 2𝑔
From the continuity equation, At after the globe : 𝑈23
𝐻4𝑏 = 𝐸4𝑏 − = 16.2549 m
𝐴1 𝑈1 = 𝐴2 𝑈2 = 𝐴3 𝑈3 valve 2𝑔
At before exit : 𝑈23
𝐻5𝑎 = 𝐸5𝑎 − = 5.0420 m
𝐴1 𝑑21 2𝑔
𝑈2 = 𝑈1 = 2 𝑈1 = 0.25𝑈1
𝐴2 𝑑2 At after exit : 𝑈2𝑜2
𝐻5𝑏 = 𝐸5𝑏 − = 5.0420 m
2𝑔
𝐴1 𝑑21
𝑈3 = 𝑈 = 𝑈1 = 𝑈1
𝐴3 1 𝑑23

June2012[Rev2.10:May2017]-pkq\dmee\afm\c2 2-16
For NON-COMMERCIAL use only.
 CHAPTER TWO [BTME2223 APPLIED FLUID MECHANICS]

Example 2.5
Water discharges from a reservoir A
through a 100 mm pipe 15 m long which
rises to its highest point at B, 1.5 m
above the free surface of the reservoir,
and discharges direst to the atmosphere
at C, 4 m below the free surface at A. The
length of pipe 𝑙1 from A to B is 5 m and
the length of pipe 𝑙2 from B to C is 10 m.
Both the entrance and exit of the pipe are
sharp and the value of 𝑓 is 0.08.
Calculate,
a) the mean velocity of the water leaving the pipe at C and
b) the pressure in the pipe at B.

a) Using Bernoulli equation, b) Again, apply the Bernoulli equation. But, at point B,
it is located before the exit. Thus, the local losses are
𝑃𝐴 𝑢𝐴2 𝑃𝐶 𝑢𝐶2 entry loss and frictional loss along pipe AB.
+ + 𝑍𝐴 = + + 𝑍𝐶 + ∑ ℎ𝑓
𝜌𝑔 2𝑔 𝜌𝑔 2𝑔
𝑃𝐴 𝑢𝐴2 𝑃𝐵 𝑢𝐵2
Based on the diagram given, the local losses are entry + + 𝑍𝐴 = + + 𝑍𝐵 + ∑ ℎ𝑓
𝜌𝑔 2𝑔 𝜌𝑔 2𝑔
loss and frictional loss along pipe AC. Noted that exit
loss is not considered in the calculation is due to the 𝑢12 4𝑓𝑙𝐴𝐶 𝑢12
water is delivered to atmosphere (atmospheric ∑ ℎ𝑓 = 0.5 + ∙
2𝑔 𝑑 2𝑔
pressure). Thus,
Same conditions are applied to eliminate the
𝑢12 4𝑓𝑙𝐴𝐶 𝑢12 unnecessary terms in the equation. Also, the velocity at
∑ ℎ𝑓 = 0.5 + ∙
2𝑔 𝑑 2𝑔 point B is same as the velocity at point C, thus, 𝑢𝐵 =
1.26 m/s.
The water pressure at point A and C is open atmosphere,
thus 𝑃𝐴 = 0 and 𝑃𝐶 = 0 ; the velocity at point A is 𝑃𝐵 𝑢𝐵2 𝑢𝐵2 4𝑓𝑙𝐴𝐵 𝑢𝐵2
stagnant, thus 𝑢𝐴 = 0. According to continuity law at 𝑍𝐴 = + + 𝑍𝐵 + 0.5 + ∙
𝜌𝑔 2𝑔 2𝑔 𝑑 2𝑔
constant cross section area, the velocity in the pipe is
equal to the velocity at point C, thus 𝑢1 = 𝑢𝐶 . Apply 𝑃𝐵 𝑢𝐵2 4𝑓𝑙𝐴𝐵
− = 𝑍𝐵 − 𝑍𝐴 + [1 + 0.5 + ]
these conditions and the Bernoulli equation can be 𝜌𝑔 2𝑔 𝑑
further simplified into the following form,
𝑃𝐵 1.585 4 × 0.08 × 5
− = 1.5 + [1 + 0.5 + ]
𝑢𝐶2 𝑢𝐶2 4𝑓𝑙𝐴𝐶 𝑢𝐶2 𝜌𝑔 2𝑔 0.1
𝑍𝐴 = + 𝑍𝐶 + 0.5 + ∙
2𝑔 2𝑔 𝑑 2𝑔
𝑃𝐵 = −28.58 kPa#
𝑢𝐶2 4 × 0.08 × 15
4= [1 + 0.5 + ] OR
2𝑔 0.1
𝑃𝐵 = 28.58 kPa (Below atmospheric pressure)
𝑢𝐶2 = 1.585

𝑢𝐶 ≈ 1.26 m/s#

June2012[Rev2.10:May2017]-pkq\dmee\afm\c2 2-17
For NON-COMMERCIAL use only.
 CHAPTER TWO [BTME2223 APPLIED FLUID MECHANICS]

Example 2.6 (self-practice)

Two reservoir A and B have a difference level of 9 m and are connected by a pipeline 200 mm in
diameter over the first part AC, which is 15 m long, and then 250 mm diameter for CB, the remaining
45 m length. The entrance to and exit from the pipes are sharp and the change of section at C is
sudden. The friction coefficient 𝑓 is 0.01 for both pipes. List the losses of head (energy per unit
weight) that occur, giving an expression for each.

June2012[Rev2.10:May2017]-pkq\dmee\afm\c2 2-18
For NON-COMMERCIAL use only.
 CHAPTER TWO [BTME2223 APPLIED FLUID MECHANICS]

Example 2.7 (Parallel Pipe network)

Two sharp-ended pipes of diameter 𝑑1 = 50 mm, and 𝑑2 = 100 mm, each of length 𝑙 = 100 m, are
connected in parallel between two reservoirs which have a difference of level ℎ = 10 m .
If the Darcy coefficient 𝑓 = 0.008 for each pipe, calculate:
a) the rate of follow for each pipe,
b) the diameter D of a single pipe 100 m long that would give the same flow if it was substituted
for the original two pipes.

a) Using Bernoulli equation, For the second pipe of its diameter of 100 mm,

𝑃𝐴 𝑢𝐴2 𝑃𝐵 𝑢𝐵2 𝑢22 4𝑓𝑙2 𝑢22 𝑢22

+ + 𝑍𝐴 = + + 𝑍𝐵 + ∑ ℎ𝑓 𝑍𝐴 = 𝑍𝐵 + 0.5 + ∙ +
𝜌𝑔 2𝑔 𝜌𝑔 2𝑔 2𝑔 𝑑2 2𝑔 2𝑔

Based on the diagram given, both point A and B are open 𝑢22 4𝑓𝑙2
[0.5 + + 1] = 𝑍𝐴 − 𝑍𝐵
atmosphere, thus 𝑃𝐴 = 𝑃𝐵 = 0 and velocity at both point 2𝑔 𝑑2
are stagnant, thus 𝑢𝐴 = 𝑢𝐵 = 0. The local losses for both
circular pipes are entry loss, frictional loss along pipe AB 𝑢22 4 × 0.008 × 100
[0.5 + + 1] = 10
(with different diameter), and exit loss. Thus, 2𝑔 0.10

𝑢2 4𝑓𝑙 𝑢2 𝑢2 𝑢2 = 2.42 m/s

∑ ℎ𝑓 = 0.5 + ∙ +
2𝑔 𝑑 2𝑔 2𝑔
Thus, the flowrate at first pipe is,
For the first pipe of its diameter of 50 mm, 𝜋
𝑄2 = 𝐴2 𝑢2 = × 0.102 × 2.42
4
𝑢12 4𝑓𝑙1 𝑢12 𝑢12
𝑍𝐴 = 𝑍𝐵 + 0.5 + ∙ +
2𝑔 𝑑1 2𝑔 2𝑔 𝑄2 ≈ 19 × 10−3 m3 /s#

𝑢12 4𝑓𝑙1
[0.5 + + 1] = 𝑍𝐴 − 𝑍𝐵
2𝑔 𝑑1

𝑢12 4 × 0.008 × 100

[0.5 + + 1] = 10
2𝑔 0.05

𝑢1 = 1.731 m/s

Thus, the flowrate at first pipe is,

𝜋
𝑄1 = 𝐴1 𝑢1 = × 0.052 × 1.731
4

𝑄1 ≈ 3.4 × 10−3 m3 /s#

June2012[Rev2.10:May2017]-pkq\dmee\afm\c2 2-19
For NON-COMMERCIAL use only.
 CHAPTER TWO [BTME2223 APPLIED FLUID MECHANICS]

Example 2.7
Two sharp-ended pipes of diameter 𝑑1 = 50 mm, and 𝑑2 = 100 mm, each of length 𝑙 = 100 m, are
connected in parallel between two reservoirs which have a difference of level ℎ = 10 m .
If the Darcy coefficient 𝑓 = 0.008 for each pipe, calculate:
b) the diameter D of a single pipe 100 m long that would give the same flow if it was substituted
for the original two pipes.

b) Now, both pipes are replaced by a single pipe with Rearrange the equation,
100 m long and provide same flowrate. Thus,
8.132 × 10−4 3.2
10 = [1.5 + ]
𝑄 = 𝑄1 + 𝑄2 = 22.4 × 10−3 m3 /s 2𝑔𝑑 4 𝑑

Using Bernoulli equation, 8.132 × 10−4 1.5𝑑 + 3.2

10 = [ ]
2𝑔𝑑4 𝑑
𝑃𝐴 𝑢𝐴2 𝑃𝐵 𝑢𝐵2
+ + 𝑍𝐴 = + + 𝑍𝐵 + ∑ ℎ𝑓
𝜌𝑔 2𝑔 𝜌𝑔 2𝑔 10 × 2 × 9.81 5
𝑑 = 1.5𝑑 + 3.2
8.132 × 10−4
Based on the diagram given, both point A and B are open
atmosphere, thus 𝑃𝐴 = 𝑃𝐵 = 0 and velocity at both point 241269𝑑5 − 1.5𝑑 − 3.2 = 0
are stagnant, thus 𝑢𝐴 = 𝑢𝐵 = 0 . These conditions are
To solve this equation, it is necessary to apply
applied and simplified the Bernoulli equation.
mathematical software to get the value of 𝑑. If not, this
equation can be solved graphically or by successive
𝑍𝐴 = 𝑍𝐵 + ∑ ℎ𝑓
approximation. An approximate answer can be obtained by
omitting the insignificant term. In this case, the
If a single pipe is used to connect both reservoir, then the
insignificant term is 𝑑. Therefore,
local losses would be entry loss, frictional loss along the
pipe, and the exit loss. 241269𝑑5 = 3.2
𝑢2 4𝑓𝑙 𝑢2 𝑢2 𝑑 ≈ 0.1058 m
∑ ℎ𝑓 = 0.5 + ∙ +
2𝑔 𝑑 2𝑔 2𝑔
The obtained value is only a start guess value which closes
To calculate the diameter of the pipe, we can rearrange the to the exact value. So,
continuity equation to eliminate the velocity term.
𝑓(𝑑) = 241269𝑑5 − 1.5𝑑 − 3.2
𝑄
𝑢=𝜋
2 If the function is equal to zero, the 𝑑 would be the solution
4𝑑 for this equation. Thus,
Substitute this term into the summation of head loss,
𝑓(0.1058) = 241269𝑑5 − 1.5𝑑 − 3.2 = −0.161
𝑢2 4𝑓𝑙
𝑍𝐴 − 𝑍𝐵 = [0.5 + + 1] The negative value of 𝑓(𝑑) suggests that the value chosen
2𝑔 𝑑
for 𝑑 was too small. So, it is necessary to increase the 𝑑
2 value, say 𝑑 = 0.107
𝑄
(𝜋 )
4𝑑
2
4𝑓𝑙 𝑓(0.107) = 241269𝑑5 − 1.5𝑑 − 3.2 = 0.022
𝑍𝐴 − 𝑍𝐵 = [0.5 + + 1]
2𝑔 𝑑
Perhaps, this 𝑑 is slightly higher. To obtain more precise
2 value of 𝑑, few more iterations are required. However, the
22.4 × 10−3 error is 0.022 which closes to zero. Thus, this 𝑑 value is
( 𝜋 2 )
4 𝑑 4 × 0.008 × 100 acceptable.
10 = [0.5 + + 1]
2𝑔 𝑑
𝑑 ≈ 0.1070 m#

June2012[Rev2.10:May2017]-pkq\dmee\afm\c2 2-20
For NON-COMMERCIAL use only.
 CHAPTER TWO [BTME2223 APPLIED FLUID MECHANICS]

Example 2.8
Water flow from a reservoir A through
a pipe of diameter d1 = 120mm and
length l1 = 120m to a junction at D,
from which a pipe of diameter d2 =
75mm and length l2 = 60m leads to
reservoir B in which the water level is
16m below that in reservoir A. A third
pipe, of diameter d3 = 60mm and
length l3 = 40m, leads from D to
reservoir C, in which the water level is
24m below that in reservoir A. Taking
f = 0.01 for all the pipes and neglecting all losses other than those due to friction, determine the
volume rates of flow in each pipe.

June2012[Rev2.10:May2017]-pkq\dmee\afm\c2 2-21
For NON-COMMERCIAL use only.
 CHAPTER TWO [BTME2223 APPLIED FLUID MECHANICS]

2.5 Incompressible Flow though Pipe Network

When the pipes are connected in series, the flow rate through the entire system remains constant
regardless of the diameters of the individual pipes in the system. The total head loss in the system,
including the sum of the head losses in individual pipe in the system, including the mirror losses.

∑ ℎ𝑓 = 𝐾1 𝑄 2 + 𝐾2 𝑄 2 + 𝐾3 𝑄 2 + 𝐾4 𝑄 2 + ⋯

For a pipe that branches out into two (or more) parallel pipes and then rejoins at a junction
downstream, the total flow rate is the sum of the flow rates in the individual pipes. The pressure drop
(or head loss) in each individual pipe connected in parallel must be the same since ∆𝑃 = 𝑃𝐴 − 𝑃𝐵 .
The governing equations for parallel pipes are
𝑄 = 𝑄1 + 𝑄2 + 𝑄3 + ⋯
and
ℎ𝑓 = 𝐾1 𝑄12 = 𝐾2 𝑄22 = 𝐾3 𝑄32 = ⋯
where 𝐾 is the loss coefficient and 𝑄 is the volume flowrate.

Extra Example:
IF this parallel pipe is now replaced by a single pipe connected between the same nodes and
transporting the same total flowrate, find the equivalent pipe loss coefficient (K) for this parallel pipe.

June2012[Rev2.10:May2017]-pkq\dmee\afm\c2 2-22
For NON-COMMERCIAL use only.
 CHAPTER TWO [BTME2223 APPLIED FLUID MECHANICS]

A pipe network is an interconnected set of pipes. The flow from a given inlet to a given outlet may
come through several different routes (loops).

Fig. 2.12 Pipe network

An attempt to apply Bernoulli's equation and the continuity of flow equation to the various elements
in the network would lead to a very large number of simultaneous equations which would be
cumbersome to solve.

The alternative approach is to use a method of successive approximations (known as Hardy-Cross

Method), assuming values for the flow in each pipe, or the heads at each junction, and checking
whether the values chosen satisfy the requirement that

a) the losses of head between any two junctions must be equal for all routes between these
junctions or the net head loss around a closed loop in the network must be zero. For
example:

b) the inflow to each junction must equal the outflow from that junction.

There are two methods to perform the analysis. The following derivation of equations are based on
the assumed flowrate and assumed head method. For assume flowrate, in general, the friction head
loss from the Darcy’s equation can be written as (refer to tutorial)
ℎ𝑓 = 𝐾𝑄 2 (2.22)
If the guessed distribution was correct (very unlikely), then
∑ ℎ𝑓 = 0 (2.23)
For ∑ ℎ𝑓 ≠ 0 the flow in each pipe is adjusted by an amount ∆𝑄 in order to bring the sum of the head
losses to zero. The correction, ∆𝑄 is found as follows. For any pipe, if the true flow is Q, and the
assumed flow is 𝑄0 , then
𝑄 = 𝑄0 + ∆𝑄 (2.24)

June2012[Rev2.10:May2017]-pkq\dmee\afm\c2 2-23
For NON-COMMERCIAL use only.
 CHAPTER TWO [BTME2223 APPLIED FLUID MECHANICS]

Substitute it into Darcy’s equation,

ℎ𝑓 = 𝐾(𝑄0 + ∆𝑄)2
ℎ𝑓 = 𝐾[𝑄02 + 2𝑄0 ∙ ∆𝑄 + (∆𝑄)2 ] (2.25)
If the guessed flow distribution was “reasonable” as suggested earlier, then ∆𝑄 is small in comparison
to 𝑄0 , and terms on the right hand side after the second can be neglected. Then for a head loss of zero
in a loop
∑ ℎ𝑓 = 0 = ∑ 𝐾 (𝑄02 + 2𝑄0 ∙ ∆𝑄)

∑ 𝐾𝑄02 + 2∆𝑄 ∑ 𝐾𝑄0 = 0 (2.26)

Therefore, solving for ∆𝑄 gives
∑ 𝐾𝑄02
∆𝑄 = − (2.27)
2 ∑ 𝐾𝑄0
To ensure that ∆𝑄 has an appropriate sign, the above equation must be sensitive to flow direction. To
do this, the denominator is made insensitive to direction, and the numerator becomes the term which
determines sign by writing the form
∑ 𝐾𝑄0 |𝑄0 |
∆𝑄 = − = 0.0000
2 ∑ 𝐾|𝑄0 |
The Hardy Cross method is known as a loop balancing procedure or successive approximation.The
error calculated ΔQ can be corrected by applying a balancing flow rate of the same magnitude but
opposite in sign to each pipe in the loop.

For the assumed head method, in general, the friction head loss from the Darcy’s equation can be
written as
ℎ𝑓 (2.28a)
ℎ𝑓 = 𝐾𝑄 2 → 𝑄=√
𝐾
Differentiating the equation with respect to volume rate of flow,
δℎ𝑓 = 2𝐾𝑄δ𝑄 (2.28b)
Rearrange the Darcy’s equation and it gives,
𝑄
δ𝑄 = δℎ (2.29)
2ℎ𝑓 𝑓
From the Fig. has shown that Flow in ab + Flow in cb = Flow in bd.
(𝑄𝑎𝑏 + δ𝑄𝑎𝑏 ) + (𝑄𝑐𝑏 + δ𝑄𝑐𝑏 ) = 𝑄𝑏𝑑 − δ𝑄𝑏𝑑 (2.30)
If hab, hcb and hbd are the assumed losses of head in pipes ab, cb and bd used to calculate 𝑄𝑎𝑏 , 𝑄𝑐𝑏 ,
and 𝑄𝑏𝑑 ,
𝑄𝑎𝑏 𝑄𝑐𝑏 𝑄𝑏𝑑
δ𝑄𝑎𝑏 = δℎ𝑓 δ𝑄𝑐𝑏 = δℎ𝑓 δ𝑄𝑏𝑑 = δℎ
2ℎ𝑎𝑏 2ℎ𝑐𝑏 2ℎ𝑏𝑑 𝑓

June2012[Rev2.10:May2017]-pkq\dmee\afm\c2 2-24
For NON-COMMERCIAL use only.
 CHAPTER TWO [BTME2223 APPLIED FLUID MECHANICS]

Substitute them into Eq. (2.30),

𝑄𝑎𝑏 𝑄𝑐𝑏 𝑄𝑏𝑑
(𝑄𝑎𝑏 + δℎ𝑓 ) + (𝑄𝑐𝑏 + δℎ𝑓 ) = 𝑄𝑏𝑑 − δℎ
2ℎ𝑎𝑏 2ℎ𝑐𝑏 2ℎ𝑏𝑑 𝑓
(𝑄𝑎𝑏 + 𝑄𝑐𝑏 − 𝑄𝑏𝑑 )
δℎ𝑓 = −
𝑄𝑎𝑏 𝑄 𝑄
+ 𝑐𝑏 + 𝑏𝑑
2ℎ𝑎𝑏 2ℎ𝑐𝑏 2ℎ𝑏𝑑
Using the sign convention that for flow towards the junction b both Q and h are positive,
∑𝑄
δℎ𝑓 = − (2.31)
𝑄
∑
2ℎ
When 𝛿ℎ𝑓 has been calculated, the new value of the head at b can be used to determine revised values
of 𝑄𝑎𝑏 , 𝑄𝑐𝑏 , and 𝑄𝑏𝑑 . The process is repeated until ∑ 𝑄 has been reduced to a negligible quantity.
∑𝑄
δℎ𝑓 = − = 0.0000
𝑄
∑
2ℎ

June2012[Rev2.10:May2017]-pkq\dmee\afm\c2 2-25
For NON-COMMERCIAL use only.
 CHAPTER TWO [BTME2223 APPLIED FLUID MECHANICS]

Example 2.9
A single looped network as shown
in the Fig. above has to be analyzed
by Hardy Cross method for given
inflow and outflow discharges. The
pipe diameters, d and lengths l are
shown in the Fig.. Use Darcy-
Weisbach head loss-discharge
relationship to determine the actual
flowrate at each pipe and assuming
the friction factor for all pipe are
constant which is 0.02.

Apply the sign convention where the flow in the loop is clockwise. First of all, assume the flow direction and its
flowrate at each pipe. Noted that the summation of flowrate in P1 and P4 must be equal to the inflow, which is
0.6 m3 /s. Also, in N2 and N4 has only one outlet which connecting P2 and P3 respectively, the flowrate must be
same in the previous pipe.

Pipe Flowrate (m3 /s) Direction

P1 0.1 Clockwise
P2 0.1 Clockwise
P3 0.5 Anti-clockwise
P4 0.5 Anti-clockwise

Now, create a table to include all the necessary information. With this, calculate the 𝐾 value for all pipes.

Pipe Pipe Length, m Assumed 𝐾 ℎ = 𝐾𝑄|𝑄| ℎ New flowrate,

= 𝐾|𝑄|
flowrate, Q, m3 / 𝑄 Q, m3 /s
s
P1 300 0.1 26337.45 263.3745 2633.7449 0.3
P2 200 0.1 17558.30 175.5830 1755.8299 0.3
P3 300 -0.5 26337.45 -6584.3621 13168.7243 -0.3
P4 200 -0.5 17558.30 -4389.5748 8779.1495 -0.3
Summation -10534.9794 26337.4486

∑ 𝐾𝑄0 |𝑄0 |
∆𝑄 = − = 0.2
2 ∑ 𝐾|𝑄0 |

If the ∆𝑄 is positive, new flowrate at each pipe has to be increased by ∆𝑄; while ∆𝑄 is negative, the new flowrate
must be decreased by ∆𝑄. The iteration is stopped when ∆𝑄 = 0. Now, second iteration:

Pipe Pipe Length, m Assumed 𝐾 ℎ = 𝐾𝑄|𝑄| ℎ New flowrate,

flowrate, Q, m3 / = 𝐾|𝑄| Q, m3 /s
𝑄
s
P1 300 0.3 26337.45 2370.3704 7901.2346 0.3
P2 200 0.3 17558.30 1580.2469 5267.4897 0.3
P3 300 -0.3 26337.45 -2370.3704 7901.2346 -0.3
P4 200 -0.3 17558.30 -1580.2469 5267.4897 -0.3
Summation 0 26337.4486

∑ 𝐾𝑄0 |𝑄0 |
∆𝑄 = − =0
2 ∑ 𝐾|𝑄0 |

Therefore, the final discharge at each pipe is 0.3 m3 /s.

June2012[Rev2.10:May2017]-pkq\dmee\afm\c2 2-26
For NON-COMMERCIAL use only.
 CHAPTER TWO [BTME2223 APPLIED FLUID MECHANICS]

Example 2.10
Reservoir A with its surface 60 m
above datum supplies water to junction
D through a 300 mm diameter pipe,
1500 m long. From the junction, a 250
mm diameter pipe, 800 m long, feeds
reservoir B, in which the surface level
is 30 m above datum, while a 200 mm
diameter pipe, 400 m long, feeds
reservoir C, in which the surface level
is 15 m above datum. Calculate the
volume rate of flow to each reservoir.
Assume that the loss of head due to friction is given by h = flQ2/3d5 for each pipe and that f = 0.01.

By inspection, point D is slightly higher than B, the flow is likely to be from D to B and C. Assume a trial value for
the head at D is 35 m. Then, initially,

Head loss at AD : ℎ𝐴𝐷 = 60 − 35 = 25 m

Head loss at DB : ℎ𝐷𝐵 = 35 − 30 = 5 m
Head loss at DC : ℎ𝐷𝐶 = 35 − 15 = 20 m

For each pipe, ℎ = 𝐾𝑄 2 , where 𝐾 = 𝑓𝑙⁄3𝑑5 . Therefore, the flow 𝑄 for any value of ℎ is
𝑄 = √ℎ⁄𝐾

For pipe AD : 0.01 × 1500

𝐾𝐴𝐷 = = 2058 𝑄𝐴𝐷 = √25⁄2058 = 0.1102 m3 /s
3 × 0.35

For pipe DB : 0.01 × 800

𝐾𝐴𝐷 = = 2730 𝑄𝐷𝐵 = −√5⁄2730 = 0.0428 m3 /s
3 × 0.255

For pipe DC : 0.01 × 400

𝐾𝐴𝐷 = = 4167 𝑄𝐷𝐶 = −√20⁄4167 = 0.0693 m3 /s
3 × 0.25

For the first iteration:

Pipe 𝐾 Assumed 𝑄, m3/s 𝑄⁄ Corrected head,

value for ℎ, m 2ℎ ℎ1 = ℎ + ∆ℎ, m
AD 2058 25 0.1102 0.00220 25.226
DB 2730 -5 -0.0428 0.00428 -4.774
DC 4167 -20 -0.0693 0.00173 -19.774
Summation -0.00186 0.00822

∑𝑄
∆ℎ = − = 0.226 m
∑ 𝑄⁄2ℎ

June2012[Rev2.10:May2017]-pkq\dmee\afm\c2 2-27
For NON-COMMERCIAL use only.
 CHAPTER TWO [BTME2223 APPLIED FLUID MECHANICS]

(continue Example 2.10)

For the second iteration:

Pipe 𝐾 Assumed 𝑄, m3/s 𝑄⁄ Corrected head,

value for ℎ, m 2ℎ ℎ1 = ℎ + ∆ℎ, m
AD 2058 25.226 0.1107 0.00219 25.225
DB 2730 -4.774 -0.0418 0.00438 -4.775
DC 4167 -19.774 -0.0689 0.00173 -19.775
Summation -0.00001 0.00832

∑𝑄
∆ℎ = − = −0.001 m
∑ 𝑄⁄2ℎ

For the third iteration:

Pipe 𝐾 Assumed 𝑄, m3/s 𝑄⁄ Corrected head,

value for ℎ, m 2ℎ ℎ1 = ℎ + ∆ℎ, m
AD 2058 25.225 0.11071 0.00219 25.225
DB 2730 -4.775 -0.04182 0.00438 -4.775
DC 4167 -19.775 -0.06889 0.00174 -19.775
Summation 0.00000 0.00832

The calculations for junction D can now be set out. After few corrections it can be seen that ∑ 𝑄 ≈ 0 (up to four
decimal places) and no further correction is necessary. The flows in pipes are

For pipe AD : 𝑄𝐴𝐷 = 0.1107 m3 /s from reservoir A.

For pipe DB : 𝑄𝐷𝐵 = 0.0418 m3 /s to reservoir B.
For pipe DC : 𝑄𝐷𝐶 = 0.0689 m3 /s to reservoir C.

If given the pressure at entrance is 500 kPa, then the pressure at junction D:

𝑃𝐷 = 200 kPa − 1000 × 9.81 × 25.225 = 25.3 kPa

June2012[Rev2.10:May2017]-pkq\dmee\afm\c2 2-28
For NON-COMMERCIAL use only.
 CHAPTER TWO [BTME2223 APPLIED FLUID MECHANICS]

Chapter Review:
 If steady flow occurs between two parallel plates, or within a pipe, then regardless if the flow
is laminar or turbulent, the shear stress within the fluid varies in a linear manner such that it
will balance the forces of pressure, gravity, and viscosity.
 Unless otherwise specified, in this text, laminar flow for pipes is Re ≤ 2000; turbulent flow
for pipes is Re > 4000.
 The frictional loss within a pipe can be determined analytically for both laminar and turbulent
flow.
 For laminar flow in smooth and rough pipes, the Darcy (frictional) factor can be determined
using
16
𝑓=
Re
 For turbulent flow in smooth pipes, the Darcy (frictional) factor can be determined using
0.079
𝑓= 1
Re4
 For turbulent flow in rough pipes, the Darcy (frictional) factor can be determined from Moody
diagram with known Reynolds number (Re) and relative roughness (𝜀).
 Once 𝑓 is obtained, then the head loss, called a “major loss”, can be determined using the
Darcy-Weisbach equation.
4𝑓𝑙 𝑢2
ℎ𝑓 = ∙
𝑑 2𝑔
 If a pipe network has fittings and transitions, then the head loss produced by these connections
must be taken into account. These losses are called “minor losses”.
 Pipe systems can be arranged in series, in which case the flow through each pipe must be the
same, and the head loss is the total loss for all the pipes.
𝑄𝐴 = 𝑄𝐵 ∑ ℎ𝑓 = ℎ𝑓1 + ℎ𝑓2 + ℎ𝑓3 + ℎminor

 Pipes can also be arranged in parallel, in which case the total flow is the accumulation of flow
from each branch pipe in the system, and the head loss for each branch pipe is the same.
𝑄𝐴 = 𝑄𝐵 = 𝑄1 + 𝑄2 ℎ𝑓1 = ℎ𝑓2

June2012[Rev2.10:May2017]-pkq\dmee\afm\c2 2-29
For NON-COMMERCIAL use only.
 CHAPTER TWO [BTME2223 APPLIED FLUID MECHANICS]

Appendix 2A

Moody Diagram

Fig. 2A.1: Moody diagram. (Source: Pao, R.H.F., 1961, Fluid Mechanics, New York: John Wiley & Sons, p.284)

June2012[Rev2.10:May2017]-pkq\dmee\afm\c2 2-30
For NON-COMMERCIAL use only.