KNS2153

HYDRAULICS
CLASS NOTES

Chapter 1 Behavior of Real Fluids

Chapter 2 Flow in Pipes and Closed Conduits
Chapter 3 Pump-Pipeline System Analyses and Design
Chapter 4 Pressure Surge in Pipelines (Water Hammer)

Instructor: Puan Rosmina Ahmad Bustami

Ext. 3337
[email protected]

Please visit KNS2153 on morhpeus.calm.unimas.my for regular updates

*This material is intended for instructions in KNS2153 class only. Further explanations
on theory/practical aspects practiced in hydraulic engineering discipline may be obtained
from other relevant resources.*
 1 - BEHAVIOUR OF REAL FLUIDS

1. Real and Ideal Fluids

A fluid is assumed ideal if it posses the following characteristics:

1. It is inviscid
2. It is incompressible
3. It has no surface tension
4. It always form a continuum

In civil engineering, surface tension problems are encountered only in the very low flows
over weirs or in hydraulic models. Compressibility phenomena are associated with surge
as in the case of water hammer.

The viscosity is the characteristics which differentiates so many aspects of real fluid
flows from ideal flows.

Figure 1 Ideal and real flow around a cylinder.

KNS2153/RB 1
 2. Viscous Flow

In Fig 2, fluids are both sandwiched between a fixed solid surface on one side and a
movable belt on the other.

Figure 2 Effect of shear force on fluid

If the belt is set in motion, experimental measurements will indicate:

a- Ideal Fluid
 The force required to move the belt is negligible
 The movement of the belt has no effect whatsoever on the ideal fluid,
which remains stationary.

b- Real Fluid
 A considerable force is required to maintain belt motion, even at low
speed.
 The whole body of fluid is deforming and continues to deform as long as
belt motion continues. Closer investigation will reveal that deformation
pattern consists in the shearing, sliding, of one layer of fluid over another.
The layer of fluid immediately adjacent to the slid surface will adhere to
that surface and, similarly, the layer adjacent to the belt adheres to the belt.
Between the solid surface and the belt the fluid velocity is assumed to vary
linearly.

KNS2153/RB 2
The stability of laminar flows and the onset of turbulence

If the speed of the belt increased, the pattern of linearly sheared flow will continue to
exist only up to a certain belt velocity. Above that velocity a dramatic transformation
takes place in the flow pattern.

Effect of a disturbance in a sheared flow

Consider a slowly moving sheared flow. If a small disturbance happened (a small local
vibration), the pathlines will be slightly deflected, bunching together at A and opening
out correspondingly at B. This implies that the local velocity at A, u A, increases slightly
compared with the upstream velocity u, while at the same time uB reduces. From
Bernoulli equation,

P u2 PA u A2 PB u B2
g 2g g 2g g 2g

Figure 3 Effects of disturbance on a viscous flow

Therefore PB>PA. It can thus be reasoned that the disturbance will produce a small
transverse resultant force acting from B towards A. The lateral components of velocity
will also produce a corresponding component of viscous shear force, which acts in the
opposite sense to the resultant disturbing force. As long as the fluid is moving slowly, the
resultant disturbing force tends to be outweighed by the viscous force. Disturbances are
therefore damped out. As the rate of shear increases, the effect of the disturbance
becomes more pronounced:

 The difference between uA and uB increases

 The pressure differences (PA-PB) increases with (uA2-uB2), so the deflection of the
pathline becomes more pronounced
 The greater shear results in a deformation of the crest of the pathline pattern.
When the shear is sufficiently great, the deformation is carried beyond the point at
which the rectilinear pattern of pathlines can cohere. The flow disintegrates into a
disorderly pattern of eddies in place of the orderly patterns of layers.

KNS2153/RB 3
The Boundary layer

Figure 4 Development of a boundary layer

Consider a rectilinear flow passing over a stationary flat plate, which lies parallel to the
flow. The incident flow has a uniform velocity, U . As the flow comes into contact with
the plate, the layer of fluid immediately adjacent to the plate decelerates (due to viscous
friction) and comes to rest. This follows from the postulate that in viscous fluids a thin
layer of fluid ‘adhere’ to a solid surface. There is then a considerable shearing action
between the layer of fluid on the plate surface and the second layer of fluid. The second
layer is therefore forced to decelerate (it is not quiet brought to rest), creating a shearing
action with the third layer of fluid and so on. As the fluid passes further along the plate,
the zone in which shearing action occurs tends to spread further outwards. This zone is
known as a ‘boundary layer’. Outside the boundary layer the flow remains effectively
free of shear, so the fluid here is not subjected to viscosity-related forces. The fluid flow
outside the boundary layer may therefore be assumed to act like an ideal fluid.

The flow within the boundary layer may be viscous or turbulent, depending on the value
of the Reynolds Number. To evaluate Re we need a ‘typical dimension’ and in boundary
layers this dimension is usually the distance in the x-plane from the leading edge of the
solid boundary. The Reynolds number thus becomes Rex= U x/ .

The structure of the boundary layer is therefore as shown in figure 5.

KNS2153/RB 4
 Figure 5 Structure of a boundary layer

Implications of the boundary layer concept

Different material exhibit different degrees of roughness. Does this have any effect on the
boundary layer?

 In laminar flow, the friction is transmitted by pure shearing action. Consequently, the
roughness of the solid surface has no effect, except to trap small ‘pools’ of stationary
fluid in the interstices, and thus slightly increase the thickness of the stationary layer
of fluid.
 In a turbulent flow, a laminar sub-layer forms close to the solid surface. If the average
height of the surface roughness is smaller than the height of the laminar sub-layer,
there will be little or no effect on the overall flow.
 Turbulent flow embodies a process of momentum transfer from layer to layer.
Consequently, if the surface roughness protrudes through the laminar region into the
turbulent region, then it will cause additional eddy formation and therefore greater
energy loss in the turbulent flow. This implies that the apparent frictional shear will
be increased.

KNS2153/RB 5
 2 – FLOW IN PIPES AND CLOSED CONDUITS

1 Introduction

The flow of water, oil and gas in pipes is of immense practical significance in civil
engineering. Water is conveyed from its sources normally in pressure pipelines, to water
treatment plants where it enters the distribution system and finally arrives at the
consumer. Surface water drainage and sewerage is conveyed by closed conducts, which
do not usually operate under pressure, to sewage treatment plants, from where it is
usually discharged to a river or the sea. Oil and gas are often transferred from their source
by pressure pipelines to refineries (oil) or into distribution network for a supply (gas).

In the following, the theories of pipe flow will be described.

2 Two kinds of flow

Osborn Reynolds demonstrated in 1880 that there were two kinds of flow

- Laminar flow
- Turbulent flow

He found that the transition from laminar to turbulent flow occurred at a critical
Reynolds number Re.

VD
Re (1)
v

where V : mean flow velocity

D : pipe diameter
v : kinetic viscosity

Re<2000 Flow is always laminar

2000<Re<4000 Flow is either laminar or turbulent
Re>4000 Flow is always turbulent

He also showed that:

For laminar flow, the frictional head loss was proportional to velocity.
For turbulent flow, the head loss was proportional to the square of the velocity.

KNS2153/RB 6
 Figure 1 Reynolds Demonstration for the different kinds of flow

Figure 2 Hydraulic gradient versus flow velocity

KNS2153/RB 7
3 Frictional loss along a pipe

A fluid moving through a pipeline is subjected to energy losses due to frictional

resistance at the pipe wall.

Figure 3 Flow along a uniform pipe

i: hydraulic gradient = slope of the pressure line

Application of Bernoulli equation between section 1 and 2 yields,

P1 P2
Z1 Z2 hf (2)
g g

Momentum equation between section 1 and 2 leads to,

( P1 P2 ) A gAL sin 0 PL (3)

A : area of cross section

P : wetted perimeter
0 : the boundary shear stress

Rearranging equation 2 and noting that L sin = Z1-Z2

P1 P2 0 PL
g Z1 Z2 (4)
gA

As
P1 P2
hf Z1 Z2 (5)
g

KNS2153/RB 8
Whence
0 PL
hf (6)
gA

or
hf
0 gR gRS f (7)
L
where R (hydraulic radius) = A/P ( = D/4 for a circular pipe of diameter D).

Rewriting (7):
D hf
0 g (8)
4 L

Figure 4 Velocity distribution across a pipe

(i) Laminar flow

hf can be obtained theoretically in the form of the Hagen-Pouiseuille equation:

32 LV
hf Hagen-Pouiseuille Equation (9)
gD 2

(ii) Turbulent flow

L V2
hf Darcy-Weisbach Formula (10)
D 2g

: Dimensionless friction factor

KNS2153/RB 9
1. Smooth pipes

In the case of turbulent flow experimental work on smooth pipes by Blasius (1913)
yielded the relationship:

0.3164
Blasius equation (11)
Re 0.25

2. Nikuradse’s experimental results

Nikuradse conducted experiment on pipes which were artificially roughened by sticking

uniform sand grains.

He defined Ks/D as the relative roughness.

His experimental results are plotted in Fig 4.

Figure 5 Nikuradse’s experimental results

This figure shows that there are 5 regions of flow, as follows:

a) Laminar Flow

The relative roughness has no influence on the friction factor.

64 64
(12)
DV Re

b) Transition from laminar to turbulent flow

An unstable region between Re=2000 and 4000. Pipe flow normally lies outside this
region.

c) Smooth turbulence

KNS2153/RB 10
The limiting line of turbulent flow, approached by all values of relative roughness as Re
decreases.

d) Transitional turbulence

The region in which varies with both Re and Ks/D. In practice, most pipe flow lies
within this region.

e) Rough turbulence

This region in which remains constant for a given Ks/D, and is independent of Re.

3. The rough and smooth laws of Van Karman and Prandtl

Van Karman and Prandtl suggested the following semi-empirical formulas:

a- For smooth pipes

1 Re
2 log (13)
2.51
b- For rough pipes

1 3.7 D
2 log (14)
Ks

The smooth law is a better fit to the experimental data than the Blasius equation.

4. The Colebrook-White formula

Colebrook and White studied the effects of non uniform roughness as found in
commercial pipes. They demonstrated that in the turbulent region, the -Re curves
exhibited a gradual change from smooth to rough turbulence in contrast to Nikuradse’s
‘S-shaped’ curves for uniform roughness. They determined an ‘effective roughness’ size
for the commercial pipes equivalent to Nikuradse’s results.

Their experimental study leads to the following formula:

1 ks 2.51
2 log (15)
3.7 D Re

This formula is applicable for the whole range of the turbulent region for commercial
pipes using an effective roughness value determined experimentally for each type of pipe,
as given in Table 1.

KNS2153/RB 11
Pipe Material Ks (mm)
Brass, copper, glass, Perspex 0.003
Asbestos, Cement 0.03
Wrought iron 0.06
Galvanised Iron 0.15
Plastic 0.03
Bitumen-lined ductile iron 0.03
Spun concrete lined ductile iron 0.03
Slimed concrete sewer 6.0

Combining the Darcy-Weisbach (10) and Colebrook-White (15) equations yields an

explicit expression for the velocity V:

Ks v
V 2 2 gDi log 2.51 (16)
3.7 D D 2 gDi

The Colebrook-White formula (15) was first plotted in the form of a -Re diagram by
Moody and hence is generally referred to as the ‘Moody Diagram’.

Due to the implicit form of the Colebrook-White equation, a number of approximations

in explicit form in have been proposed.

Moody produced the following formulation:

1/ 3
K 106
0.0055 1 20000 s (17)
D Re

which gives correct to 5% for 4x103 < Re < 1x107 and Ks/D<0.01.

More recently, Barr proposed another explicit formulation for

1 Ks 5.1286
2 log (18)
3.7 D Re 0.89

For Re>105 this equation provides a solution for I (hf/L) to an accuracy better than 1%.

KNS2153/RB 12
5. The Hazen Williams formula

Using Colebrook-White formula, Moody diagram; or Barr formula is simple to apply for
single pipeline. However, for pipes in series, parallel or for the more general case of pipe
networks, it rapidly becomes impossible to use for hand calculations. For this reason,
simpler empirical formulae are still in use. Perhaps the most notable is the Hazen-
Williams formulae, which takes the form:

0.54
0.63
hf
V 0.355CD (19)
L

or alternatively,

1.85
hf 6.78L V
(20)
L D 1.165 C

Where c is a coefficient depending on pipe diameter, material and age.

4 Resistance to flow in non-circular sections

In order to use the same form of resistance equations such as Darcy and Colebrook-
White, it is convenient to treat the non-circular sections as an equivalent hypothetical
circular section yielding the same hydraulic gradient at the same discharge.

The ‘transformation’ is achieved by expressing the diameter D in terms of the hydraulic

radius R=A/P and since for circular pipes R=D/4, equations (10) and (15) become:

LV 2
hf (21)
8g R
1 Ks 2.51v
and 2 log (22)
14.8 R 4V R

5 Local head loss

In addition to the spatially continuous head loss due to friction, local head losses occur at
changes of cross-section, at valves and at bends. These local losses are sometimes
referred to as ‘minor’ losses since in long pipelines their effect may be small in relation to
the friction loss. However, the head loss at a control valve has a primary effect in
regulating the discharge in a pipeline.

Local losses occur at pipe bends, junctions, and valves, etc. In the case of long pipelines
(e.g. several kilometers) the local losses may be negligible, but for short pipelines, they
may be greater than the frictional losses.

KNS2153/RB 13
The local losses is usually given by:

V2
hL KL (23)
2g

hL : local head loss

KL: constant for a particular fitting.

For the case of sudden enlargement (exit from a tank) an expression may be derived from
KL in terms of the area of the pipe. This result may be extended to the case of a sudden
contraction. For all other cases (bends, valves, junctions, bellmouths, etc) values of KL
must be derived experimentally.

A. Sudden enlargement

Figure 6 Sudden enlargement

Application of the equation of momentum and the equation of continuity between section
1 and 2 yields

P1 A1 P2 A2 Q(V2 V1 )
(24) and (25)
Q V1 A1 V2 A2

these two equations lead to,

P1 P2 V2
(V1 V2 ) (26)
g g

Bernoulli equation yields,

(V12 V22 ) ( P2 P1 )
hL (27)
2g g

KNS2153/RB 14
So,
(V1 V2 ) 2
hL (28)
2g

by using (25) we obtain

2
V12 A1
hL 1 (29)
2g A2

Whence,
2
A1
KL 1 (30)
A2

For the case of a pipe discharging into a tank, A2>A1, then KL=1.

B. Sudden Contraction

Figure 7 Sudden contraction

From position 1 to 1’ the flow contracts, forming a vena contracta. Experiments indicate
that the contraction of the flow area is generally about 40%. If the energy loss from (1) to
(1’) is assumed to be negligible, then the remaining head loss occurs in the expansion
from (1) to (2). Since an expansion loss gave rise to (30), that equation may now be
applicable here. As A1’=0.6A2 then:

V22
hL 0.44 (31)
2g

KNS2153/RB 15
6 Series, parallel and branched pipe systems

To determine the heads and discharges is more complex than in simple pipe problems,
and requires the use of the continuity equation in addition to the energy and frictional
head loss equations.

A. Pipes in Series

Figure 8 Pipes in Series

H hf1 hf 2
(32)
Q Q1 Q2

B. Pipes in Parallel

Figure 9 Pipes in Parallel

H hf1 hf 2
(33)
Q Q1 Q2

KNS2153/RB 16
 C. Branched Pipes

Figure 10 Branched pipes

hf z1 Hj
hf 2 z2 Hj
(34)
hf 3 Hj z3
Q3 Q1 Q2

A method of solution is as follows:

1- Guess Hj
2- Calculate Q1, Q2, and Q3
3- If Q1 + Q2 = Q3, then the solution is correct
4- If Q1 + Q2 Q3, then return to (1)

KNS2153/RB 17
7 Water distribution systems: Pipe network analyses

A water supply distribution system consists of a complex network of interconnected

pipes, service reservoir and pumps which deliver water from the treatment plant to the
consumer. Water demand is highly variable, both by day and season. The distribution
system must include a storage element and must be capable of flexible operation. Water
pressures within the system are normally kept between a maximum (about 70m head) and
a minimum (about 20m head).

The pipe network analysis involves the determination of the pipe flow rates and pressure
heads which satisfy continuity and energy conservation equations.

Two standard methods are generally used to analyze pipe networks; the loop method
(Hardy-Cross method) and the nodal method. These will be described in the following
sections:

7.1 The Loop Method (The head balance method)

This method, originally proposed by Prof. Hardy-Cross in 1936, often referred to as the
Hardy-Cross method. It is applicable to closed-loop network. The basis of the method is
as follows:

1- For a given pipe system, with junction outflows, the flows in each pipe is
estimated. At each junction, these flows must satisfy the continuity criterion.
2- Calculate the head loss for each pipe
3- If the algebraic sum of the head losses around any closed loop is zero, then the
solution is correct (the sign convention is that clockwise flows are positive).
4- If not, apply a correction factor Q to all the discharges in the loop and return to
(2).

A reasonable efficient value for Q for rapid convergence is given by:

h fi
Q (35)
2 h fi / Qi

Figure 11 Flow in pipe loop

KNS2153/RB 18
7.2 Nodal method (Quantity balance method)

This method is suited for branched network. The basis of the method is as follow:

1- Assume the head Hj at each junction

2- Calculate the discharge flowing in each pipe, Qi
3- If at each junction we verify that Qi=0 then the solution is correct.
4- If Qi 0, then apply a correction factor H to Hj and return to (2), where

2 Qi
H (36)
Qi / h fi

Figure 12 Flow in a branched network

Temperature ( C) Kinematic Viscosity, (m2/s)

0 1.787 x 10-6
5 1.519 x 10-6
10 1.307 x 10-6
20 1.004 x 10-6
30 8.009 x 10-7
40 6.580 x 10-7
50 5.534 x 10-7
60 4.745 x 10-7
70 4.134 x 10-7
80 3.650 x 10-7
90 3.260 x 10-7
100 2.940 x 10-7

KNS2153/RB 19
 3 – PUMP-PIPELINE SYSTEM ANALYSES AND DESIGN

1 Introduction

This chapter deals with the analysis and design of pipe systems which incorporate
rotodynamic pumps. The civil engineer is mostly concerned with pump selection, in the
design of river abstraction, borehole supplies from ground water and surface water and
foulwater drainage from low-lying land. Rotodynamic pumps can be sub-classified
according to the shape of impellers into 3 main categories:

(i) centrifugal (radial flow)

(ii) mixed flow
(iii) propeller (axial flow)

For the same power input and efficiency, the centrifugal pump would generate relatively
large pressure head with a low discharge, the propeller type a relatively large discharge at
a low head with the mixed flow having characteristics somewhere between the two.

Pump types may be more explicitly defined by the parameter called SPECIFIC SPEED
(Ns) expressed by:

N Q
Ns (1)
H 3/ 4

where Q: discharge
H: total head
N: rotational speed (rev/min)

This expression may be interpreted as the speed in rev/min at which a geometrically

scaled model would have to operate to deliver unit discharge (eg. 1 l/s) when generating
unit head (eg. 1m).

Pump type Ns range

Centrifugal Up to 2600
Mixed flow 2600 to 5000
Axial flow 5000 to 10 000

The total head generated by a pump is also called the manometric head (Hm) since it is
the difference in pressure head recorded by pressure gauges connected to the delivery and
inlet pipes on either side of the pump, provided that the pipes are of the same diameter.

KNS2153/RB 20
2 Head terms in pumping

For the following terms, refer to Fig 1.

State suction lift. The vertical distance from the water level in the source tank to the
centerline of the pump. If the pump is located at a lower level than the source tank (Fig
b), the static suction lift is negative.

Static discharge head. The vertical distance from the centerline of the pump to the water
level in the discharge tank.

Total static head. The sum of the static suction lift and the static discharge head, which
is equal to the difference between the water levels of discharge and source tanks.

Total dynamic head (TDH). The sum of the total static head and the friction and minor
losses. This term is also known as the total head.

Figure 1 Head terms in pumping

KNS2153/RB 21
3 System Head Curve

For any piping system, the friction loss, hf [=( L/D)(V2/2g)], and the minor loss, hm
(= KV2/2g), can be expressed in terms of the flow through the system. Thus,

LV 2 kV 2
Hp Z (2)
D2 g 2g

or,

0.81 LQ 2 KQ 2
Hp Z (3)
g D5 D4

The plot of eq (3) between Hp versus Q, as shown in Fig 2 is known as the system head
curve. This curve, representing the behavior of the piping system, is important in the
selection of a pump.

Figure 2 Typical system head curve

4 Pump Characteristics Curves

For a given pump at a given speed, there are definite relationships among the pump
discharge capacity, head, power, and efficiency. These relations are derived from actual
tests on a given pump or similar unit and are usually depicted graphically by the pump
characteristics or performance curves, comprising the following:

a- Pumping head versus discharge

b- Power input (P) versus discharge
c- Efficiency ( ) versus discharge

The energy imparted to fluid is gQHm, so that a pump efficiency may be derived as

gQH m
(4)
P

KNS2153/RB 22
 Figure 3 Pump Characteristic Curve

Fig 3 illustrates the typical characteristics curves. The general shape of the curves varies
with the size, speed, and design of a particular pump. The important feature of the curves
is that an increase in the head reduces the capacity. These curves are supplied by the
manufacturer of the pump. A set of characteristic curves represents the behavior of a
given-size pump operating at a given speed, in the same manner as a system head curve
represents the behavior of a piping system. At a given speed, a pump is rated at the head
and discharge, which gives the maximum efficiency, referred to as the best efficiency
point, shown by point A in Fig 3. The characteristic curves, particularly the head-
discharge curve, are important in pump selection.

5 Single Pump and Pipeline System

The primary requirement is to determine a suitable pump and pipe combination for the
required design discharge. Consider Fig 4 which shows a simple pumping main. It is
clear that the pump must generate a total head equal to Hst plus the pipeline head losses at
Q.

Vs : velocity in suction pipe

Vd : velocity in delivery pipe
hld : head loss in delivery pipe (friction, valves, etc.)
hls : head loss in suction pipe
hm : local losses

Manometric head is defined as the rise in total head across the pump.

Pd Vd2 Ps Vs2
Hm (5)
g 2g g 2g

KNS2153/RB 23
with:
Ps Vs2 Pd Vd2
Z1 hls ; Z2 hld
g 2g g 2g

Thus,

Hm H st hld hls (6)

The discharge is a function of both the pump and pipeline. For a given system, the head-
discharge characteristic curves for the pump may be superimposed on that for the
pipeline, as shown in Fig 5. The point of intersection of the two characteristics curves
locates the one possible combination of head and discharge for the system under steady
flow conditions. The intersection point is referred to as the operating point.

Fig 4 Simple pumping main

Figure 5 Pump and pipeline characteristics curves

KNS2153/RB 24
6 Multiple Pump Systems

A. Pumps in Parallel

Pumping stations frequently contain several pumps in ‘parallel’ arrangement. In this

configuration (Fig 6), any number of the pumps can be operated simultaneously, the
objective being to deliver a range of discharges. This is a common feature of sewage
pumping stations where the inflow rate varies during the day. By automatic switching
according to the level in the suction well, any number of the pumps can be brought into
operation.

In predicting the head vs discharge curve for parallel operation, it is assumed that the
head across each pump is the same. Thus, at any arbitrary head the individual pump
discharges are added (Fig 6b).

H np Hp Qnp nQp (7)

Fig 6 (a) Pumps operating in parallel

6 (b) Characteristic curves for identical pumps operating in parallel

KNS2153/RB 25
B. Pumps in Series

This configuration is the basis of multistage and borehole pumps; the discharge from the
first pump (or stage) is delivered to the inlet of the second pump, and so on. The same
discharge passes through each pump receiving a pressure boost in doing so. Fig 7 shows
the series configuration together with the resulting head vs. discharge characteristics
which are clearly obtained by adding individual pump manometric heads at any
discharges. All the pump in series system must be operating simultaneously.

H np nH p Qnp Qp (8)

Figure 7 Pumps operating in series

7 Variable speed pump operation

By the use of variable speed motors the discharge of a single pump can be varied to suit
the operating requirements of the system. Using dimensionless analysis and dynamic
similarity criteria, it can be shown that if the pump delivers a discharge Q1 at manometric
head H1 when running at speed N1 the corresponding values when the pump is running at
speed N2 are given by the relationships

N2
Q2 Q1 ( )
N1
(9)
N
H2 H1 ( 2 ) 2
N1

KNS2153/RB 26
 Figure 8 Effect of speed change on pump characteristics

In constructing the characteristic curve for speed N2, several pairs of values of Q1, H1
from the curve for N1 can be obtained and transformed into homologous points Q2, H2 on
the N2 curve. (Fig 8).

8 Suction lift limitations

Cavitation, the phenomenon which consists of local vaporization of a liquid and which
occurs when the absolute pressure falls to the vapour pressure of the liquid at the
operating temperature, can occur at the inlet to a pump and on the impeller blades,
particularly if the pump is mounted above the level in the suction well. Cavitation causes
physical damage, reduction in discharge and noise, and to avoid it the pressure head at
inlet should not fall below a certain minimum which is influenced by the further
reduction in pressure within the pump impeller. (Fig 9)

Figure 9 Head conditions in suction pipe

ps pv
If ps represents the pressure at inlet, then is the absolute head at the pump inlet
g
above the vapour pressure (pv) and is known as the net positive suction head (NPSH).

KNS2153/RB 27
 ps pv pa pv Vs2
Thus, NPSH = hs hls (10)
g g 2g

Where pa : ambient atmospheric pressure

hs : suction lift
hls : total head loss in suction pipe
Vs : velocity head in suction pipe
: density of liquid.

Value of NPSH can be obtained from the pump manufacturer and are derived from full-
scale or model tests; these values must not be exceeded if cavitation is to be avoided.

Thoma introduced a cavitation number:

NPSH
(11)
Hm

and from physical tests found this to be strongly related to specific speed.

In recent years electro-submersible pumps in the small to medium size range have been
widely used. This type eliminates the need for suction pipes and are provided that the
pump is immersed to the manufacturer’s specified depth, the problems of cavitation and
cooling are avoided.

9 Installation of a Centrifugal Pump

A typical centrifugal pump installation is shown in Fig 10. The pump is installed between
two valves for easy removal in case of maintenance. On the suction side, a combined
bellmouth entry and strainer are necessary, together with a non-return valve to ensure
self-priming. On the delivery side, a second non-return valve is necessary to prevent
damage from possible surge pressures. In addition, an air valve and flow meter (venturi
type) are desirable.

Figure 10 Typical centrifugal pump installation

KNS2153/RB 28
 4 – Pressure Surge in
Pipelines
Water Hammer

Introduction
 What is Pressure Surge in Pipelines
 Changes in discharge in pipelines, caused by
valve or pump operation, either closing or
opening.
 Result in acceleration (or deceleration) of flow.
 Unsteady (or ‘surge’) conditions often continue for
only a very short period.

1
Water Hammer
 Pressure surge or wave caused by kinetic energy of
a fluid in motion, when forced to stop or change
direction suddenly.

 Example:
 If a valve is closed suddenly, water hammer wave
propagates in the pipe.

 If pressure is high, pipes may collapse or implode.

 Particularly of concern if pipe is on a downhill slope.

Water Hammer - Domestic

 In homes, water hammer problems often
occur when dishwashers, washing machines
or toilets shuts off water flow.

 May result in banging or hammering sound.

2
Instant Closure of the Valve
 A rapid valve closure cause non-uniform
deceleration along the pipeline.

 The alteration generates a shock wave.

 A shock wave is a zone in which the fluid is

rapidly compressed (i.e. and P increase).

Shock wave
 Travels through a fluid at the speed (or) celerity ‘c’.

 The speed (celerity) of the shock wave is dependent

on the compressibility of the liquid and the elasticity
of the pipeline.

 The simplest case of waterhammer, due to

instantaneous valve closure, can be used to
illustrate the phenomenon (fig 2).

3
 7

 Immediately following the valve closure, the fluid in proximity of

the valve is brought to rest. This causes a local pressure
increase. Due to the increased pressure, the liquid is
compressed and the pipe walls expand. This provides a little
extra volume in which the liquid enters to come to a stop. An
instant later, the next section immediately upstream follows the
same process. Thus, a wave of increased pressure propagate
upstream towards the reservoir.

 The process which started in Fig 2a completes in Fig 2c. When

the wave reaches the reservoir, the entire pipe is expanded,
the liquid column is compressed and comes to a complete
stop. The time taken to the wave to reach the reservoir is L/C.

4
 At the end of step 1, the pressure in the pipe is
much higher than the pressure in the reservoir.
Hence, the halted water begins to flow from the pipe
into the reservoir. This relieves the pressure in the
pipe. The liquid column decompresses and the pipe
materials contracts. The process starts at the
reservoir end of the pipe and continues towards the
valve. The time in this step is L/C, and the total time
taken by the wave to return to the valve is 2L/C,
which is known as the water-hammer period.

 The wave of backwater motion cannot go past the closed valve.

The sudden stoppage (inertia) of this moving fluid at the valve
causes the pressure to drop below normal level. This send a
negative pressure again upstream towards the reservoir.

 At the end of step 3, since the pressure at the reservoir end of

the pipe is negative (less than that in the reservoir), liquid starts
flowing from the reservoir into the pipe. At the end of this step,
the conditions are identical to the beginning of step 1 (i.e. one
cycle is completed in a period of 4L/C) and the whole sequence
repeated. In practice, friction eventually damps out the
oscillations.

10

5
Causes of Water Hammer
 2 types of water hammer that commonly
occurs:

 Caused by accumulation of condensate (water)

trapped in portion of horizontal steam piping

 Cavitations

11

 Main cause of water hammer: pressure wave

Components distributing
Pressure Wave

Change in Change in
Entrained Air
Flow Rate Temperature

12

6
1. Change in flow rate
 Rapid changes in flow rate require large
forces, seen as large pressure

 Thus, cause water hammer.

13

2. Entrained air
 Cause pressure in water lines; consequently
cause water hammer

 Air trapped will compress and exert extra

pressure on the water

14

7
3. Change of temperature
 Cause water to expand or contract due to
pressure.

 Max. pressure experienced in piping system

frequently results from vapor column
separation, which is caused by the formation
of void packets of vapor when pressure drops
so low that the liquid boils or vaporizes.

15

Other causes of water hammer

 Water hammer is also influenced by length, time and velocity of
water.

 Length of Pipes
 Small pipe may not be a problem if it is of a very short length.

 Time
 Instantaneous closure causes water hammer

 Velocity of Water
 The faster the water is traveling in pipe, water hammer will be
greater.

16

8
Inadequate System Design &
Installation
 Several factors exacerbate water hammer
and are generally traceable to inadequate
system design and installation.

17

 These include:
 Improperly sized supply lines for given peak water
flow-velocity

 Excessive system water pressure and lack of

pressure-reducing apparatus

 Inadequate/ poor strapping or securing of

plumbing to structure

18

9
  Excessively long straight runs with no bends

 Lack of expansion tank or other dampening

system to reduce or absorb shockwaves

 High water pressure with no pressure-reducing

valve

19

Solutions to Water Hammer

 Surge pressure must be incorporated with
operating pressure in the design of a pipe.

20

10
Assignment 3
 The surge pressure must be incorporated
with the operating pressure in the design of
the pipe.
 Discuss at least THREE (3) tools to reduce
the effects of water hammer in pipeline
design.

21

Thank you

22

11
References:

1. Chadwick, A., Morfett, J. & Borthwick, M. (2004), Hydraulics in Civil and

Environmental Engineering. 4th Edition. Taylor & Francis.
2. Chow V.T (1959), Open Channel Hydraulics. Mc Graw Hill.
3. Hwang, N.H.C. & Houghtalen, R.J. (1996) Fundamental of Hydraulic Engineering
Systems. 3rd Edition. Prentice Hall. NJ.
4. M. Hanif Chaudhry (2007) Open-Channel Flow. 2nd Edition. Springer.
5. R.E Featherstone & C. Nalluri (2001) Civil Engineering Hydraulics. 4th Edition.
Wiley.
6. Roberson, J.A, Cassidy J.J and Chaudhry M.H (1998), Hydraulic Engineering, John
Wiley & Sons, Inc
7. http://www.msmam.com

40
 KNS2153
HYDRAULICS

SERIES

Series 1 Flow in Pipes and Closed Conduits

Series 2 Pipes in Series, Parallel and Branched Pipe System
Series 3 Pump-Pipeline System Analyses

41
 KNS 2153 Hydraulics
Department of Civil Engineering
Faculty of Engineering
UNIMAS

Series 1

Flow in Pipes

1. Crude oil of density 925 kg/m3 and absolute viscosity of 0.065 Ns/m2 at 20 C is pumped
through a horizontal pipeline 100mm in diameter, at a rate of 10 l/s. Determine the head loss
in each kilometer of pipeline and the shear stress at the pipe wall.

2. Water from a large reservoir is discharged to atmosphere through a 100 mm pipe 450 m long.
The entry from the reservoir to the pipe is sharp and the outlet is 12m below the surface level
in the reservoir. Taking =0.01 in the Darcy-Weisbach formula, calculate the discharge.

3. A uniform pipeline, 500m long, 200mm in diameter and roughness size 0.03mm, conveys
water at 15 C between two reservoirs, the difference in water level which is maintained
constant at 50 m. Take v=1.13 x 10-6 m2/s.

Determine the steady discharge between the reservoirs using:

a) Colebrook-White equation
b) The Moody diagram
c) An explicit function of

5. A pipeline 20 km long delivers water from an impounding reservoir to a service reservoir, the
minimum difference in level between which is 100m. The pipe of uncoated cast iron (K s=0.3
mm) is 400mm in diameter. Local losses amount 10 V2/2g.

a) Calculate the minimum uncontrolled discharge to the service reservoir.

b) What additional head loss would need to be created by a valve to regulate the discharge to
160 l/s?

6. A discharge of 400 l/s is to be conveyed from a headwork at 1050 m above datum to a

treatment plant at 1000 m above datum. The pipeline length is 5km. Estimates the required
diameter, assuming Ks=0.03mm.

42
7. A pipeline connecting two reservoirs having a difference of level of 6m is 720m long, and
rises to a height of 3m above the upper reservoir at a distance of 240m from the entrance
before falling to the lower reservoir. If the pipe is 1.2m in diameter and =0.01, calculate:

a) The discharge
b) The pressure at the highest point of pipeline

8. A 1.5m diameter concrete pipe of length 1390m for which Ks=1.5mm conveys water at 12 C
between two reservoir at a rate of 6.0 m3/s. What must be the difference in water surface
elevation between the two reservoirs?

9. The known outflow from a branch of a distribution system is 30l/s. The pipe diameter is
150mm, length 500m and a roughness coefficient estimated at 0.06mm.

Calculate the head loss using:

a) Colebrook-White equation
b) Moody diagram
c) Barr formula

10. Booster pumps are installed at 2km interval on a horizontal sewage pipeline of diameter
200mm and effective roughness size, when new, of 0.06mm. Each pump was found to deliver
a head of 30m when the pipeline was new. At the end of one year the discharge was found to
have decreased by 10 percent due to pipe deposits while the head at the pumps increased to
32m. Considering only friction losses, determine the discharge when the pipeline was in new
condition and the effective roughness size after one year.

43
 KNS 2153 Hydraulics
Department of Civil Engineering
Faculty of Engineering
UNIMAS

Series 2

Pipes in Series, Parallel and Branched Pipe System

1. Water is discharged from a reservoir into the atmosphere through a pipe 39m long. The
diameter is 50mm for 15m from the entrance. The pipe then enlarges suddenly to 75mm in
diameter for the remainder of its length. Taking into account the loss of head at the entry and
the enlargement, calculate the difference of level between the surface of the reservoir and the
pipe exit which will maintain a flow of 2.8 l/s. Take 1=0.0232 for the 50mm pipe and
2=0.0192 for the 75mm pipe.

d1, L1, 1

2. Two reservoirs are connected by three pipes laid in parallel, their diameters are respectively
d, 2d and 3d and they are all of the same length L. Assuming to be the same for all pipes,
what will be the discharge through the larger pipes if that through the smallest is 0.03 m3/s.
Ignore local losses.

3. A reservoir, surface level 60m above datum, supplies a junction box through a 300mm pipe,
1500m long. From the junction box two 300mm, each 1500m long, feed respectively into two
reservoirs whose surface levels are 30m and 15m above datum. Taking for all pipes being
0.04, what will be the quantity entering each reservoir. Ks = 0.015 mm.

KNS2153/RB 44
4. Reservoir A delivers to reservoir B through two uniform pipelines AJ:JB of diameters
300mm and 200mm respectively. Just upstream of the change in the section, which is
assumed gradual, a controlled discharge of 30 l/s is taken off.

Length of AJ = 3000m
JB = 4000m
Ks = 0.015mm (for both pipes)
H = 25m

Determine the discharge to B, neglecting the loss at J.

5. A 200 mm diameter pipeline, 5000 m long and of effective roughness 0.03mm delivers water
between reservoirs, with the minimum difference in water level between is 40m.

a) Determine the steady discharge between the reservoirs

b) If the discharge is to be increased to 50 l/s without increase in gross head, determine the
length of 200 mm pipeline of roughness Ks=0.015mm to be fitted in parallel. Consider only
friction losses.

KNS2153/RB 45
6. Neglecting minor losses in the pipes determine the flows in the pipes and the pressure heads
at the nodes.

Roughness size for all pipes: Ks=0.06mm

Pressure head elevation at A is 70 m.

Pipe AB BC CD DE EF AF BE
Length (m) 600 600 200 600 600 200 200
Diameter (mm) 250 150 100 150 150 200 100

Elevation of pipe nodes:

Node A B C D E F
Elevation (m) 30 25 20 20 22 25

7. Determine the discharges in the pipes of the network shown in figure P7, neglecting minor
losses.

Pipe Length (m) Diameter (mm)

AJ 10 000 450
BJ 2 000 350
CJ 3 000 300
DJ 3 000 250

Roughness size of all pipes Ks = 0.06mm

KNS2153/RB 46
8. Determine the flows in the network shown in Fig P8, neglecting minor losses.

Data:

Pipe AB BC BD BE EF EG
Length (m) 10 000 3 000 4 000 6 000 3 000 3 000
Diameter (mm) 450 250 250 350 250 200

Roughness of all pipes, Ks = 0.03 mm

9. In the network shown (Fig P9), a valve in BC is partially closed to produce a local head loss
of 10.0 VBC2/2g. Analyse the flow in the network.

Roughness of all pipes Ks=0.06 mm

Pipe AB BC CD DE BE EF AF
Length (m) 500 400 200 400 200 600 300
Diameter (mm) 250 150 100 150 150 200 250

KNS2153/RB 47
10. Determine the flows in the network illustrated in Fig P10, minor losses are given by C mV2/2g.

Pipe Length (m) Diameter (mm) Ks (mm) Cm

AB 20 000 500 0.3 20
BC 5 000 350 0.3 10
BD1 6 000 300 0.3 10
BD2 6 000 250 0.06 10

KNS2153/RB 48
 KNS 2153 Hydraulics
Department of Civil Engineering
Faculty of Engineering
UNIMAS

Series 3

Pump-Pipeline System Analyses and Design

1. Test on physical model pump indicated a cavitation number of 0.12. A homologous

(geometrically and dynamically similar) unit is to be installed where the atmospheric pressure
is 950 mb. And the vapour pressure head is 0.2 m. The pump will be situated above the
suction well, the suction pipe being 200mm in diameter of uPVC K s=0.03 mm, 10 m long, it
is vertical with 90° elbow leading into the pump inlet and is fitted with a foot-valve. The foot-
valve head loss (hv) = 4.5 V2/2g; bend loss hb = 1.0 Vs2/2g. The total head at the operating
discharge of 35 l/s is 25 m. Calculate the maximum permissible suction head and suction lift.

2. Calculate steady discharge of water between the tanks in the system shown in Figure P2, and
the power consumption. Pipe diameter Ds=Dd=200 mm; length 2000 m, Ks = 0.03mm. Losses
in valves, bends plus the velocity head amount to 6.2 V2/2g. Static lift = 10.0 m.

Pump characteristics:

Discharge (l/s) 0 10 20 30 40 50
Total head (m) 25 23.2 20.8 16.5 12.4 7.3
(%) - 45 65 71 65 45

Figure P2

KNS2153/RB 49
3. An existing pump, having the tabulated characteristics, is to be used to pump raw sewage to a
treatment plant through a static lift of 20m. A uPVC pipeline 10 km is to be used. Allowing
for minor losses totaling 10 V2/2g and taking an effective roughness of 0.15 mm because of
slimming, select a suitable commercially available pipe size to achieve a discharge of 60 l/s.
Calculate the power consumption.

Pump Characteristics:

Discharge (l/s) 0 10 20 30 40 50 60 70
Total head (m) 45 44.7 43.7 42.5 40.6 38.0 35 31
(%) - 35 50 57 60 60 53 40

4. Two identical pumps having the tabulated characteristics are to be installed in a pumping
station to deliver sewage to a settling tank through a 200mm uPVC pipeline 2.5 km long. The
static lift is 15m. Allowing for minor losses of 10.0 V 2/2g and assuming an effective
roughness of 0.15 mm, calculate the discharge and power consumption if the pumps were to
be connected:

a) in Parallel
b) in Series

Discharge (l/s) 0 10 20 30 40
Total head (m) 30 27.5 23.5 17.0 7.5
(%) - 44 58 50 18

5. A variable speed pump having the tabulated characteristics, at 1450 rev/min is installed in a
pumping station to handle variable inflows. Static lift = 15m, diameter of pipeline = 250 mm,
length 2000 m, Ks = 0.06 mm. Minor losses = 10.0 V2/2g.

Determine the total head of the pump and discharge in the pipeline at pump speeds of 1000
rev/min and 500 rev/min.

Pump characteristics at 1450 rev/min

Q (l/s) 0 10 20 30 40 50 60 70
H (m) 45.0 44.0 42.5 39.5 35.0 29.0 20.0 6.0

KNS2153/RB 50
6. Neglecting minor losses, determine the discharges in the pipes of the network illustrated in
Fig P6

a) with the pump at BC absent

b) with a pump which boosts the flow to C in operation, and calculate the power consumption.

Pump characteristics

Discharge (l/s) 0 20 40 60 80 100

Total head (m) 40 38.8 35.4 29.5 21.0 10.0
(%) - 50.0 70.0 73.0 58.0 22.0

Figure P6

7. A rotodynamic pump having the characteristic tabulated below delivers water from a river at
elevation 52.0 m to a reservoir with a water level of 85.0 m, through a 350 mm diameter
coated cast iron pipeline 2000m long (Ks=0.15 mm). Allowing 10 V2/2g for losses in valves,
etc.

Calculate the discharge in the pipeline and the power consumption.

Pump characteristics
Discharge (l/s) 0 50 100 150 200
Total head (m) 60 58 52 41 25
(%) - 44 65 64 48

If the discharge is to be increased to 175 l/s by the installation of a second identical pump;

a) determine the unregulated discharges produced by connecting the pumps

i. in parallel
ii. in series

b) calculate the power demand when the discharge is regulated (by valve control) to 175 l/s
in the case of
i. parallel operation
ii. series operation

KNS2153/RB 51
8. A pump is required to discharge 250 l/s against a calculated system head of 6.0 m. Assuming
that the pump will run at 960 rev/min, what type of pump would be most suitable?

9. The performance characteristics of a variable speed pump when running at 1450 rev/min are
tabulated below, together with the tabulated system head losses. The static lift is 8.0 m.
Determine the discharge in the pipeline when the pump runs at 1450, 1200 and 100 rev/min.

Pump characteristics

Q (l/s) 0 10 20 30 40
H (m) 20 19.2 17.0 13.7 8.7
System head loss (m) - 0.7 2.3 4.8 9.0

KNS2153/RB 52