ADDIS ABABA UNIVERSITY

FACULTY OF TECHNOLOGY
DEPARTMENT OF CIVIL ENGINEERING

CE 2602- Hydraulics II (3 Cr. Hr.) 2008/2009( 2001 E.C.) A.Y.

Course outline

1. Open channel flow

o Introduction
o Types of Open Channel
o Uniform flow
o Uniform flow equations
o channels of efficient cross sections
o specific energy
o flow in channel transitions
o The hydraulic jump

2. Dimensional Analysis and Similitude

o Dimensional homogeneity
o The Buckingham π theorem
o Similitude; Model Studies

3. Closed conduit flow

o Pipe-friction formula, laminar and turbulent flow

o Pipes in series, parallel and branching pipes
o Network of pipes
o Introduction to Water Hammer Analysis

4. Hydraulic Machinery

o Pumps
o Turbines

Text: Fluid Mechanics, V.L. Streeter

Hydraulics through Laboratory practice, Csongrady

References: Fluid Mechanics, Frank M. White

Civil Engineering hydraulics, Featherstone and Nalluri

Instructors: Ato Geremew Sahilu

1
CODE OF CONDUCT

The student is expected to attend at least 75% of class and laboratory hours. If
the student misses any required work due to absence, he/she will lose credit for that
work.

Home works and laboratory reports should be submitted on time. The student
should make an utmost effort to turn in a neat work. Late submission is
unacceptable.

Grading
The final total points of a student will be based on the results of laboratory reports,
Assignments, mid-semester exam and final exam weighed as follows:
Final Exam. 50%
Mid-semester exam 30%
Assignments & laboratory report 20%

N.B. the above distribution is by no means fixed. The instructors may take different
alternatives if necessary.

2
1 Open Channel Flow
1.1 Introduction

Open channel flow occurs where ever the flow proceeds with the liquid surface exposed to
constant pressure. In practice this pressure is the atmospheric pressure, and the flow proceeds
with free surface (exposed to the atmosphere). Thus open channel flow may occur regardless of
the type of conduit in which it is occurring i.e. an open channel flow may exist in a pipe, if it is
flowing partially full. In practice flow in sewers, canals, streams and gutters is exposed to
atmospheric pressure and hence is an example of open channel flow.

The longitudinal profile of the free surface in an open channel flow defines the hydraulic gradient
and determines the cross-sectional area of flow, as is shown in Figure 1.1. It also necessitates the
introduction of an extra variable, the stage, to define the position of the free surface at any point
in the channel.

In consequence, problems in open channel flow are more complex, and the solutions are more
varied, making the study of such problems both interesting and challenging.

Fig. 1.1 Flow in open channel

Flow classification

Recalling that flow may be steady or unsteady and uniform or non-uniform, the major
classifications applied to open channels are as follows:

Steady uniform flow, in which the depth is constant, both with time and distance. This constitutes
the fundamental type of flow in an open channel in which the gravity forces are in equilibrium
with the resistance forces.

Steady non-uniform flow, in which the depth varies with distance, but not with time. The flow
may be either (a) gradually varied or (b) rapidly varied.

Type (a) requires the joint application of energy and frictional resistance equations. Type (b)
requires the application of energy and momentum principles.

3
Unsteady flow, in which the depth varies with both time and distance (unsteady uniform flow is
very rare). This is the most complex flow type, requiring the solution of energy, momentum and
friction equations through time. The various flow types are all shown in Figure 1.2.

Flood wave (GVF)

Fig. 1.2 Types of flow

1.2 Types of Open Channel

Channels where flow occurs under free surface can either be natural, such as rivers and
streams, or artificial. Artificial channels comprise all man-made channels, including
irrigation and navigation canals, spillway channels, sewers, culverts and drainage ditches.
They are normally of regular cross-sectional shape and bed slope, and as such are termed
prismatic channels. Their construction materials are varied, but commonly used
materials include concrete, steel and earth. The surface roughness characteristics of these
materials are normally well defined within engineering tolerances. In consequence, the
application of hydraulic theories to flow in artificial channels will normally yield reasonably
accurate results. Various terms are used to refer to channels built under different
conditions.

Canal: a channel built on ground, i.e excavated to the desired shape and slope with or without
lining, usually having a mild slope. The lining could be made of concrete, stone masonry, cement,
wood or bituminous material.

A prismatic channel is characterized by unvarying cross section, constant bottom slope, and relatively
straight alignment.

4
Flume: a channel built (or supported) above the ground to convey fluid from one point to
another. In the field flumes are made of concrete, wood, sheet metal or masonry. Laboratory
flumes are usually made of wood, metal, glass or a composite of these materials.

Chute: is a channel of steep slopes. If the change in elevation in the direction of flow occurs in a
relatively short distance the channel is called a drop.

Culvert: is a relatively short and usually buried conduit that is commonly used for drainage
purposes, as in highways and embankments. Open channel prevails whenever the culvert is
flowing partially full.

In contrast, natural channels are normally very irregular in shape, and their materials are diverse.
The surface roughness of natural channels changes with time, distance and water surface
elevation. Therefore, it is more difficult to apply hydraulic theory to natural channels and obtain
satisfactory results. Many applications involve man-made alterations to natural channels (e.g.
river control structures and flood alleviation measures). Such applications require an
understanding not only of hydraulic theory, but also of the associated disciplines of sediment
transport, hydrology and river morphology.

Various geometric properties of natural and artificial channels need to be determined for
hydraulic purposes. In the case of artificial channels, these may all be expressed algebraically in
terms of the depth (y), as is shown in Table 1.1. This is not possible for natural channels, so
graphs or tables relating them to stage (h) must be used.

Figure 1.3 Definition sketch of geometric channel properties

Depth (y) - the vertical distance of the lowest point of a channel section from the free surface;

Stage (h) - the vertical distance of the free surface from an arbitrary datum;

Area (A) - the cross-sectional area of flow normal to the direction of flow;

Wetted perimeter (P) - the length of the wetted surface measured normal to the direction of flow;

Surface width (B) - the width of the channel section at the free surface;

Hydraulic radius (R) - the ratio of area to wetted perimeter (A / P);

5
Hydraulic mean depth (Dm) - the ratio of area to surface width (A / B).

Table 1.1 Geometric properties of some common prismatic channels

Velocity distribution in open channels

The measured velocity in an open channel will always vary across the channel section because of
friction along the boundary. Neither is this velocity distribution usually axisymmetric (as it is in
pipe flow) due to the existence of the free surface. It might be expected to find the maximum
velocity at the free surface where the shear force is zero but this is not the case. The maximum
velocity is usually found just below the surface. The explanation for this is the presence of
secondary currents which are circulating from the boundaries towards the section centre and
resistance at the air/water interface. These have been found in both laboratory measurements and
3d numerical simulation of turbulence.

The figure below shows some typical velocity distributions across some channel cross
sections. The number indicates percentage of maximum velocity.

Fig. 1.5 velocity

distribution in open channels
Determination of energy and momentum coefficients

6
To determine the values of α and β the velocity distribution must have been measured (or
be known in some way). In irregular channels where the flow may be divided into distinct
regions α may exceed 2 and should be included in the Bernoulli equation.

The figure below is a typical example of this situation. The channel may be of this shape
when a river is in flood - this is known as a compound channel.

Fig 1.6 compound channel with three regions of flow

If the channel is divided as shown into three regions and making the assumption that α =
1 for each then

u 
3
dA
V13 A1  V23 A2  V33 A3 Vi 3 Ai
   3
V 3A V 3 ( A1  A2  A3 ) V  Ai Where

V 
Q V1 A1  V2 A2  V3 A3
 
V Ai i

A A1  A2  A3  A i

Similarly,
u Vi 2 Ai
2
dA V12 A1  V22 A2  V32 A3
   
V 2A V 2 ( A1  A2  A3 ) V 2  Ai

Uniform flow and the Development of Friction formulae

When uniform flow occurs gravitational forces exactly balance the frictional resistance
forces which apply as a shear force along the boundary (channel bed and walls).

Fig. 1.7 forces on a channel length in

uniform flow
Considering the above diagram, the gravity force resolved in the direction of flow is
gravity force = ρgAL sin θ

and the boundary shear force resolved in the direction of flow is

7
shear force = τoPL , where P is the wetted perimeter

In uniform flow these balance, i.e. τoPL = ρgAL sin θ

Considering a channel of small slope, (as channel slopes for uniform and gradually varied
flow seldom exceed about 1 in 50) then
Sin θ ≈ tan θ = So
So
gAS o
o   gRS o
P

The Chezy equation

For a state of rough turbulent flow, which is the predominant type of flow in open
channel, it was experimentally verified that the shear force is proportional to the flow
velocity squared i.e.
τo α V2
τo = KV2
Substituting into the above equation
g
V  RS o
K
Or grouping the constants together as one equal to C
V C RS 0
This is the Chezy equation and the C the "Chezy C"
Table 1.2: Selected values of C.
Type of channel bed Mean value of C
Smooth cement 90
Well-laid brickwork 70
Cement concrete 70
Natural channel ( in good condition) 35
Natural channel ( in bad condition 25
The value of C can also be estimated using the Ganguillet and Kutter formula, which has
been developed based on measurements in open channels of various types.
0.00281 1.811
41.65  
C S n
 0.00281  n
1   41.65  
 S  R
where n is known as Kutter’s n. The above formula gives C in British units, which could
be converted into metric units (exercise)

The Manning equation

Many studies have been made on the evaluation of C for different natural and manmade
channels. Today most practicing engineers use some form of these relationships to give
C:

8
 R1 / 6
C
n

This is known as Manning's formula and the n as Manning's n.

Substituting Chezy’s equation in to the above formula gives velocity of uniform flow:

R 2/ 3So
V 
n

In terms of discharge,

1 A5 / 3
Q
n P 2 / 3 S o1 / 2

Note:
Several other names have been associated with the derivation of this formula - or ones
similar and consequently in some countries the same equation is named after one of these
people. Some of these names are; Strickler, Gauckler, Kutter, Gauguillet and Hagen.

The Manning's n is also numerically identical to the Kutter n.

The Manning equation has the great benefits that it is simple, accurate and due to it long
extensive practical use, there exists a wealth of publicly available values of n for a very
wide range of channels.

Table 1.3 a few typical values of Manning's n

Conveyance
Channel conveyance, K, is a measure of the carrying capacity of a channel. The K is
really an agglomeration of several terms in the Chezy or Manning's equation:
Q  AC RS o
Q  KS o1 / 2

A5 / 3
So K  ACR 1 / 2 
nP 2 / 3

9
Use of conveyance may be made when calculating discharge and stage in compound
channels and also calculating the energy and momentum coefficients in this situation.

Computations in uniform flow

We can use Manning's formula for discharge to calculate steady uniform flow. Two
calculations are usually performed to solve uniform flow problems.

1. Discharge from a given depth

2. Depth for a given discharge

In steady uniform flow the flow depth is know as normal depth.

As we have already mentioned, and by definition, uniform flow can only occur in
channels of constant cross-section (prismatic channels) so natural channel can be
excluded. However we will need to use Manning's equation for gradually varied flow in
natural channels - so application to natural/irregular channels will often be required.

Optimal Shape of Cross-Section

The most hydraulically-efficient shape of channel is the one which can pass the greatest
quantity of flow for any given area or, equivalently, the smallest area for a given quantity
of flow. From Manning’s formula and the corresponding expression for quantity of flow
we see that this occurs for the minimum hydraulic radius or, equivalently, for the
minimum wetted perimeter.

A semi-circle is the most hydraulically-efficient of all channel cross-sections. However,

hydraulic efficiency is not the only consideration and one must also consider, for
example, fabrication costs, excavation and, for loose granular linings, the maximum slope
of the sides.

The application of Energy Equation for rapidly varied flow

Rapid changes in stage and velocity occur whenever there is a sudden change in cross-
section, a very steep bed-slope or some obstruction in the channel. This type of flow is
termed rapidly varied flow. Typical examples are flow over sharp-crested weirs and
flow through regions of greatly changing cross-section (Venturi flumes and broad-crested
weirs).
Rapid change can also occur when there is a change from super-critical to sub-critical
flow (see later) in a channel reach at a hydraulic jump.

In these regions the surface is highly curved and the assumptions of hydro static pressure
distribution and parallel streamlines do not apply. However it is possibly to get good
approximate solutions to these situations yet still use the energy and momentum concepts
outlined earlier. The solutions will usually be sufficiently accurate for engineering
purposes.

10
The energy (Bernoulli) equation

The figure below shows a length of channel inclined at a slope of e and flowing with
uniform flow.

Fig 1.8 Channel in uniform flow

p V 2
Recalling the Bernoulli equation, g   z  cons tan t
2g

And assuming a hydrostatic pressure distribution we can write the pressure at a point on a
streamline, A say, in terms of the depth d (the depth measured from the water surface in a
direction normal to the bed) and the channel slope.

pA =ρgy2

In terms of the vertical distance

y2
d  y1 cos 
cos 
y 2  y1 cos 2 

So p A  gy1 cos 2 
So the pressure term in the above Bernoulli equation becomes
pA
 y1 cos 2 
g
As channel slope in open channel are very small (1:100 = θ = 0.57 o and cos2θ = 0.9999)
pA
so unless the channel is unusually steep  y1
g

11
 V 2
And the Bernoulli equation becomes y zH
2g

Flow over a raised hump - Application of the Bernoulli equation

Steady uniform flow is interrupted by a raised bed level as shown. If the upstream depth
and discharge are known we can use Bernoulli’s equation and the continuity equation to
give the velocity and depth of flow over the raised hump.

Applying Bernoulli’s equation

between sections 1 and 2
(assuming a horizontal
rectangular channel z1 = z2 and
taking α =1.0)
V12 V2 Apply
y1   y 2  2  z
2g 2g
Bernoullis’equation between
sections 1 and 2

Using continuity equation

V1A1 = V2A2 = Q
Q
V1 y1  V 2 y 2   q ,Where q
B
is the flow per unit width
Fig. 1.9 uniform flow interrupted by a raised hump

Substituting this into Bernoulli’s equation we have:

q2 q2
y1   y2   z
2 gy12 2 gy 22

Rearranging:
 q2 
2 gy 23  y 22  2 gz  2 gy1  2   q 2  0
 y1 


Thus we have a cubic function with the only unknown being the downstream depth, y2.
There are three solutions to this but only one is correct for this situation. We must find
out more about the flow before we can decide which it is.

Specific Energy

12
Specific energy, Es, is defined as the energy of the flow with reference to the channel bed
as the datum. The concept of
specific energy was first
introduced by Bakmenteff
(1918). With reference to the
figure below the total energy of
flow with respect to the channel
bottom is given by

p1 V12 Fig. 1.10 Specific energy (definition sketch)

E s1  ( y  y1 )  
g 2 g

V12
 y
2g

Thus the specific energy at an open channel section is equal to the sum of the flow depth and the
velocity head. In the above equation V1 denotes the velocity of flow at the point of interest, in the
figure above at point 1. In practice it is easier to use the average velocity of flow at the section
and speak about the specific energy of the flow at a section. How ever the velocity of flow
changes from point to point with in the flow and as a result the specific energy changes from
stream line to stream line. It is common to use the average velocity of flow with a correction
factor. The specific energy computed using the average velocity is taken to apply for all points in
the section, i.e. is taken as the specific energy of the section. For steady flow this can be written
in terms of discharge Q
 (Q / A) 2
Es  y 
2g

For a rectangular channel of width b, Q/A = q/y

q 2
Es  y 
2 gy 2
q 2
(E s  y) y 2   cons tan t
2g
cons tan t
(E s  y) 
y2

It can be observed that the specific energy is a function of depth of flow, y, only. If one
plots the depth of flow as ordinate against the specific energy for a constant Q, the energy
diagram is obtained, which is a very useful curve in open channel hydraulics.

Flow over a raised hump –

revisited: Application of the
Specific energy equation.

The specific energy equation may be

used to solve the raised hump

13
 problem. The figure below shows the hump and stage drawn alongside a graph of Specific energy
Es against y.

The Bernoulli equation was applied earlier to this problem and the equation from that example
may be written in terms of specify energy:
Es1 = Es2 +∆z
These points are marked on the figure. Point A on the curve corresponds to the specific energy at
section 1 in the channel, but Point B or Point B' on the graph may correspond to the specific
z energy at point 2 in the channel.

All point in the channel between point 1 and 2 must lie on the specific energy curve between
point A and B or B'. To reach point B' then this implies that E s1 - Es2 > ∆z which is not physically
possible. So point B on the curve corresponds to the specific energy and the flow depth at section
2.

Critical, Sub-critical and super critical flow

The specific energy change with depth was plotted above for a constant discharge Q, it is also
possible to plot a graph with the specific energy fixed and see how Q changes with depth. These
two forms are plotted side by side below.

From these graphs we can identify several important features of rapidly varied flow.

For a fixed discharge:

1. The specific energy is a minimum, Esc, at depth Yc, This depth is known as critical depth.
2. For all other values of Es there are two possible depths. These are called alternate depths. For
subcritical flow y > yc
supercritical flow y < yc

For a fixed Specific energy :

1. The discharge is a maximum at critical depth, Yc.
2. For all other discharges there are two possible depths of flow for a particular Es i.e. there
is a sub-critical depth and a super-critical depth with the same Es.
An equation for critical depth can be obtained by setting the differential of Es to zero:

14
  (Q / A) 2
Es  y 
2g
dE s Q 2 d  1  dA
 0  1  
dy 2 g dA  A 2  dy
Since δA=Bδy, in the limit dA/dy = B and
Q 2
0  1 B c 2 Ac3
2g
Q 2 Bc
1
gAc3

For a rectangular channel Q = qb, B = b and A = by, and taking α = 1 this equation
becomes
1/ 3
 q2 
yc   
 g 
  as Vc yc = q
Vc  gy c

Substituting this in to the specific energy equation

Vc2 y
E sc  y c   yc  c
2g 2
2
yc  E sc
3

The Froude number

The Froude number is defined for channels as:

V
FN 
gD m

Its physical significance is the ratio of inertial forces to gravitational forces squared
Inertial Force
FN2 
Gravitational Force

It can also be interpreted as the ratio of water velocity to wave velocity

Water Velocity
FN 
Wave Velocity

This is an extremely useful non-dimensional number in open-channel hydraulics.

ts value determines the regime of flow - sub, super or critical, and the direction in which
disturbances travel

15
Fr < 1 sub-critical
 water velocity > wave velocity
 upstream levels affected by downstream controls

Fr = 1 critical

Fr > 1 super-critical
 water velocity < wave velocity
 upstream levels not affected by downstream controls

The Hydraulic jump

The hydraulic jump is an

important feature in open
channel flow and is an
example of rapidly varied
flow. A hydraulic jump
occurs when a super-critical
flow and a sub-critical flow
meet. The jump is the
mechanism for the two
surfaces to join. They join
in an extremely turbulent manner which causes large energy losses.

Because of the large energy losses the energy or specific energy equation cannot be use in
analysis, the momentum equation is used instead.

Resultant force in x- direction = F1 - F2

Momentum change = M2 – M1

F1- F2 =M2 –M1

16
Or for a constant discharge

F1 +M1 =F2 +M2 = constant

For a rectangular channel this may be evaluated using

y1 y2
F1  g y1 b F2  g y2b
2 2
M 1  QV1 M 2  QV2

Q Q
 Q  Q
y1 b y2b
Substituting for these and rearranging gives
y
y 2  1  1  8 FN21  1
2  
or
y
y1  2  1  8 FN2 2  1
2  
So knowing the discharge and either one of the depths on the upstream or downstream side of the
jump the other - or conjugate depth - may be easily computed.

More manipulation with the above equation and the specific energy give the energy loss in the
jump as

 y 2  y1  3
E 
4 y1 y 2

These are useful results and which can be used in gradually varied flow calculations to determine
water surface profiles.

In summary, a hydraulic jump will only occur if the upstream flow is super-critical. The higher
the upstream Froude number the higher the jump and the greater the loss of energy in the jump.

17
2 Closed Conduit Flow
Flow in closed conduits (pipe, if conduit is circular in section, and duct otherwise) differs from
that of open channel flow in the mechanism that derives the flow. In the case of open channel
flow, flow occurs due to the action of gravity. In closed-conduit flow, however, although gravity
is important, the main driving force is the pressure gradient along the flow. The emphasis of this
section will be on pipes.
Flow in pipes is an example of internal flow, i.e., the flow is bounded by the walls, in contrast to
external flow where the flow is unbounded. For internal flows, the fluid enters the conduit at one
point and leaves at the other. At the entrance to the conduit there appears what is known as
entrance region with in which the viscous boundary layer grows and finally at the downstream
end of this region covers the entire cross section. The flow beyond the entrance region is said to
have fully developed. The fully developed flow is characterized by a constant velocity profile (for
a steady flow), a linear drop in pressure with distance, and a constant wall shear stress.

The entrance length is

a function of
Reynolds number and
is given by relations
below:
Le
 0 .06 Re
d
for laminar flow, and
Le
 4.4 Re 1 / 6
d
for turbulent flow.

Where
vd
Re 


Laminar flow in
pipes
Recall that flow can be classified into one of two types, laminar or turbulent flow (with a small
transitional region between these two). The non-dimensional number, the Reynolds number, Re,
is used to determine which type of flow occurs:
Laminar flow: Re < 2000
Transitional flow: 2000 < Re < 4000
Turbulent flow: Re > 4000

18
Derivation of basic equations of steady laminar flow in pipes
Consider a case of steady laminar flow in a circular pipe shown below:

Since the flow is steady

velocity distribution remains
the same through out the
length of the pipe. Hence
acceleration of the flow is
zero. Hence the sum of all
forces for the fluid element
shown should be zero.

 dp 
pA   p  s  A   * 2rs  W sin   0 but W  As and A  r 2
 ds 
dp
r 2 s sin   2rs  sr 2  0
ds
1  dp  z dz
     sin  r sin   and p is a function of s only
2  ds  s ds
1 d
   p  z  r
2 ds
dv
but for laminar flow    dy
Substituting this and simplifying one obtains the relationship for velocity as:
R2  r 2 d
V  ( z  p )
4  ds
Thus the velocity distribution in a circular pipe under laminar flow condition is parabolic, with
maximum value at the center.
R2 d
Vmax   ( z  p )
4  ds
For a horizontal pipe
R 2 dp
Vmax  
4  ds
The discharge through the pipe is obtained as
R
 R2  r2 d  d R 4
Q  
0
4 ds
( z  p ) 2r .dr  
 ds
( z  p )
8
The average velocity,

19
_
Q R2 d D2 d V
V   ( z  p )   ( p  z )  max
A 8  ds 32 ds 2
_
d dH H  H1 32  V
( p  z ) or  2 
ds ds L D2
_
H H 1  H 2 32  V
hf    L
  D 2
This is known as the Hagen –Poiseuille Formula for Laminar flow
This equation for head loss due to friction is commonly written as
64 L V 2
hf 
Re D 2 g
Turbulent Flow
In turbulent flow there is no longer an explicit relationship between mean stress and mean
velocity gradient u/r (because momentum is transferred more by the net effect of random
fluctuations than by viscous forces). Hence, to relate quantity of flow to head loss we require an
empirical relation connecting the wall shear stress and the average velocity in the pipe.
V 2
For turbulent flow, the boundary shear stress is taken as  o   and the derivation of the
2
equation for the friction head loss proceeds in the same way as in the case of laminar flow.
Consider a segment of an inclined circular pipe conveying a fluid of density ρ and viscosity µ,

1
P1A
θ L
2
Δz

γAL
τo
P2A

Sin θ = Δz/L
For steady uniform flow, since there is no acceleration, ΣF = m a=0
(P1 – P2)A + γAΔz – τoPL = 0 , where P is the wetted perimeter
Substituting ΔP = (P1- P2) and dividing the whole expression by A, one gets
ΔP+γΔz = τoL/R where R = A/P
Hence (ΔP +γΔz)/γL – ½ λ V2/gR
But (ΔP+γΔz)/γ = hf
hf V2
Thus  , for a pipe flowing full R= D/4
L 2 gR
L V2
hf  f ; where f = 4λ = friction factor for turbulent flow.
D 2g
The last equation for the friction loss in pipes is known as the Darcy-Weisbach equation. f is
called the Darcy coefficient. This equation also applies for laminar flow with a substitution of
64/Re for the friction factor. For a turbulent flow f is a function of the Reynolds number and the
relative wall roughness of the pipe for turbulent flow.

20
A graphical summary of past experimental results has been presented by moody. This chart,
known as the Moody diagram, is a plot of the friction factor as a function of Reynolds number
and the relative roughness of the pipe wall, i.e. ε/D where ε is the roughness in consistent units.
An empirical equation for the friction coefficient is also given by Colebrook and White,
1   2.51 
 2 log    , which applies in both smooth and rough turbulent zones.
f  3.7 D Re f 
Hazen-Williams Formula
The Hazen-Williams Formula has been developed specifically for use with water and has been
accepted as the formula used for pipe-flow problems in North America. It reads
V= 0.849CR0.63 s0.54
Where: V = average velocity of flow,(m/s)
R = hydraulic radius, m
S = slope of the energy gradient ( s = hL/L)
C = a roughness coefficient
This formula can be rearranged to give
1.852
hL  V 
 0.63  where R= D/4 for pipes
L  0.849CR 
Local Losses (Minor Losses)
In addition to head loss due to friction there are always head losses in pipe lines due to bends,
junctions, valves etc. Such losses are called Minor losses. For completeness of analysis these
should be taken into account. In practice, in long pipe lines of several kilometers their effect may
be negligible but for short pipeline the losses may be greater than those for friction.
Local losses are usually expressed in terms of the velocity head, i.e.
V2
hi  k i where ki is the minor loss coefficient
2g
Losses at Sudden Enlargement
Consider the flow in the sudden enlargement, shown in figure below, fluid flows from
section 1 to section 2. The velocity must reduce and so the pressure increases (as follows
from Bernoulli). At position 1' turbulent eddies occur which give rise to the local head
loss.

Apply the momentum equation between positions 1 and 2 to give:

P1A1 – P2 A2 = ρQ(V2 – V1)
Now use the continuity equation to remove Q. (i.e. substitute Q = A2V2)
P1A1 – P2 A2 = ρA2V2(V2 – V1)

P2  P1 V2
Rearranging gives  V1  V2 
g g

21
Now apply the Bernoulli equation from point 1 to 2, with the head loss term hL
P1 V12 P V2
  2  2  hL
g 2 g g 2 g
V12  V22 P2  P1
And rearranging gives hL  
2g g

V12  V22 V2 V1  V2 

Combining the two expressions hL  
2g g

hL 
V1  V2  2
2g
Substituting again for the continuity equation to get an expression involving the two
2
 A  V2
areas, (i.e. V2 = V1A1/A2) gives hL  1  1  1
 A2  2 g

2
 A 
This gives the expansion loss coefficient k e  1  1 
 A2 
When a pipe expands in to a large tank A1 << A2 i.e. A1/A2 ≈ 0 so ke = 1. That is, the head
loss is equal to the velocity head just before the expansion into the tank.

In other situations such as bends, junctions, sudden contractions, valves and fittings
determination analytical values for the loss coefficient is difficult. The loss coefficient is a
function of the type of obstruction in the flow and its values are given as in the subsequent figures
and tables.
The concept of equivalent pipe length
From the previous discussions it can be observed that all types of energy loss in pipes are
expressed as a coefficient times the velocity head. Hence if one is interested in the energy loss
alone, the minor losses can be expressed in terms of a friction loss over an equivalent length of
the pipe. Hence the equivalent length corresponding to a fitting having the minor loss coefficient
of k1 can be obtained from
Le V 2 V2
f  k1
D 2g 2g
k
 Le  1 D
f
In the same way, if a pipe system consists of a series of two pipes having diameters D 1 and D2,
and friction coefficients f1 and f2 , if the head loss is same in both pipe segments for the same Q,
then two pipes are said to be equivalent. Equivalently the length of, say the second pipe, that
produces the same total head loss as for the first pipe can be obtained from,
L V2 L V2
f1 1 1  f 2 2 2
D1 2 g D2 2 g
5
f  D2 
 L2  1   L1
f2  D1 

22
23
24
25
26
Multiple Pipe systems
In most practical pipe-flow problems the system constitutes multiple pipes joined in different
ways. Such complex systems can be one or a combination of the following types

i) pipes in series: here one pipe takes the fluid after

the other so that the same flow rate passes through
out the entire pipe system.

ii) pipes in parallel: in paraIle1 pipes two or more

pipes branch from a point (node) and rejoin some
distance downstream. Hence at the node the flow is
divided into the pipes whereas the pipes flow under
the same energy difference between the nodes.

iii) Branching pipes: such pipes branch off from the main
and may return to it. Typical example is pipes that
convey flow from multiple reservoirs.

iv) Pipe networks: such a system consists of pipes

interconnected in such a way that the flow makes a circuit.

Pipes in series
In such a system the same flow passes through all the pipes involved and hence the usual
problems are either:
To determine Q for a given head H, or
To determine the required head H to maintain a certain flow rate.
The latter problem is relatively simple as the friction coefficients for each pipe can easily be
computed.

For datum through B. the energy equation including the loss terms takes the form:

27
 PA V A2 PB V B2
ZA    ZB    h L ( A  B )
 2g  2g
H   hL ( A B )
V12 L1 V12 V12 L 2 V 212 V32 L3 V 32 V32
 hL( A B )  k en1  f1 k ex1  f2  k c2  f3  k et 3
2g Da 2 g 2g D2 2 g 2g D3 2 g 2g
Since the flow rate is the same through out the pipes, the above equation can be reduced to
16Q 2  k en1 L1 k ex1   L 2   k c 2 L3 k et 3 
H  4  f 1 5  4    f 2 5    4  f 3 5  4 
2 g 2  D1 D1 D1   D 2   D3 D3 D3 
To determine the flow rate, since Re is not known, assume values of the friction coefficient for
the pipes and compute the value of Q from the equation above. With this value of Q compute Re
and based on ε/D determine f for each pipe. This iterative procedure is repeated until the assumed
and computed values of the friction coefficient are closer to each other.
Pipes in parallel

In such arrangements the flow must satisfy:

i) Q = Ql + Q2 + Q3
ii) hf(A-B) = hfl = hf2 = hf3
The common types of problems and the recommended procedures are given below.
i) to determine the discharge Q for a given head difference between A and B. Since in such a case,
the head loss is known, one can write
L V2
h f 1  f1 1 1 and solve for V1 by trial
D1 2 g
Similar equations can be written for the other pipes that make up the system and solved for their
respective velocities. The total discharge is the sum of the product of the velocities and the cross-
sectional areas.
ii) the other problem is the determination of the distribution of the discharge among the pipes
involved given the total flowrate. A step by step procedure for such problems is:
o Assume a likely discharge in one of the pipes, say pipe 1, as Q l and compute the head
loss through the pipe.
o Using the computed value of the head loss in pipe 1, compute thee discharge through the
other pipes.
o Add the computed trial discharges in all the pipes and compare the sum with the given
total discharge. If the sum is not equal to the total discharge, correct the trial discharges
by adjusting them as given below
Q1' Q2' Q3'
Q1  ' Q Q2  ' Q Q3  ' Q
Q Q Q
o Compute the head losses again and check if they are equal.

28
Branching pipes
Such an arrangement of
pipes falls in neither of the
above two (i.e. parallel or
series) categories. The pipes
do not also from a network
of complete loops. A typical
example is the three-
reservoir problem shown in
the figure.

The problem is often to find

the flow rate (including the
direction) in each pipe. As
the elevation of the HGL at
the junction is not known, Datum
the flow can not be readily
computed. Hence the
procedure for solution starts by assuming a value for this head at the junction. The flow rate in
each pipe is then computed for the assumed head at the junction. The flow rates computed in such
a way are then checked if they satisfy continuity. If the sum of the discharges in the pipes is less
than zero (with flow away from the junction taken negative), then this is means the assumed head
is too high and it is reduced for the next trial. The procedure is repeated until the sum of the flow
rates is very close to zero.
Pipe networks
Pipes that are interconnected in such a way that they make loops (or circuits) form a network. In
such systems the flow in any of the pipes may come from different circuits and as such it is not
simple to know the direction of flow by observation. Pipe network problems involve the analysis
of existing systems, i.e. the determination of flow rate in each pipe, pressure at junctions (or
nodes), the head losses in the pipes and the selection of appropriate material and size.
The solution of network problems always uses iterative procedures that make use of the following
two facts:
o the flow into a junction must equal the flow out of the junction, i.e. at each node (and for
the entire system) continuity must be satisfied,
o the algebraic sum of the head losses around any circuit must add up to zero.

Below is outlined a method (commonly known as the Hardy-Cross method. after Prof.
Hardy Cross)
o by careful inspection of the network, assume a reasonable distribution of flow rate in the
pipes so that continuity is satisfied at each node,
o compute the head loss in each pipe. For this either the Darcy-Weisbach equation can be
used with the friction coefficient determined from the Moody diagram, or other methods
as discussed below. The Darcy- Weisbach equation can be reformulated as

29
 1  L  1  L 
hL  f   k V 2  f   k Q 2  rQ n
2g  D  2 gA 2
 D 
1  L 
where r  f   k ; n  2
2gA 2  D 
Industrial (commercial) pipe-friction formulas are also used in practice, which are generally of
the form:
LRQ n
hL  n
 rQ n
D
LR
where r m
D
R is a resistance coefficient, which, in the case of the Hazen-Williams formula is given as
R = l0.675/C,
n = 1.852, and m = 4.8704 and C depends upon the roughness and is given in the following table.
Pipe material and condition C

Extremely smooth, straight pipes; asbestos-cement 140

Very smooth pipes; concrete; new cast iron 130
Wood stave; new welded steel 120
Vitrified clay; new riveted steel 110
Cast iron after years of use 100
Riveted steel after years of use 95
Old pipes in bad condition 60 to 80
Thus compute ΣhL = ΣQn and check if the sum of the head losses is zero. If it is not the
case then adjust the initially assumed flow rate as given below.
o if the initially assumed discharge is Qo and the adjustment that should be made to
have sum of head losses zero is ΔQ, then head loss for the new discharge is given by
hL = rQn = r(Qo + ΔQ)n, which can be expanded to get
 n(n  1) n  2 
h L  r  Qon  nQon 1 Q  Qo Q 2  ....
 1* 2 
for a small ΔQ, higher order terms of this value can be neglected and the equation
approximated as
h L  rQon  r.nQon 1 Q ,hence
 h L   rQ on   r.n.Q on 1 Q  0
from which

Q 
 rQon
 r.n.Q on 1
Note: ΣQn is the algebraic sum of the head losses with due regard to signs whereas
Σr*n*Qn-1 is the arithmetic sum without any sign consideration.
o the procedure is repeated until the discharges in the pipes satisfy the two conditions
mentioned at the beginning, i.e. ΣrQn = 0 and continuity is satisfied at each node.
When these are satisfied t11e successive values of the corrective discharge, ΔQ
become very small.

30
4. Hydraulic machinery (Turbo Machinery)
Fluid machines either add energy in to a flow or extract from it. A machine that adds energy to a
flow is known as a pump. In common usage the term pump is used when the flow is liquid. For
gases the terms used are: fans and propellers (low pressure rise), and blowers and compressors (if
pressure rise involved is high). Machines that convert the energy of flow into some other form
(mechanical) of energy are called turbines.
The emphasis in this section shall be on pumps and turbines with the fluid being essentially water.
Types of pumps and their selection
Various types of pumps are used in practice. The choice usually depends on the flow rate, Q, the
head to be developed, H and the type of fluid. The most important characteristics of the pump are
the shape, the size and the speed.
1. Archimedean Screws
In this pump the mechanical energy of the device is converted in to an increase of the potential
energy by the continuous lifting of the fluid. There is no pressure added to the fluid. An
Archimedean screw consists of an inclined shaft to which one or more blades are helically
attached. The unit is closely fitted in a semi circular casing. With the rotation of the screw, the
fluid, enclosed by two successive blades, the shaft and casing, is lifted. The speed of this type of
this type of pump is relatively low (5-50 rpm). The most important features of the Archimedean
screw are:
 only suitable for practically constant heads
 only suitable to pump fluid from one free surface reservoir to another one
 physical size is large compared to the capacity.
 suitable for polluted fluids
 open, easily accessible construction.

Fig.4.1 Archimedean screw

2. Positive displacement pumps

In positive displacement pumps a moving boundary forces the fluid along by volume change. The
pumping proceeds in steps thereby resulting in a pulsating type of flow. The steps involve the

31
 admission of the liquid in to a cavity through an inlet (which usually is fitted with a non-return
valve). Then the liquid is squeezed so that it leaves the cavity the outlet, figure below.
PDPs have the advantage that they are suitable for any liquid (regardless of viscosity). They do
not, however, offer a continuous flow of the fluid.
A: 3-Lobe Pump B: Screw Pump C: Double Screw Pump
(Progressive Cavity)

Fig.4.2 positive displacement pumps

3. Dynamic (impeller) pumps

Such pumps deliver the fluid by adding momentum to it by means of fast moving blades or vanes.
If the motion of the blades is a rotational one the machines are known as rotodynamic pumps. The
most common types of such pumps are classified based on the direction of flow relative to the
axis of rotation. They can then be either radial exit flow (Centrifugal), or axial flow or mixed
flow (some where between axial and radial).
For the same power input and efficiency the centrifugal type would generate a relatively large
pressure head with a low discharge, the axial flow type a relatively large discharge at low head
with the mixed flow having characteristics somewhere in between the other two.

Dynamic pumps generally offer steadier and higher discharge than PDPs. The emphasis in this
section shall be on rotodynamic pumps. Impeller pumps are by far the most widely used type of
pumps in practice.

Fig. 4.3 Dynamic pumps

b. Axial flow pump

a. Centrifugal pump

c. Mixed flow pump

32
Pump- pipeline systems and basic definitions
A typical installation of a pump is shown in the figure below:

Fig.4.4 pump –pipe line

If one writes Bernoulli's equation between the pump inlet (suction side) and outlet (deliver side),
Ps V2 P V
 Zs  s  Hm  d  Zd  d
 2g  2g
P P
 Hm  d  s
 
where the differences in velocity and elevation between inlet and outlet are ignored.
By applying Bernoulli's equation between lower tank and inlet, and outlet and upper reservoir one
gets the realtion
Hm = Hst +hls + hld
Thus the pump has to deliver the static head plus the losses in the suction and delivery pipes.

Pump Characteristics
The essential hydraulic properties of a pump are laid down in the following basic pump
characteristics, refer to the figure below.
 the Q-H characteristics, which indicates the relation between the flow Q and the head H
at constant speed of the pump.
 the Q-η characteristics, which indicates the relation between the flow Q and the
hydraulic efficiency η. The hydraulic efficiency η is the relation between the hydraulic
energy absorbed by the fluid and the mechanical energy supplied by the motor via the
QH QH
draft shaft.   
T T (2n / 60)

33
 where n= speed of the shaft(1/min)
T = shaft torque (Nm)
QH
The power required for the drive is: Pd 
 d
where Pd = the power required for the drive ( watt)
ηd = efficiency of the drive
 the Q-P characteristics, which indicates the relation between the flow Q and the power P
that is supplied to the pump shaft.
 the net positive suction head (NPSH) characteristics, which indicates the relation between
the flow Q and the margin that is required between the energy level at the suction side of
the pump and the vapor pressure level of the pumped fluid, so that a certain amount of
cavitation is not exceeded.
Pumps, once installed may have to operate under different combinations of Q and H. Hence the
choice of pumps and drives always has to be based on the most unfavorable operating conditions,
both with regard to the required power, head and efficiency and with regard to NPSH (cavitation).
When a pump has to operate under a wide range of discharges a flat curve is to be preferred.
Pump characteristics of impeller pumps are generally determined by measurements at the pumps
concerned, measurements at a small scale model of the pump or by extrapolation of the
characteristics of an already existing pump that is identical in shape, with the help of hydraulic
scaling laws.
If pumps a and b are two geometrically similar pumps operating under identical flow conditions,
then the following similarity rules apply to them
3 5
Pb  b  nb   Db 
    
Pa  a  na   Da 
3
Qb nb  Db 
  
Qa n a  Da 
where Q flow rate (m3/s)
n speed
D diameter of impeller
P power
Index a: pump whose characteristics are available
Index b: pump that is identical in shape, the characteristics of which are derived from
existing pump.

Most pump applications involve a known head and discharge for the particular system, plus a
speed range dictated by electric motor speeds or cavitation requirements. The designer then
selects the best size and shape of the pump. Pumps exhibit varying efficiencies over a range of
discharges and operating conditions. One of the most important parameters that are used in the
selection of appropriate type of pump is specific speed. The specific speed (n s) is a dimensionless
ratio involving the rotational speed, the discharge (or power output for turbines) and the head the
N Q
pump delivers. It is given by the expression: n s 
 gH  3 / 4
where N is the rotational speed in rad/s. In engineering practice, however, the specific speed is
N Q
commonly defined as N s 
H 3/ 4
where N is the speed of the pump shaft in rev/min,
Q is the pump flow with the maximum efficiency (best efficiency point) in m 3/s and

34
H is the head with maximum efficiency in meters.
Such definition of specific speed does not result in a dimensionless constant; hence its unit
depends on the system of units used.
The specific speed is also used to classify pumps, the reason why the specific speed is also called
type number.
Low specific speed Ns ≤ 2600 (centrifugal pumps)
Medium specific speed 2600< N s ≤ 5000 (mixed flow pumps)
High specific speed 5000 < Ns < 10000 (axial flow pumps)
Note: the Ns is computed with Q in l/s and H in m.
A pump with low specific speed will generally render a relatively small flow with a relatively
high head; where as a pump with a high specific speed will generally render a large flow with a
relatively low head.

Suction and cavitation, the Net Positive Suction Head (NPSH)

Cavitation consists of local vaporization od a liquid due to a drop in its pressure below its vapour
pressure. In a pump installation if the pump is located above the liquid surface of the reservoir
from which it delivers then there could be a danger of cavitation. Cavitation causes physical
damage, reduction in discharge and makes undesired noise.
Thus for cavitation not to occur the suction pressure (P s) should not equal the vapour pressure
under the given conditions.
If the suction pressure head is expressed in absolute terms ( i.e. wrt absolute zero), then the Net
Positive Suction Head (NPSH) is defined as
P P P P
NPSH  s  v  a  H s  v
   
where Pa is the local atmospheric pressure,
Pv = vapor pressure of the liquid at the given temperature,
V2
Hs = the manometric suction head = hs  hls  s
2g
Vs = velocity in the suction pipe.
Values of NPSH are delivered from the manufacturer of the pumps and should not be exceeded

35
Fig 4.5 pump characteristic curves

36
Pump- pipe line systems design
For effective design of a pump - pipe line system, basic knowledge of the hydrodynamic
characteristics of pumps is required. Particularly the head curve (Q-H), the efficiency curve, (Q-
η) and the power curve, (Q-P).
In order to transport a fluid through a pipe line, energy may have to be added to the fluid to
overcome losses and to increase the potential energy of the fluid. The device that is most often
used to achieve this is the pump.
In order to reach the optimal operation, a good consideration of the following items is required:
a. Accurate calculations of the required operating point or points (Q, H).
b. Possibility of deviation of actual operating point from the calculated operating point
should be accurately assessed and catered for. This could happen as a result of bad
prediction of the hydraulic losses of the pipeline system.
c. Determine the minimum available suction pressure with in the system and the vapor
pressure of the pumped fluid, which will influence the choice of pump with regard to its
cavitation problems.
d. optimal pump arrangement according to the available capabilities
e. select the most suitable pump type according to the system requirements

Consider the system shown in fig 4.4. The total energy the pump has to supply is the sum of the
static lift (the elevation difference between the water surfaces in the two reservoirs) and the head
loss in the suction and delivery pipes (hls +hld). This can be expressed as
E = Hst + kQ2
Where the static lift is the elevation difference H st and the head losses can be expressed by kQ 2.
This equation (or its graphical form) is known as the system characteristics. It should be noted
that the term kQ2 includes all minor as well as major losses in the pipe system leading the water
from the source to the destination.
For rotodynamic pumps, the head generated is not constant but is a function of discharge, as
expressed by the Q-H curve. If such a pump is installed in a pipe system, then the two must
handle the same flow rate, i.e. the flow through the pump is the same as that through the
connected pipeline system. On the other hand the head generated by the pump must equal to the
system energy requirement at that flow rate. Hence, the solution of
f(Q) = ΔZ+kQ2
gives the actual discharge for the particular installation. The solution is usually obtained
graphically. The intersection point of the two curves is known as the operating point. Pump
matching usually means the process of selecting a pump to operate in conjunction with a given
system so that it delivers the required flow rate, operating at its best efficiency, which
corresponds to the pump’s design point.
The point on the system characteristic which corresponds to the required rate through the system
is known as the duty required. Thus, for correct matching the operating point should coincide
with the system duty required.

37
 80

P u m p H -Q C u rv e
70

60

50 P u m p O p e r a tin g P o in t
H ead, m

80
40
70
S y s te m H - Q C u rv e

E ff ic ie n c y , %
30 60

50
20 h st E ffic ie n c y 40
30
10

0
0 50 100 150 200 250 300 350
D is c h a r g e , m 3/h r
Fig. 4.6 pump operating point

Multiple pump systems

a. parallel connection
pumping stations usually contain several pumps connected in “parallel”. This arrangement allows
to deliver large range of discharges. In such a connection the head across all pumps is assumed to
be equal and the discharges in the individual pumps is added up.
Head, m

30.0 P1

25.0 P1
P2
20.0 P2
P1+P2 (Parallel)
15.0
10.0
5.0
0.0
0.0 100.0 200.0 300.0 400.0 500.0
Discharge, m 3/hr

fig.4.7 H-Q curves for two pumps connected in parallel

b. series connection
This connection is the basis for multi-stage pumping. The discharge from the first pump is
delivered to the inlet of the second pump and so on in order to lift the water over a high elevation
which would have otherwise been difficult to achieve by using a single pump. The same
discharge passes through each pump receiving a pressure boost in doing so. The resulting H-Q
curve will be obtained by adding individual pump heads for the given discharges.

38
Fig.4.7 Pumps operating in series connection

Turbines
Turbines extract energy from a flowing fluid. Hydraulic turbines extract energy of
flowing water and coupled wit generators convert it to electric power.
Hydraulic features
Classification based on head
Low head 2-15 m
Medium head 15- 50m
High head >50m
The total power driven from a turbine is given by the equation

P  QH

39
5 Introduction to water hammer analysis
The results of steady flow analysis discussed so far fail to deliver useful results when flow in a
pipe is unsteady. Such flow conditions may occur during closure and opening of flow control
valves. For instance, a gate or valve at the end of the penstock pipe (the high pressure pipe
bringing water to turbines) controls the discharge of the turbine. As soon as this automatically
operated gate opening is altered, the pipe flow has to be adjusted to the new magnitude of flow. In
doing so, there are rapid pressure oscillations in the pipe, often accompanied by hammering like
sound. Hence the phenomenon is called water hammer although it also occurs in other fluids.
Water hammer can be defined as the change in pressure above or below the normal working
pressure caused by the change in the velocity of flow. Depending on the rate and magnitude of the
velocity change, this can lead to high pressure fluctuations in such pipe lines as supply lines for
drinking water, discharge line for sewerage water, oil-transmission line, penstock pipeline, etc.
The phenomenon is described taking an example of a simple pipe fitted with a valve at its end, as
shown in the figure below:

Fig 5.1 Conditions for instantaneous pipe closure

If the valve is suddenly closed, the fluid immediately upstream of the valve is stopped but the
remaining fluid continues to move towards the valve as if nothing has happened. As the fluid in
the immediate neighborhood of the valve has been brought to rest, it is slightly compressed by the
fluid still flowing with its original velocity towards the valve. Not only will the fluid be
compressed but also the wall of the pipe surrounding this will get stretched by the excess pressure
produced. The next element of the fluid now finds an increased pressure with static condition in
front of it, therefore, it to come to rest, is itself compressed and expands the pipe wall slightly.
Each element of the fluid thus stops the element following it until all the fluid in the pipe has

40
been brought to rest, figure above. Thus the pressure gets propagate in the upstream direction at
the sonic wave speed. This speed is given by the following equation:
K
c

Where K= the bulk modulus and ρ is the mass density of the fluid, if elasticity of the material is
negligible.
If the elasticity of the pipe and that of the fluid are taken in to consideration a more accurate
expression for the pressure wave velocity (also known as water hammer wave celerity) is
Where K = bulk modulus of fluid (2100 N/mm2 for water)
E = elastic modulus of pipe material (210000 N/mm 2 for steel)
α = a coefficient which takes the values:
5/4 – ν for pipe supported at one end only
1 – ν2 for pipe with both ends fixed
1 for pipe with expansion joints, where ν is Poisson ratio
t and d are pipe thick ness and diameter respectively.
Hence at any instant after the closure of the valve before all the fluid has stopped there is a
discontinuity of conditions (such as pressures, velocity, etc) represented by line XX in the figure
above.
There are two approaches to the water hammer problem, namely the rigid eater column theory
and the elastic water column theory. In the rigid water column theory the compressibility effects
of water are neglected where as the latter one the same is taken in to consideration. The rigid
water column theory is easier top apply but is useful only for specific cases where the
compressibility of water can safely be neglected.
Thus the pressure wave takes a time of 2L/c to make a round trip between the valve and the inlet,
if the length of the pipe is L. In practice if the closure time is longer than this time, the closure is
said to be gradual and the rigid water column theory gives a fairly accurate result. This is because
for a gradual closure the resulting pressure is relatively low and hence the associated change in
volume of the flowing fluid is also small. If the time of closure is less than 2L/c, however,
compressibility effects cannot be neglected hence the rigid water column theory gives an
approximate result. In practice if the closure is rapid, i.e. time of closure is less than 2L/c, a more
accurate method is adopted which takes into consideration the compressibility of the fluid and
elastic deformation of the pipe.
When the pressure wave reaches the inlet at A, the entire fluid mass in pipe length L comes to rest
and subjected to an excess pressure throughout. There will be a transient hydraulic grade line
parallel to the original hydraulic grade line but at a distance Δp/γ, where Δp is the excess water
hammer pressure. When the pressure wave reaches the tank it gets relieved and drops to that
corresponding to the water surface in the tank. Hence an unbalanced pressure head is created,
since the water in the pipe is under excess pressure. This causes the water to flow back into the
tank. This flow will not last long, however, as it creates vacuum immediately upstream of the
valve and the flow direction is reversed. Hence the pressure just upstream of the valve swings
from positive to negative pressure, which is then propagated towards the inlet.

Rigid water column theory: It is assumed in this method of analysis that any increase in
pressure change is transmitted instantaneously throughout the fluid in the conduit, disregarding
the compressibility of the fluid and the elasticity of the conduit.
To derive the basic equation using the rigid water column theory, consider the following simple
case where pipe fitted with valve at one of its ends conveys liquid from a tank.

41
Considering the entire water mass in the pipe, which is the approach used in rigid water column
theory, the unbalanced head from the right is h. Here frictional effects are neglected, again one of
the limitations of this approach. Under the action of the unbalanced force due to the excess
pressure the motion of the liquid mass is retarded, hence one can write
From Newton’s second law,
dV
hA   AL
dt
From which
L dV
h
g dt
The minus sign is used because the velocity is decreasing with time.
This equation can be used for calculating the head rise associated wit slow deceleration of water
column. It may be used for calculating the water level variation in a surge shaft (tank) following
load rejection, starting up in pumping line, etc.

When the compressibility of the fluid and the pipe material is taken in to account (Semi-elastic
approach), if the pipe closure is uniform, and the entire mass of fluid between the inlet and the
valve reduces its velocity in time Δt, then
L V L
h But t 
g t c
c c
And h  V and if the valve closure is complete, h  V
g g
Valve closure cannot occur in reality with in zero time. Hence classification of closure time in to
rapid and slow is adopted. If the closure time is less than 2L/c, the closure is classified as rapid
and the above equation for maximum water hammer pressure can be used. If, however, the
closure time is greater than 2L/c then due to the reflected negative pressure wave the maximum
pressure is reduced. Here it is proposed to reduce the maximum pressure according to time
proportion, i.e.
h 2L

hmax c .t c

This equation gives the values of pressure at the closing point that is required. Instead the
pressure wave propagation and change in velocity are both needed. In addition practical flow
conditions are much more complex than that used in the derivation of the simple formula given
above. In such situations more accurate derivation of equations of motion is needed, which is
beyond the scope of the course.

42