c Indian Academy of Sciences

Sadhana Vol. 39, Part 5, October 2014, pp. 12271244. 

Impact of pipes networks simplification on water hammer

phenomenon
ALI A M GAD1 and HASSAN I MOHAMMED1,2,
1 Civil

Engineering Department, Assiut University, Assiut 71516, Egypt

and Environmental Engineering Department, Majmaah University,
Al Majmaah 11952, Saudi Arabia
e-mail: [email protected]; [email protected]
2 Civil

MS received 9 July 2013; revised 28 February 2014; accepted 14 March 2014

Abstract. Simplification of water supply networks is an indispensible design step
to make the original network easier to be analysed. The impact of networks simplification on water hammer phenomenon is investigated. This study uses two loops
network with different diameters, thicknesses, and roughness coefficients. The network is fed from a boundary head reservoir and loaded by either distributed or
concentrated boundary water demands. According to both hydraulic and hydraulic
plus water quality equivalence, three simplification levels are performed. The effect
of demands concentration on the transient flow is checked. The transient flow is initialized by either concentrated or distributed boundary demands which are suddenly
shut-off or released. WHAMO software is used for simulation. All scenarios showed
that both hydraulic equivalence and demands concentration simplifications increase
the transient pressure and flow rate. However, hydraulic plus water quality equivalence
simplification produces an adverse effect. Therefore, simplifications of the networks
should be done carefully. Also, it was found that pump shut-off gives the same trend
of valve shut-off or release.
Keywords.

Water hammer; pipes network; simplification; demands variations.

1. Introduction
The potable water distribution system is one of the most significant hydraulic engineering
accomplishments. Potable water can be delivered to water users through distribution systems.
However, variable water demands and water usage patterns can produce significant variations
of pressure in the distribution system, especially when the changes are sudden. Sudden changes

For correspondence

1227

1228

Ali A M Gad and Hassan I Mohammed

of water demands can create transient flow that could make so many undesirable consequences
such as backflow, negative pressure, or excessive high pressure. Therefore, it is important for
engineers to explore the various transient flow effects and to develop the emergency response
strategies in order to minimize the negative impacts (Kwon 2007). The total force acting within
a pipe is obtained by summing the steady-state and transient pressures in the line. The severity of transient pressures must be accurately determined so that water mains can be properly
designed to withstand these additional loads (Jung et al 2007). Many researchers studied the
water hammer phenomenon in the last decades with different viewpoints. For example, Abd
El-Gawad (1994), Ali et al (2010), Jnnsson (1999), Stephenson (2002), Yang (2001), and many
others.
Al-Khomairi (2005) discussed the use of the steady-state orifice equation for the computation
of unsteady leak rates from pipes through crack or rupture. It has been found that the orifice
equation gives a very good estimation of the unsteady leak rate history for normal leak openings. Fouzi & Ali (2001) studied water hammer in gravity piping systems due to sudden closure
of valves, using both the most effective numerical methods for discretizing and solving the
problem; the finite difference method using water hammer and mass oscillation (WHAMO) program and the method of characteristics with software AFT impulse. They showed that pressure
fluctuations vary dangerously especially in the case of pipes which has variable characteristics
(section changes with a divergence, a convergence or a bifurcation). Jung et al (2009) studied
the effect of pressure-sensitive demand on transient pressure. They concluded that a pressuresensitive demand formulation should be used for surge analysis to adequately evaluate both
system performance and the ultimate cost of system protection.
Mohamed (2003) introduced the effect of different parameters such as time of valve closure,
pipes material rigidity, and pipes roughness on the transient pressure damping. It was found that
the pipe friction factor and the closing time of the valve have a significant effect on the transient
pressure reduction and the elastic pipes such as PVC are better than rigid pipes in pressure
damping.
Ramos et al (2004) carried out several simulations and experimental tests in order to analyse
the dynamic response of single pipelines with different characteristics, such as pipes material,
diameters, thicknesses, lengths and transient conditions. They concluded that being the plastic
pipe with a future increasing application, the viscoelastic effect must be considered, either for
model calibration, leakage detection or in the prediction of operational conditions (e.g., start up
or trip-off electromechanical equipment, valve closure or opening).
Samani & Khayatzadeh (2002) employed the method of characteristics to analyse transient
flow in pipe networks. They applied various numerical tests to examine the accuracy of these
methods and found that the method in which the implicit finite difference was coupled with the
method of characteristics to obtain the discretized equations is the best when compared to others.
According to the aforementioned studies, water hammer in pipes networks has been studied
from different viewpoints. However, each water supply network has its own special characteristics which make it different from other networks. Also, due to the lack of field measurements
which are costly, it becomes important to use numerical models to gain an understanding about
the behaviour of networks under transient effects.
This study aims to investigate the effect of the hydraulic equivalence, hydraulic plus water
quality equivalence, and demands concentration simplifications of pipes networks on the
transient pressure head and flow rate induced from sudden demands shut-off or release.

Impact of pipes networks simplification on water hammer phenomenon

1229

2. Theoretical considerations
Because of the difficulty in solution of water hammer governing equations, engineers in pipelines
design may neglect this phenomenon. Recently a number of numerical methods which may be
used to solve these equations and suitable for digital computer analyses have been reported in
the literature (Chaudhry & Yevjevich 1981).
2.1 Governing equations for unsteady flow in pipelines
The governing equations for unsteady flow in pipeline are derived under the following assumptions including; (1) one-dimensional flow i.e., velocity and pressure are assumed constant at a
cross section; (2) the pipe is full and remains full during the transient; (3) no column separation occurs during the transient; (4) the pipe wall and fluid behave linearly elastically; and (5)
unsteady friction loss is approximated by steady-state losses.
The unsteady flow inside the pipeline is described in terms of unsteady mass balance (continuity) equation and unsteady momentum equation, which define the state of variables of
V (velocity) and P (pressure) given as Simpson & Wu (1997);

V
dA
+V
+
+
= 0,
t
x
x
A dt

(1)

V
1 P
f |V | V
V
+V
+
g sin +
= 0,
(2)
t
x
x
2D
where x = distance along the pipeline; t = time; V = velocity; P = hydraulic pressure in the
pipe; g = acceleration due to gravity; f = DarcyWeisbach friction factor; = fluid density;
D = pipe diameter; = pipe slope angle, and A = cross sectional area of the pipe.
Eq. (1) is the continuity equation and takes into account the compressibility of water and
the flexibility
of

 pipe material. Eq. (2) is the equation of motion. In Eq. (1), the terms

1/
+
V
are replaced by equivalent 1/ d/dt, where V = dx/dt, d
t
x
dt = /K dP /dt,
and
 Kis
 the bulk modulus of the fluid. Also, the fourth term in Eq. (1) can be expressed as
1 2 D/eE dP /dt, where is the poisons ratio of the pipe, e is the pipe wall thickness
and E is the Youngs modulus of elasticity of the pipe. Substitution by these abbreviations in Eq.
(1), it can be reduced to the following formula;



V
dP 1
1 2 D
+
+
= 0.
(3)
dt K
E
e
x
Wave speed can be defined as the time taken by the pressure wave generated by instantaneous
change in velocity to propagate from one point to another in a closed conduit. Wave speed (c)
can be expressed as;





1
1 2 D
1
1
Kc1 D
=
=
1+
,
(4)
+
K
E
e
K
Ee
c2


where: c1 = 1 v 2 . Substitution by Eq. (4) in (3) and dividing the result by yields;


H
H
c2 V
+
V +
= 0,
t
x
g x

(5)

1230

Ali A M Gad and Hassan I Mohammed

where H is the piezometric head, i.e., pressure head plus the elevation head. The term H /x is
small compared to H /t and it is often neglected. Thus, the simplified form of the continuity
equation in terms of discharge, Eq. (5) becomes
H
Q c2
+
= 0.
t
x gA

(6)

By the same way, the momentum equation, i.e., Eq. (2) can be simplified and written in terms of
discharge and piezometric head as follows
H
1 Q f Q |Q|
= 0.0.
+
+
x
gA t
2gDA2

(7)

2.2 Implicit finite difference solution method

The continuity and momentum equations form a pair of hyperbolic, partial differential for which
an exact solution cannot be obtained analytically. However, other methods have been developed
to solve water hammer equations. If the equations are hyperbolic, it means the solutions follow
certain characteristic pathways. For water hammer equations, the wave speed is the characteristic. The implicit finite difference method is a numerical method used for solving water hammer
equations. The implicit method replaces the partial derivatives with finite differences and provides a set of equations that can then be solved simultaneously. The computer program WHAMO
uses the implicit finite-difference technique but converts its equations to a linear form before it
solves the set of equations (Fitzgerald & Van Blaricum 1998).
The solution space is discretized into the x t plane, so that at any point on the grid (x, t)
there is a certain H and Q for that point, H (x, t) and Q(x, t) as shown in figure 1.
The momentum equation and the continuity equation can be represented in a short form by
introducing the following coefficients for the known values in a system;
j =

2tcj2
gAj xj

 (1 ) 


j Qnj Qnj+1 ,
j = Hjn+1 + Hjn +

Figure 1. The finite difference grid.

(8)

(9)

Impact of pipes networks simplification on water hammer phenomenon

1231

xj
,
(10)
2gAj t





xj fj  n

(1 )  n

n
Q
Hj Hjn+1 + j Qnj + Qnj+1
j =
Q
+
Q
Q

j
j
j +1
j +1
,

4gDj A2j
(11)
j =

where is a weighing factor included for numerical stability. All parameters for the coefficients
should be known from the properties of the pipe or the values of head and flow at the previous
time step. With the coefficients, the momentum and continuity equations of the j th segment of
the pipe become as given by Batterton (2006) as follows;


n+1
n+1
Q
+

+
Q
(12)
Momentum : Hjn+1 + Hjn+1
j
j
+1
j +1 = j ,


n+1
n+1
= j .
Continuity : Hjn+1 + Hjn+1
+1 + j Qj +1 Qj

(13)

Now, with equations for the all links and nodes in the system, the initial and boundary conditions,
a matrix of the linear system of equations can be set-up to solve for head and flow everywhere,
simultaneously, for the first time step. The process is repeated for the next time step, and again
for the next step until the specified end of the simulation.
3. Applications
The simple pipes network shown in figure 2 consisting of 11 joints (J1 J11) and 12 high density
polyethylene (HDPE) pipes (C1 C12) at the same elevation is representative of a general
parallel/series system. HDPE pipes with their common low Youngs modulus were preferred in
this application to avoid negative pressure waves to drop to the saturated vapor pressure of the
water which form a cavity in the fluid as the simulation program (WHAMO) does not allow for
the effect of cavitations. Other strong pipes networks with high Youngs modulus must employ
systems to help control increase and decrease in pressure due to water hammer. The Youngs
modulus for the HDPE material and water were taken as 0.80 and 2.20 GPa, respectively. Joint
J10 only has a boundary concentrated demand of 126 L/s and the network is fed by a reservoir

Figure 2. A Simple pipes network (the original network).

1232

Ali A M Gad and Hassan I Mohammed

with a boundary head of 59 m. Each pipe from C1 to C12 has a circular cross section. Table 1
gives lengths, diameters and DarcyWeisbach friction factors (f ) for all pipes in the network.
The thicknesses of the pipes walls were taken according to their diameters to suit for a working
pressure of 10 bars.

4. Simplification methods
4.1 The hydraulic equivalence simplification method
Using conservation of energy across a set of pipes in parallel or series, equivalent pipes relationships can be derived. Since these relationships are developed from conservation of energy,
the equivalent pipes have consistent flow and pressure losses as the original set of pipes. Typically, an equivalent diameter is determined by fixing the equivalent pipes length and roughness,
(Mohamed & Ahmed 2011). Eqs. (14) and (15) can be used for calculating the hydraulic
equivalent diameter for n pipes in series and in parallel, respectively.
 1
1
=
De5
Di5
n

i=1

fi
fe

Li
Le

(14)

and
De =

 n    0.5
 fe 0.5 Le
i=1

fi

Li

0.4
Di2.5

(15)

where fi , Di and Li are the DarcyWeisbach friction factor, diameter, and length of the pipe i in
series or parallel and fe , De and Le are the same parameters for the hydraulic equivalent system.

Table 1. Lengths, diameters and friction factors for all pipes of the original network.
Pipe ID
C1
C2
C3
C4
C5
C6
C7
C8
C9
C10
C11
C12

Length (m)

Diameter (mm)

DarcyWeisbach
friction factor (f )

305
305
305
305
305
305
305
305
305
305
305
305

300
300
300
200
150
200
150
250
200
150
300
250

0.026
0.026
0.024
0.020
0.018
0.020
0.018
0.022
0.020
0.018
0.024
0.022

Impact of pipes networks simplification on water hammer phenomenon

1233

By fixing two of the three parameters, the third can be determined using a form of the above
equations.
4.2 The hydraulic and water age equivalence simplification method
In general, water quality has an adverse relation with its age thus the travel time of water in pipes
could be used to indicate its quality. Raczynski et al (2008) developed the following equation
for computing the water age equivalent diameter, Dew .

n 

Di2 Li

i=1
Dew =

Le

 0.5

(16)

Dew ensures that the travel time in the equivalent pipe will equal that of the series or parallel pipes. However, it does not ensure that the system will be hydraulically equivalent. Since
Eq. (16) shows that Dew is independent on f , it is possible to find an equivalent hydraulic system
without affecting the travel time equivalence by modifying the pipe roughness. To do so, rather
than solving for De for a defined value of fe in hydraulic equivalence equations Eqs. (14) and
(15), De is set to Dew , and fe is solved for as an unknown term.
4.3 Demands concentration simplification method
Simplification is an important primary design step of water supply and irrigation pipes networks.
In most real networks, the demands leave through most parts of the pipe lines. However, an
indispensible design step of pipes networks is the demands concentration. The demands that
leave the network at distributed locations over its pipes lengths are replaced with equivalent
concentrated demands that reallocated at the joints of the simplified network.
4.4 Applied simplifications on the used network
The effect of the equivalence simplification methods are evaluated for the used pipes network
with different three levels of simplifications (aggregations orskeletonizations) and demands
concentration. The simplificationis loosely defined as the removing of pipes and nodes from
a network to make the model simpler. In this study, the aggregation simplifies the system
by replacing a series or parallel set of pipes with a single pipe. The first level of simplification (Level 1) as shown in figure 3a aggregates the two series pipes between nodes
(J4J9), (J3J10), and (J5J11) and removes joints J6, J7 and J8. Since there are no demands
at these nodes, no demands reallocation are required. The second level of simplification
(Level 2) is shown in figure 3b which aggregates the upper and lower series pipes between
nodes J3 and J10 to only one pipe on each upper and lower side. The third level of simplification(level 3) as shown in figure 3c replaces the three parallel pipes of level 2 with a single pipe.
Table 2 shows the calculated properties of the pipes of the simplified network for the three levels
of simplifications according to both hydraulic and hydraulic plus water quality equivalence. The
other simplification type is performed only on the original network, which includes concentrating distributed demands with a value of 14 L/s loaded on 9 nodes (J3J11) to be at the end node
(J10) with a total concentrated equivalent value of 126 L/s.

1234

Ali A M Gad and Hassan I Mohammed

Figure 3. The different three levels of simplifications for the used pipes network. (a) The first simplification level. (b) The second simplification level (c) The third simplification level.

5. Results and discussions

To show the effect of the pipes network simplification on water hammer phenomenon, three
scenarios of transient flows on simplified and original networks were simulated and compared.
The original network was simplified up to three levels according to both hydraulic and hydraulic
plus water quality equivalence, moreover the original network is loaded by either concentrated or
distributed water demands at their joints. The transient flow was initialized through linearly and
suddenly shut-off or release of concentrated or distributed water demands through a short period
of two seconds. WHAMO software which uses the implicit finite difference scheme for solving
the momentum and continuity equations at unsteady-state case was used in the simulation.
Three scenarios producing transient flows were performed for both simplified and original
networks. In the 1st scenario, a concentrated boundary demand at joint J10 is linearly decreased
from 126 to 0 L/s through two seconds period. In the 2nd scenario, a concentrated boundary

Impact of pipes networks simplification on water hammer phenomenon

1235

Table 2. Lengths, diameters and friction factors of pipes for the three simplification levels.
Pipes ID
Level 1

H. Equiv.

H. & Q. Equiv.

Level 2

Pipes ID
H. Equiv.

H. & Q. Equiv.

Level 3

Pipes ID
H. Equiv.

H. & Q. Equiv.

C1

C2

C3

C4, 5

C6, 7

C8

C9,10

C11

C12

L
D
f
L
D
f

305
300
0.026
305
300
0.026

305
300
0.026
305
300
0.026

305
300
0.024
305
300
0.024

610
164.4
0.018
610
176.8
0.026

610
164.4
0.018
610
176.8
0.026

305
250
0.022
305
250
0.022

610
164.4
0.018
610
176.8
0.026

305
300
0.024
305
300
0.024

305
250
0.022
305
250
0.022

C1
305
300
0.026
305
300
0.026

C2
305
300
0.026
305
300
0.026

C3, 4, 5, 11
1220
186.5
0.018
1220
246.2
0.072

C6,7
610
164.4
0.018
610
176.8
0.026

C8, 9, 10, 12

L
D
f
L
D
f
L
D
f
L
D
f

C1
305
300
0.026
305
300
0.026

C2
305
300
0.026
305
300
0.026

C 3 ~C12
610
251.8
0.018
610
504
0.536

183.7
0.018
1220
216.5
0.041

L = Pipe length (m), D = Pipe diameter (mm), and f = Friction factor.

-H. Equiv. = Hydraulic equivalence, H. & Q. Equiv. = Hydraulic plus quality equivalence

demand at joint J10 is linearly increased from 0 to 126 L/s through two seconds period. Both
the first and second scenarios are applied on the original network and three levels of simplified networks according to both hydraulic and hydraulic plus water quality equivalence. In the
3rd scenario, a concentrated boundary demand of 126 L/sat the endpoint (J10) was distributed
equally with a boundary value of 14 L/s on nine nodes of the original network (J3J11) and
both the concentrated and distributed demands were suddenly shut-off or released through a
short period of two seconds. The third scenario was applied only on the original network.
5.1 Effect of suddenly concentrated demand shut-off (1st scenario)
To examine the effect of network simplifications on water hammer phenomenon when a concentrated boundary demand is suddenly shut-off, the valve at node J10was assumed to be linearly
closed in a short period of 2 seconds. Before closing the valve, the flow in the network will be at
steady-state with the pressure head controlled by friction losses in the pipes, minor losses in the
fittings, and the type of the valve and its opened area. When the valve is closed instantaneously
the liquid next to the valve comes to a halt. The liquid is then compressed by the liquid upstream
which is still flowing. This compression causes a local increase in the pressure of the liquid.
The total pressure acting within the pipes equals the summing of the steady-state and the water
hammer induced pressure. The walls of the pipes around the fluid are stretched by the resulting excess pressure. A chain reaction then takes place along the lengths of the pipes with each
stationary element of fluid being compressed by the flowing fluid upstream. When the pressure
wave reaches the reservoir, the fluid in the pipes is now at rest and the pressure cannot exceed

1236

Ali A M Gad and Hassan I Mohammed

the boundary water depth in the reservoir, thus water starts to flow out of the pipes into the reservoir. An unloading pressure wave now travels back along the pipes towards the valve. When the
unloading wave reaches the valve, the water in the pipes is now flowing out of the pipe into the
reservoir, but at the closed valve the water must be at rest. This now causes a negative pressure
wave to travel back up the pipes towards the reservoir. When the pressure wave hits the reservoir the flow in the pipes will be at rest, but the pressure head is now below the reservoir level,
flow reverses in the pipes and another unloading wave travels back along the pipes towards the
valve. A cycle of pressure waves (positive unloading negative unloading) now travels up
and down the lengths of the pipes.
The pressure wave travels along the pipes network with a certain velocity, which is called
the celerity. The pressure waves celerity is affected by the modulus of elasticity of fluid and
pipes material, water density, pipes diameter, and pipes wall thickness. For instantaneous valve
closure the transient increase and decrease in water pressure due to water hammer depend mainly
on the celerity of the wave, water density, and water velocity in the pipes under the steady-state
conditions.
Figures 4 to 6 show the transient pressure head at node J10 due to its suddenly demand
shut-off for the different three levels of simplifications compared with that of the original network through a duration of simulation of 100 seconds just after the valve closure. As shown in
each figure, for both simplified and original networks the peak pressure values occur in the first

Figure 4. Transient pressure head for the simplified (level 1) and original Networks at node J10 due to its
sudden shut-off demand.
110
Original Network
Level 2 (Hydraulic)
Level 2 (Hydraulic + water age)

Pressure head (m)

100
90
80
70
60
50
40
30
20

10

20

30

40

50

60

70

80

90

100

Time (s)

Figure 5. Transient pressure head for the simplified (level 2) and original Networks at node J10 due to its
sudden shut-off demand.

Impact of pipes networks simplification on water hammer phenomenon

1237

Figure 6. Transient pressure head for the simplified (level 3) and original Networks at node J10 due to its
sudden shut-off demand.

cycle thus demonstrate the effect of friction on damping pressure waves. From the figures it is
noticeable that, the hydraulic equivalence simplification increases the peak values of the transient pressure head compared with those of the original network, however, the simplification
according to hydraulic plus water quality equivalence reduces the peak values.
In comparison between the figures, it is clear that as the level of simplification increases, the
transient pressure head of the simplified network deviates more from that of the original network.
The figures illustrate that, in the case of the hydraulic equivalence the frequency of the transient
pressure waves increases as the simplification level increases. However, the frequencies of the
transient pressure waves for the original and hydraulic plus water quality simplified networks
are the same. For the hydraulic plus water quality equivalence simplification as the water age
is a controlling parameter, the frequencies of the transient pressure waves for both original and
simplified networks should have the same trend. In all cases it is clear that, the simulation period
(100 seconds) is not sufficient to achieve the steady-state flow conditions in both simplified and
original networks.
Just after the valve is suddenly closed, a cycle of pressure waves (positive unloading
negative unloading) has been induced. Positive and negative waves start at node J10 while the
unloading waves start at the reservoir. At the time of the positive and second unloading pressure
waves the direction of the transient flow of the liquid remains towards the valve. Through the
first unloading and negative pressure waves the liquid reverses towards the reservoir (backflow).
Figures 7 to 9 show the transient flow rate at node J2, as an example, due to suddenly concentrated boundary demand shut-off at node J10 (linearly closed in a short period of 2 seconds)
through a duration of simulation of 100 seconds for the different levels of simplifications compared with that of the original network. As shown in the figures at time zero and before closing
the valve, the flow was at the steady-state with the boundary flow rate of 126 L/s. From each
figure, it can be seen that the hydraulic equivalence simplification increases the peak values of
the transient flow rate compared with those of the original network, however, the simplification
according to the hydraulic plus water quality equivalence reduces the peak values. In comparison
between these figures, it is noticeable that as the level of simplification increases the transient
flow rate of the simplified network deviates more from that of the original network.
As shown in figure 9, the third level of hydraulic equivalence simplification produces a transient flow rate with waves that have high frequency compared with those of hydraulic plus
water quality equivalence simplification, original network, and even others hydraulic equivalence
simplification levels.

1238

Ali A M Gad and Hassan I Mohammed

Figure 7. Transient flow rate at node J2 due to suddenly demand shut-off at node 10 for the simplified
(level 1) and original networks.

Figure 8. Transient flow rate at node J2 due to suddenly demand shut-off at node 10 for the simplified
(level 2) and original networks.

Figure 9. Transient flow rate at node J2 due to suddenly demand shut-off at node 10 for the simplified
(level 3) and original networks.

Impact of pipes networks simplification on water hammer phenomenon

1239

5.2 Effect of suddenly concentrated demand release (2nd scenario)

To show the effect of network simplifications on water hammer phenomenon when a concentrated boundary demand is released suddenly, the valve at node J10 was assumed to be linearly
opened in a short period of 2 seconds. When the valve at node J10 is completely closed, there
is no-flow in the network; consequently the pressure head through the network equals the water
boundary level in the reservoir (59 m). As the valve at node J10 is suddenly opened and the
demand is released from 0 to a boundary value of 126 L/s a negative pressure wave travels along
the pipes network from node J10 towards the reservoir. A cycle of pressure waves (negative
unloading -positive unloading) starts to travel up from node J10 towards the reservoir and down
from the reservoir to node J10 through the pipes in a successive manner.
Figures 10 to 12 illustrate the transient pressure head at node J10 after its suddenly demand
release for the three levels of simplifications compared with that of the original network. As
shown in each figure, the peaks of the transient pressure waves are gradually damped due to
the friction effect. From figures, it is clear that the transient pressure conditions are damped fast
within 50 seconds and the steady-state conditions prevail. From each figure, it is noticeable that
the hydraulic equivalence simplification increases the peak values of the transient pressure heads
compared with those of the original network, however, the simplification according to hydraulic
plus water quality equivalence reduces the peak values. In comparison between these figures, it
70
Original network
Level 1 (Hydraulic)
Level 1 (Hydraulic + water age)

pressure head (m)

60
50
40
30
20
10
0
-10

10

20

30

40

50

60

70

80

90

100

Time (s)

Figure 10. Transient pressure head at node J10 due to its suddenly demand Release for the simplified
(level 1) and original networks.
70
Original network
Level 2 (Hydraulic)
Level 2 (Hydraulic + water age)

Pressure head (m)

60
50
40
30
20
10
0
-10

10

20

30

40

50

60

70

80

90

100

Time (s)

Figure 11. Transient pressure head at node J10 due to its suddenly demand Release for the simplified
(level 2) and original networks.

1240

Ali A M Gad and Hassan I Mohammed

70
Original network
Level 3 (Hydraulic)
Level 3 (Hydraulic + water age)

Pressure head (m)

60
50
40
30
20
10
0
-10

10

20

30

40

50

60

70

80

90

100

Time (s)

Figure 12. Transient pressure head at node J10 due to its sudden demand release for the simplified
(level 3) and original networks.

is clear that as the level of simplification increases, the transient pressure head of the simplified
network deviates more from that of the original network.
Figure 12 illustrates that the frequency of the transient pressure waves for the third level of
the hydraulic equivalence simplification is higher than those of the hydraulic plus water quality
simplification, original network, and even others hydraulic simplification levels.
Figures 13 to 15 show the transient flow rate at node J2, as an example, after releasing the
demand at node J10 from zero to a boundary value of 126 L/s linearly in a short period of
two seconds for the different three levels of simplifications compared with that of the original
network. In comparison between these figures, it is noticeable that as the level of simplification
increases the deviation of the transient flow rate from the original case increases. Also, it can
be seen that the hydraulic equivalence simplification increases the peak values of the transient
flow rate compared with those of the original network, however, the simplification according to
hydraulic plus water quality equivalence reduces the peak values. Also, from the figures, it is
clear that the transient flow conditions are damped fast within 50 seconds and the steady-state
conditions prevail.
In general, figures 10 to 15 demonstrate that the transient flow rate at node J2 for the different
levels of simplifications has an inverse trend to that of the transient pressure head at node J10
with a short time lag which could be attributed to the locations difference between the concerned
nodes.
220
200

Flow rate (L/s)

180
160
140
120
100
80
60
Original network
Level 1 (Hydraulic)
Level 1 (Hydraulic + water age)

40
20
0

10

20

30

40

50

60

70

80

90

100

Time (s)

Figure 13. Transient flow rate at node J2 due to sudden demand releases at node J10 for the simplified
(level 1) and original networks.

Impact of pipes networks simplification on water hammer phenomenon

1241

Figure 14. Transient flow rate at node J2 due to sudden demand releases at node J10 for the simplified
(level 2) and original networks.

Figure 15. Transient flow rate at node J2 due to sudden demand releases at node J10 for the simplified
(level 3) and original networks.

5.3 Effect of water demands concentration (3rd scenario)

To show the effect of water demands concentration on water hammer phenomenon, a boundary
distributed demands loaded on nine nodes of the original network (J3J11) with a value of
14 L/s at each node were concentrated at the end node (J10) with an equivalent boundary value
of 126 L/s and both distributed and concentrated demands were suddenly and linearly shut-off
and released through a short period of 2 seconds. Figures 16 and 17 show the simulated transient
pressure head at node J10 and flow rate at node J2 for both distributed and concentrated demands
for two cases of suddenly shut-off and release, respectively. It is clear from these figures that
concentrating the demands produces bigger transient pressure head and flow rate compared with
the distributed one in case of demands shut-off as well as demands release. It is observed from
the figures that 100 seconds period after shutting-off the concentrated or distributed demands
is not sufficient to reach to the steady-state flow while it takes only around 50 seconds after
releasing the demands to reach to the steady-state conditions. Also, figures 16 and 17 illustrate
that the transient pressure head at node J10 and flow rate at node J2 for the original network
loaded by either distributed or concentrated demands have an inverse trend with a short time lag
which may be attributed to the locations difference between the two joints.
To confirm the obtained results, pump shut-off effect on transient flow at different simplification levels is shown in figure 18, where the transient head at node J2 was drawn versus time for

1242

Ali A M Gad and Hassan I Mohammed

Figure 16. Transient pressure head at node J10 and flow rate at node J2 due to concentrated and
distributed demands sudden shut-off.

Pressure head (concentrated demand)

Pressure head (distributed demands)
Flow rate (concentrated demand)
Flow rate (distributed demands)

Pressure head (m)

65
60

200
180
160

55

140

50

120

45

100

40

80

35

60

30

40

25

20

20

10

20

30

40

50

60

70

80

90

Flow rate (L/s)

70

0
100

Time (s)

Pressure head (m)

Figure 17. Transient pressure head at node J10 and flow rate at node J2 due to sudden release of
concentrated and distributed demands.

75
70
65
60
55
50
45
40
35
30
0

20

Original network

Level 1 (Hydraulics)

Level 2 (Hydraulics)

Level 3 (Hydraulics)

40

60

80

100

Time (s)
Figure 18. Transient pressure head at node J2 due to pump shut-off for different simplified levels.

Impact of pipes networks simplification on water hammer phenomenon

1243

original network and the three levels of simplification. It can be shown from this figure that the
maximum pressure increases by increasing simplification degree and also the wave frequency.
6. Conclusions
Simplification is an indispensible design step for water supply and irrigation pipes networks.
Three types of simplifications may be performed on distribution networks as; hydraulic equivalence, hydraulic plus water quality equivalence, and demands concentration. Variable water
demands and usage patterns in water distribution systems may create transient flow that could
make so many undesirable consequences. The effect of pipes networks simplification on the
transient flow must be accurately determined so that they can be properly designed. Three scenarios producing transient flow in both simplified and original networks were investigated in this
research. The transient flow was initialized by linearly and suddenly shutting-off or releasing of
either distributed or concentrated boundary demands in a short period of 2 seconds. WHAMO
software which uses the implicit finite difference scheme for solving the momentum and continuity equations at unsteady-state case was used in the simulation. The major findings of this
study can be summarized as follows.
(i) In all cases, for both simplified and original networks the peaks of the transient pressure
and flow rate occur in the first cycle thus demonstrate the effect of friction on damping the
transient flow.
(ii) Hydraulic equivalence and demands concentration simplifications increase the peak values
for the transient pressure and flow rate in the simplified network compared with the original
one. However, hydraulic plus water quality equivalence simplification results in an adverse
effect.
(iii) As the degree of simplification increases the transient pressure head and flow rate of the
simplified network deviate more from those of the original network.
(iv) In case of the hydraulic equivalence, the frequency of the transient waves increase as the
simplification level increases. However, the frequencies of the original and hydraulic plus
water quality simplified networks are found to be the same. This result is quite clear in
the 3rd level of hydraulic simplification which converts the pipes network from looped to a
single line.
(v) For the transient flow results from boundary demands shutting-off, the simulation period
(100 seconds) is not sufficient to achieve the steady-state flow conditions in both simplified
and original networks simulated in this study. However, it takes only around 50 seconds
after releasing the boundary demands to reach to the steady-state flow conditions.
References
Abd El-Gawad S M 1994 Water hammer analysis for the pipeline Ahmed Hamdi Tunnel, Abu-Radis.
Eng. Res. J. 6: 4054
Ali N A, Mohamed H I, El-Darder M E and Mohame A A 2010 Analysis of transient flow phenomenon in
pressurized pipes system and methods of protection. J. Eng. Sci. Assiut University 38(2): 323342
Al-Khomairi A M 2005 Use of the steady-state orifice equation in the computation of transient flow through
pipe leaks. The Arabian J. Sci. Eng. 30(IB): 3345
Batterton S 2006 Water hammer: An analysis of plumbing systems, intrusion, and pump operation. Thesis
submitted to the Faculty of The Virginia Polytechnic Institute and State Univ. in partial fulfillment of the
requirements for the degree of M. Sc. in Civil Eng., pp. 147

1244

Ali A M Gad and Hassan I Mohammed

Chaudhry H M and Yevjevich V 1981 Closed-Conduit flow. Water Resources Publications, Littleton, U.S.A
Fitzgerald R and Van Blaricum V L 1998 Water hammer and mass oscillation (WHAMO) 3.0 users manual
Fouzi A and Ali F 2001 Comparative study of the phenomenon of propagation of elastic waves in conduits.
Proceed. of The World Congress on Eng., Vol. III, London, U.K
Jnnsson L 1999 Hydraulic transient as a monitoring device. XXVII IAHR Congress, Graz, Austria
Jung B S, Boulos P F and Wood D J 2007 Pitfalls of water distribution model skeletonization for surge
analysis. J. AWWA 99(12): 8798
Jung B S, Boulos P F and Wood D J 2009 Effect of pressure-sensitive demand on surge analysis. J. AWWA
101(4): 100111
Kwon H J 2007 Computer simulations of transient flow in a real city water distribution system, KSCE.
J. Civil Eng. 11(1): 4349
Mohamed H I 2003 Parametric study for the water hammer phenomenon in pipelines. 1st Int. Conf. of Civil
Eng. Sci. ICCESI, Vol. 2, pp. 112, Assiut, Egypt
Mohamed H I and Ahmed S S 2011 Effect of simplifying the water supply pipe networks on water quality
simulation. Inter. Confer. for Water, Energy Environ., Sharijah, UAE, pp. 4146
Raczynski A, Kirkpatrick W, Rehnstrom D, Boulos P and Lansey K 2008 Developing hydraulic and water
quality equivalent systems. Proceed. of the 10th Annual Water Distr. Systems Conf., WDSA2008, Krugar
National Park, South Africa
Ramos H, Covas D, Borga A and Loureiro A 2004 A surge damping analysis in pipe systems: Modeling
and experiments. 42(4): 413425
Samani H M V and Khayatzadeh A 2002 Transient flow in pipe networks. J. Hyd. Res. 40(5): 637644
Simpson A R and Wu Z Y 1997 Computer modelling of hydraulic transient in pipe networks and associated design criteria. MODSIM97, International Congress on Modelling and Simulation, Modelling
and Simulation Society of Australia, Hobart, Tasmania, Australia
Stephenson D 2002 Simple guide for design of air vessels for water hammer protection of pumping lines.
J. Hyd. Eng. 128(8): 792797
Yang K 2001 Practical method to prevent liquid column separation. J. Hyd. Eng. 127(7): 620623