Analysis of Water Supply

Distribution Networks

Dr. Mohsin Siddique

Assistant Professor

1
 Schematic
Diagram of
Water Supply
System

Note: Surface water requires

more advanced treatment
then for groundwater
Component of Water Supply System

(1). Source (2).Treatment plant

(2). Storage Tanks/Reservoirs (3).Water Transmission/distribution

3
Natural Raw Water Sources
The various natural sources of water can be classified into
two categories:

Surface sources, such as

Ponds and lakes;

Streams and rivers;

Storage reservoirs.

Sub-surface sources or underground sources, such as

Springs; and

Fresh groundwater

4
 Other Water Sources
Other non-traditional water sources include:

Ocean water

Desalination of sea water

Brackish underground sources

Desalination

Wastewater reuse

Treatment and re-use of wastewater

Rainwater harvesting (i.e., houses and domes roof, storm water)

Water trading !!

import/export of fresh water

Virtual water trade

Brackish water or briny water is water that has more salinity than fresh
water, but not as much as seawater
5
Water Resources in UAE

/Seasonal floods

6
Water Quality and Treatment

Water to be used in a public water supply is required to be fit for drinking.

This implies that it poses no danger to health, and it should be colourless,
clear, odourless, sparkling and pleasant to taste.

The raw or treated water is analyzed by testing their physical, chemical and
bacteriological characteristics:

Physical Characteristics:

Turbidity; Color; Taste and odor; and Temperature

Chemical Characteristics:

pH; Acidity; Alkalinity; Hardness; Chlorides; Sulphates; Iron;
Nitrate, and Dissolved solids.

Bacteriological Characteristics:

Bacterial examination i.e., pathogenic bacteria or non pathogenic bacteria
such as E.Coli,

7
 Distribution Reservoirs/Tanks

Reservoirs in water distribution systems plays an
important role to:

Provide service storage to meet widely fluctuating demands imposed
on water supply distribution systems

Accommodate fire-fighting and emergency requirements

Equalize operating pressures

Elevated water tank Surface water tanks Fire water tank water tanks in Kuwait
Jumaira-UAE
Water Transmission

Water transmission refers to the transportation of the water from the
source to the treatment plant and to the area of distribution.

It can be realized through

free-flow conduits, (gravity flow)

pressurized pipelines (pumping system) or

a combination of the two (combination of gravity flow and pump).

For small community water supplies through pressurized pipelined

are most common, since they are not very limited by the
topography of the area to be traversed.

Free-flow conduits (canals, aqueducts and tunnels) are preferred in
hilly areas or in areas where the required slope of the conduit more
or less coincides with the slope of the terrain.
9
Methods of Water Transmission/Distribution

(pumping system)

(2. Pumping system)

(1. Gravity flow)

10
(3. Gravity flow and pumping system)
Water Distribution System

A water distribution system consists of complex interconnected
pipes, service reservoirs and/or pumps, which deliver water from
the treatment plant to the consumer.

Water demand is highly variable, whereas supply is normally

constant. Thus, the distribution system must include storage
elements, and must be capable of flexible operation.
11
Layouts of Distribution Network

12
 Design of Water Supply Networks

General Consideration:

A municipal water distribution system includes a network of mains with storage
reservoirs, booster pumping stations (if needed),fire hydrants, and service lines. Arterial
mains, or feeders, are pipelines of larger size that are connected to the transmission lines
that supply the water for distribution. A major water demand areas in a city should be
served by a feeder loop; where possible the arterial mains should be laid in duplicate.
Parallel feeder mains are cross connected at intervals of one to two kilometers, with
valves to permit isolation of sections in case of a main break. Distribution lines tie to each
arterial loop, forming a complete gridiron “close loops” system that domestic, commercial
consumers and services fire hydrants. The gridiron system illustrated in the Figure 1 is the
best arrangement for distributing water. All of the arterials and secondary mains are
looped and interconnected, eliminating dead ends and permitting water circulation such
that a heavy discharge from one main allows drawing water from other pipes.

The dead-end system shown in Figure 2 is avoided in new construction and can often be
corrected in existing systems by proper looping. Trunk lines placed in the main streets
supply sub-mains, which are extended at right angles to serve individual streets without
interconnections. Consequently, if a pipe break occurs substantial portion of the
community may be without water. Under a some conditions, the water in dead-end lines
develops tastes and odors from stagnation. To prevent this, dead ends may require frequent
flushing where houses are widely separated.
Pipe Network Analysis

Pipe network analysis involves the determination of the pipe flow
rates and pressure heads at the outflows points of the network. The
flow rates and pressure heads must satisfy the continuity and energy
equations.

ANALYSIS METHODS

(1). Hardy-Cross Method (Looped Method)

(2). Nodal Method

(3). Newton-Raphson Method
(1). The Hardy Cross Method

The earliest systematic method of network analysis (Hardy-Cross
Method) is known as the head balance or closed loop method.

This method is applicable to system in which pipes form closed

loops. The outflows from the system are generally assumed to occur
at the nodes junction.

For a given pipe system with known outflows, the Hardy-Cross

method is an iterative procedure based on initially iterated flows in
the pipes.

At each junction these flows must satisfy the continuity criterion, i.e.
the algebraic sum of the flow rates in the pipe meeting at a junction,
together with any external flows is zero.
15
 (1).The Hardy Cross Method

The method is based on

Continuity Equation:

Inflow = outflow at nodes Qb
Qa = Qb + Qc Qa A B

Qc
C D

Energy Equation:

Summation of head loss in closed loop is zero

∑ hl (loop ) = 0 ⇒ ∑ K (Q + ∆Q ) = 0
n
(1).The Hardy Cross Method

The relationship between head loss and discharge must be maintained
for each pipe

Darcy-Weisbach Equation
n 8 fL
hl ( pipe) = KQ n=2 K =
gπ 2 D5

Exponential friction formula Hazen-Williams

n 10.67
hl ( pipe) = KQ n = 1.85 K = 1.85 4.87
C d

17
(1). Hardy Cross Method (Derivation)

Q = Qa + ∆Q

∑
n
K (Qa + ∆Q ) = 0

∑ ∑ ∑
n −1
KQa n + nK (∆Q )Qa n −1 + nK (∆Q )2 Qa n − 2 +... = 0
2

∑ KQa n + ∑ nK (∆Q )Qa n −1 = 0 ∆Q = −

∑ KQa n

∑ nKQ a
n −1

18
(1). Hardy Cross Method

Problem Description

Network of pipes forming one or more closed loops

Given

Demands @ network nodes (junctions)

d, L, pipe material, Temp, and P @ one node

Find

Discharge & flow direction for all pipes in network

Pressure @ all nodes & HGL

19
 (1). Hardy-Cross Method (Procedure)
1. Divide network into number of closed d) Calculate hf / Qa for each pipe and sum
loops. for loop Σhf/ Qa.
e) Calculate correction
2. For each loop: ∆Q = −Σ hf /(nΣhf/Qa).
a) Assume discharge Qa and direction for NOTE: For common members between 2
each pipe. Apply Continuity at each node, loops both corrections have to be made.
Total inflow = Total Outflow. Clockwise As loop 1 member, ∆Q = ∆Q 1 − ∆Q 2. As
positive. loop 2 member, ∆ Q = ∆ Q 2 − ∆ Q 1.

b) Calculate equivalent resistance K for each f) Apply correction to Q , Q =Q + ∆ Q.

a new a
pipe given L, d, pipe material and water
temperature.
g) Repeat steps (c) to (f) until ∆Q becomes
very small and Σhf=0 in step (c).
c) Calculate hf=K Qan for each pipe. Retain
sign from step (a) and compute sum for
loop Σ hf. h) Solve for pressure at each node using
energy conservation.

20
 Example: 5-1

Neglecting minor losses in the pipe,

determine the flows in the pipes and
the pressure heads at the nodes

(kinematic viscosity= 1.13x10-6m2/s)

21
Example: 5-1

Solution:

Identify loops

Loop 1 and Loop 2

Allocate estimated flows in
each pipe

Compute head loss Discharge Sign Convention

coefficient of each pipe
8 fL Moody’s Diagram CW=+ve
K= f or Colebrook Eq.
gπ 2 D 5
1  e/ D 2.51 
= −2 log + 
f  3.7 Re f 
 

Compute head loss in each CCW=-ve
pipe
22 hl = h = KQ 2
 1st Iteration
Assumed Discharged
e/D f hL hL/Q

Correction

Corrected Discharge
e/D Q=Q+∆Q
120+14.23=134.23
10+14.23=24.23
-60+14.23=-45.77
-100+14.23=-85.77

23
 1st Iteration
Assumed/corrected Discharged
f hL hL/Q

Correction

Corrected Discharge
Q=Q+∆Q

50-2.23=47.77
10-2.23=7.77
-20-2.23=-23.23
-24.23-2.23=-26.46

24
2nd Iteration

f hL hL/Q

134.23-1.92= 132.31
26.46-1.92= 24.54
-45.77-1.92= -47.69
-85.77-1.92= -87.69

25
2nd Iteration

Proceed to loop 2 again and continue iterating until ∑hl=0

70-30

70-25-6.59

26
Example:
Find the flows in the loop given the inflows and outflows.
The pipes are all 25 cm cast iron (e=0.26 mm).

0.32 m3/s A B 0.28 m3/s

100 m

0.10 m3/s C D 0.14 m3/s

200 m
Example:

Identify Loop

Assign a flow to each pipe link

Flow into each junction must equal flow out of the
junction
assumed

0.32 m3/s A B 0.28 m3/s

0.32
0.00 0.04
0.10 m3/s C D 0.14 m3/s
0.10
Example:

Calculate the head loss in each pipe
h f1 = 34.7 m
h f = KQ 2
h f 2 = 0.222m
 8 fL  f=0.02 for Re>200000 h f 3 = −3.39m
K =  5 2 
 gD π  h f 4 = −0.00m
4

∑h
i =1
fi = 31.53m

 8(0.02)(200)  k1,k3=339
k1 =   = 339 s 2

 (9.8)(0.25) 5 π 2  m5
k2,k4=169
 

0.32 m3/s A 1 B 0.28 m3/s

4 2
Sign convention
+CW
0.10 m3/s C 3 D 0.14 m3/s
Example:

The head loss around the loop isn’t zero

Need to change the flow around the loop

clockwise
the ___________ flow is too great (head loss is positive)

reduce the clockwise flow to reduce the head loss

Solution techniques


optimizes correction
Hardy Cross loop-balancing (___________ _________)

Use a numeric solver (Solver in Excel) to find a change in
flow that will give zero head loss around the loop

Use Network Analysis software (EPANET, WaterCad, etc.)
 Example:

Numerical Solver

Set up a spreadsheet as shown below.

the numbers in bold were entered, the other cells are calculations

initially ∆Q is 0

use “solver” to set the sum of the head loss to 0 by changing ∆Q

the column Q0+ ∆Q contains the correct flows

∆Q 0.000
pipe f L D k Q0 Q0+∆Q hf
P1 0.02 200 0.25 339 0.32 0.320 34.69
P2 0.02 100 0.25 169 0.04 0.040 0.27
P3 0.02 200 0.25 339 -0.1 -0.100 -3.39
P4 0.02 100 0.25 169 0 0.000 0.00
Sum Head Loss 31.575
 Example
Solution to Loop Problem

Q0+ ∆Q
0.218
−0.062
−0.202
−0.102

0.32 m3/s A 1 B 0.28 m3/s

4
0.218 2

0.102 0.062

0.202
0.10 m3/s C 3 D 0.14 m3/s
You must be able to make a hand calculation before opting for a
better solution using software with a GUI.
(2). Nodal Method

Fig shows a branched pipe system delivering water from impounding
reservoir A to the service reservoirs B, C and D. F is known direct out flow
from J.

Eq. (1)

33
(2). Nodal Method

Eq. (2)

Eq. (1) can be written as

Eq. (3)

Iteration of ZJ can be performed such that QIJ from Eq. (3)

satisfies Eq. (2).

If (∑QIJ-F)≠0 then a correction, ∆ZJ, is applied to ZJ such that

34
Example 5-4

(kinematic viscosity= 1.13x10-6m2/s)

f
35
Example 5-4

36 f
Example 5-4

37
Newton Raphson Method

Newton–Raphson method is a powerful numerical method for solving
systems of nonlinear equations.

Suppose that there are three nonlinear equations F1(Q1, Q2, Q3) = 0,
F2(Q1, Q2, Q3) = 0, and F3(Q1, Q2, Q3) = 0 to be solved for Q1, Q2,
and Q3.

Adopt a starting solution (Q1, Q2, Q3).

Also consider that (Q1 + ∆Q1, Q2,+ ∆Q2, Q3 + ∆Q3) is the solution
of the set of equations. That is

Eq. (1)

38
 Newton Raphson Method

Eq. (2)

Eq. (3)

Solving Eq (3)., we get

39
Eq. (4)
Newton Raphson Method

Eq. (5)

40
Procedure

The overall procedure for looped network analysis by the Newton–
Raphson method can be summarized in the following steps:

Step 1: Number all the nodes, pipe links, and loops.

Step 2: Write nodal discharge equations as

where Qjn is the discharge in nth pipe at node j, qj is nodal

withdrawal, and jn is the total number of pipes at node j.

Step 3: Write loop head-loss equations as

where kn is total pipes in kth loop.

41
Procedure

Step 4: Assume initial pipe discharges Q1, Q2, and Q3., . . . satisfying
continuity equations.
Step 5: Determine friction factors, fi, in all pipe links and compute

corresponding Ki using

Step 6: Find values of partial derivatives ∂Fn / ∂Qi and functions Fn,
using the initial pipe discharges Qi and Ki.

Step 7: Find ∆Qi. The equations generated are of the form Ax = b,
which can be solved for ∆Qi.

Step 8: Using the obtained ∆Qi values, the pipe discharges are
modified and the process is repeated again until the calculated ∆Qi
values are very small.

42
Example

The pipe network of two loops as shown in Fig. has to be analyzed by the
Newton Raphson method for pipe flows for given pipe lengths L and pipe
diameters D.

The nodal inflow at node 1 and nodal outflow at node 3 are shown in the
figure. Assume a constant friction factor f = 0.02.

43
Single looped network
Example

Solution: Write nodal and loop equations
Q1

Q2
Q4

Q3

Node1: Q1+Q4-0.6=0
hL ( pipe) = KQ n

Node 2: Q1+Q2=0
8 fL

Node 3: Q2+Q3-0.6=0 n=2 K
gπ 2 D 5

Node 4: !!!

and loop Eq.

K1Q12+ K2Q22+ K3Q32+ K4Q42=0
44
Example

Assume initial pipe discharges based on continuity euqtion:
Q1 = 0.5 m3/s
= 0.5
In = +
Out = -ve
0.5
0.1

Apply signs to flowrate in eqns 0.1

Nodal discharge functions, F, are

Q1+Q4-0.6=F1 Apply signs (+ or -)

- Q1+Q2=F2

Q2+Q3-0.6=F3

Loop head-loss function

K1Q12+ K2Q22 - K3Q32 - K4Q42=F4
45
Example

The final nodal discharge functions, F, become

and loop head-loss function is

F4=6528Q12+ 4352Q22 - 6528Q32 - 4352Q42=0

Take derivatives of equations:

46
Example

Assembled equation in the following matrix form:

Substitute the derivatives

47
 Example

Now, calculate values of F1 to F4 using previously assumed flows i.e.,

Q1 = 0.5 m3/s

F4=6528Q12+ 4352Q22 - 6528Q32 - 4352Q42=2611.2

48
Example

Substitute values of F1 to F4 and Q1 to Q4 in the matrix form :

Solve the matrix, we get following flowrate corrections:

49
Example

Using discharge corrections, the revised pipe flowrates become:

The process is repeated with the new pipe flowrates. Revised values
of F and derivative ∂F/∂Q values are obtained. Substituting the
revised values, the following new solution matrix is generated:

50
Example

As the right-hand side is operated upon null vector, all the
discharge corrections ∆Q = 0.

Thus, the final flowrates are

51
 Problem

Write solution matrix for analysis using Newton Raphson Method

Node Eqns.=No. of nodes-1

Loop Eqns. =3

52
Problem: Solution

Nodal and loops equation in their general form are given below

Discharge_in=+ve

Discharge_out=-ve

CW=+ve

CCW=-ve

53
 Thank You

54