Water Supply

Simulation of Water Distribution Networks

The Use of EPANET

Mohammad N. Almasri

1 Optimal Design of Water Distribution Networks – Mohammad N. Almasri, PhD An-Najah National University
 Introduction

This p
presentation focuses on two issues:

ƒ The simulation of water distribution networks

(WDNs) using EPANET
ƒ The optimal design of water distribution networks

2 Optimal Design of Water Distribution Networks – Mohammad N. Almasri, PhD An-Najah National University
 Definitions
ƒ A WDN is comprised of a number of links
connected together to form loops or branches

ƒ These
h links
l k contain pumps, fittings,
f valves,
l etc..

ƒ Links: A link is a segment of the network that

has a constant flow and no branches. Each link
may contain one or more pipes (with
( h different
d ff
diameters) connected in series

ƒ Pipes: A pipe is a segment of a link that has a

constant flow,
fl constant diameter,
di andd no
branches
3 Optimal Design of Water Distribution Networks – Mohammad N. Almasri, PhD An-Najah National University
 Pipes and Links

D1 A pipe

D3 D2 D1 A link
L3 L2 L1
L

4 Optimal Design of Water Distribution Networks – Mohammad N. Almasri, PhD An-Najah National University
 Definitions

ƒ Nodes: They are the end points of the pipe

sections where
h two or more links
l k are joined.
d
Water can enter or leave the network at these
nodes

ƒ LLoops: The
Th loop
l is
i a closed
l d circuit
i i consists
i off a
series of links in which the demand nodes are
supplied from more than one pipe

ƒ Th
The path:
h It
I represents the
h way or the
h route
through which the demand nodes are reached
from the source nodes
5 Optimal Design of Water Distribution Networks – Mohammad N. Almasri, PhD An-Najah National University
 Example of a Water Distribution Network

6 Optimal Design of Water Distribution Networks – Mohammad N. Almasri, PhD An-Najah National University
 Main Principles of Network Analysis

Continuity:
y The algebraic
g sum of the flow rates
in the pipes meeting at a node together with any
external flows is zero

D
Q3 Q1 + Q2 = Q3 + D
Q1
D = Q1 + Q2 - Q3
Q2

A demand node
The node is connected to three supplying pipes
7 Optimal Design of Water Distribution Networks – Mohammad N. Almasri, PhD An-Najah National University
 Main Principles of Network Analysis

Energy
gy conservation: For all p
paths around closed
loops and between fixed grade nodes, the
accumulated energy
gy loss including
g minor losses
minus any energy gain or heads generated by
pumps
p p must be zero
hf2
A part of a looped network
Closed loop
Given total headloss for each link (pipe) as hf
hf1
hf3

Assume counterclockwise to be positive

hf4 -hf
hf1 – hf4 + hf3 + hf2 = 0

8 Optimal Design of Water Distribution Networks – Mohammad N. Almasri, PhD An-Najah National University
 Branched WDNs

Q2 [1] 1
ƒ In branched WDNs, there
is onlyy one p
path ((route))
Q3 2
[2] from the source node to
each node
Q5 Q4 3
[3]
5 [4]
4 ƒ How can we compute the
Q6 fl
flow in
i each
h link?
li k?
Q7 [5]
7 [6]
6

9 Optimal Design of Water Distribution Networks – Mohammad N. Almasri, PhD An-Najah National University
 Branched WDNs

What is the flow

1 in each link?
20 [1]

30 2
[2]

50 40 3
[[3]]
5 [4]
4
70 60 [5]
7 [6]
6

10 Optimal Design of Water Distribution Networks – Mohammad N. Almasri, PhD An-Najah National University
 Looped WDNs

ƒ More than one path to

each node

ƒ In looped network, how

can we compute the flow
in each link?

11 Optimal Design of Water Distribution Networks – Mohammad N. Almasri, PhD An-Najah National University
 Main Principles of Network Analysis

Link Head Loss (m)

1 6.753
2 13 238
13.238
3 4.504
4 18 728
18.728
5 3.737
6 4.998
7 9.994
8 9.992

12 Optimal Design of Water Distribution Networks – Mohammad N. Almasri, PhD An-Najah National University
 Energy Loss in Pipelines – Hazen Williams
Formula
1.852
⎛ Q ⎞
h f = 162.5⎜⎜ ⎟ D − 4.87
L
⎝ HW ⎠
C ƒ Cast Iron:
hf: head loss (m); ƒ New: 130
L: pipe length (m) ƒ 5 year old:
120
D: pipe diameter (in)
ƒ 10 year old:
Q: is flowrate in the pipe (m3/h) 110
CHW: Hazen Williams coefficient (-) ƒ Plastic: 150

13 Optimal Design of Water Distribution Networks – Mohammad N. Almasri, PhD An-Najah National University
 What is EPANET?
ƒ EPANET is a program for analyzing the hydraulic and
water
t quality
lit behavior
b h i off WDNs
WDN

ƒ D
Developed
l d by
b the
th US Environmental
E i t l Protection
P t ti
Agency

ƒ It is a public domain software that may be freely

copied and distributed

ƒ A complete Users Manual as well as full source code

and other updates can be downloaded from:
www.epa.gov/ORD/NRMRL/wswrd/epanet.html
p g / / / / p

14 Optimal Design of Water Distribution Networks – Mohammad N. Almasri, PhD An-Najah National University
 What Does EPANET Do?

ƒ Analyzing WDNs. This means mainly the

following:
ƒ Determination of the flow in each link
ƒ Determination of p
pressure head at each node

ƒ Additional outcome includes the simulation of

chlorine concentration in each link and at each
node

15 Optimal Design of Water Distribution Networks – Mohammad N. Almasri, PhD An-Najah National University
 Elements

ƒ Source reservoir
ƒ p
Pumps
ƒ Pipes
ƒ Nodes
ƒ Tanks
ƒ Valves

16 Optimal Design of Water Distribution Networks – Mohammad N. Almasri, PhD An-Najah National University
 EPANET Elements – Reservoirs

ƒ Reservoirs are nodes that represent an external source

or sink of water to the network. They are used to model
lakes, rivers, and groundwater aquifers. Reservoirs can
also serve as water quality source points

ƒ The primary input properties for a reservoir are its

hydraulic head and initial water quality

ƒ Because a reservoir is a boundary point to a network, its

head and water quality cannot be affected by what
happens within the network. Therefore it has no
computed output properties. However its head can be
made to vary with time by assigning a time pattern to it
17 Optimal Design of Water Distribution Networks – Mohammad N. Almasri, PhD An-Najah National University
 EPANET Elements – Tanks
ƒ Tanks are nodes with storage capacity, where the volume of
stored water can vary with time during a simulation

ƒ The primary input properties for tanks are:

ƒ Bottom elevation
ƒ Diameter (or shape if non-cylindrical)
ƒ Initial,
I iti l minimum
i i and d maximum
i water
t levels
l l
ƒ Initial water quality

ƒ The principal computed outputs are:

ƒ Total head (water surface elevation)
ƒ water quality

ƒ T
Tanks
k are required
i d to
t operate
t within
ithi their
th i minimum
i i and
d
maximum levels
18 Optimal Design of Water Distribution Networks – Mohammad N. Almasri, PhD An-Najah National University
 EPANET Elements – Pipes
ƒ Pipes convey water from one point in the network to
another

ƒ EPANET assumes that all pipes are full at all times

ƒ The principal hydraulic input parameters for pipes

are: Diameter, Length, Roughness coefficient, and
Initial status (open, closed, or contains a check
valve)

ƒ The water quality inputs for pipes consist of:

ƒ Bulk reaction coefficient
ƒ Wall reaction coefficient

19 Optimal Design of Water Distribution Networks – Mohammad N. Almasri, PhD An-Najah National University
 EPANET Elements – Pipes
ƒ Computed outputs for pipes include: Flow rate,
V l it Headloss,
Velocity, H dl Friction
F i ti factor,
f t Reaction
R ti rate,
t
and Water quality

ƒ The hydraulic head lost by water flowing in a pipe

d to friction
due f with
h the
h pipe walls
ll can be
b computed d
using three different formulas: Hazen-Williams
Fo m la Darcy-Weisbach
Formula, Da c Weisbach FoFormula,
m la and Chezy-
Che
Manning Formula

ƒ Minor losses caused by bends and fittings can also

b accounted
be t d for
f b by assigning
i i theth pipe
i a minor
i loss
l
coefficient
20 Optimal Design of Water Distribution Networks – Mohammad N. Almasri, PhD An-Najah National University
 EPANET Elements – Pumps

ƒ The principal input parameter for a pump is its

pump curve

ƒ Pumps can be turned on and off at preset times

ƒ Variable speed pumps can be considered

ƒ EPANET can also compute

p the energy
gy
consumption and cost of a pump

21 Optimal Design of Water Distribution Networks – Mohammad N. Almasri, PhD An-Najah National University
 EPANET Elements – Valves
ƒ Valves are used to control the pressure or flow at a specific
point in the network

ƒ The main different types

yp of valves considered in EPANET
include:

ƒ PRV (Pressure Reducing Valve): Used to limit pressure

ƒ PSV (Pressure Sustaining Valve): To maintain pressure at
a certain value
ƒ PBV (Pressure Breaker Valve): Forces a specified pressure
loss across the valve
ƒ FCV (Flow Control Valve): Used to control flow
ƒ GPV (General Purpose Valve): Can be used to represent a
link where the flow – headloss relationship
p is supplied
pp byy
the user

22 Optimal Design of Water Distribution Networks – Mohammad N. Almasri, PhD An-Najah National University
 Network Layout

23 Optimal Design of Water Distribution Networks – Mohammad N. Almasri, PhD An-Najah National University
 Time Pattern of Demands

ƒ Y
You can address
dd
the variability in
demands through
multipliers of the
“Base
Base Demand”
Demand at
each node

ƒ This is called in
EPANET Time
Pattern

24 Optimal Design of Water Distribution Networks – Mohammad N. Almasri, PhD An-Najah National University
 General Results of EPANET
Nodes

25 Optimal Design of Water Distribution Networks – Mohammad N. Almasri, PhD An-Najah National University
 General Results of EPANET
Nodes

26 Optimal Design of Water Distribution Networks – Mohammad N. Almasri, PhD An-Najah National University
 General Results of EPANET
Pipes

27 Optimal Design of Water Distribution Networks – Mohammad N. Almasri, PhD An-Najah National University
 General Results of EPANET
Pipes

28 Optimal Design of Water Distribution Networks – Mohammad N. Almasri, PhD An-Najah National University
 General Results of EPANET
Time series

ƒ Note the negative and positive values of flow

ƒ What does this indicate?
29 Optimal Design of Water Distribution Networks – Mohammad N. Almasri, PhD An-Najah National University
 General Results of EPANET
Time series

30 Optimal Design of Water Distribution Networks – Mohammad N. Almasri, PhD An-Najah National University
 General Results of EPANET
Time series (Pump)

31 Optimal Design of Water Distribution Networks – Mohammad N. Almasri, PhD An-Najah National University