Overview

of
BRANCH and EPANET

Tools for design and analysis of piped water
supply networks


CTARA
2014
September


 Rural piped water supply schemes
Good design is criAcal
What is good design?
Village/habitaAon(s) evenly and adequately covered
Minimum pressure head of 5-8 m at demand nodes
Minimized piping and pumping energy costs
Feasible operaAon (e.g. pumping hours chosen based
on load shedding paQern)
BRANCH and EPANET are tools to help reach a
good (enough) design
 BRANCH input/output
For ElevaAon DESIGN output:
every
node Demand Lowest pipe diameters
Cost per pipe and total
Pipe Length piping cost
For every
Pipe roughness
pipe BRANCH
ExisAng/Planned

SIMULATION output:
Network layout
Source HGL Flow in each pipe
Pressure at each node
Min. residual pressure
For the Head-loss in each pipe
network Diameter opAons
Units (SI or Imperial)
Head loss formula
 EPANET input/output
For ElevaAon
every
node Demand

For Pipe length

every SIMULATION output:
pipe Pipe roughness
EPANET Pressure at each node
Network layout Velocity in the pipe
Head-loss in each pipe
Source details and
scheduling
Extended >me
For the Details of pumps and simula>on
network valves
Pipe roughness
Units (SI or Imperial)
Head loss formula
 BRANCH vs. EPANET
Note: Network layout required for both.

In general
Use BRANCH for design: it opAmizes pipe
diameters (but only if the network is branched
and gravity-fed)
Use EPANET for simulaAon if the system has
pumps, valves, loops, and Ame-variaAons in
demand or supply
 Example network layout
Junc>on Elev - 105 m
1 Pipe # 1 Pipe # 2 4 Demand Node
2
L 1000 m L 300 m Elev - 107 m
Village Pop. 1200
ESR
Elev -- 110 Pipe # 3
HGL 120 m L 400 m
3
Demand Node Elev - 103 m
Village Pop. 600
 Demand calculaAon
Rural supply norm: 55 lpcd
Assume the source is an ESR which will
supply the full days water in 6 hours

Demand (lps) = pop.*55 lpcd/(6 hr * 3600 s/hr)
Node 3 => 600 * 55/(6*3600) = 1.5 lps
Node 4 => 1200 * 55/(6*3600) = 3.0 lps

 BRANCH input - 1
Pipe data

Node data
 BRANCH input - 2
Source HGL Hydraulic grade
line
(e.g. eleva>on + staging height
of ESR = 110+10 = 120)

Pipe diameter op>ons and cost

per meter
(taken from Maharashtra Jeevan
Pradhikarans Schedule of Rates,
accessible online)
 HGL vs. Total pipe cost
and Pipe 1 diameter
1000
900 140/180 mm
800
700
600
Total Pipe Cost 125 mm
( x Rs. 1000 )
500
400
300 110 mm 110 mm
200
100
0
112 114 116 118 120 122 124 126
HGL of source (m)

Increasing the source HGL can (some>mes) reduce total piping cost
ElevaAon vs Pressure head at Node 3
14

12

10
Residual 8
Pressure head
at Node 3 (m) 6
4

0
101 102 103 104 105 106 107
Eleva>on at Node 3 (m)

If the minimum pressure head is not being met at a point, one solu>on
would be to move that node to a lower eleva>on point
 EPANET
What does EPANET do?

Public domain sohware for simulaAon of water distribuAon networks
EPANET analyses the ow of water in each pipe, the pressure at each node, the height of
water in a network.

Advantages:

1. Extended period hydraulic analysis for any system size.
2. SimulaAon of varying water demand, constant or variable speed pumps, and the minor
head losses for bends and kngs.
3. EPANET can compute the energy consumpAon and cost of a pump.
4. Can model various valve types - pressure regulaAng, and ow control valves
5. Provides a good visual depicAon of the hydraulic network
6. Data can but imported in several ways the network can be drawn and data can be
imported from Google Earth.
7. Water quality-SimulaAon of chlorine concentraAon in each pipe and at each node.
 EPANET slide 1 ( set up)

Pipe 1 Node 2 Pipe 2

Node 4
Source
Reservoir
Pipe 3

Node 3
EPANET Output le- Nodes

EPANET Output le- Pipes

 Extended period analysis
EPANET Time PaQern: To make our network more realisAc for analyzing an extended period
of operaAon we will create a Time PaQern that makes demands at the nodes vary in a
periodic way over the course of a day.
The variability in demands can be addressed through mulApliers of the Base Demand at
each node.
Nodal demands, reservoir heads, pump schedules can all have Ame paQerns associated with
them.
As an example of how Ame paQerns work consider a juncAon node with an average demand
of 3 lps. Assume that the Ame paQern interval has been set to 4 hours and a paQern with
the following mulApliers has been specied for demand at this node-
Time Period 1 2 3 4 5 6
Mul>plier 0.5 0.8 1.0 1.2 0.9 0.7

Then during the simulaAon the actual demand exerted at this node will be as follows:

Hours 0-4 4-8 8-12 12-16 16-20 20-24

Demand 1.5 2.4 3.0 3.6 2.7 2.1
 Manual Contents
InstallaAon
GIS
BRANCH
EPANET
References and exercise