MIKE URBAN

Water Distribution

MIKE 2019
2 MIKE URBAN - © DHI
PLEASE NOTE

COPYRIGHT This document refers to proprietary computer software which is pro-

tected by copyright. All rights are reserved. Copying or other repro-
duction of this manual or the related programs is prohibited without
prior written consent of DHI. For details please refer to your 'DHI
Software Licence Agreement'.

LIMITED LIABILITY The liability of DHI is limited as specified in your DHI Software
Licence Agreement:

In no event shall DHI or its representatives (agents and suppliers)

be liable for any damages whatsoever including, without limitation,
special, indirect, incidental or consequential damages or damages
for loss of business profits or savings, business interruption, loss of
business information or other pecuniary loss arising in connection
with the Agreement, e.g. out of Licensee's use of or the inability to
use the Software, even if DHI has been advised of the possibility of
such damages.

This limitation shall apply to claims of personal injury to the extent

permitted by law. Some jurisdictions do not allow the exclusion or
limitation of liability for consequential, special, indirect, incidental
damages and, accordingly, some portions of these limitations may
not apply.

Notwithstanding the above, DHI's total liability (whether in contract,

tort, including negligence, or otherwise) under or in connection with
the Agreement shall in aggregate during the term not exceed the
lesser of EUR 10.000 or the fees paid by Licensee under the Agree-
ment during the 12 months' period previous to the event giving rise
to a claim.

Licensee acknowledge that the liability limitations and exclusions

set out in the Agreement reflect the allocation of risk negotiated and
agreed by the parties and that DHI would not enter into the Agree-
ment without these limitations and exclusions on its liability. These
limitations and exclusions will apply notwithstanding any failure of
essential purpose of any limited remedy.

3
4 MIKE URBAN - © DHI
1 EPANET Input Description . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
1.1 Network Component Editors . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
1.1.1 Junction Editor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
Emitter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
Air-Valve . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
Graphical Placement and Editing of Junction Nodes . . . . . . . . . . 17
1.1.2 Pipe Editor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
Warning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
Flow Direction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
Graphical Placement and Editing of Pipes . . . . . . . . . . . . . . . 23
Snap Tolerance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
Defining Curved Pipes . . . . . . . . . . . . . . . . . . . . . . . . . 24
1.1.3 Pump Editor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
Pump Energy Settings . . . . . . . . . . . . . . . . . . . . . . . . . 29
Pump Water Hammer Settings (optional) . . . . . . . . . . . . . . . . 30
Pump Variable Speed Settings (optional) . . . . . . . . . . . . . . . 31
Variable Pumps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
Flow Direction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
Graphical Placement and Editing of Pumps . . . . . . . . . . . . . . 33
1.1.4 Valve Editor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
Valve Water Hammer Settings (optional) . . . . . . . . . . . . . . . . 37
Variable Valves . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
Check Valves . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
Flow Direction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
Graphical Placement and Editing of Valves . . . . . . . . . . . . . . 38
1.1.5 Tank Editor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
Graphical Placement and Editing of Tanks . . . . . . . . . . . . . . . 44
1.1.6 Air-Chamber Editor . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
Graphical Placement and Editing of Air-Chambers . . . . . . . . . . . 47
1.2 Other Editors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
1.2.1 Project Options Editor . . . . . . . . . . . . . . . . . . . . . . . . . 48
1.2.2 Pressure Zone Editor . . . . . . . . . . . . . . . . . . . . . . . . . . 52
1.2.3 Energy Editor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
1.2.4 Report Editor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
1.2.5 Curves and Relations Editor . . . . . . . . . . . . . . . . . . . . . . 56
1.2.6 Boundary Conditions Editor . . . . . . . . . . . . . . . . . . . . . . . 58
1.3 Network Demand . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
1.3.1 Distributed Demands . . . . . . . . . . . . . . . . . . . . . . . . . . 59
1.3.2 Demand Allocation Editor . . . . . . . . . . . . . . . . . . . . . . . . 64
Geocoding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
Aggregation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
1.3.3 Multiple Demand Editor . . . . . . . . . . . . . . . . . . . . . . . . . 69
1.3.4 Demand Statistics . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
1.4 Extended Period Simulations . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
1.4.1 Simple Control Editor . . . . . . . . . . . . . . . . . . . . . . . . . . 74
Multiple Controls . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
Pipes with Check Valves . . . . . . . . . . . . . . . . . . . . . . . . 77
Available Operational Controls . . . . . . . . . . . . . . . . . . . . . 77

5
 Time Based Controls . . . . . . . . . . . . . . . . . . . . . . . . . . .77
Storage Tank Water Level Based Controls . . . . . . . . . . . . . . .77
Junction Node Pressure Based Controls . . . . . . . . . . . . . . . . .77
1.4.2 Rule Based Controls . . . . . . . . . . . . . . . . . . . . . . . . . . 78
Multiple Pumps, Valves (WD-Tools) . . . . . . . . . . . . . . . . . . .81
1.4.3 Repetitive Profiles (Pattern) Editor . . . . . . . . . . . . . . . . . . . 83
Diurnal Patterns . . . . . . . . . . . . . . . . . . . . . . . . . . . . .83
Cyclic Profiles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .86
1.4.4 Time Settings Editor . . . . . . . . . . . . . . . . . . . . . . . . . . 87
Time Units . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .88
Diurnal Curves and Patterns . . . . . . . . . . . . . . . . . . . . . . .88
1.5 Water Quality Simulations . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
1.5.1 Water Quality Analysis Selection . . . . . . . . . . . . . . . . . . . . 89
1.5.2 Water Quality Analysis Parameters . . . . . . . . . . . . . . . . . . 90
1.5.3 Initial Water Quality Editor . . . . . . . . . . . . . . . . . . . . . . . 91
1.5.4 Point Constituent Source Editor . . . . . . . . . . . . . . . . . . . . 93
Point Constituent Source Time Patterns . . . . . . . . . . . . . . . . .94
1.5.5 Reaction Rate Editor . . . . . . . . . . . . . . . . . . . . . . . . . . 95
Global Settings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .95
1.5.6 Local Settings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96
Tank Reaction Rate Assignments . . . . . . . . . . . . . . . . . . . .96
Pipe Reaction Rate Assignments . . . . . . . . . . . . . . . . . . . .96
Note . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .97
1.5.7 Source Tracing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97
1.6 Tools . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98
1.6.1 Send Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98
1.6.2 Engineering Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . 98
1.6.3 Convert Asset Valves/Pumps to Model Valves/Pumps . . . . . . . 100
1.6.4 Duplicate Pipe Attributes . . . . . . . . . . . . . . . . . . . . . . . 101

MIKE URBAN WD User Guide . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105

2 Using the Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107
2.1 Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108
2.1.1 EPANET . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108
Pipe Flow Module . . . . . . . . . . . . . . . . . . . . . . . . . . . 108
Water Quality Module . . . . . . . . . . . . . . . . . . . . . . . . . 108
2.1.2 WD-Tools . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109
Fire Flow Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . 109
Pipe Roughness Calibration . . . . . . . . . . . . . . . . . . . . . . 109
PID Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109
Variable Speed Drive . . . . . . . . . . . . . . . . . . . . . . . . . . 110
Extended Rule-Based Controls . . . . . . . . . . . . . . . . . . . . 110
Demand Adjusted Analysis . . . . . . . . . . . . . . . . . . . . . . . 110
Pressure Zone Mapping . . . . . . . . . . . . . . . . . . . . . . . . 110
Water Hammer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110
2.2 Application Basics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111
2.2.1 Defining the Model . . . . . . . . . . . . . . . . . . . . . . . . . . 111

6 MIKE URBAN - © DHI

 Constructing Complex Networks . . . . . . . . . . . . . . . . . . . . 113
Roughness Coefficient Data . . . . . . . . . . . . . . . . . . . . . . 113
2.2.2 Input Data Requirements . . . . . . . . . . . . . . . . . . . . . . . 113
Project Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113
Network Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113
Extended Period Data . . . . . . . . . . . . . . . . . . . . . . . . . 115
Water Quality Data . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
Fire Flow Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116
Water Hammer Data . . . . . . . . . . . . . . . . . . . . . . . . . . 116
2.3 Entering Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116
2.3.1 Interactive Data Entry . . . . . . . . . . . . . . . . . . . . . . . . . 116
2.3.2 Graphical Input . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117
2.3.3 Importing Graphical Data . . . . . . . . . . . . . . . . . . . . . . . 118
Hiding and Showing Graphical Data . . . . . . . . . . . . . . . . . . 119
2.4 Demand Processing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119
2.4.1 Distributed Demands . . . . . . . . . . . . . . . . . . . . . . . . . 119
Developing Pipe Demand Coefficients . . . . . . . . . . . . . . . . . 122
Demand Allocation . . . . . . . . . . . . . . . . . . . . . . . . . . . 124
Demand Editing and Demand Scenarios . . . . . . . . . . . . . . . . 124
2.5 Performing an Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125
2.5.1 Executing the Analysis . . . . . . . . . . . . . . . . . . . . . . . . 125
2.5.2 Batch simulations . . . . . . . . . . . . . . . . . . . . . . . . . . . 126
2.6 Viewing Analysis Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127
2.6.1 Comparing Alternative Solutions . . . . . . . . . . . . . . . . . . . 128
Different Demands . . . . . . . . . . . . . . . . . . . . . . . . . . . 129
Different Pipe Roughness . . . . . . . . . . . . . . . . . . . . . . . 129
Different Pumping Scenarios . . . . . . . . . . . . . . . . . . . . . . 129
Effect of Pressure Reducing Valves . . . . . . . . . . . . . . . . . . 129
Water Quality Parameters . . . . . . . . . . . . . . . . . . . . . . . 129
2.6.2 Viewing the EPANET Analysis Results . . . . . . . . . . . . . . . . 129
2.6.3 Result Browser Window . . . . . . . . . . . . . . . . . . . . . . . 130
2.6.4 Results Table . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130
2.6.5 Thematic Mapping . . . . . . . . . . . . . . . . . . . . . . . . . . 130
2.6.6 Profile Plots . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131
2.6.7 Time Series Plots . . . . . . . . . . . . . . . . . . . . . . . . . . . 131
2.6.8 Link Q-H Plots . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132
2.6.9 Fire Hydrant Q-H Plots . . . . . . . . . . . . . . . . . . . . . . . . 132
2.6.10 Current Operating Pump Q-H Point . . . . . . . . . . . . . . . . . 132
2.6.11 Check for Pipes and Nodes . . . . . . . . . . . . . . . . . . . . . . 132
2.6.12 Animation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132
2.6.13 EPANET Error and Warning Messages . . . . . . . . . . . . . . . 132
2.6.14 Store Results to Database . . . . . . . . . . . . . . . . . . . . . . 134
3 Water Distribution Tools (WD-Tools) . . . . . . . . . . . . . . . . . . . . . . 135
3.1 Fire Flow Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135
3.1.1 Browsing Results . . . . . . . . . . . . . . . . . . . . . . . . . . . 138
3.1.2 Tabular Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138
3.1.3 Thematic Maps . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138

7
 3.2 Pipe Roughness Calibration . . . . . . . . . . . . . . . . . . . . . . . . . . 139
3.2.1 Evolutionary Algorithms - Theoretical Background . . . . . . . . . . 139
3.2.2 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141
3.2.3 Pipe Roughness Calibration . . . . . . . . . . . . . . . . . . . . . 141
3.3 PID Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145
3.3.1 Architecture of RTC Systems . . . . . . . . . . . . . . . . . . . . 145
3.3.2 MIKE URBAN vs. Real Life . . . . . . . . . . . . . . . . . . . . . 147
3.3.3 The PID Algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . 149
Control Settings . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151
Set Point Settings . . . . . . . . . . . . . . . . . . . . . . . . . . . 152
PID Settings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152
3.4 Demand Adjusted Anlaysis . . . . . . . . . . . . . . . . . . . . . . . . . . 152
3.4.1 Low Pressure Systems . . . . . . . . . . . . . . . . . . . . . . . . 155
3.4.2 Intermittent Water Supply Systems or Modelling System Shutdown and
Maintenance 157
3.5 Variable Speed Drive Pump . . . . . . . . . . . . . . . . . . . . . . . . . . 157
3.6 Extended Rule-Based Controls . . . . . . . . . . . . . . . . . . . . . . . . 157
3.7 Pressure Zone Mapping . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157
3.8 Advisor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159
3.8.1 Pipe Criticality . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159
3.8.2 Sustainability Analysis . . . . . . . . . . . . . . . . . . . . . . . . 162
3.8.3 Cost Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165
Pump Energy Table . . . . . . . . . . . . . . . . . . . . . . . . . . 165
Pump Energy Graphs . . . . . . . . . . . . . . . . . . . . . . . . . 166
3.8.4 Alarms and Violations . . . . . . . . . . . . . . . . . . . . . . . . 167
4 MIKE URBAN Water Hammer . . . . . . . . . . . . . . . . . . . . . . . . . . 171
4.1 Water Hammer Calculation . . . . . . . . . . . . . . . . . . . . . . . . . . 171
4.2 Theoretical Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . 172
4.2.1 Description of Water Hammer Model . . . . . . . . . . . . . . . . 172
Wave Speed . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 172
Rigid Conduit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 173
Thick-Walled Elastic Conduit (D/e<=10) . . . . . . . . . . . . . . . . 173
Thin-Walled Elastic Conduit (D/e>10) . . . . . . . . . . . . . . . . . 173
Tunnels Through Solid Rock, Parmakian 1963 . . . . . . . . . . . . 174
Reinforced Concrete Pipe . . . . . . . . . . . . . . . . . . . . . . . 174
Diagrams . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 175
4.3 Numerical Scheme and Algorithm . . . . . . . . . . . . . . . . . . . . . . . 176
4.3.1 Coefficients for the Numerical Scheme . . . . . . . . . . . . . . . 177
4.3.2 Looped Network Solution Algorithm . . . . . . . . . . . . . . . . . 178
4.3.3 Hydraulic Structures . . . . . . . . . . . . . . . . . . . . . . . . . 180
4.4 Water Hammer Calculations . . . . . . . . . . . . . . . . . . . . . . . . . . 181
Water Hammer Data Preparation . . . . . . . . . . . . . . . . . . . 181
Water Hammer Result Presentation . . . . . . . . . . . . . . . . . . 181
4.4.1 Running Water Hammer Simulations . . . . . . . . . . . . . . . . 182
Initial Conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . 182
Boundary Conditions . . . . . . . . . . . . . . . . . . . . . . . . . . 182
Computational Parameters . . . . . . . . . . . . . . . . . . . . . . . 182

8 MIKE URBAN - © DHI

 4.4.2 Definition of Network Layout . . . . . . . . . . . . . . . . . . . . . 183
Branches . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 184
Nodes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 184
Grid Points . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 185
Computational Grid and Hydraulics Structures . . . . . . . . . . . . . 185
Specific Pipe Data . . . . . . . . . . . . . . . . . . . . . . . . . . . 185
Junction Node Demands . . . . . . . . . . . . . . . . . . . . . . . . 185
Control Rules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 185
Specific Pump Data . . . . . . . . . . . . . . . . . . . . . . . . . . . 185
Specific Valve Data . . . . . . . . . . . . . . . . . . . . . . . . . . . 191
List of Supported Components . . . . . . . . . . . . . . . . . . . . . 196
List of Unsupported Component . . . . . . . . . . . . . . . . . . . . 196
4.4.3 List of Additional Components . . . . . . . . . . . . . . . . . . . . 196
4.4.4 Components Located in Nodes . . . . . . . . . . . . . . . . . . . . 197
Tanks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 197
Air-Chambers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 197
Vented Air Chambers . . . . . . . . . . . . . . . . . . . . . . . . . . 199
Air-Valve . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 201
4.4.5 Tutorial . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 202
5 EPANET . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 205
5.1 History . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 206
5.1.1 Analysis Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . 206
Steady State Hydraulics . . . . . . . . . . . . . . . . . . . . . . . . 206
Extended Period Hydraulics . . . . . . . . . . . . . . . . . . . . . . 207
Water Quality Analysis . . . . . . . . . . . . . . . . . . . . . . . . . 207
5.1.2 Applications of MIKE URBAN WATER . . . . . . . . . . . . . . . . 207
Planning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 208
Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 208
Operation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 208
5.1.3 Skeletonization . . . . . . . . . . . . . . . . . . . . . . . . . . . . 208
5.2 The Water Distribution Network . . . . . . . . . . . . . . . . . . . . . . . . 210
5.2.1 Pipes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 211
Modeling Parallel Pipes . . . . . . . . . . . . . . . . . . . . . . . . . 213
5.2.2 Pumps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 214
Pumping Rate Limits . . . . . . . . . . . . . . . . . . . . . . . . . . 216
Controlling Pumps . . . . . . . . . . . . . . . . . . . . . . . . . . . 217
Modeling Pumps in Parallel and Series . . . . . . . . . . . . . . . . . 217
5.2.3 Valves . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 220
Pressure Reducing Valves (PRV) . . . . . . . . . . . . . . . . . . . 220
Pressure Sustaining Valves (PSV) . . . . . . . . . . . . . . . . . . . 221
Pressure Breaker Valves (PBV) . . . . . . . . . . . . . . . . . . . . 221
Flow Control Valves (FCV) . . . . . . . . . . . . . . . . . . . . . . . 221
Throttle Control Valves (TCV) . . . . . . . . . . . . . . . . . . . . . 221
General Purpose Valves (GPV) . . . . . . . . . . . . . . . . . . . . 221
5.2.4 Minor Losses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 221
5.2.5 Nodes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 222
Modeling Pumped Groundwater Wells . . . . . . . . . . . . . . . . . 223
Modeling Hydraulically Adjacent Storage Tanks . . . . . . . . . . . . 225

9
 5.3 Time Patterns . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 225
5.4 Hydraulic Simulation Model . . . . . . . . . . . . . . . . . . . . . . . . . . 226
5.4.1 Water Age and Source Tracing . . . . . . . . . . . . . . . . . . . 233
5.5 Water Quality Simulation Model . . . . . . . . . . . . . . . . . . . . . . . . 234
5.5.1 Advective Transport in Pipes . . . . . . . . . . . . . . . . . . . . . 234
5.5.2 Mixing at Pipe Junctions . . . . . . . . . . . . . . . . . . . . . . . 234
5.5.3 Mixing in Storage Facilities . . . . . . . . . . . . . . . . . . . . . . 235
5.5.4 Bulk Flow Reactions . . . . . . . . . . . . . . . . . . . . . . . . . 235
5.5.5 Pipe Wall Reactions . . . . . . . . . . . . . . . . . . . . . . . . . 237
5.5.6 Lagrangian Transport Algorithm . . . . . . . . . . . . . . . . . . . 238
5.5.7 Tank Mixing Models . . . . . . . . . . . . . . . . . . . . . . . . . 240
5.6 Model Calibration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 243
5.6.1 Accuracy Concerns . . . . . . . . . . . . . . . . . . . . . . . . . 243
Pressure Differential Method . . . . . . . . . . . . . . . . . . . . . . 244
Percent Pressure Differential Method . . . . . . . . . . . . . . . . . 244
Head Loss Differential Method . . . . . . . . . . . . . . . . . . . . . 244
5.6.2 Reasons for Calibrating a Model . . . . . . . . . . . . . . . . . . . 244
Model Credibility . . . . . . . . . . . . . . . . . . . . . . . . . . . . 245
Benchmark . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 245
Predict Potential Problems . . . . . . . . . . . . . . . . . . . . . . . 245
Understand System Operations . . . . . . . . . . . . . . . . . . . . 245
Uncover Errors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 245
5.6.3 Calibration Model Data Requirements . . . . . . . . . . . . . . . . 246
Simulation Type Considerations . . . . . . . . . . . . . . . . . . . . 246
Data Acquisition . . . . . . . . . . . . . . . . . . . . . . . . . . . . 247
Operational Review . . . . . . . . . . . . . . . . . . . . . . . . . . 247
Recalibration Frequency . . . . . . . . . . . . . . . . . . . . . . . . 247
5.6.4 Calibration Simulations . . . . . . . . . . . . . . . . . . . . . . . . 248
Initial Simulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 248
Comparing Model Output . . . . . . . . . . . . . . . . . . . . . . . 248
5.6.5 Model Adjustments . . . . . . . . . . . . . . . . . . . . . . . . . . 250
Operational Data Adjustments . . . . . . . . . . . . . . . . . . . . . 251
Consumption Data Adjustments . . . . . . . . . . . . . . . . . . . . 251
Network Data Adjustments . . . . . . . . . . . . . . . . . . . . . . . 252
Pipe Roughness Adjustments . . . . . . . . . . . . . . . . . . . . . 252
Number of Simulation Iterations . . . . . . . . . . . . . . . . . . . . 252

10 MIKE URBAN - © DHI

Network Component Editors

1 EPANET Input Description

This chapter provides an overview of the MIKE URBAN WATER data input
dialog boxes. For a detailed overview on using this program, see Chapter 2.

The Model Manager User Guide will provide information on how to work with
the various editing tools, loading of results etc.

1.1 Network Component Editors

The following sections describe the various network component editors (i.e.,
junctions, pipes, pumps, reservoirs, tanks, and values). These network com-
ponent editors allow you to define each component individually using a
spearheads like data entry dialog box. This enables you to view and edit the
attributes of all of the components that have been defined. These component
editors also allow you to define a network model without a graphical layout, if
that is necessary.

The network component editors are all available from the EPANET Menu, as
shown in Figure 1.1.

Figure 1.1 The EPANET Menu allow access to the network component editors

11
 EPANET Input Description

1.1.1 Junction Editor

The first step in defining a water distribution network is to define the junction
nodes that define the interconnection between the pipes, pumps, valves,
tanks, and reservoirs that make up the network. Junction nodes are also
placed at points of water consumption or inflow, at points where specific anal-
ysis values (e.g., pressure, concentration, etc.) are desired, and at points
where pipe attributes (e.g., diameter, roughness, etc.) change.

Junction nodes are either defined interactively on the graphical Map window
using the Junction tool, or by manual data entry using the Junction Editor dia-
log box as shown in Figure 1.2. The Junction Editor allows you to define the
junction’s ID, location, any external demand, and a description. The Junction
Editor dialog box is reached by selecting EPANET | Junctions

Figure 1.2 The Junction Editor allows you to define the junction nodes that define
the inter connectivity between the water distribution network compo-
nents

A list of the Junction Editor data entries for Figure 1.2 follows, with a short
description given for each entry.

ASSET ID (optional)
This data entry is used to specify a corresponding asset junction ID, which
uniquely identifies the junction node in the asset management system (such
as GIS, for example).

12 MIKE URBAN - © DHI

Network Component Editors

JUNCTION ID (mandatory)
This data entry is used to specify an ID which uniquely identifies the junction
node. The junction ID acts as a unique look up key that identifies the node
from all other nodes. A node can be a junction, reservoir, or tank. Therefore,
no two nodes may have the same ID. However, a node and a link (i.e., pipe,
pump, or valve) can have the same ID. The node ID value can be any string
value (up to 40 characters).

A new junction ID is automatically suggested by MIKE URBAN whenever a

new junction node is placed into the list by pressing «Insert». When defining
the junction nodes graphically on the Map window using the Add Junction
tool, the junction ID is automatically defined.

X and Y COORDINATE (optional)

The X and Y data entries are used to define the physical (map) location of the
junction node, in units of ft. or m. This location definition is optional. In some
cases, the actual location of the junction node is not known—especially in
older, legacy networks. However, if the location is defined, then the junction
node will be displayed in the Map window. When defining the junction nodes
graphically on the Map window using the Add Junction tool, the X, Y location
is automatically entered.

DESCRIPTION (optional)
This data entry allows you to enter a description identifying the junction node
being entered. This description can be optionally displayed on the Map win-
dow and in reports generated by the Report Generator.

DATA SOURCE (optional)

This data entry is used to specify a corresponding asset data source, which
uniquely identifies the junction node location (such as database table or a
database file name) in the asset management system.

STATUS (optional)
This drop down selection list data entry allows you to define whether the junc-
tion node is imported (i.e existing node was imported from the external data
source), or is inserted, modified, GIS, calibrated or similar. By default, junc-
tion node status is undefined.

PRESSURE ZONE ID (optional)

This data entry allows you to define the ID of the pressure zone that the junc-
tion node lies within. The pressure zone ID can be used when distributing the
zone demand as well as to graphically display different zones on the Map.

Selecting «...» allows you to display the Pressure Zone selection dialog box,
where the appropriate pressure zone ID can be selected.

13
 EPANET Input Description

TYPE (mandatory)
This drop down selection list data entry allows you to define whether the junc-
tion node is a simple junction, or is an emitter, or an air-valve. By default, any
junction nodes added are junctions.

DEMAND COEFFICIENT
The new field was added into the Junction editor. Demand coefficient allows
you to define the share from the whole network demand, which is taken by
the node. This filed is used only by the Demand Distribution function.

Example: the network has 3 nodes, where the demand coefficient is defined;
these values are 10, 10, and 30. For each node, the weighted coefficient is
calculated and based on it; the total network demand is distributed. The node,
where the demand coefficient is not defined will get no demand from the total
network demand: Q

QT
q i = ----------------- Ci (1.1)
 i C
i = 0,n

where:
qi = node demand
Qt = total network demand
ci = demand coefficient

MINIMAL PRESSURE (optional)

This data entry defines the estate height above the junction node elevation.
This data entry is used to calculate Tap Pressure at the junction node and is
used to verify the minimal pressure at the node.

STATE (optional)
This drop down selection list data entry allows you to define whether the junc-
tion node is unmarked (i.e an existing node already contained in the water
distribution network), or is marked (i.e. one that is being considered for future
expansion, etc.). By default, any junction nodes added are unmarked.

ELEVATION (mandatory)
This data entry defines the elevation above a common datum for the junction
node, in units of ft. or m. This value is used to determine the pressure head at
the node during a simulation. The default elevation is zero.

Junction nodes should have their elevation specified above zero datum (i.e.,
sea level) so that pressure computations can be carried out.

SURFACE ELEVATION (optional)

This data entry defines the surface elevation above a common datum for the
junction node, in units of ft. or m. This value is used to display the surface ele-
vation in the Longitudinal Profile Plot.

14 MIKE URBAN - © DHI

Network Component Editors

MULTIPLE DEMAND (optional)

Junctions may have more than one demand assigned to them. This function
is particularly useful if the demand patters of multiple water users are known
for a given junction. It is also possible to assign separate patterns to the
demands assigned to a given junction. The demands and demand patterns
can be assigned from the Junction Editor, by selecting the Multiple option for
base demand.

If no water is entering or leaving the network system at this junction node

location, this entry should be specified as zero. If flow is leaving the network
system at this junction node, then a positive value should be specified. If an
inflow into the network system occurs at this junction node (e.g., from a
groundwater extraction well), then a negative value should be specified.

This demand value will be adjusted during an extended period simulation

according to the assigned demand pattern ID. Caution should be exercised if
a groundwater extraction well is defined using this entry, since the groundwa-
ter extraction rate will be adjusted according to the assigned demand pattern
ID.

Note that the baseline demand at this junction node can be (optionally) com-
puted by globally defining the demand for the entire network and then having
MIKE URBAN distribute this demand to each of the network nodes using the
Distributed Demand dialog box. See the section titled “Distributed Demands”
on page 59 for more information on computing distributed demands.

For extended period simulations, the baseline demand can be adjusted dur-
ing the simulation due to a predefined time pattern factor. See the section
titled “Repetitive Profiles (Pattern) Editor” on page 83 for more information.

DEMAND DESCRIPTION (optional)

This data entry allows you to enter a description identifying the demand being
entered.

DEMAND CATEGORY (optional)

This data entry allows you to enter a description identifying the demand being
entered. The demand category can be used when using the Distributed
Demand tool.

DEMAND PATTERN ID (optional)

This data entry allows you to define the ID of the demand pattern to be
applied to the junction node demand values during an extended period simu-
lation. This demand pattern will be applied to the defined baseline demand.

If a groundwater well is associated with this node, then a demand pattern

should not be assigned—otherwise the groundwater extraction rate will be
adjusted according to the assigned demand pattern.

15
 EPANET Input Description

By default this data entry is blank and default demand pattern ID of 1 is

assigned. See the section titled “Repetitive Profiles (Pattern) Editor” on
page 83 for more information on demand patterns.

Emitter
The following data entries defines parameters of an emitter located at the
junction node. Emitters are needed to model flow through sprinkler systems
and irrigation networks. They can also be used to simulate leakage in a pipe
connected to the junction (if a discharge coefficient for the leading crack or
joint can be estimated).

A list of the Emitter data entries for Figure 1.3 follows, with a short description
given for each entry.

Figure 1.3 Emitter data entries

FLOW COEFFICIENT (mandatory)

This data entry allows you to define the flow coefficient of the emitter. Flow
out of the emitter equals the product of the flow coefficient and the junction
pressure raised to a power. The flow coefficient is defined in flow units at 1 psi
(1 meter) pressure drop

Air-Valve
The following data entries defines parameters of an air-valve located at the
junction node and are used only in case of a water hammer analysis. A list of
the Air-Valve data entries for Figure 1.3 follows, with a short description given
for each entry. See the section titled “Water Hammer Calculation” on
page 171 for more information on demand patterns.

16 MIKE URBAN - © DHI

Network Component Editors

Figure 1.4 Air-Valve data entries

VALVE DIAMETER (mandatory)

This data entry allows you to define the diameter of the air-valve. See the
section titled “Specific Valve Data” on page 191 for more information on
demand patterns.

POLYTROPIC EXPANSION FACTOR (mandatory)

This data entry allows you to define the polytropic exponent in the polytropic
gas equation.

By default this data entry is k=1.2. See the section titled “Specific Valve Data”
on page 191.

DUAL-ACTING VALVE CHARACTERISTICS CURVE (mandatory)

This data entry allows you to define the ID of the dual-acting valve character-
istics curve. This is the relation between gauge pressure and the air flow
through the dual-acting valve. See the section titled “Specific Valve Data” on
page 191.

Graphical Placement and Editing of Junction Nodes

From the Map window, the Add Junction tool can be used to add a junction
node. As nodes are added graphically to the Map window, their X, Y locations
are stored into the MIKE URBAN database. While placing a new node, if the
user clicks on an existing node MIKE URBAN will display an error message
informing the user that an invalid location had been selected.

If desired, from the Map window, the user can move an existing junction node
using the ‘select’ tool. By selecting the node and holding down the left mouse
button, the node can be dragged to a new location. As the node is dragged,
its connecting links (i.e., pipes, pumps, and valves) rubber band along with it.

To delete an existing junction node graphically, select the node using the
‘Select’ tool and then press «Delete». The selected node and all its connect-

17
 EPANET Input Description

ing links will then be deleted. To edit a junction node, double click on the node
using the Select tool. The Junction Editor will then be displayed allowing you
to edit the attributes of the selected node.

1.1.2 Pipe Editor

Pipes are used to transport water from one node to another. Pipes must
always begin and end at a node.

Pipes are either defined interactively on the Map window using the ‘Add pipe’
tool, or by manual data entry using the Pipe Editor dialog box as shown in
Figure 1.5

.
Figure 1.5 The Pipe Editor allows you to define the pipes that make up the water
distribution network

The Pipe Editor allows you to define the pipe’s ID, diameter, loss coefficients,
nodal connectivity, description, and other attributes. The Pipe Editor dialog
box is reached by selecting EPANET | Pipes.

A list of the Pipe Editor data entries for Figure 1.5 follows, with a short
description given for each entry.

18 MIKE URBAN - © DHI

Network Component Editors

ASSET ID (optional)
This data entry is used to specify a corresponding asset pipe ID, which
uniquely identifies the pipe link in the asset management system (such as
GIS, for example).

PIPE ID (mandatory)
This data entry is used to specify an ID which uniquely identifies the pipe link.
The pipe ID acts as a unique look up key that identifies the link from all other
links. A link can be a pipe, pump, or valve. Therefore, no two links may have
the same link ID. However, a link and node (i.e., junction, reservoir, or tank)
can have the same ID. The node ID value can be any string value (up to 40
characters).

A new pipe ID is automatically suggested by MIKE URBAN whenever a new

pipe is placed into the list by pressing «Insert». When defining the pipes
graphically on the Map window using the Add Pipe tool, the pipe ID is auto-
matically defined.

When importing (or merging) multiple water distribution network models into a
single network model, MIKE URBAN will check for collisions between identi-
cal link IDs and will automatically assign a new link ID value for any link being
imported that contains the same link ID value as what already exists in the
network model. MIKE URBAN will automatically assign a new pipe ID when-
ever any pipe is split into more pipes.

FROM NODE, TO NODE (mandatory)

These data entries define the ID of the pipe’s starting (upstream) and ending
(downstream) nodes. These IDs define the pipe connectivity of the network.

The Node Type pull-down selection list allows the user to specify what type of
node is connected to the end of the pipe. Then, choosing «...» will display the
Select Node selection dialog box from which the user can select the appropri-
ate node. Or, choosing «Pick» allows the user to graphically select the node
from the Map window. To reverse the order of the nodes, choose «Advanced
- Swap Nodes».

If the computed flow is moving from the starting node to the ending node, the
computed flow value will be positive. If the computed flow is moving from the
ending node to the starting node, the computed flow value will be negative.

DESCRIPTION (optional)
This data entry allows you to enter a description identifying the pipe being
entered. This description can be optionally displayed on the Map window and
in reports generated by the Report Generator.

DATA SOURCE (optional)

This data entry is used to specify a corresponding asset data source, which
uniquely identifies the pipe location (such as database table or a database file
name) in the asset management system.

19
 EPANET Input Description

STATUS (optional)
This drop down selection list data entry allows you to define whether the
pump is imported (i.e existing node was imported from the external data
source), or is inserted, modified, GIS, calibrated or similar. By default, pump
status is undefined.

PRESSURE ZONE ID (optional)

This data entry allows you to define the ID of the pressure zone that the pipe
lies within. The pressure zone ID can be used when distributing the zone
demand as well as to graphically display different zones on the Map.

Selecting «...» allows you to display the Pressure Zone selection dialog box,
where the appropriate pressure zone ID can be selected.

LENGTH (mandatory)
This data entry defines the pipe length, in the unit of your choice. It is possible
to define a specific pipe length—independent of the pipe network layout. Oth-
erwise, MIKE URBAN recomputes the pipe length based upon the pipe lay-
out.

A global recomputing of pipe lengths can be performed by using the function-

ality under the Advanced button on the Pipe Editor.

DIAMETER (mandatory)
This data entry defines the internal diameter of the pipe, in inches or mm. The
second field (read-only) displays the pipe diameter as it would be used for the
hydraulic analysis. The pipe diameter is automatically adjusted when the pipe
wall is defined.

MATERIAL (optional)
This option allows the user to define the material of pipe construction. The
Pipe Material is defined by a “string” and the pipe roughness can be globally
assigned based upon the pipe material and pipe construction year, for exam-
ple.

CONSTRUCTION YEAR (optional)

This option allows the user to define the age of the pipe. Pipe age is defined
by a float number, meaning that alpha characters are not allowed to be
entered in this field. There are two common ways the user can define pipe
age, either by year of construction year, or number of years the pipe has been
in service. The choice is left to the user.

WALL THICKNESS (optional)

This field is used to define the wall thickness of a pipe. The pipe diameter is
automatically adjusted by the program when the pipe thickness is defined.

WAVE SPEED (water hammer mandatory)

This field is used to define the celerity of the pressure wave. This field is used
only in case of water hammer calculations.

20 MIKE URBAN - © DHI

Network Component Editors

FORMULATION (read-only)
This field is used to display the current friction loss formulation type, such as
Darcy-Weisbach, Hazen-Williams, or Manning. To change the friction formu-
lation, use the Project Options dialog box.

ROUGHNESS (mandatory)
This data entry defines the roughness of the interior surface of the pipe.
Based upon which roughness type loss coefficient has been specified, this
value is unit less for Hazen-Williams or Chezy-Manning headloss formulas,
and in millifeet or mm for the Darcy-Weisbach (or Colebrook-White) formula-
tion. The roughness type is specified by the user within the Project Options
dialog box. Choosing «...» will display the Select Pipe Roughness Coefficient
selection dialog box, allowing the user to select the appropriate roughness
value to use.

LOSS COEFFICIENT (optional)

This data entry defines the sum of all the minor (or local) loss coefficients for
the pipe, which are unitless. Choosing «...» will display Select Minor Loss
Coefficient selection dialog box, allowing the user to select the appropriate
minor loss coefficient to use. If more that one minor loss component exists
along the pipe, then the sum of the corresponding minor loss coefficients
should be entered.

PRESSURE NORMAL (optional)

This field is used to define the normal operating pressure for a pipe. This is
an optional field and can be used when the calculated pressures are distrib-
uted to pipes.

BULK COEFF. and WALL COEFF. (optional)

Locally defined reaction rates. Please refer to section on reaction rates for
further.

DEMAND COEFFICIENT 1,2 (optional)

MIKE URBAN allows the user to recalculate the nodal water demands based
upon the total network demand using two methods: the Method of Pipe
Lengths and the Method of Two Coefficients. This feature is useful for auto-
matically assigning the nodal water demand to a large network, since the
software will automatically proportion the total network demand based upon
one of these two methods. These methods are used to mimic the amount of
actual demand along a pipe, based upon the pipe length or pre-defined
demand coefficients.

These two data entries define the demand coefficients, k1i and k2i which are
used in these two computational methods. Each method proportions the total
network demand along the pipe by splitting the computed demand between
the two nodes at each end of the pipe.

21
 EPANET Input Description

The Method of Pipe Lengths computes the total water demand assigned to
the current pipe (which is then split between the starting and ending nodes)
as:

 Q l i k li
q pi = -----------------
- (1.2)
  k li l i 

The Method of Two Coefficients computes the total water demand assigned
to the current pipe (which is then split between the starting and ending nodes)
as:

 Q k li k 2i
q pi = ---------------------
- (1.3)
  k li k 2i 

where:

qpi=total water demand applied to the pipe, split between the two end nodes.

Q= total network water demand. Note that the total network water demand is
defined in the Distributed Demands dialog box, available from the Edit Menu.

li = pipe length

See the section titled “Distributed Demands” on page 59 for more information
on computing distributed demands.

Note that there is also Demand Coefficient 3, 4 which is available only in the
table grid.

STREET NAME (optional)

This field is used to define the street name. This is an optional field and can
be used for better navigation through the pipe network and for reporting pur-
poses.

STATE (optional)
This drop down list box data entry allows you to define whether the pipe is
unmarked (i.e. a pipe already contained in the water distribution network), or
is marked (i.e. one that is being considered for future expansion, etc.). By
default, any pipes added are unmarked.

CHECK VALVE (optional)

This check box data entry defines the presence of a check valve in the pipe. If
a check valve exists, then water is only allowed to flow from the starting to
ending node. This is commonly used to prevent a flow reversal through the
pipe. If conditions exist for flow reversal, the valve shuts and the pipe carries
no flow. By default, no check valve is present in a pipe.

22 MIKE URBAN - © DHI

Network Component Editors

CLOSED (optional)
This check box data entry allows the user to toggle the OPEN and CLOSED
status of the pipe. Choosing CLOSED effectively removes the pipe from the
network system.

Warning
Note that you cannot set the pipe status of a pipe containing a check valve.
Pipes with a check valve are initially open, and close only if flow within the
pipe attempts to reverse (move from the ending downstream node to the
starting upstream node).

Flow Direction
Note that the flow direction for pipes is assumed to be from the starting
(upstream) node to the ending (downstream) node. The order in which these
nodes are specified for each pipe is critical for a proper network representa-
tion. To reverse the order of the nodes, choose «Swap·Nodes».

If the computed flow is moving from the starting node to the ending node, the
computed flow value will be positive. If the computed flow is moving from the
ending node to the starting node, the computed flow value will be negative.

Graphical Placement and Editing of Pipes

From the Map window, the Add Pipe tool can be used in different ways to
place a pipe. If the user clicks on or near an existing node, MIKE URBAN will
snap to the node and treat this as a starting node. A rubber-banding line will
then be drawn from this node, representing the pipe, while the user selects
the ending node. Then, if the user clicks on or near an existing node, it will
snap to that node and treat the selected node as the ending node and then
draw the pipe. If the user clicks anywhere else, it will place an ending node at
the selected location.

While selecting the starting node, if the user clicks with the ’Add pipe’ tool in
the Map window somewhere else other than a node or a pipe, MIKE URBAN
will place a starting node at the clicked location.

If the user selects an existing pipe while selecting either the starting or ending
nodes, MIKE URBAN will display a query dialog, as shown in Figure 1.6, ask-
ing the user whether to split the pipe and insert a node at the selected loca-
tion. If the user selects to insert a node into the pipe, the pipe will be broken
into two segments at the point of selection with a new node added. The
length of the two pipe segments will be adjusted accordingly so that the total
length of the two segments equals that of the original pipe.

23
 EPANET Input Description

Figure 1.6 When clicking on an existing pipe with the Add Pipe tool, MIKE URBAN
will ask whether to insert a junction node into the existing pipe for con-
necting the new pipe

To delete an existing pipe graphically, select the pipe using the ‘Select’ tool
and then select «Delete Selected Features». The selected pipe will then be
deleted. To edit a pipe, double click on the pipe using the ‘Select’ tool. The
Pipe Editor will then be displayed allowing you to edit the attributes of the
selected pipe.

Snap Tolerance
As you graphically layout pipes in the Map window by clicking on existing
nodes, MIKE URBAN will snap to the nearest node if it is within a specified
snap tolerance. However, if the cursor is not placed directly on top of the
node or within the specified snap tolerance, MIKE URBAN may not snap to
the node but instead will add a new junction node.

A snap tolerance is specified in the Configuration dialog box, available from

the Tools Menu. See the section titled “Snap Tolerance” on page 24. This
snap tolerance is specified in screen pixels, and defines how close the cursor
must be to the node in order to snap to the node. Anywhere outside of this
snap tolerance and a new junction node is created.

Defining Curved Pipes

MIKE URBAN uses multi-segmented polylines to represent curved pipes.
Intermediate vertices are used to define the curvature of the pipe. A curved
pipe is defined graphically in the Map window. To define a curved pipe:

1. Select the ‘Add Pipe’ tool from the Components toolbar and click on the
starting point in the Map window.
2. After selecting the starting point, construct the curve of the pipe by click-
ing in the Map window at the points the curvature changes to create
intermediate vertices so that an arc or a curved line is formed. Double-
click on the last position to end the curved pipe. If the last position is an

24 MIKE URBAN - © DHI

Network Component Editors

existing node, the end point of the curved pipe will snap to the existing
node.

To calculate the pipe length based upon the curved layout, press the
Advanced button on the Pipe Editor and select either ‘Recompute selected
pipe length’ or ‘Recompute all pipes length’. MIKE URBAN will then recom-
pute the pipe length for either the selected pipes or all the pipe based upon
the pipe layout.

1.1.3 Pump Editor

Pumps are used to raise the hydraulic head of water. Pumps are represented
as short links of negligible length. MIKE URBAN analysis engine will automat-
ically prevent flow reversal though a pump, and will issue warning messages
when a pump operates outside of its normal operating range.

Pumps are either defined interactively on the graphical Map window using the
Add Pump tool, or by manual data entry using the Pump Editor dialog box as
shown in Figure 1.7.

The Pump Editor allows you to define the pump’s ID, pump power curve, sta-
tus, nodal connectivity, description, and other attributes. The Pump Editor
dialog box is reached by selecting EPANET | Pumps.

Figure 1.7 The Pump Editor allows you to define the pumps used in the water dis-
tribution network

25
 EPANET Input Description

A list of the Pump Editor data entries for Figure 1.7 follows, with a short
description given for each entry.

ASSET ID (optional)
This data entry is used to specify a corresponding asset pump ID, which
uniquely identifies the junction node in the asset management system (such
as GIS, for example).

PUMP ID (mandatory)
This data entry is used to specify an ID which uniquely identifies the pump
link. The pump ID acts as a unique lookup key that identifies the link from all
other links. A link can be a pipe, pump, or valve. Therefore, no two links may
have the same link ID. However, a link and node (i.e., junction, reservoir, or
tank) can have the same ID. The node ID value can be any string value (up to
40 characters).

A new pump ID is automatically suggested by MIKE URBAN whenever a new

pipe is placed into the list by pressing «Insert». When defining the pipes
graphically on the Map window using the Add Pump tool, the pump ID is
automatically defined.

FROM NODE, TO NODE (mandatory)

These data entries define the ID of the pump’s starting (upstream) and ending
(downstream) nodes. These IDs define the pump connectivity of the network.

The Node Type pull-down selection list allows the user to specify what type of
node is connected to the end of the pump. Then, choosing «...» will display
the Select Node selection dialog box from which the user can select the
appropriate node. Or, choosing «Pick» allows the user to graphically select
the node from the Map window.

Pumped flow is always assumed to move from the starting (upstream) node
to the ending (downstream) node. To reverse the order of the nodes, choose
«Swap·Nodes».

DESCRIPTION (optional)
This data entry allows you to enter a description identifying the pump being
entered. This description can be optionally displayed on the Map window and
in reports generated by the Report Generator.

DATA SOURCE (optional)

This data entry is used to specify a corresponding asset data source, which
uniquely identifies the pump location (such as database table or a database
file name) in the asset management system.

STATUS (optional)
This drop down selection list data entry allows you to define whether the
pump is imported (i.e existing node was imported from the external data
source), or is inserted, modified, GIS, calibrated or similar. By default, pump
status is undefined.

26 MIKE URBAN - © DHI

Network Component Editors

PRESSURE ZONE ID (optional)

This data entry allows you to define the ID of the pressure zone that the pump
lies within. The pressure zone ID can be used when distributing the zone
demand as well as to graphically display different zones on the Map.

Selecting «...» allows you to display the Pressure Zone selection dialog box,
where the appropriate pressure zone ID can be selected.

PUMP TYPE (mandatory)

This option button data entry defines the pump operating characteristics.
There are four options available to define the pump specifications. By default,
a constant power pump is selected.

Constant Power pump is used when the pump characteristic curve is

unknown and a constant power output is assumed. The data entry specifies
the pump power rating, in hp or kw. The default power rating is zero.

1xQ-H Point Pump Curve pump, as shown in Figure 1.8, is used for a stand-
ard pump curve with no extended flow range, where the cutoff head is 133%
of the design head and the maximum flow is twice the design flow.

The Design Head is in units ft. or m, and are by default zero. The Design Flow
is in the user-specified units, and by default zero.

Figure 1.8 A 1-point pump curve contains no extended flow range

3xQ-H Point Pump Curve, as shown in Figure 1.10 can be used to describe
the flow-head relationship of the pump.

The Shutoff Head is the head value at zero flow. The Design Head is the
standard operating head. The Design Flow is corresponding flow rate. The

27
 EPANET Input Description

High End Head is the head at the upper end of the normal operating flow
range. The High End Flow is the corresponding flow rate. The Maximum Flow
is the flow rate for the extended flow range.

All heads are in units of ft. or m, and flows are in the user-specified units.

Q-H Pump Curve can alternatively. The Q-H Pump Curve is created by pro-
viding either a pair of head-flow points, or four or more such points. MIKE
URBAN creates the pump curve by connecting the points with straight line
segments. The Q-H pump curve should be created in the Curve Editor found
in the Edit menu. Once the Q-H pump curve is created, it can then be
assigned to a pump in the Pump Editor.

Figure 1.9 A Q-H Table used to define the multi-point pump curve

28 MIKE URBAN - © DHI

Network Component Editors

Figure 1.10 A 3-point pump curve with no extended flow range.

RELATIVE SPEED (optional)

Note that the Relative Speed entry field allows the user to adjust the initial
setting of the pump at the start of the simulation. For example, entering a
value of 1.2 specifies that the pump operates at 1.2 times its normal speed at
the start of the simulation.

PUMP STATE (optional)

This drop down list box data entry allows you to define whether the pump is
unmarked (i.e. a pump already contained in the water distribution network), or
is marked (i.e. one that is being considered for future expansion, etc.). By
default, any pumps added are unmarked.

CLOSED (optional)
This check box data entry allows the user to toggle the OPEN and CLOSED
status of the pump. Choosing CLOSED effectively removes the pump from
the network system.

Pump Energy Settings

MIKE URBAN is capable of modelling the cost of operating pumps. Within the
Pump Energy Editor, the user can define a method for cost calculation.

29
 EPANET Input Description

ENERGY PRICE (optional)

The user defines an energy price ($/kw-hour) to be used. In this method,
MIKE URBAN determines the energy consumed by the pump in kw-hours
and multiplies the energy consumption by the price. Leave blank if not appli-
cable or if the global value supplied with the project's Energy Options will be
used

ENERGY PRICE PATTERN (optional)

The ID label of the time pattern used to describe the variation in energy price
throughout the day. Each multiplier in the pattern is applied to the pump's
Energy Price to determine a time-of-day pricing for the corresponding period.
Leave blank if not applicable or if the global pricing pattern specified in the
project's Energy Options will be used.

EFFICIENCY CURVE (optional)

The ID label of the curve that represents the pump's wire-to-water efficiency
(in percent) as a function of flow rate. This information is used only to com-
pute energy usage. Leave blank if not applicable or if the global pump effi-
ciency supplied with the project's Energy Options will be used.

Pump Water Hammer Settings (optional)

MIKE URBAN is capable of modelling pumps under the water hammer flow
conditions. Within the Pump Water Hammer Editor, the user can define the
required pump parameters as shown in Figure 1.11.

Figure 1.11 A pump water hammer settings

OPERATIONAL SCHEDULE TYPE (mandatory)

This list box data entry defines the pump operating schedule. There are three
options available to define the pump schedule (rotational speed of pump N
versus time).

a. The pump is controlled by a pump operation schedule (N-time) curve.

b. The pump is controlled by a pump operation schedule until time of the

simulation is equal time of the power failure, then pump shutdown is
applied and pump remains stopped till the end of the computation run.

30 MIKE URBAN - © DHI

Network Component Editors

c C) The pump is primarily stopped (N equals zero) until time of pump

start-up is reached, then pump start-up equation is applied.
OPERATIONAL SCHEDULE (mandatory)
This data entry is used to select the pump operation schedule ID. To define
the pump schedule, use the Curves and Relations dialog box.

ROTATIONAL PUMP SPEED (mandatory)

This data entry is used to specify the initial rotational speed of a pump in rpm
(rotation per minute).

MOMENT OF INERTIA (optional)

The moment of inertia is the resistance of a rotating pump body to the change
of its rotational speed. The moment of inertia is defined in [kg*m^2].

PUMP TORQUE (optional)

This data entry is used to specify the ID of the pump torque curve. The pump
torque is the relation between the pump torque (Newton meter) versus pump
flow (flow units).

MOTOR TORQUE (optional)

This data entry is used to specify the ID of the motor torque curve. The motor
torque is the relation between the motor torque (Newton meter) versus pump
rotational speed (rotation per minute).

PUMP START-UP TIME (mandatory)

This data entry is used to define the time level at which the pump will start-up.

PUMP TRIP-OFF TIME (mandatory)

This data entry is used to define the time level at which the pump will trip-off.
The pump is operated based on its operational schedule till that time.

Q-H CURVE FOR RPM = 0 (optional)

This data entry is used to define the Q-H curve for the rotational speed N=0
i.e. the relation between the flow and pump head when the pump is stopped.
This data entry can help to define what happens for the positive and negative
flow through the pump.

Pump Variable Speed Settings (optional)

MIKE URBAN is capable of modelling the variable pump speed during the
extended period analysis. Within the Pump Variable Speed Editor, the user
can define the required pump parameters as shown in Figure 1.12.

31
 EPANET Input Description

Figure 1.12 A variable pump speed settings

CONTROL TYPE (mandatory)

This list box data entry defines the variable pump speed control mode. There
are two options available.

A) The pump is controlled by a pressure at the pump downstream node.

B) The pump is controlled by a pressure at any remote node (junction node or

a tank with a variable water level).

Please note that it is possible to use this control also for the downstream
node by specifying the downstream node ID as the "Control Node".

The difference between "downstream" and "remote" node control is in the

numerical solution of the variable pump speed. In case of "downstream" node
control, the control pressure settings are built into the main matrix solver
while in the case of "remote" node control the program is using an iterative
approach to find the pump speed.

CONTROL NODE (mandatory)

This data entry is used to define the control node ID.

CONTROL PRESSURE (mandatory)

This data entry is used to define the control node pressure.

MINIMUM PUMP SPEED (mandatory)

This data entry is used to define the minimum relative pump speed. Please
note that this data entry is not used by in case of "downstream node control".

MAXIMUM PUMP SPEED (mandatory)

This data entry is used to define the maximum relative pump speed. Please
note that this data entry is not used by in case of "downstream node control".

Variable Pumps
Many times, when performing an extended period simulation, it is desirable to
model a variable pump. A variable pump can vary its speed setting and/or
change its status to open or closed during a simulation. A variable pump is
modelled by defining the pump’s initial settings using the Pump Status previ-
ously described and then modifying the pump’s operation during the

32 MIKE URBAN - © DHI

Network Component Editors

extended period simulation using the Control Editor. The Control Editor is
available from the Extended Menu. See the section titled “Simple Control Edi-
tor” on page 74 for additional information.

A special case if a variable pump is VSD (variable speed drive) pump where
the rotational speed of a pump is automatically adjusted by the program in
order to maintain the control node pressure at the required level. MIKE
URBAN WATER can model VSD pump using VSD settings in the Pump Edi-
tor or using the PID (Proportional Integral Differential Control) algorithm in the
RTC (Real-Time Control) module.

Flow Direction
Note that the flow direction for pumps is assumed to be from the starting
(upstream) node to the ending (downstream) node. The order in which these
nodes are specified is critical for a proper network representation of the
pump. To reverse the order of the nodes, choose «Swap·Nodes».

Graphical Placement and Editing of Pumps

From the Map window, the ‘Add Pump’ tool can be used to place a pump.
When you want to insert a pump onto an existing pipe, you must insert it as a
new component.

To insert a pump onto the existing pipe:

1. Split the pipe by using Split tool from the Components toolbar.
2. Split the pipe once more to create a sections for 3 pipes.
3. Select the middle section by Select Feature tool and delete it by using
the Delete Selected tool.
4. Select the Pump tool from the Components floating toolbar and click on
the beginning pipe node in the Horizontal Plan window where you want
to place the valve and double click on the ending node. This will create a
new pump.

Another method of using the ‘Add Pump’ tool to add a pump is, from the Map
window, select a starting node and then an ending node. Clicking on or near
an existing node, MIKE URBAN will then snap to the selected node and a
rubber-banding line will then be drawn from this node, representing the start-
ing node, while the user selects the ending node. After selecting the ending
node, MIKE URBAN will then place a pump between the two selected nodes.

In addition, if the user clicks with the ‘Add Pump’ tool in the Map window
somewhere else other than a node or a pipe, MIKE URBAN will place a start-
ing node at the clicked location. A rubber-banding line will then be drawn from
this node, representing the pump, while the user selects the ending node. If
the user clicks on or near an existing node, it will treat the selected node as
the ending node. If the user clicks anywhere else, it will place an ending node

33
 EPANET Input Description

at the selected location. If the user selects an existing pipe, an error message
will be displayed stating that an invalid ending node was selected.

To delete an existing pump graphically, select the pump using the Select tool
and then select «Delete Selected». The selected pump will then be deleted.
To edit a pump, click on the pump using the Open Editor tool. The Pump Edi-
tor will then be displayed allowing you to edit the attributes of the selected
pump.

1.1.4 Valve Editor

Valves control the flow or pressure of water from one junction node to
another. Valves are represented as short links of negligible length. Note that
valve pressure settings are pressures (e.g., psi or m) and not total head (or
hydraulic gradeline elevation). Also, valves cannot be directly connected to
reservoir or storage tank nodes.

Valves are either defined interactively on the Map window using the
Add Valve tool, or by manual data entry using the Valve Editor dialog box as
shown in Figure 1.13. The Valve Editor allows you to define the valve’s ID,
type, status, nodal connectivity, description, and other attributes. The Valve
Editor dialog box is reached by selecting EPANET|Valves.

Figure 1.13 The Valve Editor allows you to define the valves used in the water distri-
bution network

A list of the Valve Editor data entries for Figure 1.13 follows, with a short
description given for each entry.

34 MIKE URBAN - © DHI

Network Component Editors

ASSET ID (optional)
This data entry is used to specify a corresponding asset valve ID, which
uniquely identifies the valve in the asset management system (such as GIS,
for example).

VALVE ID (mandatory)
This data entry is used to specify an ID which uniquely identifies the valve
link. The valve ID acts as a unique lookup key that identifies the link from all
other links. A link can be a pipe, pump, or valve. Therefore, no two links may
have the same link ID. However, a link and node (i.e., junction, reservoir, or
tank) can have the same ID. The valve ID value can be any string value (up to
40 characters).

A new valve ID is automatically suggested by MIKE URBAN whenever a new

valve is placed into the list by pressing «Insert». When defining the valves
graphically on the Map window using the Add Valve tool, the valve ID is auto-
matically defined.

FROM NODE, TO NODE (mandatory)

These data entries define the ID of the valve’s starting (upstream) and ending
(downstream) nodes. These IDs define the valve connectivity of the network.

The Node Type pull-down selection list allows the user to specify what type of
node is connected to the end of the pump. Then, choosing «...» will display
the Select Node selection dialog box from which the user can select the
appropriate node. Or, choosing «Pick» allows the user to graphically select
the node from the Map window.

Controlled flow is always assumed from the starting (upstream) node to the
ending (downstream) node. To reverse the order of the nodes, choose
«Swap·Nodes».

DESCRIPTION (optional)
This data entry allows you to enter a description identifying the valve being
entered. This description can be optionally displayed on the Map window and
in reports generated by the Report Generator.

DATA SOURCE (optional)

This data entry is used to specify a corresponding asset data source, which
uniquely identifies the valve location (such as database table or a database
file name) in the asset management system.

STATUS (optional)
This drop down selection list data entry allows you to define whether the
valve is imported (i.e existing node was imported from the external data
source), or is inserted, modified, GIS, calibrated or similar. By default, valve
status is undefined.

35
 EPANET Input Description

PRESSURE ZONE ID (optional)

This data entry allows you to define the ID of the pressure zone that the valve
lies within. The pressure zone ID can be used when distributing the zone
demand as well as to graphically display different zones on the Map.

Selecting «...» allows you to display the Pressure Zone selection dialog box,
where the appropriate pressure zone ID can be selected.

VALVE TYPE (mandatory)

This list box data entry defines the valve operating characteristics. There are
six options available to define the valve operating characteristics. By default a
pressure reducing valve (PRV) is selected.

A Pressure Reducing Valve (PRV) limits the pressure at the downstream

node to not exceed a preset value when the upstream node pressure is
above the PRV setting. If the upstream pressure is below the setting, flow
through the valve is unrestricted. Should the pressure at the downstream
node exceed the pressure at the upstream node, the valve closes to prevent
reverse flow. Note that PRVs cannot be placed directly in series.

A Pressure Sustaining Valve (PSV) attempts to maintain a minimum pressure

at the upstream node when the downstream node pressure is below the PSV
setting. If the downstream pressure is above the setting, flow through the
valve is unrestricted. Should the downstream nodal pressure exceed the
upstream nodal pressure, then the valve closes to prevent reverse flow. Note
that PSVs cannot be placed directly in series.

A Pressure Breaker Valve (PBV) forces a specified pressure loss to occur

across the valve. Flow can be in either direction through the valve.

A Flow Control Valve (FCV) limits the flow through a valve to a specified
amount. The program will produce a warning message if this flow cannot be
maintained without having to add additional head at the valve.

A Throttle Control Valve (TCV) is used to simulate partially closed valves by

adjusting the minor head loss coefficient of the valve. A relationship between
the degree to which the valve is closed and the resulting head loss coefficient
is typically available from the valve manufacturer.

A General Purpose Valve (GPV) provides the capability to model devices and
situations with unique headloss - flow relationships, such as reduced pres-
sure backflow prevention valves, turbines, and well drawdown behaviour. The
valve setting is the ID of a Headloss Curve.

SETTING (mandatory)
This data entry defines the pressure setting for PRVs, PSVs, and PBVs,
whose units are in psi or m. Or, this data entry defines the flow settings (in
user-defined flow units) for FCVs, or loss coefficients for TCVs.

36 MIKE URBAN - © DHI

Network Component Editors

When defining a pressure setting, the value specified is pressure (e.g., psi or
m) and not total head (or hydraulic gradeline elevation).

VALVE CURVE (water hammer mandatory)

This data entry is used to specify an valve characteristics curve for the water
hammer analysis. The valve characteristics curve is the relation between the
valve flow coefficient (Cd or Kv) versus the valve opening (%). See the sec-
tion titled “Specific Valve Data” on page 191 for additional information.

DIAMETER (mandatory)
This data entry is used to specify an valve diameter (inch or mm).

LOSS COEFFICIENT (optional)

This data entry specifies the minor loss coefficient for a fully opened valve.
The default loss coefficient is 0. Choosing «...» will display Select Minor Loss
Coefficient selection dialog box, allowing the user to select the appropriate
minor loss coefficient to use.

VALVE STATE (optional)

This drop down list box data entry allows you to define whether the valve is
unmarked (i.e. a valve already contained in the water distribution network), or
is marked (i.e. one that is being considered for future expansion, etc.). By
default, any valves added are unmarked.

CLOSED (optional)
This option check box data entry allows the user to toggle the OPEN and
CLOSED status of the valve. Choosing CLOSED effectively removes the
valve from the network system.

Valve Water Hammer Settings (optional)

MIKE URBAN is capable of modelling valves under the water hammer flow
conditions. Within the Valve Water Hammer Editor, the user can define the
required valve parameters as shown in Figure 1.11.

Figure 1.14 A valve water hammer settings

OPERATIONAL SCHEDULE (mandatory)

This data entry is used to select the valve operation schedule ID. The opera-
tional schedule is defined as valve opening (%) versus time. To define the
valve schedule, use the Curves and Relations dialog box.

37
 EPANET Input Description

Variable Valves
Many times, when performing an extended period simulation, it is desirable to
model a variable valve. A variable valve can vary its status to open or closed
during a simulation. A variable valve is modelled by defining the valve’s initial
status using the Valve Editor and then modifying the valve’s status during the
extended period simulation using the Control Editor. The Control Editor is
available from the Extended Menu.

See the section titled Control Editor or Rule Based Control Editor on -74 for
additional information.

Check Valves
Note that check valves are defined within the Pipe Editor.

Flow Direction
Note that the flow direction for valves is assumed to be from the starting
(upstream) junction node to the ending (downstream) junction node. The
order in which junction nodes are specified for each valve is critical for a
proper network representation of the valve. To reverse the order of the junc-
tion nodes, choose «Swap·Nodes».

Graphical Placement and Editing of Valves

From the Map window, the Add Valve tool can be used to place a Valve.
When you want to insert a Valve onto an existing pipe, you must insert it as a
new component. To insert a Valve onto the existing pipe:

1. Split the pipe by using Split tool from the Components toolbar.
2. Split the pipe once more to create a sections for 3 pipes.
3. Select the middle section by Select Feature tool and delete it by using
the Delete Selected tool.
4. Select the Valve tool from the Components floating toolbar and click on
the beginning pipe node in the Horizontal Plan window where you want
to place the valve and double click on the ending node. This will create a
new Valve.

Another method of using the Add Valve tool to add a Valve is, from the Map
window, select a starting node and then an ending node. Clicking on or near
an existing node, MIKE URBAN will then snap to the selected node and a
rubber-banding line will then be drawn from this node, representing the start-
ing node, while the user selects the ending node. After selecting the ending
node, MIKE URBAN will then place a Valve between the two selected nodes.

In addition, if the user clicks with the Add Valve tool in the Map window some-
where else other than a node or a pipe, MIKE URBAN will place a starting

38 MIKE URBAN - © DHI

Network Component Editors

node at the clicked location. A rubber-banding line will then be drawn from
this node, representing the Valve, while the user selects the ending node. If
the user clicks on or near an existing node, it will treat the selected node as
the ending node. If the user clicks anywhere else, it will place an ending node
at the selected location. If the user selects an existing pipe, an error message
will be displayed stating that an invalid ending node was selected.

To delete an existing Valve graphically, select the Valve using the Select tool
and then select «Delete Selected». The selected Valve will then be deleted.
To edit a Valve, click on the Valve using the ‘Open Editor’ tool. The Valve Edi-
tor will then be displayed allowing you to edit the attributes of the selected
Valve.

1.1.5 Tank Editor

Tank nodes are also placed at points in the water distribution model where a
water storage tank is located. Storage tanks can be defined as tanks with the
variable or fixed water level. The tank with the variable water level are mod-
elled as tanks where the water surface level changes with time as water flows
into and out of the tank. The tanks with the fixed water level represent places
(reservoir) within the water distribution model where a infinite source of water
(for the sake of the modelling simulation) is available. Hence, the reservoir
water level remains constant during the course of the simulation.

Tank nodes are either defined interactively on the graphical Map window
using the Add Tank tool, or by manual data entry using the Tank Editor dialog
box as shown in Figure 1.15. The Tank Editor allows you to define the reser-
voir’s ID, location, pressure zone, water surface elevation, and a description.
The Tank Editor dialog box is reached by selecting EPANET | Tanks.

39
 EPANET Input Description

Figure 1.15 The Tank Editor allows you to define the storage tank nodes that supply
water to the water distribution network

A list of the Tank Editor data entries for Figure 1.15 follows, with a short
description given for each entry.

ASSET ID (optional)
This data entry is used to specify a corresponding asset tank ID, which
uniquely identifies the tank node in the asset management system (such as
GIS, for example).

TANK ID (mandatory)
This data entry is used to specify an ID which uniquely identifies the tank
node. The tank ID acts as a unique lookup key that identifies the node from all
other nodes. A node can be a junction, reservoir, tank, or air-chamber. There-
fore, no two nodes may have the same ID. However, a node and a link (i.e.,
pipe, pump, or valve) can have the same ID. The node ID value can be any
string value (up to 40 characters).

A new tank ID is automatically suggested by MIKE URBAN whenever a new

tank node is placed into the list by pressing «Insert». When defining the tank
nodes graphically on the Map window using the Add Tank tool, the tank ID is
automatically defined.

When importing (or merging) multiple water distribution network models into a
single network model, MIKE URBAN will check for collisions between identi-
cal node IDs and will automatically assign a new node ID value for any node

40 MIKE URBAN - © DHI

Network Component Editors

being imported that contains the same node ID value as what already exists
in the network model.

X and Y COORDINATE (optional)

The X and Y data entries are used to define the physical (map) location of the
tank node, in units of ft. or m. This location definition is optional. In some
cases, the actual location of the tank node is not known—especially in older,
legacy networks. However, if the location is defined, then the junction node
will be displayed in the Map window. When defining the junction nodes graph-
ically on the Map window using the Add Tank tool, the X, Y location is auto-
matically entered.

DESCRIPTION (optional)
This data entry allows you to enter a description identifying the tank node
being entered. This description can be optionally displayed on the Map win-
dow and in reports generated by the Report Generator.

DATA SOURCE (optional)

This data entry is used to specify a corresponding asset data source, which
uniquely identifies the tank node location (such as database table or a data-
base file name) in the asset management system.

STATUS (optional)
This drop down selection list data entry allows you to define whether the tank
node is imported (i.e existing node was imported from the external data
source), or is inserted, modified, GIS, calibrated or similar. By default, tank
node status is undefined.

PRESSURE ZONE ID (optional)

This data entry allows you to define the ID of the pressure zone that the tank
node lies within. The pressure zone ID can be used when distributing the
zone demand as well as to graphically display different zones on the Map.

Selecting «...» allows you to display the Pressure Zone selection dialog box,
where the appropriate pressure zone ID can be selected.

TANK TYPE (mandatory)

This drop down selection list data entry allows you to define whether the tank
is modelled as reservoir (fixed HGL), or is tank (variable HGL).

TANK MIXING (optional)

MIKE URBAN allows the user to choose between four different types of tank
mixing, completely mixed, two compartment mixing, Last In First Out (LIFO)
and First In First Out (FIFO).

The Completely mixed model assumes that all water that enters a tank is
instantaneously and completely mixed with the water already in the tank. It is
the simplest form of mixing behaviour to assume, requires no extra parame-

41
 EPANET Input Description

ters to describe it, and seems to apply quite well to a large number of facilities
that operate in fill-and-draw fashion.

The Two-Compartment mixing model divides the available storage volume in

a tank into two compartments, both of which are assumed to be completely
mixed. The inlet/outlet pipes of the tank are assumed to be located in the first
compartment. New water that enters the tank mixes with the water in the first
compartment. If this compartment is full, then it sends its overflow to the sec-
ond compartment where it completely mixes with the water already stored
there. When water leaves the tank, it exits from the first compartment, which if
full, receives an equivalent amount of water from the second compartment to
make up the difference. The first compartment is capable of simulating short
circuiting between inflow and outflow while the second compartment can rep-
resent dead zones. The user must supply a single parameter which is the
fraction of the total tank volume devoted to the first compartment.

The First-In-First-Out (FIFO) Plug Flow mixing model assumes that there is
no mixing of water at all during its residence time in a tank. Water parcels
move through the tank in a segregated fashion where the first parcel to enter
is also the first to leave. Physically speaking, this model is most appropriate
for baffled tanks that operate with simultaneous inflow and outflow. There are
no additional parameters needed to describe this mixing model.

The Last-In-First-Out (LIFO) Plug Flow mixing model assumes that there is
no mixing between parcels of water that enter a tank. However in contrast to
FIFO Plug Flow, the water parcels stack up one on top of another, where
water enters and leaves the tank on the bottom. Physically speaking this type
of model might apply to a tall, narrow standpipe with an inlet/outlet pipe at the
bottom and a low momentum inflow. It requires no additional parameters be
provided.

REACTION COEFF. (optional)

Locally defined reaction rates. Please refer to section on reaction rates for
further.

BASE ELEVATION (mandatory)

This data entry defines the bottom elevation, in units of ft. or m, of the storage
tank where the water level in the storage tank is zero, as shown in
Figure 1.16. The default elevation is zero.

FIXED HGL (optional)

This data entry allows you to define the constant water head in case that the
tank is modelled as a reservoir with fixed water level. The water head is
defined in ft or m.

HGL PATTERN (optional)

The ID label of a time pattern used to model time variation in the tank’s (res-
ervoir's) total head. Leave blank if none applies. This property is useful if the

42 MIKE URBAN - © DHI

Network Component Editors

reservoir represents a tie-in to another system whose pressure varies with

time.

TANK SIZE (mandatory)

This option button data entry selects the type of storage tank being defined.
By default, a circular tank is defined. A the elevation-volume relationship for a
tank of variable geometry can also be defined. A Volume Curve determines
how storage tank volume (Y in cubic feet or cubic meters) varies as a function
of water level (X in feet or meters). It is used when it is necessary to accu-
rately represent tanks whose cross-sectional area varies with height. The
lower and upper water levels supplied for the curve must contain the lower
and upper levels between which the tank operates.

DIAMETER or WIDTH and LENGTH (mandatory)

This data entry allows you to define the tank chamber size (ft or m).

VOLUME CURVE (mandatory)

A Volume Curve determines how storage tank volume (Y in cubic feet or
cubic meters) varies as a function of water level (X in feet or meters). It is
used when it is necessary to accurately represent tanks whose cross-sec-
tional area varies with height.

MINIMUM LEVEL (mandatory)

This data entry defines the minimum level (or depth), in units of ft. or m, that
the water can drop to within the storage tank. The corresponding elevation is
equal to the base elevation plus the minimum level, as shown in Figure 1.16.

INITIAL LEVEL (mandatory)

This data entry defines the initial water surface level (or depth), in units of ft.
or m, that is used at the start of the simulation. The corresponding elevation is
equal to the base elevation plus the initial level, as shown in Figure 1.16.

MAXIMUM LEVEL (mandatory)

This data entry defines the maximum level (or depth), in units of ft. or m, that
the water can rise to within the storage tank. The corresponding elevation is
equal to the base elevation plus the maximum level, as shown in Figure 1.16.

INACTIVE VOLUME (optional)

This data entry defines the volume of inactive water contained between the
minimum level and the base elevation, in units of ft3 or m3, of the storage
tank, as shown in Figure 1.16.

43
 EPANET Input Description

Figure 1.16 Definition of storage tank levels

Graphical Placement and Editing of Tanks

From the Map window, the Add Tank tool can be used in different ways to
place a tank. If the user clicks on or near an existing junction node, MIKE
URBAN will replace the selected node by a new tank.

If the user clicks with the Add Tank tool in the Map window somewhere else
other than a node or a pipe, MIKE URBAN will place a tank node at the
clicked location. If the user selects an existing pipe, a program message will
be displayed allowing to split the pipe at the cursor location.

If desired, from the Map window, the user can move an existing tank using
the Move Selected tool. By selecting the tank and holding down the left
mouse button, the tank can be dragged to a new location. As the tank is
dragged, its connecting links (i.e., pipes, pumps, and valves) rubber band
along with it.

To delete an existing tank graphically, select the tank using the Select tool
and then select «Delete Selected». The selected tank will then be deleted. To
edit a tank, click on the tank using the Open Editor tool. The Tank Editor will
then be displayed allowing you to edit the attributes of the selected tank.

1.1.6 Air-Chamber Editor

Air-Chamber nodes are also placed at points in the water distribution model
where a water storage Air-Chamber is located. Storage Air-Chambers can be
defined as Air-Chambers with the variable or fixed water level. The Air-Cham-
ber with the variable water level are modelled as Air-Chambers where the

44 MIKE URBAN - © DHI

Network Component Editors

water surface level changes with time as water flows into and out of the Air-
Chamber. The Air-Chambers with the fixed water level represent places (res-
ervoir) within the water distribution model where a infinite source of water (for
the sake of the modelling simulation) is available. Hence, the reservoir water
level remains constant during the course of the simulation.

Air-Chamber nodes are either defined interactively on the graphical Map win-
dow using the ‘Add Air-Chamber’ tool, or by manual data entry using the Air-
Chamber Editor dialog box as shown in Figure 1.15. The Air-Chamber Editor
allows you to define the reservoir’s ID, location, pressure zone, water surface
elevation, and a description. The Air-Chamber Editor dialog box is reached
by selecting EPANET | Air-Chambers.

Figure 1.17 The Air-Chamber Editor allows you to define the storage tank nodes
that supply water to the water distribution network

A list of the Air-Chamber Editor data entries for Figure 1.15 follows, with a
short description given for each entry.

AIR-TANK ID (mandatory)
This data entry is used to specify an ID which uniquely identifies the air-
chamber node. The tank ID acts as a unique lookup key that identifies the
node from all other nodes. A node can be a junction, reservoir, tank, or air-
chamber. Therefore, no two nodes may have the same ID. However, a node

45
 EPANET Input Description

and a link (i.e., pipe, pump, or valve) can have the same ID. The node ID
value can be any string value (up to 40 characters).

A new air-chamber ID is automatically suggested by MIKE URBAN whenever

a new air-chamber node is placed into the list by pressing «Insert». When
defining the tank nodes graphically on the Map window using the Add Air-
Chamber tool, the air-chamber ID is automatically defined.

When importing (or merging) multiple water distribution network models into a
single network model, MIKE URBAN will check for collisions between identi-
cal node IDs and will automatically assign a new node ID value for any node
being imported that contains the same node ID value as what already exists
in the network model.

X and Y COORDINATE (optional)

The X and Y data entries are used to define the physical (map) location of the
air-chamber node, in units of ft. or m. This location definition is optional. In
some cases, the actual location of the air-chamber node is not known—espe-
cially in older, legacy networks. However, if the location is defined, then the
junction node will be displayed in the Map window. When defining the junction
nodes graphically on the Map window using the Add Air-Chamber tool, the
X, Y location is automatically entered.

DESCRIPTION (optional)
This data entry allows you to enter a description identifying the air-chamber
node being entered. This description can be optionally displayed on the Map
window and in reports generated by the Report Generator.

STATE (optional)
This drop down selection list data entry allows you to define whether the air-
chamber node is imported (i.e existing node was imported from the external
data source), or is inserted, modified, GIS, calibrated or similar. By default,
tank node status is undefined.

PRESSURE ZONE ID (optional)

This data entry allows you to define the ID of the pressure zone that the air-
chamber node lies within. The pressure zone ID can be used when distribut-
ing the zone demand as well as to graphically display different zones on the
Map.

Selecting «...» allows you to display the Pressure Zone selection dialog box,
where the appropriate pressure zone ID can be selected.

POLYTROPIC EXPANSION (mandatory)

This data entry allows you to define the polytropic exponent in the polytropic
gas equation.

By default this data entry is k=1.2. See the section titled “Specific Valve Data”
on page 191 for more information.

46 MIKE URBAN - © DHI

Network Component Editors

BASE ELEVATION (mandatory)

This data entry defines the bottom elevation, in units of ft. or m, of the storage
tank where the water level in the storage tank is zero, as shown in
Figure 1.16. The default elevation is zero.

TANK SIZE (mandatory)

This option button data entry selects the type of storage tank being defined.
By default, a circular tank is defined. A the elevation-volume relationship for a
tank of variable geometry can also be defined. A Volume Curve determines
how storage tank volume (Y in cubic feet or cubic meters) varies as a function
of water level (X in feet or meters). It is used when it is necessary to accu-
rately represent tanks whose cross-sectional area varies with height. The
lower and upper water levels supplied for the curve must contain the lower
and upper levels between which the tank operates.

DIAMETER or WIDTH and LENGTH (mandatory)

This data entry allows you to define the tank chamber size (ft or m).

VOLUME CURVE (mandatory)

A Volume Curve determines how storage tank volume (Y in cubic feet or
cubic meters) varies as a function of water level (X in feet or meters). It is
used when it is necessary to accurately represent tanks whose cross-sec-
tional area varies with height.

MINIMUM LEVEL (mandatory)

This data entry defines the minimum level (or depth), in units of ft. or m, that
the water can drop to within the storage tank. The corresponding elevation is
equal to the base elevation plus the minimum level, as shown in Figure 1.16.

INITIAL LEVEL (mandatory)

This data entry defines the initial water surface level (or depth), in units of ft.
or m, that is used at the start of the simulation. The corresponding elevation is
equal to the base elevation plus the initial level, as shown in Figure 1.16.

MAXIMUM LEVEL (mandatory)

This data entry defines the maximum level (or depth), in units of ft. or m, that
the water can rise to within the storage tank. The corresponding elevation is
equal to the base elevation plus the maximum level, as shown in Figure 1.16.

Graphical Placement and Editing of Air-Chambers

From the Map window, the Add Tank tool can be used in different ways to
place a air-chamber. If the user clicks on or near an existing junction node,
MIKE URBAN will replace the selected node by a new air-chamber.

If the user clicks with the Add Air-Chamber tool in the Map window some-
where else other than a node or a pipe, MIKE URBAN will place a air-cham-
ber node at the clicked location. If the user selects an existing pipe, a

47
 EPANET Input Description

program message will be displayed allowing to split the pipe at the cursor
location.

If desired, from the Map window, the user can move an existing air-chamber
using the Move Selected tool. By selecting the air-chamber and holding down
the left mouse button, the air-chamber can be dragged to a new location. As
the tank is dragged, its connecting links (i.e., pipes, pumps, and valves) rub-
ber band along with it.

To delete an existing air-chamber graphically, select the air-chamber using

the Select tool and then select «Delete Selected». The selected tank will then
be deleted. To edit a air-chamber, click on the air-chamber using the Open
Editor tool. The Tank Editor will then be displayed allowing you to edit the
attributes of the selected air-chamber.

1.2 Other Editors

To help develop and setup a model, MIKE URBAN provides the user with
several other editors including Project Options Editor, Report Editor, Water
Source Editor, Energy Editor, Pressure Zone Editor, and others. This section
discusses how to use these editors.

1.2.1 Project Options Editor

The project options dialog allows you to specify project configuration informa-
tion for the pipe network system to be modelled. Project configuration data
includes analysis type, friction loss formulation, simulation options, analysis
options, and flowrate units. Alternatively, to display the Project Options dialog
box, select EPANET | Project Options.

Figure 1.18 Project Options dialog box

The Project Options Editor dialog is used to define the project parameters.

48 MIKE URBAN - © DHI

Other Editors

TITLE (Optional)
This data entry allows you to enter a describe the current project. This
description can be output in reports generated by the Report Generator.

PROJECT TYPE (Mandatory)

The project type is used to define the simulation type of the project.

 Steady State Hydraulics.

EPANET based steady state analysis.

 Extended Period Hydraulics.

EPANET based extended period simulation.

 Extended Period Water Quality.

 EPANET based extended period simulation with the water quality option
such as No Water Quality Analysis (no quality analysis), Chemical con-
centrations (compute chemical concentration), Water Age (compute
water age), Source Tracing (trace flow from a specific node).

 Water hammer
UNITS (Mandatory)

This selection data entry allows you to select one of the predefined project
units environments such as CFS (cubic feet per second) or LPS (liter per sec-
ond), for example.

HEAD LOSSES (Mandatory)

This selection data entry allows you to specify the formula used to compute
headloss as a function of flow rate in a pipe. Choices are Hazen-Williams,
Darcy-Weisbach, and Chezy-Manning.

PROPERTIES (Optional)
These data entries allow you to determine the hydraulic and water quality
behaviour of the pipe network should be analysed.

 Specific Gravity
This data entry specifies the specific gravity of the fluid at the tempera-
ture condition being simulated. This data entry allows fluids other than
water to be simulated. Gravity is the weight per unit volume of the fluid
being modelled relative to water. Specific gravity is the ratio of the den-
sity of the fluid being modelled to that of water at 4 deg. C. (unitless).

 Viscosity
This data entry specifies the kinematic viscosity of the fluid at the tem-
perature condition being simulated. The units of viscosity are ft2/sec (or
m2/sec for SI units). The viscosity is the kinematic viscosity of the fluid

49
 EPANET Input Description

being modelled relative to that of water at 20 deg. C (1.0 centistoke). The

default value is 1.0.

 Molecular Diffusivity
This data entry specifies the molecular diffusivity of the chemical being
tracked. The diffusivity is the molecular diffusivity of the chemical being
analysed relative to that of chlorine in water. The default value is 1.0. Dif-
fusivity is only used when mass transfer limitations are considered in
pipe wall reactions. A value of 0 will cause MIKE URBAN to ignore mass
transfer limitations.

 Emitter Exponent
Power to which pressure is raised when computing the flow through an
emitter device. The textbook value for nozzles and sprinklers is 0.5. This
may not apply to pipe leakage.

CONVERGENCE (Optional)
These data entries allow you to determine the hydraulic and water quality
behaviour of the pipe network should be analysed.

 Maximum numbers of trials

 Accuracy
Convergence criterion used to signal that a solution has been found to
the nonlinear equations that govern network hydraulics. Trials end when
the sum of all flow changes divided by the sum of all link flows is less
than this number. Suggested value is 0.001.

 Water Quality Tolerance

Smallest change in quality that will cause a new parcel of water to be
created in a pipe. A typical setting might be 0.01 for chemicals measured
in mg/L as well as water age and source tracing. The Quality Tolerance
determines when the quality of one parcel of water is essentially the
same as another parcel. For chemical analysis this might be the detec-
tion limit of the procedure used to measure the chemical, adjusted by a
suitable factor of safety. Using too large a value for this tolerance might
affect simulation accuracy. Using too small a value will affect computa-
tional efficiency.

 Maximum Number of Segments

Maximum number of segments, which could be generated for a pipe dur-
ing the water quality analysis.

 Unbalanced System
Action to take if a hydraulic solution is not found within the maximum
number of trials. Choices are STOP to stop the simulation at this point or

50 MIKE URBAN - © DHI

Other Editors

CONTINUE to use another 10 trials, with no link status changes allowed,

in an attempt to achieve convergence.

 CHECKFREQ
This sets the number of solution trials that pass during hydraulic balanc-
ing before the status of pumps, check valves, flow control valves and
pipes connected to tanks are once again updated. The default value is 2,
meaning that status checks are made every other trial. A value equal to
the maximum number of trials would mean that status checks are made
only after a system has converged. (Whenever a status change occurs
the trials must continue since the current solution may not be balanced.)
The frequency of status checks on pressure reducing and pressure sus-
taining valves (PRVs and PSVs) is determined by the DAMPLIMIT option
(see below).

 MAXCHECK
MAXCHECK is the number of solution trials after which periodic status
checks on pumps, check valves flow control valves and pipes connected
to tanks are discontinued. Instead, a status check is made only after con-
vergence is achieved. The default value is 10, meaning that after 10 tri-
als, instead of checking status every CHECKFREQ trials, status is
checked only at convergence.

 DAMPLIMIT
This is the accuracy value at which solution damping and status checks
on PRVs and PSVs should begin. Damping limits all flow changes to
60% of what they would otherwise be as future trials unfold. The default
is 0 which indicates that no damping should be used and that status
checks on control valves are made at every iteration. Damping might be
needed on networks that have trouble converging, in which case a limit
of 0.01 is suggested.

WATER HAMMER (Optional)

These data entries allow you to determine the water behaviour of the pipe
network should be analysed.

 Theta
Numeric parameter used to centre the high order finite difference
scheme in time.

 Alpha

 Gravity
Gravity acceleration.

 Atmospheric pressure
The pressure of the atmosphere outside the pressurized pipeline system.

51
 EPANET Input Description

 Vapour cavity pressure

The pressure under which the gas particles starts to develop inside the
water pipeline.

 Temperature
Temperature of water.

1.2.2 Pressure Zone Editor

Pressure zones are service areas defined by the hydraulic gradeline value of
the sources that supply them. A pressure zone has one or more sources of
supply and may have a set of closed valves that separate it from other pres-
sure zones. The Pressure Zone Editor dialog box is reached by selecting
EPANET | Pressure Zones.

Figure 1.19 Pressure Zone Editor dialog box

The Pressure Zone Editor dialog is used to define the pressure zones for the
pipe network system.

PRESSURE ZONE ID (mandatory)

This data entry is used to define a unique positive integer ID value that speci-
fies the network pressure zone. Pressure zones are defined at junction
nodes, storage tanks and reservoirs. Note that, by default, all nodes are
defined as belonging to pressure zone 1. Therefore, pressure zone 1 is
always defined and not listed within the Pressure Zone Editor.

DESCRIPTION (Optional)
This data entry allows you to enter a description identifying the pressure zone
being defined. This description can be output in reports generated by the
Report Generator.

DEMAND (optional)
This data entry is used to specify the zone demand, which can be used for
automatic demand distribution. To distribute the zone demand to junction
nodes located within the zone, use Demand Distribution dialog.

52 MIKE URBAN - © DHI

Other Editors

See “Distributed Demands” on page 59 for more details.

1.2.3 Energy Editor

Computation of pumping energy and cost can be performed in MIKE URBAN.

Users can either choose pump-specific efficiency curves and electricity rate
schedules or can select a set of default values for these computations. The
Energy Editor dialog box is reached by selecting EPANET | Energy.

Figure 1.20 Energy Editor dialog box

The user can compute pumping energy and cost using Edit Energy Editor.
The following parameters can be defined in Energy Editor:

GLOBAL PRICE (optional)

Average cost per kW/hour.

GLOBAL PRICE PATTERN (optional)

ID label of time pattern describing how energy price varies with time.

PUMP EFFICIENCY (optional)

Either a single percent efficiency for global setting or the ID label of an effi-
ciency curve for a specific pump.

DEMAND CHARGE (optional)

Added cost per maximum kW usage during the simulation period.

1.2.4 Report Editor

Detailed report can be generated by EPANET engine when performing

steady-state, extended period, and water quality analysis. The report param-
eters and settings can be specified in the Report Editor. Users can either
choose reporting simulation status including error and warning messages or
include the full simulation results of hydraulic and water quality parameters
for each node, link, and time level. The Report Editor dialog box is reached by
selecting EPANET | Reports.

53
 EPANET Input Description

Figure 1.21 Reports Editor dialog box

The following parameters can be defined in Reports Editor:

CONDITION (optional)
The condition data entry is used to describe the contents of the output report
produced from a simulation. The following keywords can be used:

STATUS (YES/NO/FULL)
STATUS determines whether hydraulic status messages are written to the
Report file. If YES is selected the messages will identify those network com-
ponents that change status during each time step of the simulation. If FULL is
selected, then convergence information will also be included from each trial
of each hydraulic analysis. This level of detail is only useful for de-bugging
networks that become hydraulically unstable.

SUMMARY (YES/NO)
SUMMARY determines whether a summary table of number of network com-
ponents and key analysis options is generated. The default is YES.

MESSAGES (YES/NO)
MESSAGES determines whether error and warning messages generated
during a hydraulic/water quality analysis are written to the Report file. The
default is YES.

ENERGY (YES/NO)
ENERGY determines if a table reporting average energy usage and cost for
each pump is provided. The default is NO.

NODES (NONE/ALL/node1 node2...)

NODES identifies which nodes will be reported on. You can either list individ-
ual node ID labels or use the keywords NONE or ALL. Additional NODES
lines can be used to continue the list. The default is NONE.

54 MIKE URBAN - © DHI

Other Editors

LINKS (NONE/ALL/link1 link2...)

LINKS identifies which links will be reported on. You can either list individual
link ID labels or use the keywords NONE or ALL. Additional LINKS lines can
be used to continue the list. The default is NONE.

Variable (YES/NO) or Variable (BELOW/ABOVE/PRECISION, value)

This reporting option is used to identify which variables are reported on, how
many decimal places are displayed, and what kind of filtering should be used
to limit output reporting. Node variables that can be reported on include:

 Elevation
 Demand
 Head
 Pressure
 Quality

Link variables include:

 Length
 Diameter
 Flow
 Velocity
 Headloss
 LinkQuality
 LinkStatus
 Setting (Roughness for pipes, speed for pumps, pressure/flow setting for
valves)
 Reaction (reaction rate)
 F-Factor (friction factor)

The default reporting variables are Demand, Head, Pressure, and Quality for
nodes and Flow, Velocity, and Headloss for links. The default precision is two
decimal places.

Example:
The following example reports on nodes N1, N2, N3, and N17 and all links
with velocity above 3.0. The standard node variables (Demand, Head, Pres-
sure, and Quality) are reported on while only Flow, Velocity, and F-Factor
(friction factor) are displayed for links.

NODES N1 N2 N3 N17

LINKS ALL

55
 EPANET Input Description

FLOW YES

VELOCITY PRECISION 4

F-FACTOR PRECISION 4

VELOCITY ABOVE 3.0

1.2.5 Curves and Relations Editor

The user can define data curves and their X, Y coordinate points in The
Curves and Relations Editor. The Curves and Relations Editor dialog box is
reached by selecting EPANET | Curves and Relations.

The following curves can be used to represent relations:

 Pump Efficiency. Efficiency versus flow for pumps.

 Valve Head Loss. Head Loos versus flow for GPV General Purpose
Valve.
 Pump Q-H Curve. Head versus flow for pumps.
 Tank Depth-Volume Curve. Volume versus depth for tanks.
 Water Source Price. Production water costs versus produced volume.
 Transient Q-Boundary. Inflow/outflow at the boundary node versus time
(only for Water Hammer Analysis).
 Transient H-Boundary. Hydraulic Grade Line at the boundary node ver-
sus time (only for Water Hammer Analysis).
 Valve Operation Schedule. Valve opening versus time (only for Water
Hammer Analysis).
 Valve Characteristics Cd. Valve flow coefficient Cd versus time (only for
Water Hammer Analysis).
 Valve Characteristics Kv. Valve flow coefficient Kd versus time (only for
Water Hammer Analysis).
 Dual-acting Valve Characteristics. Volume of air versus pressure differ-
ence (only for Water Hammer Analysis).
 Pump Operational Schedule. Pump speed versus time (only for Water
Hammer Analysis).
 Pump Torque. Pump torque versus flow (only for Water Hammer Analy-
sis).
 Motor Torque. Motor torque versus pump speed (only for Water Hammer
Analysis).
 PID Set Point Value Curve. Set point setting versus fraction of a day
(only for RTC Real-time Control Analysis).

56 MIKE URBAN - © DHI

Other Editors

Figure 1.22 Curve Editor Window, Definition of data curves

Figure 1.23 Preview of a defined curve in Curve Editor

The points of a curve must be entered in order of increasing X-values (lowest

to highest).

57
 EPANET Input Description

1.2.6 Boundary Conditions Editor

Boundary conditions are used to define water level and flow changes during
the water hammer analysis.

There are in principle two types of boundary conditions, namely the piezo-
metric head, H, above a specified datum, e.g., in tanks, and the discharge, Q,
e.g., water demand. Both H and Q are given under selected names as time
series in the Curves and Relations Editor and stored in the database. These
boundary conditions may be assigned to any node in the network. The
boundary set point value for each time step is assigned from the correspond-
ing time series curve. If time step used by water hammer computation is
smaller than appropriate neighbouring values in boundary conditions time
series then linear interpolation is applied. There are nodes of the following
types:

 H (Water head) boundary.

 Q (Flow) boundary.
 Compatibility (default type).
 Hydraulic structure (storage tank, air-chamber).

The water hammer analysis ignores the diurnal patterns defined for the
steady state and extended period simulation. Therefore, the boundary condi-
tions using time series must be specified at nodes where the flow or water
head changes need to be defined during the water hammer simulation. The
Boundary Conditions Editor dialog box is reached by selecting EPANET |
Water Hammer | Boundary Conditions.

Figure 1.24 Boundary Conditions Editor dialog box

The following parameters can be defined in Boundary Conditions Editor:

NODE TYPE (optional)

This pull-down selection list allows the user to select what type of node (i.e.,
junction or tank) the point constituent is being specified for.

58 MIKE URBAN - © DHI

Network Demand

NODE ID (mandatory)
This data entry is used to define the ID of the boundary node the boundary
condition is being assigned to. Choosing «...» will display the Select Node
selection dialog box from which the user can select the appropriate node type
and ID. Or, choosing «Pick» allows the user to graphically select the node
from the Map window.

BOUNDARY TYPE (mandatory)

This data entry is used to define the boundary type. Head versus time and
flow versus time boundary types are available for selecting from. Note that
constant head or flows (demands) do not need to be defined.

BOUNDARY CURVE (mandatory)

This data entry is used to define the boundary curve. Choosing «...» will dis-
play the Select Curve selection dialog box from which the user can select the
appropriate boundary curve.

DESCRIPTION (optional)
This data entry allows you to enter a description identifying the boundary con-
dition being defined. This description can be output in reports generated by
the Report Generator.

1.3 Network Demand

Network demand for water is assigned at junction nodes, on a node by node
basis. To help develop a model, MIKE URBAN WATER allows the user to
automatically define the nodal demand at all of the nodes within a model, or
within a pressure zone, based upon the total demand on the system or pres-
sure zone. This section discusses how MIKE URBAN can automatically dis-
tribute this demand to the network system.

Typically in large network systems, the pipe network is broken up into differ-
ent pressure zones (or distribution zones). Since pressure is related to
ground elevation, a network system covering hilly or mountainous terrain will
have more pressure zones than one covering fairly flat terrain. The section
also discusses how MIKE URBAN defines pressure zones.

1.3.1 Distributed Demands

Network demands are defined at junction nodes, on a node by node basis.

For large network systems, assigning this demand data can be a very tedious
job. Since many times the total demand is known for a particular network
pressure zone or for the entire network system, MIKE URBAN provides the
capability to distribute this total demand among the applicable junction nodes.

Pipe Demand Coefficients

MIKE URBAN computes the water demands for each node in the network
system based upon the total network demand using two methods: the Method

59
 EPANET Input Description

of Pipe Lengths and the Method of Two Coefficients. This is useful when
assigning the nodal water demand for a large network, since the software will
automatically proportion the total network demand based upon one of these
two methods. These methods are used to mimic the amount of actual
demand along a pipe, based upon the pipe length or a pre-defined demand
coefficient.

The Distributed Demands dialog box, as shown in Figure 1.25, is used to

automatically assign the demands at the appropriate junction nodes. The Dis-
tributed Demands dialog box is reached by selecting EPANET | Distributed
Demands.

Figure 1.25 The Distributed Demands dialog box computes the demand at each
junction node automatically based upon the total demand within a pres-
sure zone or entire network system

A list of the Distributed Demands data entries for Figure 1.25 follows, with a
short description given for each entry.

TOTAL NETWORK WATER DEMAND (mandatory)

This data entry is used to specify the total network demand for a particular
network pressure zone or the entire network system. The flow units are user-
specified.

Note that this total demand represents the total demand regardless into which
multiple junction demand is distributed. Multiple demands are specified in the
Junction Editor or in the Multiple Demand Editor.

NODE PRESSURE ZONE ID, PIPE PRESSURE ZONE ID (optional)

This check box allows you to select whether the total network water demand
corresponds to the entire network or a single pressure zone depending on

60 MIKE URBAN - © DHI

Network Demand

whether the pressure zone ID is assigned to nodes or pipes. Checking this

box applies the specified water demand to a single specified pressure zone.
Unchecking this box applies the specified water demand to the entire water
distribution network.

The pressure zone must be specified in the provided data entry field. Select-
ing «...» displays the Pressure Zone selection dialog box, where the appropri-
ate pressure zone ID can be selected.

METHOD OF TWO COEFFICIENTS or METHOD OF REDUCED

PIPE LENGTHS or METHOD OF EQUIVALENT PIPE LENGTHS
MIKE URBAN allows the user to compute the nodal water demands based
upon the total network demand using two methods: the Method of Pipe
Lengths and the Method of Two Coefficients. This radio button group allows
the user to select the method to be used.

Selecting the Method of Two Coefficients, MIKE URBAN computes the total
water demand assigned to each pipe (which is then split between the starting
and ending nodes) as:

 Q k li k 2i
q pi = ---------------------
- (1.4)
  k li k 2i 

Selecting the Method of Reduced Pipe Lengths, MIKE URBAN computes the
total water demand assigned to each pipe (which is then split between the
starting and ending nodes) as:

 Q l i k li
q pi = -----------------
- (1.5)
  k li l i 

Selecting the Method of Equivalent Pipe Lengths, MIKE URBAN computes

the total water demand assigned to each pipe (which is then split between the
starting and ending nodes) as:

 Q l i k Di
q pi = -------------------
- (1.6)
  k Di l i 

where:

qpi =Total water demand applied to the pipe, split between the two end nodes.

Q = Total network water

li = Pipe length

k1i, k2i = pipe demand coefficient

61
 EPANET Input Description

kDi = pipe demand coefficient is calculated by the program as a factor, calcu-

lated as pipe diameter/diameter_normal (where diameter normal is 150mm or
6inch). This helps to scale the pipes based on their diameter i.e. perimeter;
this method is recommended when the distributed demand corresponds to
the amount of leakage.

These demand coefficients are defined for each pipe using the Pipe Editor, as
described on page -18. The computed demands, which are assigned once
«Compute» is selected, are stored at each individual node. These demands
are stored in the Junction Editor, as described on page -12. Selecting
«Reset» causes all of the nodal demand entries to be set to zero for the
entire network, but it will leave the additional demand entries that are defined
unaffected.

SELECT PIPE DEMAND COEFFICIENT 1,2 (mandatory)

This list box data entry allows you to specify the demand coefficient, which
will be used for the demand coefficient 1 or 2. There are four possible pipe
demand coefficients, which can be defined for each pipe.

TARGET DEMAND (mandatory)

This data entry allows you to specify the multiple demand category, which will
be used as a target demand for the distributed demand. If the multiple
demand with the specified category does not exist, the program will create it
and it will override the existing values in case that such demand category
already exists for each node used in the demand distribution.

Node Demand Coefficients

Node demand coefficient allows you for each node to define the share from
the whole network demand, which is taken by that node. The total network
demand is then distributed to the corresponding junction nodes by Demand
Distribution function.

This option will only assign demand to nodes with Demand Coefficient
applied (different from 0 or NULL). In the case of an equal distribution, the
node demand coefficients have to be equal and different from zero.

TOTAL NETWORK WATER DEMAND (mandatory)

This data entry is used to specify the total network demand for a particular
network pressure zone or the entire network system. The flow units are user-
specified.

Note that this total demand represents the total demand regardless into which
multiple junction demand is distributed. Multiple demands are specified in the
Junction Editor or in the Multiple Demand Editor.

NODE PRESSURE ZONE ID (optional)

This check box allows you to select whether the total network water demand
corresponds to the entire network or a single pressure zone. Checking this
box applies the specified water demand to a single specified pressure zone.

62 MIKE URBAN - © DHI

Network Demand

Unchecking this box applies the specified water demand to the entire water
distribution network.

The pressure zone must be specified in the provided data entry field. Select-
ing «...» displays the Pressure Zone selection dialog box, where the appropri-
ate pressure zone ID can be selected.

EQUAL NODE DEMAND DISTRIBUTION (optional)

This check box allows you to distribute the network (or zone) demand equally
to each node within the zone or the network.

Selecting Equal Node Demand Distribution computes the junction demand:

Q
q ni = ---- (1.7)
N

where:

Q= Total network water demand (or zone demand)

qni = calculated demand at each junction node

N = junction nodes count with the selected zone or a total network

TARGET DEMAND (mandatory)

This data entry allows you to specify the multiple demand category, which will
be used as a target demand for the distributed demand. If the multiple
demand with the specified category does not exist, the program will create it
and it will override the existing values in case that such demand category
already exists for each node used in the demand distribution.

Selecting the Method of Node Demand Coefficients, MIKE URBAN computes

the total water demand assigned to each junction node as:

QT
-C
q i = ----------------- (1.8)
i
 i C
i = 0,n

where:

qi = node demand

Qt = total network demand

Ci = demand coefficient

63
 EPANET Input Description

1.3.2 Demand Allocation Editor

Junction node demands can be edited either within the junction editor for
each particular junction node or in the Multiple Demand editor, which allows
the user to display and edit all multiple demands. In addition to that, MIKE
URBAN provides another way of developing junction node demands, which is
based on geocoding the consumption data to the appropriate junctions or
pipes and aggregating their set point demand values to the junction
demands. This allows simplifies the demand development process and
allows you, for example, to import consumption data from the consumption
database systems and geocoded it based on its X, Y geographical coordi-
nates.

The Demand Allocation Editor dialog box is reached by selecting EPANET |

Demand Allocation.

Figure 1.26 Demand Allocation Editor dialog box is used to store and edit the con-
sumption data defined as consumption points and link it to he appropri-
ate network junctions or pipes

A list of the Demand Allocation Editor data entries for Figure 1.26 follows,
with a short description given for each entry.

REFERENCE ID (optional)
This data entry is used to specify an ID which uniquely identifies the demand
point. The reference ID acts as a unique lookup key that identifies the
demand point from all other demand points. The reference ID value can be
any string value (up to 40 characters). It is recommended that this reference

64 MIKE URBAN - © DHI

Network Demand

ID corresponds to the asset ID, which uniquely identifies the demand point in
the customer information or billing database system.

PRESSURE ZONE ID (optional)

This data entry allows you to define the ID of the pressure zone that the
demand point lies within. The pressure zone ID can be used when distributing
the zone demand as well as to graphically display different zones on the Map.
Selecting «...» allows you to display the Pressure Zone selection dialog box,
where the appropriate pressure zone ID can be selected.

X, Y COORDINATE (mandatory)
The X and Y data entries are used to define the physical (map) location of the
demand point, in units of ft. or m.

JUNCTION ID or PIPE ID (mandatory)

This data entry is used to specify the node or pipe ID, to which the demand
point is linked to. The user can specify this junction or pipe ID by editing or to
use Geocode function, which will automatically geocode (link) each demand
point to the appropriate junction node or a pipe (such as nearest node, for
example).

DEMAND (mandatory)
This data entry is used to specify the demand value, which will be then used
in the process of the demand aggregation. This demand set point value can
be imported from the external database systems (such as CIS Customer
Information System, for example) or it can be developed from the minimum,
average, or maximum demand values.

MINIMUM DEMAND (optional)

This data entry is used to specify the minimum demand value, which can be
then assigned to demand for the process of the demand aggregation. The
minimum demand value can be imported from the external database systems
(such as CIS Customer Information System, for example) or can be devel-
oped within MU by applying a minimum hour factor to the average demand,
for example.

AVERAGE DEMAND (optional)

This data entry is used to specify the average demand value, which can be
then assigned to demand for the process of the demand aggregation. The
average demand value can be imported from the external database systems
(such as CIS Customer Information System, for example) and can be used
within MU for development of actual demand, minimum demand, or maxi-
mum demand.

MAXIMUM DEMAND (optional)

This data entry is used to specify the maximum demand value, which can be
then assigned to demand for the process of the demand aggregation. The
maximum demand value can be imported from the external database sys-
tems (such as CIS Customer Information System, for example) or can be

65
 EPANET Input Description

developed within MU by applying a maximum hour factor to the average

demand, for example.

DESCRIPTION (optional)
This data entry allows you to enter a description identifying the consumption
point being defined. This description can be output in reports generated by
the Report Generator.

DEMAND CATEGORY (optional)

This data entry is used to specify the demand category such as residential,
commercial, or leakage, for example. The demand category can be then
used in the process of demand aggregation when demands belonging to the
same junction node or a pipe are aggregated based on their demand cate-
gory.

CATEGORY TYPE (optional)

This data entry allows you to enter a category type identifying the consump-
tion point being defined.

ADDRESS (optional)
This data entry allows you to enter an address identifying the consumption
point being defined. This field can be output in reports generated by the
Report Generator.

OWNER (optional)
This data entry allows you to enter an owner name identifying the consump-
tion point being defined. This field can be output in reports generated by the
Report Generator.

EQUIVALENT PERSON(optional)
This data entry is used to specify the equivalent person value for each
demand point. The EP value can be used when calculating the actual junction
demand, which will be then used in the process of demand aggregation.

EQUIVALENT TENEMENT (optional)

This data entry is used to specify the equivalent tenement value for each
demand point. The EP value can be used when calculating the actual junction
demand, which will be then used in the process of demand aggregation.

Geocoding
Demand geocoding is an automatic process, where each demand point is
linked to the appropriate junction node or a pipe (such as nearest node, for
example). The user can also specify the linked junction or pipe ID manually
by editing or importing from external database systems, such as GIS or CIS.
There are several different geocoding methods available. The geocoding is
done from a wizard. Press the ‘Geocode’ button on the Demand Allocation
dialog to activate the wizard.

66 MIKE URBAN - © DHI

Network Demand

Figure 1.27 Demand geocoding wizard

Different options are available for the geocoding process.

Figure 1.28 The different geocode options

 To nearest node

 To node by nearest pipe

The nearest pipe is located first and its pipe diameters can be optionally
checked. The nearest of the pipe beginning or ending nodes is used as a
demand node.

 To node by Pipe ID
The nearest pipe is already specified and the nearest of the pipe begging
or ending nodes is used as a demand node.

 To nearest pipe

67
 EPANET Input Description

MAXIMUM PIPE DIAMETER

It is possible to restrict the maximum diameter of pipes, used during the
demand geocoding. This allows us to ignore long-distance pipes, for exam-
ple.

Aggregation
Demand aggregation is used to develop node demands for junction nodes
based on the geocoded node or link ID/

Figure 1.29 Demand Aggregation dialog box

AGGREGATE DEMANDS TO NODE DEMANDS or ASSIGN

DEMANDS TO MULTIPLE DEMANDS or AGGREGATE DEMANDS
TO PIPE DEMAND COEFFICIENTS (mandatory)

 Aggregate demands to node demands

Multiple node demands are developed by aggregating demand points to
the appropriate junction nodes based on the matching junction ID. This
option only works for demand allocations connected to nodes, not to
pipes.

 Assign demands to multiple demands

Multiple node demands are created for demand points. For each
demand point, the corresponding record is created in the multiple
demands editor. Use this method is you want to populate multiple
demands for demand points. This option only works for demand alloca-
tions connected to nodes, not to pipes.

 Aggregate demands to pipe demand coefficients

Pipe demand coefficient is calculated by aggregating demand points to
the appropriate pipes based on the matching pipe ID. This option only

68 MIKE URBAN - © DHI

Network Demand

works for demand allocations connected to pipes. Note that this does not
add to the Multiple Demands table (as is done for allocation assigned to
nodes), but sets the demand coefficients of the pipe.

SELECT PIPE DEMAND COEFFICIENT (optional)

This data entry is used to specify the pipe demand coefficient, which will be
used by Aggregate demands to pipe demand coefficient method.

RESET EXISTING NODE DEMAND (optional)

Existing junction node demands are deleted prior to demand aggregation.

RESET EXISTING PIPE DEMAND COEFFICIENTS (optional)

Existing pipe demand coefficient values are deleted prior to demand aggre-
gation.

USE DEMAND CATEGORY (optional)

This option data entry is used if the user wants to aggregate demand points
based on their category to the appropriate junction nodes.

TARGET (mandatory)
This data entry allows you to specify the multiple demand category, which will
be used as a target demand for the aggregated demand. If the multiple
demand with the specified category does not exist, the program will create it
and it will override the existing values in case that such demand category
already exists for each node used in the demand distribution.

1.3.3 Multiple Demand Editor

Multiple demands can be edited either within the junction editor for each par-
ticular junction node or in the Multiple Demand editor, which allows the user
to display and edit all multiple demands. The Multiple Demand Editor dialog
box is reached by selecting EPANET | Multiple Demands.

69
 EPANET Input Description

Figure 1.30 Multiple Demand Editor dialog box

JUNCTION ID (mandatory)
The Junction ID identifies the selected Junctions which multiple demands are
assigned to.

DEMAND (mandatory)
The demand field shows all the demand values that are assigned to junctions
with multiple demands. The demand values must be manually entered in the
demand field.

DEMAND COEFFICIENT (optional)

This data entry allows you to define the demand coefficient, which can be
used in the automatic demand distribution. See “Distributed Demands” on
page 59 for more details.

DESCRIPTION (optional)
This data entry allows you to enter a description identifying the multiple
demand being defined. This description can be output in reports generated by
the Report Generator.

DEMAND CATEGORY (optional)

Pattern category is not editable but is automatically displayed based on the
category defined in the pattern Editor for the particular Pattern ID. It is possi-
ble to import and export multiple demands from the ASCII text files, which
allows easy data exchange with other programs.

Cyclic Profile ID (optional)

This field displays the cyclic profile ID associated with the Junction. The cyclic
profiles are defined using the curve editor and can be assigned to selected
junctions in the multiple demand editor by right clicking in the Cyclic ProfileID
field.

70 MIKE URBAN - © DHI

Network Demand

1.3.4 Demand Statistics

MIKE URBAN can generate statistical information for junction node demands.
Demand statistics is generated for each pressure zone as well as for the
complete network. Additionally, demand statistics dialog box allows the user
to redistribute node demands by changing the calculated statistical results.
The Demand Statistics dialog box is reached by selecting EPANET |
Demand Statistics.

Figure 1.31 Demand statistics dialog box

A list of the Demand Statistics information for Figure 1.31 follows, with a short
description given for each entry.

 Type
This data entry is used to distinguish between data (demand statistics for
the selected category), zone (demand statistics for the pressure zone),
and network (demand statistics for the whole network).

 Pressure Zone
Pressure zone ID identifies the pressure zone for which the demand sta-
tistics is generated.

 Category
This data entry identifies the category within the current pressure zone
for which the demand statistics is generated.

 Minimum Demand
This data entry represents the minimum demand per category per pres-
sure zone. The minimum demand is calculated as minimum demand at
junction nodes at specific time level

 Maximum Demand
This data entry represents the maximum demand per category per pres-
sure zone. The maximum demand is calculated as maximum demand at
junction nodes at specific time level

71
 EPANET Input Description

 Average Demand
This data entry represents the average demand per category per pres-
sure zone. The average demand is calculated as average demand at
junction nodes during the entire simulation period.

 Sum Demand
This data entry represents the total demand per category per pressure
zone. The total demand is calculated as total demand at junction nodes
during the entire simulation period.

 New Average Demand

This data entry allows the user to specify the new average demand for
selected category, pressure zone, or a network. All corresponding junc-
tion node demands will automatically be adjusted (scaled) in order to fit
the new average demand value.

 New Sum Demand

This data entry allows the user to specify the new total demand for
selected category, pressure zone, or a network. All corresponding junc-
tion node demands will automatically be adjusted (scaled) in order to fit
the new total demand value.

 Category Type
This option data entry allows you to specify the category type such as
fixed or scaled, which will be used in the demand redistribute process.
Junction node demand with fixed demand category will not be scaled
during the demand redistribution.

REFRESH
Select Refresh to re-generate the demand statistics.

REDISTRIBUTE
Select Redistribute to redistribute node demands based on the new values of
Average or Total demand (zone or network). This powerful feature provides
the user with the option of specifying the new zone or a network demand from
within the Demand Statistics window and redistribute the node demand
accordingly. The process of the demand distribution is based on using the
existing node demands as coefficients - weights to calculate the new demand
values.

72 MIKE URBAN - © DHI

Extended Period Simulations

Figure 1.32 Demand statistics dialog box: Redistribute allows you to define the new
average or total zone (or network) demand and redistribute it to junction
nodes

1.4 Extended Period Simulations

Network models are analyzed either as steady state (static) simulations or
extended period (continuous or dynamic) simulations. The data requirements
for each type of model are basically the same; however, additional data is
required for extended period simulations. Extended period data is defined as
status, controls, patterns, and times.

Status represents the setting of various network components, which include

the initial status of pipes, pumps, and valves. For example, a pipe’s status
could be defined as being initially closed (as if it were defined as a valve).

Controls (sometimes called switches) allow the settings (or status) of various
network components to change at a particular time, when a specific pressure
is reached, or when a specific water tank level occurs. For example, a pump
could be turned off once a particular water tank level has been reached.

Repetive profiles are used to describe demand fluctuations. Profile data

(sometimes called hydrograph or curve data) is applied to nodal demands
and include a multiplicative factor that are applied to be nodal demand base
value.

Times are used define various time-step parameters, such as:

 The total duration of the modelling simulation. This is typically 24 hours,

since the demand pattern is nearly always daily.
 The time step (sometimes called the time interval), which is used to
model the simulation in steps. This is typically 1 hour.
 The starting time, which establishes the time at which the simulation
begins.

The following sections describe available editors that allow you to define an
extended period simulation.

73
 EPANET Input Description

1.4.1 Simple Control Editor

Typically during an extended period simulation, the pipes, pumps, and valves
(links) contained in a network will change their status (i.e., open or close) as
storage tanks fill and empty and pressures change throughout the network
system. Also, for a steady state simulation, network components may change
their state as the analysis model iterates to a valid solution.

Using the Control Editor, as shown in Figure 1.33, the user can specify the
operational controls for the pipe network system on a one-on-one component
basis. The Control Editor can be used for both extended period simulations
and steady state simulations.

The following situations exemplify the types of operational controls that can
be specified:

 A pipe can be opened at a given time (based upon the beginning of the
network simulation). This type of operational control has no effect in a
steady state simulation.
 A pump can be turned on or off depending on the water level in a speci-
fied tank.
 A pump’s baseline operating speed can be adjusted using a multiplier,
thereby increasing or decreasing the pump’s output.
 A valve can be opened or closed based upon the pressure in an adjacent
node.

The initial status of the network’s pipes, pumps, and valves default to an open
status, and can be changed to a closed status using the Control Editor based
upon some condition. Additionally, pumps have a default pump speed ratio of
1.0 (indicating that they are operating on their original characteristic opera-
tional curves), but which can be modified based upon a particular condition.

The Control Editor dialog box, as shown in Figure 1.33, is used to define the
operational controls for running the pipe network system. The Control Editor
dialog box is reached by selecting EPANET | Extended Period|Simple Con-
trols ....

In order to export simple controls into ASCII file, open Simple Controls Editor,
select Advanced and select "Export to ASCII file…". This will allow you to
define the name of the ASCII file and export the content of the Simple Control
Editor into the selected ASCII file. You can edit the controls within the ASCII
file and import them back to MIKE URBAN by selecting Load from ASCII file.
This is convenient in cases when you use Excel or other tools to create the
list of controls for the model.

74 MIKE URBAN - © DHI

Extended Period Simulations

Figure 1.33 The Control Editor allows the user to define the network operational
controls

A list of the Control Editor data entries for Figure 1.33 follows, with a short
description given for each entry.

LINK TYPE (optional)

This pull-down selection list allows the user to select what type of link (i.e.,
pipe, pump, or valve) a control rule is being specified for.

LINK ID (mandatory)
This data entry is used to define the ID of the link to control. Choosing «...»
will display the Select Link selection dialog box from which the user can
select the appropriate link type and ID. Or, choosing «Pick» allows the user to
graphically select the link from the Map window.

DESCRIPTION (optional)
This data entry allows you to enter a description identifying the control rule
being defined. This description can be optionally included in reports gener-
ated in MIKE URBAN.

SETTING (mandatory)
This radio button selection entry is used to specify the OPEN or CLOSED sta-
tus of the link being controlled, or VALUE which then allows the user to spec-
ify a multiplier to a pump’s operating speed if a pump is being controlled. If
VALUE is selected, then an additional data entry field will be presented to
allow the user to specify the pump speed ratio.

CONDITION (mandatory)
This radio button selection entry is used to specify the control condition that
then applies the operational rule onto the link being controlled.

75
 EPANET Input Description

If the user selects either IF NODE BELOW or IF NODE ABOVE, then a Con-
trol Node ID and a Control Level must be specified. Choosing «...» will dis-
play the Select Node selection dialog box from which the user can select the
appropriate node type and ID. Or, choosing «Pick» allows the user to graphi-
cally select the node from the Map window. Note that reservoirs are not
allowed to be selected as a Control Node type.

If a junction node is selected as the controlling node, then a trigger pressure

at the junction node must be specified in the Control Level data entry. If a
storage tank node is selected as the controlling node, then a trigger level (not
elevation) must be specified in the Control Level data entry.

If the user selects AT TIME, then a trigger time (since the start of the simula-
tion) must be specified in the adjacent data entry field and a time unit
selected from the pull-down selection list. Note that this type of condition has
no effect for steady state simulations, although can be defined should the
user want to perform an extended period simulation at a later time.

AT CLOCK TIME allows you to specify a trigger time, which periodicaly

repeats each day, such as at 10.00 a.m., for example.

CONTROL NODE ID (optional)

This data entry is used to define the ID of the control node. Choosing «...» will
display the Select Node selection dialog box from which the user can select
the appropriate node type and ID. Or, choosing «Pick» allows the user to
graphically select the link from the Map window.

CONTROL LEVEL (optional)

This data entry is used to define control node level. The control node level
represents pressure at junction nodes (psi or m) or water depth above a
datum in storage tanks (feet or m).

ADVANCED - Import from Ascii File

This button allows the user to load simple controls that have been saved as a
text file. This allows the user to use external editors to prepare and modify the
control settings for the active project.

ADVANCED - Export to Ascii File

Allows the user to export the simple controls into an ASCII file. The user can
define the file name and path that it is saved to.

Multiple Controls
During a simulation, a link (pipe, pump, or valve) can be operated by more
than one operational control rule. For example, a link can be opened at a
given time and then closed if a given nodal pressure is exceeded.

Not that it is necessary to re-apply the link setting (such as control valve set-
ting or pump speed) after the link is re-open by the control rules.

76 MIKE URBAN - © DHI

Extended Period Simulations

Pipes with Check Valves

Note that you cannot set the status of a pipe containing a check valve. Pipes
with a check valve are initially open, and close only if flow within the pipe
attempts to move from the ending (downstream) node to the starting
(upstream) node.

Available Operational Controls

The following operational controls can be defined by the user:

 A pipe, pump, or valve can be opened or closed.

 A pump’s speed ratio can be changed.
 The pressure setting for PRV (pressure reducing), PSV (pressure sus-
taining), and PBV (pressure breaker) valves can be changed.
 The flow setting for FCV (flow control) valves can be changed.
 The minor loss coefficient for TCV (throttle control) valves can be
changed.

Time Based Controls

The user can specify a specific time from the start of the simulation at which a
change in a link’s status is to occur. For example, pipe 15 can be opened 3
hours into the simulation and then closed at 4.2 hours into the simulation.
This example is shown below:

LINK 15 OPEN AT TIME 3.0

LINK 15 CLOSED AT TIME 4.2

Storage Tank Water Level Based Controls

The user can specify a storage tank level (not elevation) at which a control
will occur. For example, pump 296 can be opened when the water level in
storage tank 40 drops below 26 feet, and can be closed when the level rises
above 40 feet. This example is shown below:

LINK 296 OPEN IF NODE 40 BELOW 26.0

LINK 296 CLOSED IF NODE 40 ABOVE 40.0

Junction Node Pressure Based Controls

The user can specify a pressure level at which a control will occur if the spec-
ified pressure is above or below the pressure level. For example, valve 55
can be opened if the pressure at junction node 47 is below 50 psi, and pump
12 operating speed can be dropped to one-half of its normal speed when the
pressure at junction node 12 goes above 75 psi. This example is shown
below:

77
 EPANET Input Description

LINK 55 OPEN IF NODE 47 BELOW 50.0

LINK 12 0.5 IF NODE 12 ABOVE 75.0

1.4.2 Rule Based Controls

Rule-Based Controls allow link status and settings to be based on a combina-

tion of conditions that might exist in the network over an extended period sim-
ulation. Rule based controls will be either in the form of an action clause or a
condition clause. The Rule Based Controls Editor dialog box is reached by
selecting EPANET | Extended Period| Rule Based Controls.

Figure 1.34 The Control Editor allows the user to define the network operational
controls Rule Based Controls Editor Dialog Box

In order to rule based controls into ASCII file, open Rule Based Controls Edi-
tor, select Advanced and select "Export to ASCII file…". This will allow you to
define the name of the ASCII file and export the content of the Rule Based
Controls Editor into the selected ASCII file. You can edit the rules within the
ASCII file and import them back to MIKE URBAN by selecting Load from
ASCII file. This is convenient in cases when you use Excel or other tools to
create the list of rules for the model.

A condition clause in a Rule-Based Control takes the form of:

object id attribute relation value

where

object = a category of network object

id = the object's ID label

attribute = an attribute or property of the object

relation = a relational operator

value = an attribute value

78 MIKE URBAN - © DHI

Extended Period Simulations

Some example conditional clauses are:

JUNCTION 23 PRESSURE > 20

TANK T200FILLTIME BELOW 3.5

LINK 44 STATUS IS OPEN

SYSTEM DEMAND >= 1500

SYSTEM CLOCKTIME = 7:30 AM

The Object keyword can be any of the following:

NODE LINK SYSTEM

JUNCTION PIPE

RESERVOIR PUMP

TANK VALVE

When SYSTEM is used in a condition no ID is supplied.

The following attributes can be used with Node-type objects:

DEMAND

HEAD

PRESSURE

The following attributes can be used with Tanks:

LEVEL

FILLTIME (hours needed to fill a tank)

DRAINTIME (hours needed to empty a tank)

These attributes can be used with Link-Type objects:

FLOW

STATUS (OPEN, CLOSED, or ACTIVE)

SETTING (Pump speed or Valve setting)

LIKE (See Multiple Pumps, Valves for more details)

The SYSTEM object can use the following attributes:

DEMAND (total system demand)

79
 EPANET Input Description

TIME (hours from the start of the simulation)

CLOCKTIME (24-hour clock time with AM or PM appended)

Relation operators consist of the following:

= IS

<> NOT

< BELOW

> ABOVE

<=> =

An action clause in a Rule-Based Control takes the form of:

object id STATUS/SETTING IS value

where

object = LINK, PIPE, PUMP, or VALVE keyword

id = the object's ID label

value = a status condition (OPEN or CLOSED), pump speed setting, or valve

setting

Some example action clauses are:

LINK 23 STATUS IS CLOSED

PUMP P100 SETTING IS 1.5

VALVE 123 SETTING IS 90

LIKE A special case of action clause it the LIKE setting.:

Setting Value = A (another link setting) B (multiplier) C (increment)

The default values for the B (multiplier) and C (increment) are B=1, C=0 and
they do not need to be provided.

The setting value is calculated as: Setting Value = Setting Value (link A) * B +
C

See Multiple Pumps, Valves for more details.

80 MIKE URBAN - © DHI

Extended Period Simulations

Multiple Pumps, Valves (WD-Tools)

Note that the LIKE setting allows you to control multiple pumps or valves in
efficient way. It is possible to set a pump speed to x% of another pump, for
example. Such an option could also be used when the new value is the value
of the object it self. (Set a pump speed to increase with 20%, or valve to open
10% and so on).

Example 1:
IF SYSTEM CLOCKTIME = 8 AM

THEN PUMP 3 SETTING LIKE PUMP 4 1 0

The pump 3 setting will be set equal to the settings of the pump 4 (multiplier =
1 and increment = 0) at time 8 am.

Example 2:
IF SYSTEM CLOCKTIME = 8 AM

THEN PUMP 3 SETTING LIKE PUMP 4 1.10 0

The pump 3 setting will be higher by 10% than the settings of the pump 4
(multiplier = 1.10 and increment = 0) at time 8 am.

Example 3:
IF SYSTEM TIME >= 12 AM

THEN VALVE 10 SETTING LIKE VALVE 20 1.0 -10

The valve 10 setting (PRV setting, for example) will be lower by 10 (pressure
units) than the settings of the valve 20 (multiplier = 1 and increment = -10) at
any time (12 AM is the simulation start).

Example 4:
IF SYSTEM TIME >= 12 AM

THEN PUMP 3 STATUS LIKE PUMP 4

The pump 3 status (OPEN, CLOSED) will be set equal to the status of the
pump 4 at any time (12 AM is the simulation start).

Example 5:
IF PUMP 3 SETTING LIKE PUMP 4

THEN …

If pump 3 setting is equal to the settings of the pump 4 (default multiplier = 1

and increment = 0) then …

81
 EPANET Input Description

Control of valves
This set of rules opens and closes a Valve based on different water levels in
tanks.

RULE PIT-010_Close ;BUNKER RD NORTH CV CLOSE

IF TANK 173 LEVEL > 7.315

OR TANK 139B LEVEL < 8.56

THEN VALVE PIT-010 SETTING IS CLOSED

RULE PIT-010_Open ;BUNKER RD NORTH CV OPEN

IF TANK 173 LEVEL <= 7.315

OR TANK 139B LEVEL >= 8.56

THEN VALVE PIT-010 SETTING IS OPEN

Control of pumps
This set of rules opens and closes a pump based on different water levels in
tanks.

RULE AH_PS1_Start ;ALEXHILL PS1 - AUTOMODE1

IF TANK 172 LEVEL < 5.39

AND TANK 170 LEVEL >= 0.75

AND TANK 171 LEVEL >= 0.75

THEN PUMP 14146 STATUS IS OPEN

RULE AH_PS1_Stop ;ALEXHILL PS1 - AUTOMODE1

IF TANK 172 LEVEL < 6.55

OR TANK 170 LEVEL < 0.36

OR TANK 171 LEVEL < 0.36

THEN PUMP 14146 STATUS IS CLOSED

82 MIKE URBAN - © DHI

Extended Period Simulations

1.4.3 Repetitive Profiles (Pattern) Editor

MIKE URBAN assumes that water usage rates, external water supply rates,
and constituent source concentrations (for water quality analysis) at nodes
remain constant over a fixed period of time, but that these quantities can
change from one time period to another. The default time period interval is
1 hour, but this can be set to any value. The value of any of these quantities
in a time period equals a baseline value multiplied by a time pattern factor for
that period.

The definition of repetitive profiles consists of two main steps:

Step 1: Definition of Diurnal Profiles (24 hrs. duration)

Step 2: Definition of Cyclic Patterns, which are combining 1 or more diurnal

profiles together.

Diurnal Patterns
The diurnal profiles are used to define a series of multipliers (multiplication
factors applied to a baseline value of junction node demand, constituent
source concentration, storage tank water level). The duration of such diurnal
profile curve is 1 day i.e. 24 hours. The Diurnal Profiles Editor dialog box is
reached by selecting EPANET | Extended Period|Repetitive Profiles |
Diurnal Patterns.

83
 EPANET Input Description

Figure 1.35 Diurnal Profiles Editor, the Multipliers grid allows the user to define
demand and/or constituent concentration multipliers for the current pat-
tern ID

PATTERN ID (mandatory)
This data entry is used to specify the ID of the component being defined. The
pattern ID value can be any string value (up to 40 characters). There is no
limit to the number of demand patterns that can be defined.

TIME STEP (mandatory)

The pattern time specifies the length of time between each pattern change
(i.e., the period of time over which water demands and constituent source
strengths remain constant). To change the pattern time step, use Time Set-
tings Editor. The default value is 1 hour.

CATEGORY (optional)
This data entry allows you to enter a description identifying the demand pat-
tern being defined. This description can be optionally included in reports.

DESCRIPTION (optional)
This data entry allows the user to enter a category that further define the
demand pattern. For example, a demand might have the description of resi-
dential, and a category of either high density, medium-density or low density
to further define what is meant by residential. This description can optionally
be included in reports generated by the Report Generator.

84 MIKE URBAN - © DHI

Extended Period Simulations

MULTIPLIERS (mandatory)
Selecting «Define» will display the Multipliers dialog box, as displayed in
Figure 1.35. This dialog box allows you to define the baseline demand factors
and constituent concentration levels (multipliers) for the current pattern ID.
The complete pattern is then applied to the baseline demand (or source con-
centration) specified at each node (that corresponds with this pattern ID) over
the extended period simulation.

Within each time period, the demand (or source concentration) remains con-
stant—at a level equal to the multiplier times the baseline value.

There is no limit on the number of multipliers that can be defined for a pattern.
However, since the diurnal profile duration is exactly 24 hours, the number of
multipliers is determined by the pattern time step. The typical time period
length is 1 hour, 24 multipliers are required to completely define the pattern
for the entire simulation. For more information on defining the time step inter-
val and analysis period, see the section titled Time Setting Editor which fol-
lows this section.

DISTRIBUTE
Select Distribute to let the program calculate the number of multipliers to fill
entirely the diurnal pattern duration of 24 hours with the current pattern time
step.

GRAPH
It is possible to create a GRAPH for pattern multipliers directly in the PAT-
TERN EDITOR window, not only in the MULTIPLIERS window. This function
allows the user to display the pattern graph quickly.

Figure 1.36 Graphical representation of the diurnal pattern multipliers

85
 EPANET Input Description

It is possible to import and export the diurnal pattern multipliers from an ASCII
file. Select Advanced - Import or Advanced - Export to import or export the
diurnal pattern multipliers from an ASCII text file.

Cyclic Profiles
The cyclic pattern editor is used to combine 1 or more diurnal profiles into one
cyclic pattern. This allows the user to compose the cyclic patterns of two diur-
nal profiles one for e.g. working days and one for the weekend. If the total
duration of the defined cyclic pattern is less than the total duration of the sim-
ulation, then the defined pattern will be repeated. For example, a 5 day simu-
lation whose hourly demands repeat themselves on a daily cycle, only
24 multipliers would need to be defined for the cyclic pattern diurnal profile.

It is possible to display the graph for the cyclic profile, consisting of several
diurnal patterns. In order to display the graph, open Cyclic Profile Editor,
select the profile and click "Graph".

The Diurnal Profiles Editor dialog box is reached by selecting EPANET |

Extended Period| Repetitive Profiles | Cyclic Profiles.

Figure 1.37 Cyclic Patterns Editor

CYCLIC PROFILE ID (mandatory)

This data entry is used to specify the ID of the cyclic pattern. The cyclic profile
ID value can be any string value (up to 40 characters).

DIURNAL PATTERN ID (mandatory)

This data entry is used to specify the ID of the diurnal pattern. Choosing «...»
will display the Select Diurnal Profile selection dialog box from which the user
can select the appropriate diurnal pattern ID.

86 MIKE URBAN - © DHI

Extended Period Simulations

1.4.4 Time Settings Editor

For an extended period simulation, a simulation duration and time step must
be specified. Using the Time Editor, as shown in Figure 1.38, the user can
specify the extended period simulation time parameters. Only those time
parameters that differ from the default values must be specified. The default
time units are hours.

The Time Editor dialog box is reached by selecting EPANET | Extended

Period| Time Settings.

Figure 1.38 The Time Settings Editor allows the user to define the extended period
simulation time parameters

A list of the Time Editor data entries for Figure 1.38 follows, with a short
description given for each entry.

ANALYSIS DURATION (optional)

This data entry must be specified in order to perform an extended period sim-
ulation, and specifies the length of the entire simulation for both hydraulic
extended period simulations and water quality simulations. The default value
for this data entry is 0 hours, which implies that a steady state simulation will
be performed. Note that a water quality analysis cannot be performed from a
steady state simulation.

HYDRAULIC TIME STEP (optional)

This data entry must be specified in order to perform an extended period sim-
ulation, and specifies how often a new hydraulic computation of the pipe net-
work system is to be computed. The default value is 1 hour.

PATTERN TIME STEP (optional)

This data entry is optional, and specifies the length of time between each pat-
tern change (i.e., the period of time over which water demands and constitu-
ent source strengths remain constant). If necessary, the software will adjust
the specified Hydraulic Time Step so that it is not greater than the specified
Pattern Time Step. The default value is 1 hour.

87
 EPANET Input Description

REPORT TIME STEP (optional)

This data entry is optional, and specifies the interval of time between which
network conditions are reported. If necessary, the software will automatically
reduce the specified value for the Hydraulic Time Step so that it is no greater
than the Report Time Step. The default value is 1 hour.

REPORT START TIME (optional)

This data entry is optional, and specifies at what time into the simulation the
analysis results should start to be reported. The default value is 0 hour.

QUALITY TIME STEP (optional)

This data entry is used for water quality analysis, and specifies the time step
to be used to track water quality changes in the pipe network system. If this
entry is left blank, the program then uses an internally computed time step
based upon the smallest time of travel through any pipe in the network.

START CLOCK TIME (optional)

Start clock time is the time of day (e.g., 3:00 PM) at which the simulation
begins. The default is 12:00 AM midnight.

STATISTICS (optional)
Type of statistical processing used to summarize the results of an extended
period simulation. Choices are:

NONE (results reported at each reporting time step)

AVERAGE (time-averaged results reported)

MINIMUM (minimum value results reported)

MAXIMUM (maximum value results reported)

RANGE (difference between maximum and minimum results reported)

Statistical processing is applied to all node and link results obtained between
the Report Start Time and the Total Duration

Time Units
Note that seconds, minutes, hours, and days can be used as time units, but
that hours are generally used. It is recommended that the same time units be
used for all entries within the Time Settings Editor—that you do not mix time
units (i.e., hours and seconds).

Diurnal Curves and Patterns

Note that changing the pattern time step changes the number pattern multipli-
ers, which define the 24 hrs. duration diurnal curve. MIKE URBAN will display
a query dialog, as shown in Figure 1.39, asking the user whether to redistrib-

88 MIKE URBAN - © DHI

Water Quality Simulations

ute the pattern multipliers. If the user selects yes, the pattern number of pat-
tern multipliers will be adjusted to fit the diurnal curve duration of 24 hrs.

Figure 1.39 When changing the pattern time step, MIKE URBAN will ask whether to
redistribute the pattern multipliers

1.5 Water Quality Simulations

MIKE URBAN allows you to perform water quality simulations. In order to per-
form a water quality simulation, an extended period simulation must also be
specified. Defining an extended period simulation was discussed in the previ-
ous section.

The following sections describe how to perform a particular type of water

quality simulation, and the various water quality editors used to define each
type of water quality simulation.

1.5.1 Water Quality Analysis Selection

The first step in defining a water quality simulation is to select the type of
water quality analysis to be performed. This is accomplished using the Pro-
ject Options dialog box Project Type tab, as shown in Figure 1.40. Note that
MIKE URBAN can only perform one type of water quality analysis during a
simulation.

The Project Options dialog box is reached by selecting EPANET |

Project Options

Figure 1.40 The Project Options dialog box allows the user to define the type of
water quality analysis to be performed

89
 EPANET Input Description

A list of water quality options are available from the Project Options dialog
box is shown in Figure 1.40, along with a short description given for each
option.

NO WATER QUALITY ANALYSIS (default)

By selecting this option button entry, no water quality analysis will be per-
formed. This is the default selection.

CHEMICAL CONCENTRATIONS
This option button entry is used to select that a chemical concentration analy-
sis is to be performed. MIKE URBAN will report the concentration of the spec-
ified chemical at each time step for each network node. This type of analysis
is typically performed to determine the amount of chlorine in the network to
maintain safe drinking water standards.

WATER AGE
This option button entry is used to select that a water age analysis is to be
performed. MIKE URBAN will report the water age at each time step for any
network node, assuming an initial age of zero at the start of the simulation.
This type of analysis is typically performed to determine dead ends (locations
of stagnant water) within a pipe network design.

SOURCE TRACING
This option button entry is used to select that a source tracing analysis is to
be performed. MIKE URBAN will report, for each time step, the percentage of
water reaching each node from a selected source (origin) node. This type of
analysis is typically performed for constituent tracking.

1.5.2 Water Quality Analysis Parameters

The Project Options dialog box Properties tab, as shown in Figure 1.41,
allows you to specify the analysis parameters used to perform a water quality
simulation.

The Project Options dialog box is reached by selecting EPANET |

Project Options.

90 MIKE URBAN - © DHI

Water Quality Simulations

Figure 1.41 The Project Options dialog box allows the user to define the analysis
parameters used in a water quality simulation

A list of the water quality analysis parameters from the Project Options dialog
box is shown in Figure 1.41, along with a short description given for each
parameter. These water quality modelling parameters are used when model-
ling the pipe-wall reaction mechanism.

SPECIFIC GRAVITY
This data entry specifies the specific gravity of the fluid at the temperature
condition being simulated. This data entry allows fluids other than water to be
simulated. Gravity is the weight per unit volume of the fluid being modelled
relative to water. Specific gravity is the ratio of the density of the fluid being
modelled to that of water at 4 deg. C. (unitless).

KINEMATIC VISCOSITY
This data entry specifies the kinematic viscosity of the fluid at the temperature
condition being simulated. The units of viscosity are ft2/sec (or m2/sec for SI
units). The viscosity is the kinematic viscosity of the fluid being modelled rela-
tive to that of water at 20 deg. C (1.0 centistoke). The default value is 1.0.

MOLECULAR DIFFUSIVITY
This data entry specifies the molecular diffusivity of the chemical being
tracked. The diffusivity is the molecular diffusivity of the chemical being ana-
lyzed relative to that of chlorine in water. The default value is 1.0. Diffusivity is
only used when mass transfer limitations are considered in pipe wall reac-
tions. A value of 0 will cause MIKE URBAN to ignore mass transfer limita-
tions.

1.5.3 Initial Water Quality Editor

The initial water quality at the start of a simulation can be assigned to individ-
ual nodes or to groups of nodes. The initial water quality can represent one of
the following:

 Initial concentration for chemical constituents in a chemical propagation

analysis.

91
 EPANET Input Description

 Initial hour for water age determination.

 Initial percentage of water originating at a specified source node for
source tracing.

By default, all nodes are assigned with an initial water quality of zero. The Ini-
tial Water Quality Editor, as shown in Figure 1.42, is used to assign the initial
water quality levels for the pipe network system. The Initial Water Quality Edi-
tor dialog box is reached by selecting EPANET | Water Quality |
Initial Water Quality.

Figure 1.42 The Initial Water Quality Editor is used to define the initial water quality
conditions of the pipe network system

A list of the Initial Water Quality Editor data entries for Figure 1.42 follows,
with a short description given for each entry.

NODE ID
This data entry is used to specify an ID which uniquely identifies a node in
which the initial water quality is being specified for.

A new entry is automatically inserted into the list by pressing «Insert». Choos-
ing «...» will display the Select Node selection dialog box from which the user
can select the appropriate node Or, choosing «Pick» allows the user to
graphically select the node from the Map window.

QUALITY
This data entry is used to specify the initial water quality at the node (or group
of nodes) being defined.

If a chemical concentration water quality simulation is being performed, then

this entry denotes the water quality in mg/litres.

If a water age water quality simulation is being performed, then this entry
denotes the initial age of the water in hours.

If a source tracing water quality simulation is being performed, then this entry
denotes the initial percentage of water from the source (origin) node in per-
cent.

92 MIKE URBAN - © DHI

Water Quality Simulations

1.5.4 Point Constituent Source Editor

The Point Constituent Source Editor, as shown in Figure 1.43, allows you to
specify at which nodes an external chemical constituent enters the network
system. At least one node in the network must be specified as a point source
of chemical constituent when performing a chemical concentration analysis.

The Point Constituent Source Editor dialog box is reached by selecting

EPANET | Water Quality | Quality | Point Constituent Sources . Note that
a Chemical Concentrations water quality simulation must be specified in the
Projects Options dialog box (see Figure 1.40) in order access the Point Con-
stituent Source Editor dialog box.

Figure 1.43 The Point Constituent Source Editor is used to specify at which nodes
an external chemical constituent enters the pipe network system

A list of the Point Constituent Source Editor data entries for Figure 1.43 fol-
lows, with a short description given for each entry.

NODE TYPE
This pull-down selection list allows the user to select what type of node (i.e.,
junction, reservoir, or tank) the point constituent is being specified for.

NODE ID
This data entry is used to define the ID of the node the point constituent is
being assigned to. Choosing «...» will display the Select Node selection dia-
log box from which the user can select the appropriate node type and ID. Or,
choosing «Pick» allows the user to graphically select the node from the Map
window.

CONCENTRATION
This data entry is used to specify the baseline concentration (in mg/liter) of
the constituent entering the node as an external source.

93
 EPANET Input Description

For a junction node, if there is no external inflow assigned to the node (such
as a well—denoted by a negative demand), then the resulting water quality at
the node always equals the specified concentration. This allows the user to
simulate chlorine booster stations at a node, such as is used in satellite treat-
ment in a network.

SOURCE TYPE
Water quality sources are nodes where the quality of external flow entering
the network is specified. They can represent the main treatment works, a
well-head or satellite treatment facility, or an unwanted contaminant intrusion.
Source quality can be made to vary over time by assigning it a time pattern.
MIKE URBAN can model the following types of sources:

 A concentration source fixes the concentration of any external inflow

entering the network at a node, such as flow from a reservoir or from a
negative demand placed at a junction.

 A mass booster source adds a fixed mass flow to that entering the
node from other points in the network.

 A flow paced booster source adds a fixed concentration to that result-

ing from the mixing of all inflow to the node from other points in the net-
work.

 A setpoint booster source fixes the concentration of any flow leaving

the node (as long as the concentration resulting from all inflow to the
node is below the setpoint).

The concentration-type source is best used for nodes that represent source
water supplies or treatment works (e.g., reservoirs or nodes assigned a nega-
tive demand). The booster-type source is best used to model direct injection
of a tracer or additional disinfectant into the network or to model a contami-
nant intrusion.

CYCLIC PROFILE ID
This data entry allows you to define the ID of the constituent pattern to be
applied to the specified baseline concentration entering the node. If a pattern
ID is omitted for the specified source node, then there is no variation in the
source strength of the constituent.

Selecting «Button» allows you to display the Select Pattern selection dialog
box, where the appropriate pattern ID can be selected.

Point Constituent Source Time Patterns

Point constituent source concentration time patterns are similar in concept to
demand patterns. Each concentration time pattern consists of a set of multi-
pliers that are multiplied to the specified baseline concentration over the
extended period simulation. This allows the user to model changes in the

94 MIKE URBAN - © DHI

Water Quality Simulations

amount of constituent applied at a node over an extended period simulation.

See the section on Cyclic Profiles for further information on time patterns.

1.5.5 Reaction Rate Editor

The Reaction Rate Editor, as shown in Figure 1.44, allows you to specify the
rate at which a constituent decays (or grows) by reaction as the constituent
travels through the pipe network. Reaction rates can be defined at a global
level where the same reaction rate applies to the entire network, and at a
local level allowing the user to define reaction rates based upon a particular
pipe type, etc. Note that locally defined reaction rates override globally
defined reaction rates.

The Reaction Rate Editor dialog box is reached by selecting EPANET | Water
Quality | Reaction Rates. Note that a Chemical Concentrations water quality
simulation must be specified in the Projects Options dialog box (see
Figure 1.41) in order access the Reaction Rate Editor dialog box.

Global Settings

Figure 1.44 The Reaction Rate Editor is used to specify constituent reaction rates at
both a global and local level

A list of the Reaction Rate Editor data entries for Figure 1.44 follows, with a
short description given for each entry.

BULK REACTION RATE COEFFICIENT

This data entry defines the bulk reaction rate that is applied to all flow in the
pipe network system. Units for bulk reaction rates are in days–1. Note that this
reaction rate coefficient is applied globally to the entire pipe network.

PIPE WALL REACTION RATE COEFFICIENT

This data entry defines the pipe wall reaction rate that is applied to all flow in
the pipe network system. Units for pipe wall reaction rates are in ft/day (or
m/day). Note that this reaction rate coefficient is applied globally to the entire
pipe network.

95
 EPANET Input Description

One method that can be used to compare the relative magnitude of the pipe
wall reaction rate with the bulk reaction rate is to divide the pipe wall reaction
rate coefficient by the hydraulic radius of the pipe (i.e., 1/2 the pipe radius).
The resulting quantity will have the same units as the bulk reaction rate coef-
ficient, days–1.

BULK REACTION ORDER

Power to which concentration is raised when computing a bulk flow reaction
rate. Use 1 for first-order reactions, 2 for second-order reactions, etc. Use
any negative number for Michaelis-Menton kinetics. If no global or pipe-spe-
cific bulk reaction coefficients are assigned then this option is ignored.

PIPE WALL REACTION ORDER

Power to which concentration is raised when computing a bulk flow reaction
rate. Choices are 1 for first-order reactions or 0 for constant rate reactions. If
no global or pipe-specific wall reaction coefficients are assigned then this
option is ignored

LIMITING POTENTIAL
LIMITING POTENTIAL specifies that reaction rates are proportional to the
difference between the current concentration and some limiting potential
value.

ROUGHNESS CORRELATION
ROUGHNESS CORRELATION will make all default pipe wall reaction coeffi-
cients be related to pipe roughness in the following manner

1.5.6 Local Settings

Tank Reaction Rate Assignments

The locally defined tank reaction rate is specified in the Tank Editor.

Pipe Reaction Rate Assignments

The locally defined pipe wall and pipe bulk reaction rates are specified in the
Pipe Editor. This allows the use of the field calculator to specify the reaction
rate based on the pipe material, pipe diameter, and pipe construction year, for
example.

Table 1.1 Roughness for Head Loss Equations

Head loss Equation Wall reaction Formula

Hazen-Williams F/C

96 MIKE URBAN - © DHI

Water Quality Simulations

Table 1.1 Roughness for Head Loss Equations

Head loss Equation Wall reaction Formula

Darcy-Weisbach -F/log(e/D)

Chezy-Manning F*n

where F = roughness correlation, C = Hazen-Williams C-factor, e = Darcy-

Weisbach roughness, D = pipe diameter, and n = Chezy-Manning roughness
coefficient. The default value computed this way can be overridden for any
pipe by using the wall reaction rate format to supply a specific value for the
pipe.

Note
Remember to use negative signs on all reaction coefficients that are to repre-
sent constituent decay (e.g., chlorine decay). Otherwise, a positive value
denotes constituent growth.

1.5.7 Source Tracing

The Trace Node dialog box, as shown in Figure 1.45, allows you to track over
time what percent of water reaching any node in the network had its origin
from a specified node (i.e., junction, tank, or reservoir). Source tracing is a
useful tool for analyzing a network distribution system that draws water from
two or more different raw water supplies. It can be used to show to what
degree water from a given source blends with that from other sources, and
how the spatial pattern of this blending changes over time.

The Trace Node dialog box is reached by selecting EPANET|Water Quality |

Trace Node. Note that a Source Tracing water quality simulation must be
specified in the Projects Options dialog box (see Figure 1.41) in order access
the Trace Node dialog box.

Figure 1.45 The Trace Node dialog box is used to specify a single node that acts
like a tracer in determining what percent of its water reaches any other
node in the network

A list of the Trace Node dialog box data entries for Figure 1.45 follows, with a
short description given for each entry.

97
 EPANET Input Description

NODE TYPE
This pull-down selection list allows the user to select what type of node (i.e.,
junction, reservoir, or tank) the trace node is being specified for.

NODE ID
This data entry is used to define the ID of the node the trace node is being
assigned to. Choosing «...» will display the Select Node selection dialog box
from which the user can select the appropriate node type and ID. Or, choos-
ing «Pick» allows the user to graphically select the node from the Map win-
dow.

1.6 Tools
The following section describes additional capabilities in MIKE URBAN that
allow the user to work effectively with the program.

1.6.1 Send Data

You can transfer data between the asset and the model database of MIKE
URBAN. See the section titled Transferring data between ‘Asset’ and ‘Model’
in the MIKE URBAN Model Manager book, chapter Modelling Water Distribu-
tion networks for more details on Send Data.

1.6.2 Engineering Tables

Included with MIKE URBAN are predefined engineering tables that include
specified pipe roughness coefficients and minor loss coefficients, as shown in
the Engineering Tables dialog box as shown in Figure 1.47. These values are
used as lookup coefficient values when defining the pipe and valves that
make up the network system. The Engineering Tables dialog box is reached
by selecting Tools | Engineering Tables. Note that additional coefficients
can be added to the provided engineering table values using this dialog box.

98 MIKE URBAN - © DHI

Tools

Figure 1.46 The Engineering Tables define lookup table values for minor loss coeffi-
cients

Pipe roughness coefficients are specified as Hazen-Williams C, Darcy-Weis-

bach , and Manning’s n values for various pipe material. Minor loss coeffi-
cients are used in pipe and valve definitions that contain additional
headlosses due to pipe fittings and/or flow obstructions. When defining minor
losses for a network component and more than one minor loss applies, then
the sum of all the minor loss coefficients should be used to account for the
total additional headloss.

Figure 1.47 The Engineering Tables define lookup table values for pipe roughness
coefficients

Roughness coefficient and minor loss coefficient lookup table dialog boxes
are displayed from either the Pipe Editor and Valve Editor by clicking on the
«...» button, adjacent to the Roughness and Minor Loss data entries.

99
 EPANET Input Description

1.6.3 Convert Asset Valves/Pumps to Model Valves/Pumps

The Convert Asset Valves (or Pumps) to Model Valves (or Pumps) dialog box
is tailored to convert asset valves/pumps (point feature class mwa_-
valve/mwa_pump) to the model valves/pumps (polyline feature class mw_-
valve/mw_pump). This tool is available when running MIKE URBAN in the
WD Water Distribution mode.

The Convert Asset Valves (or Pumps) to Model Valves (or Pumps) dialog box
is reached by selecting Tools | Toolbox | Convert Asset Valves (Pumps) to
Model Valves (Pumps). This tool will start converting asset valves to model
valves by finding the nearest model pipe for each asset valve (or selected
asset valve), which is located within a snap tolerance radius to the nearest
model pipe. In order to import GIS valves to the asset valves (mwa_valve fea-
ture class) use File Import and Export dialog box; it is also possible to create
asset valves by graphical editing.

Figure 1.48 Convert asset valves to model valves dialog box

A list of the Convert Asset Valves to Model Valves dialog box data fields for
Figure 34.57 follows, with a short description given for each entry.

SNAP TOLERANCE RADIUS

This data entry is used to specify the snap tolerance radius, which is used to
convert asset valves close to the model pipes.

MODEL VALVE LENGTH/MODEL PUMP LENGTH

This data entry is used to specify the length of the model valve to be created.

CONVERT SELECTED ASSET VALVES/PUMPS or CONVERT ALL

ASSET VALVES/PUMPS
This data entry choice is used to specify whether all asset valves will be con-
verted to the model valves os whether only selected asset valves will be con-
verted. Note that the snap tolerance radius will be applied in addition to this
primary source data filter.

100 MIKE URBAN - © DHI

Tools

Figure 1.49 Conversion of the asset valve to the model valve; note that the pipe is
split and two junction nodes are created nearby the asset valve loca-
tion.

1.6.4 Duplicate Pipe Attributes

It is very common that you import the model based on the CAD files, for
example, and that the pipe attributes (imported as CAD labels) were not
available for every pipe. Similar situation will also happen when you start edit-
ing pipes manually and you do not really want to assign the pipe attributes
one by one.

MIKE URBAN provides a tool, which allows you to select pipes from the Map
and duplicate their attributes. Pipe material, as illustrated in the example
below, is automatically duplicated to all pipes, which are adjacent to the
selected pipe(s) until "T" or other complex junction exists.

Figure 1.50 Step1: Select pipes which attribute(s) you want to duplicate.

Now, select attributes you would like to duplicate, such as pipe material,
diameter, and others. Click "OK" where ready.

101
 EPANET Input Description

Figure 1.51 Step 2: Select what attributes you want to duplicate

Figure 1.52 Step 3: Pipe attributes are duplicated

"Generate tag" tool allows you to generate unique identifier "Tag" for every
stretch of pipes within the model. This is very useful for generating unique
"GIS" tag for such "chain of pipes", which are linked together by just 1 single
node.

102 MIKE URBAN - © DHI

Tools

Figure 1.53 llustration: GIS tag is generated for each stretch of pipes

103
 EPANET Input Description

104 MIKE URBAN - © DHI

MIKE URBAN WD
User Guide

105
106 MIKE URBAN - © DHI
2 Using the Model
MIKE URBAN™ is the most advanced, powerful, and comprehensive water
distribution modelling package available. MIKE URBAN WD (Water Distribu-
tion) can analyse an entire water distribution system, or selected portions,
under steady state, extended period and water hammer flow conditions, with
water quality analysis if needed. You can enter model data suited for your
choice of numerical engine EPANET (public domain EPANET engine devel-
oped by EPA, US), Water hammer engine (developed by DHI and a part of
the WD-Tools) and/or WD-Tools (DHI suit of tools such as fire flow analysis,
pipe roughness calibration, variable speed drive pump, extended controls
based on the modified EPANET engine). You can also enter ‘Asset WD’ data
suited for a water supply and water distribution system. The public domain
EPANET engine can be run from the Model Manager, while running WD-
Tools requires you to have the license for the MIKE URBAN Water Distribu-
tion system with WD-Tools.

MIKE URBAN allows extreme flexibility when developing a water distribution

model. The user can develop a model from scratch using a variety of input
methods including importation of data files from a GIS database or pre-exist-
ing water distribution model, schematically drawing the pipe network, or by
direct data entry using the program editors.

If a map of the water distribution system is available, MIKE URBAN can

import this map and display it as a background image—allowing the user to
interactively construct and layout the pipe network system. Network compo-
nents can be selected from a component toolbar, and then graphically placed
on the screen at the precise location of each component.

Many times, existing water distribution systems do not have a detailed map
that can be used to graphically construct a network model. For these situa-
tions, MIKE URBAN allows the user to develop a model by simply defining
water distribution components (i.e., pipes, junction nodes, pumps, values,
tanks, and reservoirs) in interactive, easy-to-use dialog boxes. This allows
the user to define a model when an accurate map is not available for the pipe
network model.

After the pipe network has been defined, a hydraulic analysis of the network
can be performed. Detailed reports can be generated from the analysis
results and printed out. In addition, if a graphical representation of the pipe
network has been created, the computed pipe flowrates and junction node
pressures can be displayed and printed out.

This User Guide consists of the following chapters:

1. Using the Model: provides the fundamental understanding of the basic

operation of MIKE URBAN Water Distribution modeling.
2. EPANET Input Descriptions: provides detailed descriptions of the data
input dialog boxes used by this application.

107
 Using the Model

3. Water Distribution Tools: provides detailed description of the additional

modeling tools including fire flow analysis, pipe roughness calibration,
and others.
4. Water Hammer: provides detailed descriptions of the water hammer sim-
ulations.
5. Program Methodology: presents the theory and methodology that was
used in developing this program.

2.1 Overview
MIKE URBAN Water Distribution (WD) consists of the following ‘modules’.

2.1.1 EPANET

Pipe Flow Module

MIKE URBAN applies the widely used and universally accepted "numerical
engine" developed by US EPA - EPANET 2.0 - for the computation of steady
state and quasi steady state flow, pressure conditions and water quality in
pipe networks. The method used by EPANET to solve the system of node
and link equations is known as the "gradient algorithm" and has several
attractive features. First, the system of linear equations to be solved at each
iteration of the algorithm is sparse, symmetric, and positive-definite. This
allows highly efficient sparse matrix techniques to be used for their solution.
Second, the method maintains flow continuity at all nodes after its first itera-
tion. And third, it can readily handle pumps and valves without having to
change the structure of the equation matrix when the status of these compo-
nents changes.

The most important MIKE URBAN Pipe Flow features are:

 Steady state analysis.

 Extended period analysis.
 Rule based control.
Water Quality Module
Water quality analysis is used to compute the age of water, perform source
tracking, calculate the fate of a dissolved substance, or determine the growth
or decay of a substance. The water quality solver is based on the efficient
Lagrangian Time Driven method. Water quality sources are defined in terms
of mass inflow rate in addition to concentration level. Bulk flow reactions can
be modeled with n-th order kinetics, pipe wall reactions can be modeled with
either zero- or first-order kinetics, a water quality constituent's growth or
decay can be modeled up to a limiting potential, and pipe wall reaction coeffi-
cients can be correlated to pipe roughness coefficients.

108 MIKE URBAN - © DHI

Overview

The most important MIKE URBAN Water Quality features are:

 Water age analysis.

 Chlorine concentrations analysis.
 Path and concentrations of pollutants analysis.

2.1.2 WD-Tools

Fire Flow Analysis

Calculating water availability for fire protection requirements is one of the
most frequent modeling tasks of water utility. The Fire Flow Analysis module
allows you to calculate the available flow for the design pressure or to calcu-
late the residual pressure for the design flow.

There are two basic ways to model a fire flow:

 Specify a design fire flow rate and compute the available fire flow pres-
sure.
 Specify a design fire flow pressure and compute the available fire flow
rate.

Pipe Roughness Calibration

The Pipe Roughness Calibration module automatically adjusts pipe rough-
ness coefficients to best match field pressure observations and best reflect
what is actually occurring in your system. Both English and Metric units are
fully supported in the Pipe Roughness Calibration. Pipe roughness coeffi-
cients can be calculated for the Hazen-Williams, Darcy-Weisbach (Cole-
brook-White), or Manning friction loss equations.

PID Control
The PID (Proportional-Integral-Differential) Control module provides
advanced Real-Time Control simulation capabilities for water supply and
water distribution systems. It permits description of various controllable
devices and makes the definition of complex operational logic for interde-
pendent regulators fully transparent and time efficient.

The following controllable devices can be specified:

 Control valve (PRV-Pressure Reducing Valve, PSV-Pressure Sustaining

Valve, PBV-Pressure Breaker Valve, TCV-Throttle Control Valve, FCV-
Flow Control Valve)
 Pump

109
 Using the Model

Variable Speed Drive

A variable pump can vary its speed setting and/or change its status to open or
closed during a simulation. VSD pump extension allows modeling of a pump
where the rotational speed of a pump is automatically adjusted by the pro-
gram in order to maintain the control node pressure at the required level. VSD
pump using can also be modeled using the PID control module.

Extended Rule-Based Controls

Rule-Based Controls allow link status and settings to be based on a combina-
tion of conditions that might exist in the network over an extended period sim-
ulation. In order to allow for modeling of chains of pumps or valves chains in
efficient way, the EPANET Rule-based control syntax was extended by add-
ing LIKE keyword, for example.

Demand Adjusted Analysis

Traditionally, water demands are defined prior to the simulation and thus
independent of the actual pressure. With the Demand Adjusted Analysis, the
Wagner equation can be used to adjust the node demands based on the
available pressure.

Pressure Zone Mapping

This tool helps to visualise how different network parts are hydraulically inter-
connected and where the HGL line breaks.

Water Hammer
Water hammer simulates transient (unsteady) flow in any fully pressurized
system carrying liquids. MIKE URBAN Water Hammer provides a cost effec-
tive tool for engineers seeking fast answers to questions about rapid opera-
tion of piping systems. Water hammer is based on the high-order implicit
scheme solving the continuity and momentum equation using the finite differ-
ence method. The initial conditions are modeled using MIKE URBAN
EPANET based steady state flow module.

Water Hammer allows you to model:

 Pump start-up and pump trip-off.

 Water hammer protection of pumping stations.
 Power failure events.

110 MIKE URBAN - © DHI

Application Basics

2.2 Application Basics

The following subsections discuss the application basics of MIKE URBAN
Water Distribution. These subsections describe the steps involved in defining
a water distribution model and what data is required for the model definition.

2.2.1 Defining the Model

It is generally more efficient to gather and organize the data required to define
your pipe network model before you begin input of data into MIKE URBAN
Water Distribution.

MIKE URBAN Water Distribution views the water distribution network as a

collection of links connected together at their end-points by nodes. Links and
nodes are identified with ID numbers and can be arranged in any fashion.
Figure 2.1 shows an example water distribution network and its related com-
ponents.

Figure 2.1 The network components that make up a typical water distribution
model

For large networks, collecting the input data to define a pipe network model
can involve considerable effort. Also, entering this input data into the com-
puter can, at times, become tedious. This process can be prolonged by errors
introduced while entering the data. To reduce the possibility of error, the fol-
lowing procedure is recommended for preparing and entering the input data
and performing the network analysis.

111
 Using the Model

1. Obtain a map, or create a schematic diagram of the pipe network to be

modelled. Number all pipes, junction nodes, and control components in
the network. These numbers will then be used as the numerical or string
IDs by MIKE URBAN Water Distribution in its analysis. These IDs can be
used in checking your work. In addition, pipes, nodes, pumps, valves,
storage tanks, and all other network components can also be given text
descriptions for reference purposes.

2. Prepare the input data for each pipe and junction node. The data for
these components should be complete. For example, pipe data includes
the pipe length, diameter, roughness coefficient, and minor loss coeffi-
cient for each pipe in the network. If the roughness coefficient is not
available, the pipe material and age can be used to interpolate a rough-
ness coefficient value using the roughness coefficient tables provided in
MIKE URBAN Water Distribution. Junction node data includes the water
demand and node elevation for each node in the network.

3. If there are control components in the network, prepare the input data for
each component. Control components include pumps, check valves, reg-
ulating valves, sustaining valves, flow control valves, and storage tanks.

4. Enter the input data which defines the network model using the MIKE
URBAN Water Distribution data input dialog boxes. The program will
report an error if invalid data is entered. This immediate checking of input
data helps prevent errors from being introduced into the pipe network
model.

5. After the input data has been entered, MIKE URBAN Water Distribution
can perform a complete check of the input data and a geometric verifica-
tion of the network connectivity. In addition, it can print a summary of the
input data to allow the user to verify the entered data. Many of the input
data dialog boxes, such as the and Pipe Editor dialog boxes, can be
placed side by side to allow the user to more easily verify model consist-
ency.

6. After the input data has been verified, the pipe network can then be ana-
lysed. Analysis results should be carefully examined to make certain that
the input data accurately defines the model and that the analysis results
appear reasonable. Of particular importance are the junction node pres-
sures. These values should lie within a reasonable range.

7. Calibrate the network model to make certain that the analysis provides
accurate modelling results. For more information on how to calibrate
your network model, see the section titled “Model Calibration” on
page 243.

112 MIKE URBAN - © DHI

Application Basics

Constructing Complex Networks

When constructing a complex pipe network model, begin by first constructing
a simplified version of the network. For example, leave out the network con-
trol components. Then, analyse the network and verify that the modelling
results appear reasonable. If they are, then add the control components to
the network.

Roughness Coefficient Data

Model calibration should be performed to properly adjust the pipe roughness
and minor loss coefficients so that the defined computer model pipe network
accurately models the actual physical pipe network.

However, as a pipe network ages, pipe roughness change due to corrosion

and deposition. During a pipe network's design stage, it is important to know
how age can affect the pipe roughness coefficients. Roughness coefficient
data from textbooks and engineering handbooks may not accurately reflect
the aged roughness coefficients. Generally, accurate aged roughness coeffi-
cients can be obtained from the pipe manufacturer. Otherwise, laboratory
experiments or field measurements may need to be performed to estimate
these values.

2.2.2 Input Data Requirements

Defining a pipe network model requires a variety of complex and interrelated

input data. The input data required to define a pipe network model is catego-
rized and described in this section.

Project Data
General data describing the pipe network model, such as project and output
specification settings, is categorized as project data.

Project Specification Data

Included in the project specification data are general settings, such as friction
loss formulation, units used in the analysis, and the type of analysis to be
conducted (i.e., steady state, extended period, water quality, etc.). If the anal-
ysis is an extended period simulation, this data also defines the total simula-
tion time and time step size.

Network Data
Network data consists of all components that make up the network model,
such as pipes and junction nodes. This subsection lists all components that
can be used in a steady state analysis.

113
 Using the Model

Pipe Data
Pipe data includes the physical characteristics of each pipe, such as pipe
length, diameter, roughness coefficient, and minor loss due to fittings along
the pipe. In addition, check valves can be defined along a pipe to prevent flow
reversal. The pipe data provides the basic information for the hydraulic analy-
sis, and should be prepared carefully.

A text description can be defined for each pipe. Because pipes are associ-
ated with the roads and streets that they lie under, a pipe is generally labeled
by its associated street name.

Junction Data
Junction data describes the physical characteristics of each junction node,
such as external water demand and node elevation.

A text description can be defined for each junction node. Junction nodes are
often labelled with names corresponding to the buildings they are located
near.

Reservoir Data
Reservoir data specifies the reservoir's water surface elevation. In addition, a
text description can be defined for each reservoir.

Storage Tank Data

Storage tank data includes the minimum, maximum, and initial water surface
elevations for each storage tank. MIKE URBAN Water Distribution distin-
guishes between cylindrical, rectangular, and any shape storage tanks. The
geometry of a cylindrical tank can be fully described by its diameter, while
rectangular tanks are defined by a width and length. A text description can be
defined for each storage tank.

Emitters Data
Emitters are devices associated with junctions that model the flow through a
nozzle or orifice. In these situations the demand (i.e. the flow rate through the
emitter) varies in proportion to the pressure at the junction raised to some
power. The constant of proportionality is termed the "discharge coefficient".
For nozzles and sprinkler heads the exponent on pressure is 0.5 and the
manufacturer usually states the value of the discharge coefficient as the flow
rate in gpm through the device at a 1 psi pressure drop. Emitters are used to
model flow through sprinkler systems and irrigation networks. They can also
be used to simulate leakage in a pipe connected to the junction (if a dis-
charge coefficient and pressure exponent for the leaking crack or joint can be
estimated) and compute a fire flow at the junction (the flow available at some
minimum residual pressure).

Pump Data
Pump data is defined by selecting one of four pump types: constant power,
standard pump curve, custom pump curve with no extended flow range, or
custom pump curve with extended flow range. Each of these pump types

114 MIKE URBAN - © DHI

Application Basics

require specific data. If a pump is a constant power pump or a constant head

pump, the pump is described by the power or head setting of the pump. If a
pump has a pump curve associated with it, the pump is described by a pump
characteristic curve consisting of pump head and discharge values. A text
description can be defined for each pump.

Valve Data
Valve data includes the valve type, the HGL setting of the valve, and the ref-
erence junction node. In MIKE URBAN Water Distribution, there are six valve
types: pressure reducing valves (PRV), pressure sustaining valves (PSV),
pressure breaker valves (PBV), flow control valves (FCV), throttle control
valves (TCV) and general purpose valves (GPV). The GPV allows the user to
define the flow-headloss relationship through the valve. Most pipe networks
use valves to regulate flowrates and pressures. Valves greatly affect the anal-
ysis results. Therefore, valves should be defined carefully. A text description
can be defined for each valve. A check valve is not defined with the valve
data, but is defined with the pipe data.

Consumption Data
Consummation data includes junction node demands and time patterns to
additional categories of water users at a junction as well as demand alloca-
tion data, and pipe demand coefficients. For more information on how to
develop node demands, see the section titled “Distributed Demands” on
page 119.

Extended Period Data

If an extended period analysis is to be performed, extended period (dynamic)
components such as pressure switches can be incorporated into the pipe net-
work.

Control Data

Control data allows pipes, pumps, and valve settings to change at specific
times or when specific pressures or tank levels are reached in the network. A
text description can be defined for each control.

Water Quality Data

If a water quality analysis is to be performed, then specialized data is
required.

Point Source Data

Point constituent source data includes the reservoir, storage tank, or junction
node where the constituent (e.g., chlorine) is introduced into the pipe net-
work. In addition, the substance's decay rate, current age, and the amount
introduced is specified. A text description can be defined for each point
source.

115
 Using the Model

The decay/growth rate is defined using water quality reaction equations.

MIKE URBAN Water Distribution can model either bulk flow reactions or reac-
tions with the pipe wall. Bulk flow reactions are reactions that occur in the
main flow stream of a pipe or in a storage tank, unaffected by any processes
that might involve the pipe wall. Pipe wall reactions are reactions that may
involve the pipe material or occur close to the pipe wall out of the main flow
stream.

Fire Flow Data

MIKE URBAN Water Distribution can be used to perform a fire flow analysis
for any junction node within the network system. Please refer to the Tutorials
Guide for a complete discussion on how to perform a fire flow analysis with
MIKE URBAN Water Distribution.

Water Hammer Data

MIKE URBAN Water Distribution can be used to perform a water hammer
analysis for the network system. See the Tutorials Guide for a tutorial on how
to run a Water Hammer analysis as well as Water Hammer Analysis Chapter
for a complete discussion on how to perform a water hammer analysis with
MIKE URBAN Water Distribution.

2.3 Entering Data

MIKE URBAN is extremely flexible in how a water distribution model can be
developed. The user can develop a model from scratch using a variety of
input methods, including importation of data files from a GIS database or pre-
existing water distribution models, schematically drawing the pipe network, or
by simple data entry.

2.3.1 Interactive Data Entry

MIKE URBAN has been designed to make developing a water distribution

model easy and flexible by providing a variety of interactive input methods.
For example, the user can graphically trace out an existing water distribution
pipe network on top of a scanned-in aerial map. Or, the user can manually
specify the pipe network system through the program’s interactive network
component editors. And any of these methods can be used simultaneously
with any other, allowing complete flexibility while defining the network system.
This translates into greater productivity for the user.

Most data used to describe a water distribution model is defined using inter-
active dialog boxes, an example of which is shown in Figure 2.2. These dia-
log boxes allow the user to quickly comprehend what data input requirements
are needed to define the model.

116 MIKE URBAN - © DHI

Entering Data

Figure 2.2 This interactive dialog box illustrates how data is entered to define a
water distribution network model

The MIKE URBAN Water Distribution data input requirements, the methods
available to describe this data, and the dialog boxes used to enter this data
are discussed in detail in Chapter “EPANET Input Description” on page 11.

2.3.2 Graphical Input

As was discussed in the previous section, MIKE URBAN allows you to enter
data interactively into the pipe network using interactive dialog boxes. This
allows you to quickly define a model if a schematic layout of the pipe network
is unavailable, or if developing a schematic layout is too expensive or time-
consuming to construct. However, a powerful capability of MIKE URBAN is to
allow the user to graphically construct the pipe network model by simply
pointing and clicking with a mouse. This graphical construction of a pipe net-
work system is shown in Figure 2.3.

117
 Using the Model

Figure 2.3 MIKE URBAN allows the user to interactively develop a pipe network
model by simply pointing and clicking to graphically add the compo-
nents to construct the water distribution network

Using the tools in the Component toolbar, for example, the user can select
the junction node component tool (Add Junction icon) and then interactively
locate and place junction nodes on the Map window. A background aerial
image can be displayed in the Map window to aid in the placement of network
components. While the user is moving the cursor in the Map window, a pre-
cise X,Y location is displayed on the status line at the bottom of the program.

To delete a network component, the select component tool can be used inter-
actively to select network elements. Then, selecting «Delete Selected» will
delete the selected elements. Note that you can undelete the deleted ele-
ments.

2.3.3 Importing Graphical Data

MIKE URBAN allows you to import network data, such as ESRI Shapefiles,
DWG, DGN, DXF files and raster images (such as scanned-in quad maps or
aerial photographs), and display them as a background layer in the Map win-
dow. This enables you to import a graphical background layer to facilitate lay-
ing out the pipe network. You can also import a DXF file containing lines and
polylines representing the pipe network and MIKE URBAN will convert the
lines and polylines to equivalent pipes.

To import a graphical data file, select Add Data tool. MIKE URBAN will then
display the Add Data dialog box, as shown in Figure 2.4. Select the file format
to import and then choose «Add».

118 MIKE URBAN - © DHI

Demand Processing

Figure 2.4 The Import dialog box provides support for importing vector files and
raster image files

Hiding and Showing Graphical Data

Once graphical data has been imported and displayed as a background layer,
it may be necessary to hide (or show) this layer or other layers. To do so,
select View | Table of Contents. This will display the Table of Contents dia-
log box.. This dialog box allows you to control what layers are to be displayed
in the Horizontal View window. Select the layer to hide (or show) and then
choose «Close». MIKE URBAN will the update then display.

2.4 Demand Processing

MIKE URBAN Water Distribution provides the user with a wide range of tools
suitable for network demand processing including individual node demand
definition, multiple demands, diurnal demand curves, automatic demand dis-
tribution based on the node and pipe demand coefficients, and reusing data
from the customer information system.

2.4.1 Distributed Demands

For large network systems, assigning demand data can be very tedious job.
Since many times the total demand is known for a particular network pres-
sure zone or for the entire network system, MIKE URBAN Water Distribution
provides the capability to distribute this total demand among the applicable
junction nodes.

The Distributed Demands dialog box, as shown in Figure 2.5, is used to auto-
matically assign the demands at the appropriate junction nodes. The Distrib-
uted Demands dialog box is reached by selecting EPANET | Distributed
Demands.

119
 Using the Model

Pipe Demand Coefficients

MIKE URBAN computes the water demands for each node in the network
system based upon the total network demand using two methods: the Method
of Pipe Lengths and the Method of Two Coefficients. This is useful when
assigning the nodal water demand for a large network, since the software will
automatically proportion the total network demand based upon one of these
two methods. These methods are used to mimic the amount of actual
demand along a pipe, based upon the pipe length or a pre-defined demand
coefficient.

Figure 2.5 The Distributed Demands dialog box computes the demand at each
junction node automatically based upon the total demand within a pres-
sure zone or entire network system

See the lesson in Tutorials Guide on Distributed Demands for a complete

description on how to perform a demand distribution with MIKE URBAN
Water Distribution.

METHOD OF TWO COEFFICIENTS or METHOD OF REDUCED

PIPE LENGTHS or METHOD OF EQUIVALENT PIPE LENGTHS
MIKE URBAN allows the user to compute the nodal water demands based
upon the total network demand using two methods: the Method of Pipe
Lengths and the Method of Two Coefficients. This radio button group allows
the user to select the method to be used.

120 MIKE URBAN - © DHI

Demand Processing

Selecting the Method of Two Coefficients, MIKE URBAN computes the total
water demand assigned to each pipe (which is then split between the starting
and ending nodes) as:

 Q k li k 2i
q pi = ---------------------
- (2.1)
  k li k 2i 

Selecting the Method of Reduced Pipe Lengths, MIKE URBAN computes the
total water demand assigned to each pipe (which is then split between the
starting and ending nodes) as:

 Q l i k li
q pi = -----------------
- (2.2)
  k li l i 

Selecting the Method of Equivalent Pipe Lengths, MIKE URBAN computes

the total water demand assigned to each pipe (which is then split between the
starting and ending nodes) as:

 Q l i k Di
q pi = -------------------
- (2.3)
  k Di l i 

where:

qpi =Total water demand applied to the pipe, split between the two end nodes.

Q = Total network water

li = Pipe length

k1i, k2i = pipe demand coefficient

kDi = pipe demand coefficient is calculated by the program as a factor, calcu-

lated as pipe diameter/diameter_normal (where diameter normal is 150mm or
6inch). This helps to scale the pipes based on their diameter i.e. perimeter;
this method is recommended when the distributed demand corresponds to
the amount of leakage.

Node Demand Coefficients

Node demand coefficient allows you, for each node, to define the share of the
whole network demand, which is taken by a node. The total network demand
is then distributed to the corresponding nodes by the Demand Distribution
function.

This option will only assign demand to nodes with Demand Coefficient
applied (different from 0 or NULL). In the case of an equal distribution, the
node demand coefficients have to be equal and different from zero.

121
 Using the Model

EQUAL NODE DEMAND DISTRIBUTION

This check box allows you to distribute the network (or zone) demand equally
to each node within the zone or the network.

Selecting Equal Node Demand Distribution computes the junction demand:

Q
q ni = ---- (2.4)
N

where:

Q= Total network water demand (or zone demand)

qni = calculated demand at each junction node

N = junction nodes count with the selected zone or a total network

NODE DEMAND DISTRIBUTION

Selecting the Method of Node Demand Coefficients, MIKE URBAN computes
the total water demand assigned to each junction node as:

QT
-C
q i = ----------------- (2.5)
i
 i C
i = 0,n

where:

qi = node demand

Qt = total network demand

Ci = demand coefficient

Developing Pipe Demand Coefficients

There are many ways of developing pipe demand coefficients. Typically, pipe
demand coefficients correspond to the amount of invoiced water along the
specified pipe, which is the most accurate data source for the Two Coefficient
Method. Alternatively, the number of inhabitants can be supplied. If such data
is unavailable, it is also possible to classify pipes by the residential type (fam-
ily houses, commercial, city centre) and to use such classification in the Pipe
Length Method.

In cases, when node demands are retrieved from the Customer Information
Systems, it is possible to calculate pipe demand coefficients in the form of
aggregated demands for streets or the counted lots

122 MIKE URBAN - © DHI

Demand Processing

Counted Lots
The pipe demand coefficient is derived from the X,Y position of the counted
lots. In order to use X,Y position of a point-attributes for the geocoding pro-
cess, it is necessary to create a database file, such as Dbase.DBF, Microsoft
Access.MDB where X,Y and other attributes are stored, or create the X,Y
position for each point in ArcGIS, for example. Use the import features of
MIKE URBAN to import such data to the Demand Allocation editor.

Example: the pipe demand coefficient COEFF1 was updated (aggregated)

from the X, Y points (demand lots), which are also displayed in the horizontal
plan window for each corresponding demand allocation point.

Figure 2.6 The demand coefficients (represented by circles) are assigned to the
nearest pipe within the snapping radius defined in the demand alloca-
tion dialog box

Accumulated Street Demands-Polygons

In some cases, it is possible to create polygons, covering streets and to
assign the aggregated demand to them. This solution assumes, that each
street included in the model has its corresponding covering polygon in
ArcView, and the name of a street is included in the model database together
with its sum demand value from the billing database.

123
 Using the Model

Figure 2.7 Polygons created to cover streets

Demand Allocation
Junction node demands can be edited either within the junction editor for
each particular junction node or in the Multiple Demand editor, which allows
the user to display and edit all multiple demands. In addition to that, MIKE
URBAN provides another way of developing junction node demands, which is
based on geocoding the consumption data to the appropriate junctions or
pipes and aggregating their set point demand values to the junction
demands. This allows simplifies the demand development process and
allows you, for example, to import consumption data from the consumption
database systems and geocoded it based on its X, Y geographical coordi-
nates.

See the Tutorials Guide for the tutorial on demand allocation for a complete
description on how to perform a demand allocation with MIKE URBAN Water
Distribution.

Demand Editing and Demand Scenarios

In addition to the automated demand processing; node demands can be
edited individually within the Junction Editor and/or the Multiple Demand Edi-
tor. Each node can have unlimited number of demands and each demand

124 MIKE URBAN - © DHI

Performing an Analysis

can be linked to its diurnal curve. For more details, see Junction Editor and
Multiple Demand Editor.

Global Editing
Using the Field Calculator tool can change node demands. This allows us to
select the part of the network and to increase the node demand by 20%, for
example.

To do this, open the Multiple Demand Editor, select the column Demand and
select the Field Calculator from the right-mouse click pop-up menu. Define
the following statement:

DEMAND=DEMAND * 1.2

To increase node demands associated with the specific diurnal curve (Pat-
tern), use the following statement:

DEMAND=DEMAND * 1.2 where PATTERN =P1

Convenient way of handling different scenarios is to change the Total Net-

work Demand within the Distributed Demand dialog and to increase or
decrease the node demands by redistributing another total consumption.

2.5 Performing an Analysis

Once you have defined a water distribution input model, you are ready to per-
form an analysis. The following tasks are required in performing an analysis.

First, you need to check the model for errors. MIKE URBAN contains a pro-
ject check tool that can verify the data used to define an EPANET model (see
more in the Model Manager User Guide).

Second, after the program has checked the model for errors, you can then
execute the EPANET network hydraulic analysis. The EPANET network anal-
ysis program will read the input data and will then compute the flow rates,
pressures, and water quality for the defined pipe network.

The following subsections discuss these tasks in detail.

2.5.1 Executing the Analysis

Once the project check tool has successfully checked the model for potential
problems, you are ready to perform an analysis of the water distribution net-
work model. To execute the analysis, select Simulation | Run Simulation,
as shown in Figure 2.8.

125
 Using the Model

Figure 2.8 The File Menu is used to select PERFORM ANALYSIS to analyze the
water distribution network

Please note that EPANET engine does not handle special characters such as
< > ? " : | \ / * . Make sure that neither the database nor the scenario ID con-
tain such characters.

After completing the analysis, the program will report the total number of
warnings generated during the analysis and whether the analysis ended with
an error. You can then view the analysis results. If the analysis ended with an
error or if the warnings are serious, you can then make the appropriate cor-
rections to the input data.

2.5.2 Batch simulations

The ‘Simulation|Run Batch Simulation’ allows you to select any number of

available scenarios and use them for running the simulation. You can use ">",
">>", "<", and "<<" commands to select or deselect available scenarios. Once
done, select "OK" to start the batch analysis.

126 MIKE URBAN - © DHI

Viewing Analysis Results

Figure 2.9 The LF1_N and LF2_N scenarios are selected for the batch analysis

2.6 Viewing Analysis Results

Once the analysis has been run successfully, the analysis results can be dis-
played. To load the computed analysis results, select Model Results| Load
Simulation Results. MIKE URBAN will then load the computed analysis
results.

The water distribution analysis results are written to an output file with the
same filename (but with a RES file extension) as the input data file.

To load EPANET results, go to EPANET | Load Results | EPANET Results.

A dialog with choices of which data to load from the result then appears as
shown in

127
 Using the Model

Figure 2.10 EPANET result file selection dialog

In addition to the standard EPANET analysis results, MIKE URBAN can dis-
play the output results in a summary table, generate custom output reports,
and display many different graphical representations of the analysis results.
The following sections describe these capabilities.

2.6.1 Comparing Alternative Solutions

MIKE URBAN can display multiple analysis result files for a network. In order
to compare two analysis result files for a network (for example, to compare
design alternatives of the same network), MIKE URBAN can subtract the two
analysis result files from each other. MIKE URBAN will then display the differ-
ence between these two analysis results for both node and pipe values (such
as node pressure, hydraulic gradeline, water quality, demand, flow, velocity,
and hydraulic headloss). The result difference can be displayed in the same
format as the standard analysis results.

To subtract two analysis result files, select Model Results| Results Compari-
son. Note that it is only possible to subtract two analysis result files if the
number of nodes, pipes, and time-steps (if performing an extended period
simulation) are the same. The following comparisons can be performed.

128 MIKE URBAN - © DHI

Viewing Analysis Results

Different Demands
Network demands can be different (such as comparing a work and weekend
day), or local nodal demands can increase (such as simulating fire flows or
higher demands from what was predicted from the original master plan).

Different Pipe Roughness

Model calibration requires many changes in pipe roughness coefficients. Pipe
aging can also be modeled (where the pipe roughness increases due to
aging).

Different Pumping Scenarios

Different pump schedule scenarios can be compared. Network optimization
may be based on reducing pumping to save money used in powering pump
stations. By displaying the original and improved network results, it is easy to
show that the new network pumping schedule meets the same criterion (flow
and pressure specifications) with less pumping and/or pumping at more off-
peak hours.

Effect of Pressure Reducing Valves

Decreasing the maximum pressure within the network has the positive effect
of decreasing water leakage at pipe, pump, and valve joints. Optimizing pres-
sure reducing valves (PRVs) settings can sometimes be a primary task of
network optimization. Displaying the pressure differences between the origi-
nal and optimized network provides a very good overview of potential water
losses.

Water Quality Parameters

Modeling different water quality scenarios (such as different chlorine supply
patterns from water treatment booster stations) allow the user to quickly
examine water quality for the network.

When subtracting the two analysis result files, a positive resulting value
means that the first analysis result file has a larger value than the second
analysis result file.

2.6.2 Viewing the EPANET Analysis Results

When EPANET performs its analysis of the water distribution network, it gen-
erates two ASCII output files containing the analysis results—a summary out-
put file and a complete output file. The summary output file contains a short
description of the analysis results. The complete output file contains the con-
tents of the summary output file in addition to output results for every compo-
nent within the water distribution network.

129
 Using the Model

To view the EPANET summary analysis results, open the.SUM file using
Notepad, for example or select Summary from within Run Simulation dialog.

2.6.3 Result Browser Window

The Result Browser window allows you to graphically select any network
component in the Map window, by simply clicking on it with the mouse, and
the program will then display that component’s input attributes and analysis
results. This allows you to quickly examine the pipe network system at the
component level (i.e., pipe, junction node, valve, pump, tank, and reservoir),
check what is defined for the model, and determine the computed analysis
results. For example, selecting a pipe from the Map window will display in the
Browser window the pipe’s ID, diameter, length, roughness coefficient, minor
loss coefficient, reaction rate coefficient, open/close status, and flowrate.

Please also refer to the Model Manager User Guide for further on presenting
simulation results.

2.6.4 Results Table

The Results Table allows you to display the tabular list of analysis results for
junctions, tanks, pipes, pumps, and valves. It is possible to display the results
for all elements or for the selection set. The selection se can be defined
based on the rules (constraints) such as ”pressure below 15”, “velocity above
1.5”, and similar. It is also possible to display the tabular results for nodes and
links, which are graphically selected from the Map.

2.6.5 Thematic Mapping

The Map window allows you to graphically plot the analysis results directly
onto the pipe network schematic. To add a layer to the Map window, select
Model Results | Add Results Layer and pick up the selected result item, such
as Link:Flow, from the list.

The user can change most of the display options. In the Map window, com-
plete thematic mapping of the analysis results is available, including node
elevation, HGL, pressure, demand, and any water quality constituent. This
allows you to quickly interpret the modeling results and identify any trouble
areas. And, directional flow arrows can be plotted on top of the pipes to show
the flow direction for any time-step. In addition, MIKE URBAN provides auto-
matic color-coding of pipes and nodes based upon any input or output prop-
erty, allowing the network to be color-coded based upon pipe sizes, flowrates,
velocities, headlosses, nodal pressures, nodal demands, hydraulic grades,
elevations, water age, percent source contributions, water quality concentra-
tions, and any other attribute. Numerical ranges for colors can be specified.
Furthermore, pipes can be plotted with variable width and nodes with variable

130 MIKE URBAN - © DHI

Viewing Analysis Results

radius, allowing you to quickly identify those areas of the network experienc-
ing the most flow, headloss, water quality constituent concentration, etc.

The link arrows corresponding to the link orientation, pipe slope and/or to the
flow direction can be drawn by the selected size. The node and link labels
can be displayed for all elements or only for selected elements. It is possible
in this way to display the labels just for filtered nodes and pipes, for nodes
and pipes along a specific profile or a flow path, for the individually selected
nodes etc. This is useful especially when we want to display very quickly
results for several elements on the plan. The values of pipe diameter and
pipe length can be displayed together with the prefix.

Please also refer to the Model Manager User Guide for further on presenting
simulation results.

2.6.6 Profile Plots

Profile plots allow you to graphically plot the analysis results along any pipe-
line path. To display a profile plot, a profile path must first be defined from the
pipe network horizontal plan. Once the profile path has been defined and the
profile plot displayed, the path can be saved for later re-use.

Profile plots allow you to graphically plot the analysis results along any pipe-
line path. Profile plots can have two separate vertical axes to allow plotting of
variables from two separate unit families, such as flow and pressure. Profile
plots can be plotted along any user-specified route. Profile plots can be gen-
erated as line graphs, bar graphs, or mixed—along with complete graph cus-
tomization. For example, profile plots can be plotted with an envelope to
show the minimum and maximum values reached during an extended period
simulation.

Please also refer to the Model Manager User Guide for further on presenting
simulation results.

2.6.7 Time Series Plots

Time series plots allow you to graphically display the analysis results for any
network element for an extended period simulation. Multiple time series plots
can be generated for the various network elements, such as pipe flow, veloc-
ity, headloss, nodal demand, pressure, hydraulic grade, water age, water
quality constituent concentration, pump characteristic operating curve, tank
water level, total and net system demand, etc.

Please also refer to the Model Manager User Guide for further on presenting
simulation results.

131
 Using the Model

2.6.8 Link Q-H Plots

Time series scatter graphs can be used to display the Q-H relation for any
selected link for an extended period simulation.

2.6.9 Fire Hydrant Q-H Plots

The Q-H relation plots allow you to graphically display the fire flow analysis
results (calculate hydrant curve) for any fire hydrant node.

2.6.10 Current Operating Pump Q-H Point

The position of the current operating points can be displayed for any pump for
an extended period simulation. In order to display this graph, load the simula-
tion results, open the Pump Editor and display the graph.

2.6.11 Check for Pipes and Nodes

It is possible to detect pipes where the flow direction has changed during the
hydraulic simulation. The same tool can also be used for detecting links and
nodes, where the results (pressures, for example) changed more than a
specified limit.

2.6.12 Animation

MIKE URBAN will automatically generate animations of extended period sim-

ulations for both horizontal plan plots and profile plot plots. animations show
values that change with respect to time for extended time period simulations.
To display the Animate dialog box, select Animate from the Model Results
menu. Note that in order to succeed, either map window or other plot must be
displayed. Also we recommend setting the options of the animated plot on
color or grade coding to get better results.

See the Model Manager book for more information on viewing simulation
results.

2.6.13 EPANET Error and Warning Messages

In order to display EPANET error and warning messages, generated during

the steady state or extended period simulation, select Model Results |
EPANET Errors and Warnings.

The EPANET Errors and Warnings dialog box allows you to read the error
and warning messages from the .SUM file, generated by EPANET and store
this information in the database manner. All messages are classified based
on their type, which makes it possible to filter out errors or warning related to

132 MIKE URBAN - © DHI

Viewing Analysis Results

pumps or to filter out errors and warning for a specific time level or an model
element, for example.

Figure 2.11 The EPANET Errors and Warnings dialog box is used to process simu-
lation messages

A list of the EPANET Errors and Warnings dialog box data fields for Figure
33.29follows, with a short description given for each entry.

MESSAGE TYPE
This list box classifies the error and warning message, such as FCV, Pump,
Pipe, Negative Pressures, System Unbalanced and similar.

LINK/NODE ID
This data entry contains the model element ID, where applicable.

LINK/NODE TYPE
This data entry contains the model element type, where applicable, such as
Junction, Tank, Pipe, Pump, or Valve.

MESSAGE
This is the error or warning message generated by EPANET engine.

SIMULATION TIME
Simulation time of the error or warning message is available in
[days:hrs:min:sec] format as well as the number of elapsed seconds from the
simulation start.

ADVANCED...
It is possible to Select or Select All messages and map highlight the corre-
sponding network elements on the Map; it is also possible to open Select by
Attribute dialog box and filter out messages based on their type, simulation
time, model element identification or a type. HTM based report can also be
generated.

133
 Using the Model

2.6.14 Store Results to Database

It is possible to store the complete model results (for all time steps) and for
the selected part of the network into the database. This is useful when you
want to use Microsoft Access or Excel tools for analysing the model results
and for using user-defined tools for the additional results processing. In order
to store the model results to the database, load the simulation results into
MIKE URBAN, and select nodes and links you want to use for storing the sim-
ulation results to the database. You can also store all nodes and links in case
you do not use the selection set.

134 MIKE URBAN - © DHI

Fire Flow Analysis

3 Water Distribution Tools (WD-Tools)

Water Distribution Tools provides the user with modules and extensions
allowing for automatic fire flow analysis, pipe roughness calibration, PID con-
trol and demand adjusted analysis. Apart from that, several extensions to the
standard EPANET such as VSD (variable speed drive pump) or extended
syntax for Rule-Based Controls are available from within standard program
dialogs. For information on Water Hammer, please refer to the separate
chapter.

3.1 Fire Flow Analysis

The Fire Flow Analysis module allows you to calculate the available flow for
the design pressure or to calculate the residual pressure for the design flow.
Fire flow requirements are one of the most common design requirements
when designing the new or evaluating the existing water supply and water
distribution system.

A fire flow is the maximum flow rate available at a specific minimum pressure,
typically 20 psi (15m). There are two basic ways to model a fire flow:

1. Specify a design fire flow rate and compute the available fire flow pres-
sure.

2. Specify a design fire flow pressure and compute the available fire flow
rate.

A unique feature of MIKE URBAN is its capability of modelling a fire hydrant

connecting pipe with the user defined resistance and other properties. This
pipe connects the fire flow node with a fictitious reservoir whose water level
equals the required residual pressure. This concept helps to calculate more
accurate or realistic results.

The Fire Flow Analysis dialog box is reached by selecting EPANET| WD

Tools| Fire Flow Analysis. Note that to run the fire flow analysis, you need to
select “Fire Flow Analysis“ from within the Run Simulation dialog box.

135
 Water Distribution Tools (WD-Tools)

Figure 3.1 The Fire Flow dialog box is used to specify fire flow analysis parameters

A list of the Fire Flow Analysis dialog box data entries for Figure 3.1 follows,
with a short description given for each entry.

Calculate available flow for design pressure, Calculate available

pressure for design flow, Calculate fire hydrant Q-H curve for selected
node
This radio button selection allows you to select the fire flow analysis type. It is
possible to specify a design fire flow rate and compute the available fire flow
pressure or to specify a design fire flow pressure and compute the available
fire flow rate. In addition to this, it is also possible to calculate the Q-H curve
for the selected junction node.

Design fire pressure

This data entry is used to define the design (required) residual fire pressure
for which the available fire flow will be calculated.

136 MIKE URBAN - © DHI

Fire Flow Analysis

Maintain minimum residual pressure

This data entry allows you to define the minimum residual pressure for the fire
flow analysis. The fire flow analysis is then iterated as long as the minimum
residual pressure is maintained in the appropriate nodes (as specified in the
‘report critical nodes and pipes section’).

Maintain maximum velocity

This data entry allows you to define the maximum acceptable velocity in
pipes during the fire flow analysis. The fire flow analysis is then iterated as
long as the maximum velocity is maintained in the appropriate pipes (as
specified in the ‘report critical nodes and pipes section’).

Design fire flow

This data entry is used to define the design (required) fire flow for which the
available (residual) fire pressure will be calculated.

Node demand multiplier

This data entry allows you to specify the node demand multiplier which will be
used to define the required fire flow demand.

Simulate fire hydrant

This data entry allows you to model a fire hydrant connecting pipe at the junc-
tion node. Note that the fire flow results may significantly change with or with-
out such pipe.

Fire hydrant local loss coefficient

This data entry allows you to define a local loss coefficient (sum of) repre-
senting all local losses at the fire hydrant connecting pipe. A typical value is
6-10.

Connecting pipe diameter

This data entry allows you to define a diameter of the fire hydrant connecting
pipe.

Connecting pipe length

This data entry allows you to define a length of the fire hydrant connecting
pipe.

Connecting pipe roughness

This data entry allows you to define a roughness coefficient of the fire hydrant
connecting pipe.

Report critical nodes and pipes

This data entry allows you to report the critical junction nodes, where the min-
imum residual pressure is less than the critical pressure.

Critical node pressure

This data entry allows you to define the critical residual pressure.

137
 Water Distribution Tools (WD-Tools)

Maximum velocity
This data entry allows you to define the maximum velocity in pipes

Critical node/pipe search radius

This data entry allows you to define the search node and pipe radius, which
will be used to identify nodes where the minimum residual pressure is less
than a critical pressure or pipe where the maximum velocity is higher than the
maximum velocity limit.

Search within the same zone

This data entry allows you to define that the search for the nodes should hap-
pen within the same pressure zone.

Use only selected junction nodes, Use all junction nodes

This radio button selection allows you to define whether the fire flow analysis
is performed for the selected junction nodes or for all junction nodes.

Select time level

This data entry allows you to define a time level at which the fire flow analysis
is performed. Note that the default value is 0 i.e. the first time level corre-
sponding with the steady state analysis.

3.1.1 Browsing Results

Select Model Results | Results Browser to display the Results Browser win-
dow and click the junction node to view the simulated fire flow results.

3.1.2 Tabular Results

The simulated fire flow results for “Calculate available flow for design pres-
sure“ and “Calculate available pressure for design flow“ are reported to the
*.FFA text file. Open this text file using any file processing program such as
Windows Notepad to see the detailed list of the fire flow results for each sim-
ulated junction node. The FFA text file format is suitable for importing into
Microsoft Excel, for example. The simulated fire flow results for “Calculate fire
hydrant Q-H curve“ are reported to the *.FQH text file.

3.1.3 Thematic Maps

The simulated fire flow results can be displayed using the colour coded maps.
Select Add Results Layer and select one of the fire flow result items to create
a colour coded map with the fire flow results.

List of available fire flow result items:

 Static pressure: steady state pressure at the fire flow node

 Static flow: steady state demand at the fire flow node

138 MIKE URBAN - © DHI

Pipe Roughness Calibration

 Residual pressure: simulated or given residual pressure during the fire

flow simulation at the fire flow node
 Fire flow: simulated or given fire flow at the node
 Critical events: the number n means that the pressure at the node was n
times below the critical pressure during the fire flow simulation (for the
surrounding nodes).
 Status: error code (“OK”, “Cannot find minimum pressure limit“, “Maxi-
mum pressure is lower then residual pressure“, “Cannot iterate flow for
the residual pressure“).

3.2 Pipe Roughness Calibration

The Pipe Roughness Calibration module automatically adjusts pipe rough-
ness coefficients to best match field pressure observations and best reflect
what is actually occurring in your system.

One of the main calibration parameters in the pipe network hydrodynamic

model is the roughness coefficients. Pipe roughness values may be esti-
mated in two ways: using values from literature or directly from field measure-
ments. To obtain initial estimates of pipe roughness through field testing, it is
a good practice to divide the water distribution system into composite zones
that contain pipes of like material and age. Additionally, several pipes of dif-
ferent diameters should be tested in each zone to obtain individual pipe
roughness estimates. The process of calibration ideally requires simulation
over an extended period of time, such as a time range for the maximum day -
not just the maximum and minimum hour for the maximum day.

Both English and Metric units are fully supported in the Pipe Roughness Cali-
bration. Pipe roughness coefficients can be calculated for the Hazen-Wil-
liams, Darcy-Weisbach (Colebrook-White), or Manning friction loss
equations.

3.2.1 Evolutionary Algorithms - Theoretical Background

Evolutionary algorithms (EAs) are engines simulating grossly simplified pro-

cesses occurring in nature and implemented in artificial media -- such as a
computer. The fundamental idea is that of emulating the Darwinian theory of
evolution. According to Darwin, evolution is best depicted as the process of
the adaptation of species to their environment as one of "natural selection".
Perceived in this way, all species inhabiting our planet are actually results of
this process of adaptation.

Evolutionary algorithms effectively provide an alternative approach to prob-

lem solving - one in which solutions to the problem are evolved rather than
the problems being solved directly. The family of evolutionary algorithms
today is divided into four main streams: Evolution Strategies (Schwefel,

139
 Water Distribution Tools (WD-Tools)

1981), Evolutionary Programming (Fogel, 1966), Genetic Algorithms (Hol-

land, 1975) and Genetic Programming (Koza, 1992).

Although different and intended for different purposes, all EAs share a com-
mon conceptual base (schematized in Figure 3.2). In principle, an initial pop-
ulation of individuals is created in a computer and allowed to evolve using the
principles of inheritance (so that offspring resemble parents), variability (the
process of offspring creation is not perfect -- some mutations occur) and
selection (more fit individuals are allowed to reproduce more often and less fit
less often so that their "genealogical" trees disappear in time). One of the
main advantages of EAs is their domain independence. EAs can evolve
almost anything, given an appropriate representation of evolving structures.
Similarly to processes observed in nature, one should distinguish between an
evolving entity's genotype and its phenotype. The genotype is basically a
code to be executed (such as a code in a DNA strand), whereas the pheno-
type represents a result of the execution of this code (such as any living
being).

Although the information exchange between evolving entities (parents)

occurs at the level of genotypes, it is the phenotypes in which one is really
interested.

Figure 3.2 Schematic illustration of an evolutionary algorithm. The population is ini-

tialized (usually randomly). From this population, the most fit entities are
selected to be altered by genetic operators exemplified by crossover
(corresponding to sexual reproduction) and mutation. Selection is per-
formed based on certain fitness criteria in which the more 'fit' are
selected more often. Crossover simply combines two genotypes by
exchanging sub-strings around randomly selected points. In the illustra-
tion above, parental genotypes are indicated as either all 1s or all 0s, for
the sake of clarity. Mutation simply flips the randomly selected bit

The phenotype is actually an interpretation of a genotype in a problem

domain. This interpretation can take the form of any feasible mapping. For
example, for optimization and constraint satisfaction purposes, genotypes are
typically interpreted as independent variables of a function to be optimized.
Along these lines, one can employ mapping in which genotypes are inter-
preted as roughness coefficients in a free surface pipe flow model with the
genetic algorithms (GAs) directed towards the minimization of the discrepan-
cies between model output and measured water level and discharge values.

140 MIKE URBAN - © DHI

Pipe Roughness Calibration

Resulting GA represents an automatic calibration model of hydrodynamic

systems.

3.2.2 Introduction

The pipe roughness calibration will be demonstrated on the example of a

gravity network consisting of approximately 600 nodes and 650 pipes. The
water distribution network is supplied by water from the upstream reservoir
with a fixed HGL = 257 m. The total network demand is approximately 140 l/s,
additional 327 l/s supplies adjacent pressure zones from the main pipeline.
Flow and pressure measurements were conducted on the network. The pipe
age is ranging between 20-80 years and the main pipe material is cast-iron.

Figure 3.3 Map with the network layout

3.2.3 Pipe Roughness Calibration

The calibration of pipe roughness coefficients consists of several steps:

1. Definition of pipe roughness groups and pipe group assignment.

2. Definition of targeted pressure values.
3. Definition of targeted flow values.
4. Detection of closed pipes (optional).
5. Automated calibration of pipe roughness coefficients by Genetic Algo-
rithms.
6. Assignment of calibrated pipe roughness values to the pipes.

141
 Water Distribution Tools (WD-Tools)

Figure 3.4 Network calibration and optimization dialog box

Pipe Groupings
Pipes can be lumped together in separate logical groups based on their
known physical characteristics such as pipe material, age, and diameter. It is
assumed that all pipes within a calibration group (where a single pipe may
constitute a group) will possess an identical roughness coefficient. Any com-
bination of pipe calibration groups can be specified and their fitness evalu-
ated to match field observations. Selected pipes can be excluded from these
groups and their roughness coefficients will remain unchanged during the
calibration process. The Group ID can be assigned for each pipe the Pipe
Editor. Pipes with undefined Group ID will be excluded from the automatic
pipe roughness calibration.

Figure 3.5 Pipes with fixed value of roughness coefficient (k=3mm)

Group Definition
We can create as many pipe groups as necessary and define their minimum
and maximum roughness coefficients. Minimum and maximum pipe rough-
ness coefficient is defined for each pipe. Any group can be also displayed on
the Map.

142 MIKE URBAN - © DHI

Pipe Roughness Calibration

Figure 3.6 Pipe roughness groups define the roughness coefficient limits

Pressure and Flow Measurements

Pressure and flow measurements (fire flow tests, SCADA) can be graphically
designated for any set of junction nodes, pipes in the system. Pipe roughness
coefficients are automatically adjusted so that the model pressure/flow pre-
dictions correlate well with the targeted junction nodes/pipes.

Figure 3.7 Measured pressure in five selected points

The measured flow and pressure values can be entered from within the cor-
responding editors.

Genetic Algorithms Calibration

Genetic Algorithms pipe roughness calibration can be run when the pipes are
assigned to the roughness groups and when one of the flow and pressure tar-
geted values is defined. Several parameters can be defined for the pipe cali-
bration:

 Initial Population Size: 25 (the default).

 Additional Population Growth: 125 (the default).
 Maximum Generation Count: 20 (the default).

143
 Water Distribution Tools (WD-Tools)

The initial population size is the number of initial generation members (the
number of networks), which is used by the automated calibration.

The additional population growth is the number of generation members (the

number of networks), which are used during the automated population selec-
tion and calibration.

Figure 3.8 Calibration parameters

The maximum generation count is the maximum length of the generation

cycle.

Each generation cycle consists of:

1. Initial population estimate.

2. Additional population is created from the initial population size.
3. The hydraulic analysis is performed for each population member (the
network).
4. Initial population for the next generation is created from the best popula-
tion members.

After step 4 has been completed, the cycle is repeated. Three thousand
hydraulic analyses will be executed in this example.

Figure 3.9 Summary file with then calibrated pipe roughness coefficients

The calibration summary is available by selecting Summary from Simulation

dialog, Figure 3.9. If the results are satisfactory, the calibrated pipe rough-
ness coefficients can be loaded and assigned to pipes by selecting Update
Pipes from Pipe Roughness Calibration window. To assign the calibrated pipe
roughness coefficients to pipes, select Advanced-Assign Pipe Roughness

144 MIKE URBAN - © DHI

PID Control

Coefficients to Pipes; note that this will assign the accepted pipe roughness
to each pipe based on the calibration pipe GroupID.

3.3 PID Control

The MIKE URBAN WD-Tools PID Control module features advanced Real-
Time Control (RTC) simulation capabilities for water supply and water distri-
bution systems. It permits description of various controllable devices and
makes the definition of complex operational logic for interdependent regula-
tors fully transparent and time efficient. The following controllable devices can
be specified:

 Control Valve (PRV-Pressure Reducing Valve, PSV-Pressure Sustaining

Valve, PBV-Pressure Breaker Valve, TCV-Throttle Control Valve, FCV-
Flow Control Valve).
 Pump.

The devices may be specified as settings or PID-controlled, with control func-

tion selection based on a global system analysis. Each regulator or pump
operates under the control logic encapsulated into a set of simple logical
rules and control functions. The system allows a schematized definition of
any form and size of decision tree, featuring logical operators AND, OR, NOT
and NOR, in any associate combination.

The process of selecting an appropriate control function under the current

operational situations relies on the evaluation of system state conditions
including measurable and derived hydraulic and WQ variables (e.g. water
level, flow, pollutant concentration, level difference), devices statuses (e.g.
valve position, pump settings) and the current control function.

The control functions range from the simplest constants for the operational
variables (e.g. constant water level) to dynamic controlled variables, set in a
continuous functional relation with any of the measurable variables in the sys-
tem (e.g. pump START/STOP levels as functions of water level at strategic
location in the system).

3.3.1 Architecture of RTC Systems

An RTC system includes monitors/sensors, which generate measurement

values characterising states of the system. To be useful for RTC, the meas-
urements must be available with the relatively insignificant time leg (delay).
The sensors must be accurate and reliable.

The active control is performed by regulators - controllable movable devices

(valves) and pumps. Regulators may take various forms and sizes, and the
regulation may be continuous within the functional range, step-wise, or dis-
continuous (e.g. ON/OFF, OPEN/CLOSED). The regulators may be powered
mechanically, hydraulically or pneumatically.

145
 Water Distribution Tools (WD-Tools)

Controllers on the basis of a pre-programmed operational strategy determine

the regulator movements (the control actions). The operational strategy may
consist of two parts: the control function(s) and, if more control functions are
specified, the control logic (rules), responsible for the selection of an appro-
priate control function. A control function establishes a relation between a
control variable and a controlled variable. A controlled variable can be a regu-
lator setting (e.g. valve opening, pump START or STOP level) or some of the
flow variables (e.g. water level, flow).

In the latter case, the control decisions are derived by evaluating (comparing)
the current value of the controlled flow variable and the pre-defined set-point
value. The control algorithm is based on the numerical solution of the "contin-
uous control problem" equation and is usually termed as PID (Proportional-
Integral-Differential) control. The actuation signal for the regulator is gener-
ated by a PID controller, which usually appears as part of the operational
strategy programmed in a Programmable Logical Controller (PLC).

Selection of a controlled variable is, however, subject to limits set by the vari-
able's "controllability". Therefore, a controlled variable is usually selected
among the flow variables (flow, water level), preferably in the vicinity of the
regulator. As a controlled variable becomes more distant from the regulator, it
becomes more difficult to control, due to time lags, diffusion and uncontrolla-
ble interference. Control of relatively distant controlled variables is difficult
and often cannot give satisfactory results.

When a regulator setting is used as a controlled variable, the control algo-

rithm is reduced to an explicit functional relation between the control variable
and the regulator setting, which controls the system response indirectly. This
is much simpler than PID control, but in turn, the control results are in many
cases inherently inexact and only a rough flow control can be achieved. This
type of control is most suitable for regulators of the ON/OFF (or
OPEN/CLOSE) type, while the application to continuously controllable regu-
lators should be carefully considered.

If the operational strategy is based on conditions local to the regulated device

(for example the ON/OFF-control of a pump based on the water level in a
groundwater well) it is called local control. A PLC receives signals (measure-
ments) from local sensors and sends the control decisions (actuation signals)
to the regulators.

If the operational logic is based on global conditions, it is then called global

control. In such a situation, a global controller is required. A global controller
is a computer program that makes the overall system state analysis in real
time and provides additional input to the local controllers, which overrides or
supplements the local logic with e.g. actuator signals, or by modified set-point
values.

Additional component needed is then a data transmission system (UHF radio,

leased or dialled telephone lines, GSM, etc.) to transfer data between sen-
sors, controllers and the global controller. In connection with the global con-

146 MIKE URBAN - © DHI

PID Control

troller function, an RTC system is usually equipped with the data

management and storage facilities (databases) and the user interface. This is
usually termed as SCADA (Supervisory Control And Data Acquisition) sys-
tem.

Both local and global RTC based on sensor measurements is termed as

reactive RTC.

The global control can be extended also to include forecast data in addition to
real-time data, which is then called predictive RTC. The most comprehensive
way to obtain forecast data is to include a model in the control system. Pre-
dictive control brings additional benefits in relatively inert systems, i.e. where
the response time of an operational variable is long compared to the change
of relevant disturbance (external input or control action).

3.3.2 MIKE URBAN vs. Real Life

MIKE URBAN RTC simulates reactive local and global RTC systems in water
supply and water distribution networks. The software implementation is inher-
ently a conceptualisation of real life, of which the user must be fully aware.
Some conceptualisations applied in MIKE URBAN RTC are listed below.

 The program does not distinguish explicitly between local and global
RTC. Per default, all elements of a modelled RTC system are assumed
available for global control.
 Sensors are specified as operational devices with definition of sensor
type and position in the MIKE URBAN network. Sensors with multiple
functionality must be specified individually.
 When devices (valves and pumps) are specified as controllable in the
MIKE URBAN interface, a number of additional physical parameters
about the behaviour of the structure is required to describe e.g. the
allowed change rates for the state of the structure.
 The actual controllers are not specified explicitly as physical devices, but
their function (i.e. operational logic as a combination of operational con-
ditions and control functions) is associated with the respective devices.
 MIKE URBAN RTC uses sampling and actuation (control loop) frequency
identical to the simulation time step.
 Sensor readings are simulated as perfectly accurate and with 100%
availability.
 Low-level logic of the pump START/STOP operation is built into the pro-
gram and is controlled by the START and STOP levels.

The PID control algorithm is built into the program and is controlled by the
PID constants and by factors for weighting the terms of the numerical solution
of the control equation.

147
 Water Distribution Tools (WD-Tools)

Figure 3.10 through Figure 3.12 show examples of how the actual variable
(flow or water level) can fluctuate around the set-point as consequence of
various choices of the PID constants. Each figure has three different graphs
depending on whether the constant is too high, too low or adequate.

Figure 3.10 Fluctuations around the set point depending on the size of the propor-
tionality factor, K

Figure 3.11 Fluctuations around the set point depending on the size of the deriva-
tion time, TD

148 MIKE URBAN - © DHI

PID Control

Figure 3.12 Fluctuations around the set point depending on the size of the integra-
tion time, Ti

3.3.3 The PID Algorithm

At each simulation time step, the set-point is evaluated against the actual
value of the control variable (flow or water level, depending on the set-point
type).

The actual value of the control variable is determined from the following
equation:

 1 i
T
de 

u = K e +  e dt + T d 
 Ti 0 dt 
 (3.1)

where:

 u=the output signal i.e. value of the control variable,

 e =the error to be minimised,
 K =the proportionality factor,
 Td=the derivation time,
 Ti=the integration time.

149
 Water Distribution Tools (WD-Tools)

For numerical solution, a discrete form of this equation is required. By use of

the Laplace transformation one obtains:

 1 
U(s) = K  1 + + T d s  E(s)
 Ti s  (3.2)

where:

 U(s)=the Laplace transform of u,

 E(s)=the Laplace transform of e.
 By use of the backward Euler transformation:

z -1
s=
Ts z (3.3)

where Ts is the sampling period, one obtains the following discrete rep-
resentation of the PID equation:

 

u(n) = K  1 + T s + T d  y ref (n) - y(n) - 
 Ti Ts 

 2 
 
- K  1 + T d  y ref (n - 1) - y(n - 1) +
 Ts 

 
+ K y ref (n - 2) - y(n - 2) T d +
Ts

u(n - 1) (3.4)

where:

 u=value of the control variable

 K=the factor of proportionality
 Ti=the integration time
 Td=the derivation time
 Ts=sampling period, i.e. simulation time step
 yref=the set point for the regulated variable

150 MIKE URBAN - © DHI

PID Control

 a1..a3=factors for weighing each time step (index n, n-1 and n-2)
 y=actual value of the regulated variable

Indexes n, n-1 and n-2 denote the current, the previous and the second previ-
ous time step, respectively.

The weight factors a1... a3 have been added to the PID algorithm in order to
give the user more ways of stabilising the algorithm in case of instability prob-
lems.

Figure 3.13 The PID Control Editor

The following is the list of parameters that are used to specify the PID based
control.

Control Settings
Controlled element settings define the element, such as a pump or a valve
which is controlled by the PID algorithm.

 Controlled element type: type of the controlled element, such as pump

speed and control valve settings.
 Controlled element variable
 Controlled element ID: ID of the controlled element.
 Description: description of the controlled element (Optional).
 Minimum value: the minimum value constraint.
 Maximum value: the maximum value constraint.
 Maximum increase rate: the maximum rate at which the variable can
increase. In units of seconds-1.

151
 Water Distribution Tools (WD-Tools)

 Maximum decrease rate: the maximum rate at which the variable can
decrease. In units of seconds-1.

Set Point Settings

Set point settings define the PID algorithm constants.

 Set point type: type of the set point node; such as junction, tank.
 Set point variable
 Set point ID: ID of the set point node.
 Set point value: the constant value.
 Set point value curve: pattern ID defining how the set point value
changes in time.

PID Settings
 Maximum value: the proportionality constant of the PID algorithm (KP
proportionality constant)
 Max. increase rate: the integration time used in the PID algorithm (TI
integration time)
 Max. decrease rate: the derivation time used in the PID algorithm (TD
derivation time)

The speed of the valve or pump control depends on the PID control settings.
The PID controls and the Control Element Settings must reflect the physical
limitations or restrictions of the real-life controlled element and the PID algo-
rithm.

3.4 Demand Adjusted Anlaysis

Demand Adjusted Analysis is an alternative computational method based on
pressure driven analysis comparing to the traditional demand driven analysis.
Node demands are automatically adjusted based on the available pressure.
This approach can be used to model intermittent water supply, low pressure
situations, and it is also suitable for modelling system shutdown and mainte-
nance.

The Wagner equation is used to adjust the node demand based on the avail-
able pressure.

152 MIKE URBAN - © DHI

Demand Adjusted Anlaysis

where:

 Qnew = adjusted node demand

 Pactual = actual pressure
 Prequired = required pressure (such as 35 m), node demand is equal to
the original demand if the pressure is above the required pressure
 Pminimum = minimum pressure (such as 5 m), node demand is 0 if the
pressure drops below the minimum pressure
 n = coefficient with recommended values between 1.7 and 2.0

Note that nodes with negative demand i.e. inflow nodes are excluded from
this demand adjustment. In order to use Demand Adjusted Analysis, select
EPANET | WD-Tools | Demand Adjusted Analysis and define the simulation
settings.

Figure 3.14 The simulation settings for the Demand Adjusted Analysis

A list of the Demand Adjusted Analysis entries with a short description given
for each entry:

 Minimum pressure: minimum pressure (such as 5m), node demand is 0 if

the pressure drops below the minimum pressure
 Maximum pressure: required pressure (such as 35m), node demand is
equal to the original demand if the pressure is above the required pres-
sure
 Selection set file: file with selected nodes (selection file from MIKE
URBAN)
 Coefficient n: coefficient "n" for the exponent in Wagner equation (Exp =
1/n)

153
 Water Distribution Tools (WD-Tools)

In order to run the Demand Adjusted Analysis, select "Demand Adjusted

Analysis" simulation option from Run Simulation dialog.

Figure 3.15 Run the Simulation with the Demand Adjusted Analysis choice

The simulation will run for any steady state or extended period model and the
simulation results will be presented in the same way as standard results. In
case that the node demands were adjusted, simulation report.INF file will pro-
vide more details. The INF-file is found in the project directory.

Figure 3.16 Content of the *.INF file.

Where:

 Time: time level (hrs)

 NoOfAllNodes: number of junction nodes
 NoOfMin: number of nodes where minimum pressure was exceeded

154 MIKE URBAN - © DHI

Demand Adjusted Anlaysis

 NoOfMax: number of nodes where maximum pressure was exceeded

 NoOfAdj: number of node where the original node demand was adjusted
 SumDemOrig: original network demand (total for the whole network)
 SumDemNew: new network demand (total for the whole network)

Figure Figure 3.17 illustrates how the original network demand (as defined in
the Multiple Demand Editor) was adjusted based on the actual pressures; this
graph was created based on the .INF file.

Figure 3.17 Comparison of the original network demand versus adjusted demands
during 24 hours

The Demand Adjusted Analysis can be used to model systems with intermit-
tent water supply or systems with low pressure conditions. We can also use
this concept in modelling system breakdowns, such as closure of parts of the
water distribution systems.

3.4.1 Low Pressure Systems

The example below illustrates a water distribution network with 3 water

sources that cannot deliver the required water demands. As a result of that,
low pressures (even "negative" in this case) are reported.

155
 Water Distribution Tools (WD-Tools)

Figure 3.18 Low pressure scenario, total network demand ~ 300 l/s (given demand)

In order to see what is the network capacity, we can run the Demand
Adjusted Analysis with the settings of minimum pressure = 5 m, for example.
We can see that the network capacity is 83 l/s in order to provide minimum
pressure of 5 m, see Figure 3.19.

Figure 3.19 Pressure dependent demands, total network demand ~ 83 l/s (given
demand)

156 MIKE URBAN - © DHI

Variable Speed Drive Pump

3.4.2 Intermittent Water Supply Systems or Modelling System Shutdown and

Maintenance

It is also possible to use the Demand Adjusted Analysis to analyze intermit-

tent water supply systems or system shutdown and maintenance. Standard
demand driven analysis will report negative pressures once a portion of the
water distribution system has not water supply while the Demand Adjusted
Analysis will automatically adjust the demands and turn them eventually off
during non-supply periods.

3.5 Variable Speed Drive Pump

A variable pump can vary its speed setting and/or change its status to open or
closed during a simulation. VSD pump extension allows modelling of a pump
where the rotational speed of a pump is automatically adjusted by the pro-
gram in order to maintain the control node pressure at the required level. VSD
pump using can also be modelled using the PID control module.

See the section titled Pump Editor on -18 for additional information.

3.6 Extended Rule-Based Controls

Rule-Based Controls allow link status and settings to be based on a combina-
tion of conditions that might exist in the network over an extended period sim-
ulation. In order to allow for modelling of chains of pumps or valves chains in
efficient way, the EPANET Rule-based control syntax was extended by add-
ing LIKE keyword, for example.

See the section Rule Based Control Editor.

3.7 Pressure Zone Mapping

MIKE URBAN can graphically display different "pressure zones" based on
the network topology and geometry, closed pipes, closed valves and pumps.
Pressure zone connectivity breakdown rules can be defined. This tool helps
to visualise how different network parts are hydraulically interconnected and
where the HGL line breaks. In order to generate pressure zones, select
EPANET | WD Tools | Pressure Zone Mapping… from the main program
menu. The Pressure Zone Mapping wizard will guide you through the set-
tings.

The Pressure Zone Mapping wizard allows you to define the separators,
merge small groups of pipes into the same group, and it can automatically
generate the colour legend to display the pressure zones.

 Separators: separators are links, which will be used to separate one

zone from another.

157
 Water Distribution Tools (WD-Tools)

 Merge zones smaller than: in case that there are many small zones (a
typical example would be small pipes located in pumping stations and
storage tanks), they will be all merged into the same zone for graphical
display instead of creating a separate zone for each of them.

This tool help you to understand the hydraulic behaviour of the network prior
to running the hydraulic simulation and it also helps you to detect possible
errors in the network connectivity.

Figure 3.20 Pressure Zone mapping wizard

Figure 3.21 Pressure zones are automatically mapped based on the connectivity
breakdown rules.

158 MIKE URBAN - © DHI

Advisor

3.8 Advisor
Advisor is a set of tools designed to support the analysis of the water supply
and water distribution systems. It helps the user to check the results of the
hydraulic simulations, understand system deficiencies, improve design and
operation. Advisor is available under EPANET|WD-Tools|Advisor and con-
tains the following tools:

 Pipe criticality
 Sustainability analysis
 Cost analysis

3.8.1 Pipe Criticality

Pipe criticality modeling is required to predict the water distribution system

response to pipe breaks situations, planned reconstructions, and other sce-
narios of limited water supply. Pipe criticality allows also the develop a pipe
ranking based on the importance for the water supply and such importance
can be then taken into account for the planning of pipe rehabilitation and
reconstructions.

The pipe criticality is determined based on evaluation of several (four) perfor-

mance indicators including:

1. Water demand criteria (PI-1)

2. Service pressure criteria (PI-2)
3. Pipe flow criteria (PI-3)
4. Pipe length criteria (PI-4)

The combined pipe criticality is computed as an average of all above perfor-

mance indicators, i.e. C (pipe i) = Average (PI-1+PI-2+PI-3+PI-4) (pipe i)

Water demand criteria (PI-1) is computed as water (in flow units) that cannot
be delivered through the pipe. The value of 1 corresponds to the total flow.

Service pressure criteria (PI-2) is computed as number of nodes, where the

service pressure is below the required level e.g. 15 m or 20 psi, for example.
The value of 1 corresponds to the total number of nodes.

Pipe flow criteria (PI-3) is computed as the value of average hour demand
Q(hour), the value of 1 corresponds to the total value of average hour
demand.

Pipe length criteria (PI-4) is computed as a total length of pipes affected by

the respective pipe. The value of 1 corresponds to the total pipe length. Simi-
larly, it is possible to use the number of population disconnected from the
water supply or number of disconnected residences or houses.

159
 Water Distribution Tools (WD-Tools)

It is recommended to use the maximum peak demand for the above simula-
tions.

In order to use Pipe Criticality, select WD-Tools from the main program menu
and then select Advisor | Pipe Criticality. The Pipe Criticality form where you
define the simulation settings will then display.

Figure 3.22 Pipe criticality setting (Global settings)

The interface allows you to define the following parameters:

 Time level: in case of Extended Period Simulation, this is the time level at
which the pipe criticality analysis will be performed
 Minimum service pressure: minimum acceptable service pressure within
the network

The pipe reliability settings can be defined on the level of the whole model
area or per each specific zone. In case of Global settings:

 Total demand: network demand

 Total demand (Qhr): maximum hour network demand

160 MIKE URBAN - © DHI

Advisor

Figure 3.23 Pipe criticality setting (Local settings)

In case of the local settings, the network demands are defined per each zone,
please note that the zone demands are defined in the Pressure Zone Editor.

 Q: network demand
 Qhr: maximum hour network demand
 Selected: use this switch to active or deactivate zones; only selected
zones will be used for the pipe criticality analysis.

In order to run the pipe criticality analysis or to load the results of the previous
pipe criticality results, use the following controls

 Run: run pipe criticality analysis

 Results: load pipe criticality results and store them into the database

161
 Water Distribution Tools (WD-Tools)

Figure 3.24 Pipe criticiality results, table view

Figure 3.25 PIpe criticality results, Map view (pipes are colour coded by the pipe
criticality C)

3.8.2 Sustainability Analysis

Sustainability analysis helps you to understand the simulation results and

analyze them for possible problems, anomalies, critical areas, and similar.
Various predefined thematic maps are available including:

 Unit headloss to determine pipe size problems

 Reverse flows to identify possible water quality issues
 Service pressures
 Pressure, velocity, and other anomalies

Detailed HTML reports can be generated to understand operation of pumps,

storage tanks, and other facilities.

In order to use Sustainability Analysis, select WD-Tools from the main pro-
gram menu and then select Advisor | Sustainability Analysis. Sustainability
Analysis dialog where you define the simulation settings will display.

162 MIKE URBAN - © DHI

Advisor

Please note, that the simulation results needs to be loaded in order to start
the sustainability analysis.

Figure 3.26 Sustainability Analysis dialog

You can define the following settings:

 Results file: name of the results file

 Flow threshold: minimum flow criteria used for reverse flow calculation

 Settings: unit selection

 Map layers: select what layers will be added to the Map

– Service pressure
– Unit headloss
– Pipe flow

 Report: select what sections will be used for reporting

– Storage tanks
– Pumps
– Unit headloss
– Service pressure
– Pipe flow
– Report each time level: please note that selecting each time level
can result in excessive processing time

 Create: perform the sustainability analysis

 Report: create a report

The program will provide a detailed analysis of the simulation results and it
will create the following layers and reports:

 Service pressures
– Pressure anomalies e.g. pressure too low or to high

163
 Water Distribution Tools (WD-Tools)

– Pressure range: pressure distribution

– Pressure fluctuation: difference between the minimum and maxi-
mum pressure at every node during the simulation.
– Unit headloss: head loss per 1000
– Pipe flows
– Reverse flow: The layer will show how many times the flow direction
has changed in every pipe. Note that a threshold value needs to be
specified e.g. "0.1"meaning that if the absolute flow is smaller than
that the pipe is not considered for reporting i.e. if the flow must be
smaller than -0.1 and greater than 0.1 to be considered as a pipe
with flow.
– Flow velocity: velocity distribution
– Flow velocity fluctuation

In addition to the Map layers, the HTML report creates a detailed report for
the whole network as well as for the particular zone.

 Storage tanks: reports tanks that are either drained or overflows during
the simulation, reports if the tanks are balanced within the simulation.
Balanced tank is a tank where the water level at the beginning of the sim-
ulation is the same as at the end of the simulation.
 Pumps: reports pump that are operated near their maximum capacity
 Unit Headloss: reports pipes with too high unit head loss
 Service Pressures: reports excessive pressures
 Pipe flows: reports reverse flows
 Flow velocity: reports excessive flow velocity

Figure 3.27 Sustainability analysis, Map view of reverse flows

164 MIKE URBAN - © DHI

Advisor

Figure 3.28 Sustainability analysis, report

3.8.3 Cost Analysis

Cost analysis allows you to review the energy consumption results on more
details, create tabular outputs, and great graphs of pump utilization, average
power consumption, and costs.

In order to use the Cost Analysis, load the simulation results and select
EPANET | WD Tools | Advisor | Cost Analysis…

Pump Energy Table

The pump energy table is displayed on the first tab of Cost Analysis. It reports
pump utilization, average efficiency, average power consumption, peak
power consumption, and cost per day.

Figure 3.29 Pump energy table

165
 Water Distribution Tools (WD-Tools)

Pump Energy Graphs

Pump energy including pump utilization, power consumption, and costs can
be displayed for every pump in the model. Such graphs help to analyze pump
energy and evaluate or compare pumps.

Figure 3.30 Pump energy graphs

The following fields are calculated and reported:

 Utilization: Percent utilization i.e. percent of the time that the pump was
operating (%)
 Efficiency: Average efficiency of the pump (%)
 kW-hr/m3 or kW-hr/gallon: Average power consumption (kW per hour)
per million gallons (or cubic meters) pumped
 Average kW: Average rate of energy usage of the pump (kW)
 Peak kW: Peak rate of energy usage of the pump operation (kW)
 Cost/day: total cost of the pump operation per day (monetary units)

Note that the total cost and demand charge (if any) are listed at the bottom of
the table.

To define energy price, price pattern, pump efficiency use EPANET | Energy
Editor and EPANET | Pump Editor where you can define the following fields:

 PRICE = average cost per kW-hour,

 PATTERN = ID label of time pattern describing how energy price varies
with time,
 EFFIC = either a single percent efficiency for global setting or the ID
label of an efficiency curve for a specific pump,

166 MIKE URBAN - © DHI

Advisor

 DEMAND CHARGE = added cost per maximum kW usage during the

simulation period.

3.8.4 Alarms and Violations

Alarms and Violations allows you to define critical values for various results
items anywhere within the model network such as maximum velocity, mini-
mum or maximum pressure, low or high level, high water age and let the
hydraulic model to evaluate them based on the actual simulation results.

In order to use the Alarms and Violations, select

EPANET | WD Tools | Alarms and Violations …

Alarms and Violations dialog

The Alarms and Violations dialog is used to define critical values and validate
simulation results.

Figure 3.31 Alarms and Violations table

The dialog table allows you to define or display the following parameters:

 Element Type: select from the following types:

– Junction
– Tank
– Pipe
– Pump
– Valve

 Element MUID: enter the MUID of the junction node, pipe, pump, valve,
or a tank

 Description: enter user defined description (optional)

 Variable: select the result item from the following types:

– Node: Demand
– Node: Head
– Node: Pressure

167
 Water Distribution Tools (WD-Tools)

– Link: Flow
– Link: Absolute Flow
– Node: Water Quality
– Link: Velocity
– Link: Unit Headloss
– Link: Water Quality
– Link: Status Code
– Link: Setting
– Link: Reaction Rate
– Link: Friction Factor
– Link: Start End Head

 Criteria: select the value criteria from the following types:

– <
– <=
– =
– >
– >=
– <>

 Alarm Value: enter the critical (alarm) value

 Actual Value: this is the highest or lowest of the actual values resulting in
the alarm

 Alarm at Time: this is the time of the simulation corresponding to the

“Actual value”

 Status: this is the status of the results validation, “OK” or “Failed”

 Status Message: this is the message from the results validation

 Comment: user defined comment

Note, that the alarm will be triggered if the criteria is fulfilled i.e. if the criteria
is defined for a “tank level < 2” then the alarm will be reported once the com-
puted level is below “2”.

Similarly, if you want to alarm for a pump operation during which the pump
flow exceeds 300 l/s, enter the following criteria “pump flow > 300”.

The dialog interface allows you to use the following commands:

 Insert: add a new record to the table

 Delete: delete the highlighted record from the table
 Show: highlight nodes and links used in alarms in the Map
 Series: displays time series for the selected alarms (rows)
 Clipboard: copy the content of the table into a clipboard

168 MIKE URBAN - © DHI

Advisor

 Validate: perform the results validation for the selected results file
 Clear: clear the results of the last validation
 On Top: forces the dialog to remain on top of other windows
 Close: close the dialog window

169
 Water Distribution Tools (WD-Tools)

170 MIKE URBAN - © DHI

Water Hammer Calculation

4 MIKE URBAN Water Hammer

Water Hammer (a part of the WD-Tools module) simulates transient
(unsteady) flow in any fully pressurized system carrying liquids. MIKE
URBAN Water Hammer provides a cost effective tool for engineers seeking
fast answers to questions about rapid operation of piping systems. Water
hammer is based on the high-order implicit scheme solving the continuity and
momentum equation using the finite difference method. The initial conditions
are modeled using MIKE URBAN Steady State Pipe Flow Module.

Water Hammer allows you to model:

 Pump start-up and pump trip-off.

 Water hammer protection of pumping stations.
 Power failure events.

4.1 Water Hammer Calculation

MIKE URBAN Water Hammer computes hydraulic transients in pipe net-
works. The computations are based on the continuity equation:

Q g A H
 0
x a2 t
(4.1)

and the equation of motion:

Q H f
 gA  Q Q 0
t  x 2 DA (4.2)

in which Q is the discharge, H - the piezometric head above arbitrary datum, f

- the Darcy-Weisbach friction factor, D - the internal pipe diameter, A - the
cross-sectional area of the pipe, g - gravitational acceleration, a - wave
speed, x - distance along the pipe axis and t - time.

In the governing equations the acceleration terms which are very small com-
pared to the other terms have been disregarded.

The general expression for the wave speed (only important for water hammer
computations) presented by Halliwell (1963) has been used

K
a
ρ 1  K/E ψ  (4.3)

171
 MIKE URBAN Water Hammer

in which E is the Young's modulus of elasticity of the conduit walls, K - the

bulk modulus of the fluid, r - the density of the fluid and y - a non dimensional
parameter. For more details see Section 5.

An implicit finite difference scheme described by Verwey and Yu (1993) has

been implemented for water distribution, slow transient and water hammer
simulations. The scheme uses only two adjacent grid points in space on a
non-staggered grid and is defined on three time levels. The elimination of the
most important phase error allows the simulation of both water hammer and
slow transients.

4.2 Theoretical Background

The following section describes the MIKE URBAN water hammer numerical
engine.

4.2.1 Description of Water Hammer Model

The water hammer computation is based on the Continuity equation

Q g A H
  0
x a2  t (4.4)

and the Momentum equation

Q H f
 gA  Q Q  0
t x 2 DA (4.5)

where Q is the discharge, H - the piezometric head above arbitrary datum, f -

the Darcy-Weisbach friction factor, D - internal pipe diameter, A - the area of
pipe, g - the gravitational acceleration, a - the wave speed, x - the distance
along pipe axis and t - the time.

In the governing equations the acceleration terms which are very small com-
pared to the other terms have been disregarded.

Wave Speed
For pure liquids Halliwell (1963) presented the general expression for the
wave speed

K
a 
 1   K / E    (4.6)

172 MIKE URBAN - © DHI

Theoretical Background

in which E is the Young's modulus of elasticity of the conduit walls, K is the

bulk modulus of the fluid,  is the density of the fluid and  is a nondimen-
sional parameter.

Rigid Conduit

 = 0 (4.7)

Thick-Walled Elastic Conduit (D/e<=10)

 anchoring at both ends = full restraint

 Ro 2  Ri 2 2  Ri 
2

  2 1     2  
 Ro  Ri Ro 2  Ri 2 
2

(4.8)

in which  is the Poison’s ratio, Ro is an external diameter, Ri is an

internal diameter.

 upstream anchoring = upper restraint

 Ro 2  15
. Ri 2   Ro 2  3 Ri 2  
  2 2  
 Ro  Ri Ro 2  Ri 2
2
 (4.9)

 frequent expansion joints = expansion joints

 Ro 2  Ri 2 
  2 2  
 Ro  Ri
2
 (4.10)

Thin-Walled Elastic Conduit (D/e>10)

 anchoring at both ends = full restraint

D
 
e
1   2 
(4.11)

in which D is the conduit diameter and e is the wall thickness

173
 MIKE URBAN Water Hammer

 upstream anchoring = upper restraint

D
 
e
 1  0.5  
(4.12)

 frequent expansion joints = expansion joints

D
 
e (4.13)

Tunnels Through Solid Rock, Parmakian 1963

 Unlined tunnel

 = 1 E  G (4.14)

where G is the modulus of rigidity of the rock.

 Steel - lined tunnel

DE
 
GD  Ee (4.15)

in which e is the thickness of the steel liner and E is the modulus of elasticity
of steel.

Reinforced Concrete Pipe

This pipe can be replaced by an equivalent steel pipe having equivalent
thickness.

As
ee  E r ec 
Ls (4.16)

in which ec is the thickness of the concrete pipe, As - the cross-sectional area

of steel bars, Ls - the spacing of steel bars, Er - the ratio of the modulus of
elasticity of concrete to steel (0.06 - 0.1), but 0.05 for cracks.

174 MIKE URBAN - © DHI

Theoretical Background

Diagrams
The following diagrams can be used in order to estimate the wave speed.

Figure 4.1 Fluid water

Values of Young's Modulus of Elasticity and Poisson's Ratio for a range of

common materials are available in the following table.

Table 4.1 Values of Young's Modulus of Elasticity and Poisson's Ratio for a range
of common materials
Material Young’s Modulus Poisson’s Ration (-)
(10E9 N/m2)

Aluminum 70 0.3

Cast Iron 80-110 0.25

Concrete 20-30 0.1-0.3

Copper 107-130 0.34

Glass 68 0.24

GRP 50 0.35

Polyethylene 3.1 -

PTFE Plastic 0.35 -

PVC Plastic 2.4-2.8 -

Reinforced Concrete 30-60 0.15

Rubber 0.7-7.0 0.46-0.49

Steel 200-24 0.3

Titanium 103.4 0.34

175
 MIKE URBAN Water Hammer

Typical values of Bulk Modulus:

 K = 2.05 x 10E9 N/m2 for water

 K = 1.62 x 10E9 N/m2 for oil.

4.3 Numerical Scheme and Algorithm

The numerical solution is based on the approach suggested by Verwey and
Yu (1993). An implicit, space-compact finite difference scheme has been
implemented for simulation in pipe networks including a variety of control ele-
ments. The same numerical scheme can be used for simulation of both
hydraulic transients and water distribution problems. The inertia terms in the
governing equations can be manipulated to produce relatively fast conver-
gence for steady state problems.

The implicit finite difference formulation is based on a non-staggered grid in

time and space, where at each grid point the independent variables Q and H
are to be computed. The friction term in the governing equations has been
expressed as

f
2 DA
Q Q
1 f
2 2 DA
 Q nj 1 Q nj 11  Q nj Q nj 1  (4.17)

The coefficients for the water hammer model have been derived and have the
following form:

176 MIKE URBAN - © DHI

Numerical Scheme and Algorithm

4.3.1 Coefficients for the Numerical Scheme

gA
  2
a (4.18)


fric 
2AD (4.19)

a t
Cr 
x (4.20)

    3a t
2
2 2 2 x
 c  6  6  1  Cr 6  6 1 2 2
(4.21)

    3a t
2
2 2 2 gA x
 m  6  6  1  Cr 6  6 1 2 2
(4.22)

A1    1    x (4.23)

B1   t   c (4.24)

C1    x (4.25)

D1   t   c (4.26)

E1  (1   ) t (Q nj 1  Q nj 11  Q nj  Q nj 1 )  t (Q nj 1  Q nj ) 

 c (Q nj 1  2Q nj  Q nj 11  2Q nj 1 )   c (1   ) xH nj 11   c xH nj 1
(4.27)

A2   g a  t   m (4.28)

177
 MIKE URBAN Water Hammer

B 2  1    x  fric  t  x Q nj1
(4.29)

C 2  g a  t   m (4.30)

D 2   x  frict x Q nj
(4.31)


E2  1  xQ nj11  xQ nj1  gA1   t H nj1  H nj  H nj11  H nj1  
  
 gA t H nj1  H nj   m H nj1  2H nj  H nj11  2H nj1  (4.32)

4.3.2 Looped Network Solution Algorithm

The main algorithm generates a set of grid points using a finite difference
scheme, see Cunge, Holly, Verwey (1980). The grid is introduced in time and
space, where at every point the values of H and Q are defined as the
unknown variables. Between the two successive grid points in time and
space both the continuity and the momentum equation are applied. Together
with the necessary boundary data, a sufficient number of equations are
obtained to solve H and Q at every grid point.

The general form of the governing equations is

A1j H nj 11  B1j Q nj 11  C1j H nj 1  D1j Q nj 1  E 1j

(4.33)

A2 j H nj 11  B2 j Q nj 11  C 2 j H nj 1  D2 j Q nj 1  E 2 j
(4.34)

where coefficients A1,B1,C1,D1,E1 for the continuity equation and

A2,B2,C2,D2, E2 for the momentum equation are derived from the high-order
scheme.

The looped algorithm is based on the fact that a looped network contains ele-
ments known as nodes which represent the confluence of several flow paths,
some of which originate from other nodes, some from boundary points. A sys-
tem of simultaneous linear equations is developed where the piezometric
head changes at each node are the only unknowns. Solution of this system
by any matrix elimination technique yields the piezometric heads at each
node.

178 MIKE URBAN - © DHI

Numerical Scheme and Algorithm

Suppose that there are three links, 2-1,2-3 and 2-4 and that there are b grid
points along branch 2-3 and c grid points along a link 2-4, see Figure 4.2. For
any computational grid point, equations (4.35), (4.36) may be written as

Hi  L1i H1  M 1i H jj  N 1i
(4.35)

Qi  L2i H1  M 2i H jj  N 2i
(4.36)

where L,M,N are functions of coefficients A, B, C, D, E, found through a dou-

ble sweep elimination

Figure 4.2 Part of a looped pipe network

These equations express the partial dependence of the unknown variables Q

and H at any grid point in a branch on the value of H in the two adjacent
nodes.

At internal nodes a compatibility condition must be satisfied. The simplest

condition is node continuity and common piezometric head.

m
Q
k 1
n 1
1k 0
(4.37)

h11n1  h12n1 K  h1nk1 K  h1nm1 (4.38)

where n+1 indicates the (n+1)Dt time level in the solution, k is the index of the
links emanating from node 2, and m is the number of such links. These rela-
tions can be written for each from M nodes, and this leads to a system of M

179
 MIKE URBAN Water Hammer

linear equations having as unknowns the piezometric head changes H at

each node.

 S  h  TL  (4.39)

where [S] is a coefficient's matrix, M x M elements, {h} is a vector of

unknowns, M elements; {T(L)} is a vector of the free terms.

This system of linear equations may be solved by any matrix inversion tech-
niques. Once the increments of piezometric head H are known at the nodes,
it is possible to recompute Q(i) and H(i) values for all intermediate grid points
through equations (4.37) and (4.38).

The looped algorithm may be described by the following steps:

 The coefficients of the high-order scheme discretize the governing equa-

tions between two successive grid points on a branch.
 The local elimination method is used to express Q and H grid point val-
ues on each branch in terms of H at the branch ends (nodes).
 One equation for each node leads to the system of linear equations that
is solved by the matrix elimination method.
 Substitutions inside the branches yield the Q(i) and H(i) values for all
intermediate grid points from the known values of H at the branch ends.

4.3.3 Hydraulic Structures

The implementation of a hydraulic structure in the domain of the solution may

be solved by replacement of the governing equations by another set of equa-
tions that characterise the particular hydraulic structure. Every time the main
algorithm comes to the location of such a structure, it must switch between
the governing equations. Any hydraulic structure can be implemented into
such a numerical scheme in the following way. The hydraulic structure is
placed between the two successive grid points, and we can assume that.

Another way of implementing a hydraulic structure is to handle it in the similar

way to a node. The hydraulic structures are not located between the two suc-
cessive grid points but in the node. Instead of modifying coefficients A,B,C,D,
and E, we increase the number of linear equations.

The main algorithm is designed in such a way, that, after a process of lineari-
sation and discretization of the governing equations, it solves them on a pre-
scribed set of grid points using an appropriate numerical scheme. If a
hydraulic structure is present in the domain of the solution, the algorithm must
replace the governing equations by other equations defining a hydraulic
structure in order to provide the numerical solution. Various hydraulic struc-
tures can be coupled together, e.g., the closing of one valve can determine

180 MIKE URBAN - © DHI

Water Hammer Calculations

the operating of another valve. In cases where this link exists between
hydraulic structures, communication must be maintained and controlled by
this main algorithm. This message has to be attached to the object in such a
way that it represents the reality. Object-oriented design has been applied to
create a safer interface to the numerical algorithm, since the low level opera-
tions that remain the same are hidden inside objects.

4.4 Water Hammer Calculations

Water hammer simulates transient (unsteady) flow in any fully pressurized
system carrying liquids. MIKE URBAN Water Distribution Water Hammer pro-
vides a cost effective tool for engineers seeking fast answers to questions
about rapid operation of piping systems. Water hammer is based on the high-
order implicit scheme solving the continuity and momentum equation using
the finite difference method. The initial conditions are modeled using MIKE
URBAN Water Distribution Steady State Pipe Flow Module.

Water Hammer allows you to model:

 Pump start-up and pump trip-off.

 Water hammer protection of pumping stations.
 Power failure events.

Water Hammer Data Preparation

allows you to create all the input files interactively and save them for compu-
tation. The Data Preparation provides interactive data input, editing and error
checking. Graphical facilities enable the display of data on a plan plot and use
the Query-By Examples (QBE) facilities of the database.

The present version will handle any number of pipes, nodes, and loops in
complex networks with various components

Water Hammer Result Presentation

enables you to view results generated from the calculation as thematic maps,
graphs or as text in ASCII format. The results can be displayed using different
plots, namely time series plots of the variables, time series plots of the varia-
bles for the current time in the longitudinal profile, and colour-coded plan
plots. The last two choices can be used for a time animation. The use of col-
our-coded plan plots allows you to define what numerical ranges of variables
between grid points correspond to a particular colour. Zoom facilities enable
to magnify interesting portions of drawings. The presentation includes the
ability to save screen displays into Windows AVI animation files or send
screen displays to printers. It is also possible to generate hydraulic grade
lines and pressures.

181
 MIKE URBAN Water Hammer

4.4.1 Running Water Hammer Simulations

In order to be able to start water hammer simulations you have to prepare the
steady state model and obtain satisfactory results. In the next step, you need
to specify Water Hammer Analysis type in the Project Options dialog, define
transient boundary conditions and computational parameters. In addition to
the specific pipe data, you need to know the wave speed in each pipe.

Initial Conditions
Initial conditions are computed with the use of the Steady State Model. The
results of the initial state are saved in the file as H, Q values at the beginning
and end of the pipes respectively and in the vicinity of hydraulic structures
such as valves, pumps, etc. There is a direct connection between the result
file from initial conditions and the water hammer execution, in spite of the fact
that the two models use different computational grids.

Boundary Conditions
There are in principle two types of boundary conditions, namely the piezo-
metric head, H, above a specified datum, e.g., in tanks, and the discharge, Q,
e.g., water demand. Both H and Q are given under selected names as time
series in the Curve Editor and stored in the database. These boundary condi-
tions may be assigned to any node in the network. Boundary for each time
step is assigned from given time series specified by the user. If time step
used by water hammer computation is smaller than appropriate neighbouring
values in boundary conditions time series then linear interpolation is applied.
There are nodes of the following types: H - boundary, Q - boundary, compati-
bility and structure (hydraulic component) description. It should be pointed
out that time patterns, used in the Steady State Model, are ignored by the
Water Hammer simulations.

For the Initial State for Water Hammer Model, the water level and/or dis-
charges are constant in time. The boundary conditions using time series must
be specified for a sufficiently long time interval.

Computational Parameters
General parameters consist of fluid density, fluid bulk modulus, absolute tem-
perature, vapour pressure and gravitational acceleration. The most important
numerical parameter is a time step. Since a numerical solution must be stable
and as accurate as possible, you have to choose a proper value of Dt. The
stability condition is given by the Courant number

a t
Cr 
x (4.40)

182 MIKE URBAN - © DHI

Water Hammer Calculations

in which a is the wave speed and Dx is the distance between two successive
grid points. In principle, an implicit, space-compact scheme is unconditionally
stable, with exact solutions generated for the Courant numbers Cr = 0.5 and
Cr = 1.0, respectively. The scheme enables us to vary the Courant number
over pipes while maintaining its high accuracy. Accurate results are produced
in the range 0< Cr < 1.1. You should try to maintain the Courant numbers
below unity, but as close as possible to Cr = 1. If you select the menu item
Geometry and Branch, you can control the values of Courant numbers. The
question how to choose the time step is dictated by the nature of the hydrau-
lic transient itself and by the shortest pipes in the system. The time step can
vary from the order of 10-3 to 10+1 seconds. The time steps must be small
enough in order to describe very fast changes of variables. It is recom-
mended to start with the shortest pipe section and to calculate the time step,
considering Cr = 1. Pipe sections with high Courant numbers are numerically
treated in MIKE URBAN Water Distribution as rigid pipelines. This simplifica-
tion enables a user to deal with a very short pipe section which would not be
important within the water hammer simulation. Once the time step has been
prescribed, you have to input the simulation time. MIKE URBAN Water Distri-
bution calculates a number of time levels, which you need to prescribe in the
Check level item. In the Project parameters' window you can also change
throughout network whether you intend to use a friction factor and/or an
absolute roughness.

The last group of parameters is referred to as advanced parameters. For an

experienced user there is direct access to a weighting coefficient q which has
a default value of 0.5. For special cases you can use weak forward centring
of the scheme and hence activate the diffusive part of the truncation error,
see Verwey and Yu (1993).

4.4.2 Definition of Network Layout

An example of a topological representation of a network is shown in Fig.3.1.

The solution domain consists of branches connected one to another by
means of nodes. Grid points are generated along branches and they repre-
sent the place where we are looking for the solution of the governing equa-
tions. Different hydraulic structures can be included later at selected places in
the network.

Figure 4.3 Definition Network Layout

183
 MIKE URBAN Water Hammer

For model construction, we can define a range of model elements such as

nodes, branches, grid points and hydraulic structures.

Branches
can be used to represent pipes of constant properties. In the pipe network,
branches may include hydraulic elements, for example, valves, pumps.
Nodes represent the applicable boundary conditions at the end of branches.

Nodes
are elements that represent free branch ends, branch connections or a spe-
cific storage. At nodes with one simple pipe connected, boundary conditions
are usually defined by specifying the values of piezometric head or discharge
as a constant value or as a function of time. Flow continuity and a piezometric
level compatibility is assumed at nodes connecting several branches
together.

Generally, there are these three different types of nodal boundary conditions:

 H (pressure (m) is given).

 Q (discharge (l/s) is given).
 Compatibility (common H).

Other types of nodes can be given as:

Table 4.2 Node boundary conditions

Node Type Meaning Variable

H-Boundary Given HGL H=f(t)

Q-Boundary Given Demand Q=f(t)

(inflow/outflow)

Continuity Continuity None

Junction node without demand Continuity None

Junction node with demand Given Demand Q=const

Tank Calculated HGL H=f(t), H=const

Air-Chamber Calculated HGL H=f(t)

Vented Air-Chamber Calculated HGL H=f(t)

Air-Valve Calculated HGL H=f(t)

Emitter Calculated Demand Q=f(t)

Shaded VARIABLE types are set automatically by the program.

184 MIKE URBAN - © DHI

Water Hammer Calculations

Grid Points
are generated automatically by Water Hammer along the branches and they
represent the computational grid where the values of piezometric head and
discharge are solved and the input and/or output data are required. The sys-
tem requires a different computational grid for steady state and water ham-
mer computations.

Computational Grid and Hydraulics Structures

The hydraulic components are located either in nodes or on branches. An
example of grid-generation in a water distribution application with a valve
illustrates the procedure of implementation of the hydraulic components. For
water hammer applications the grid is defined as a function of the length of
the pipe elements, the wave speed of water hammer and the speed of system
operation.

Specific Pipe Data

Input of pipes is the same as in the case of steady state analysis. Then you
have to specify the wave speed. Wave speed (celerity of the pressure wave)
is the only one specific (and mandatory) parameter for the water hammer cal-
culations.

Junction Node Demands

Until specified as Water Hammer Boundary conditions, node demands are
kept constant through out the water hammer simulation period. Junction
demands i.e. multiple demands and their patterns - diurnal curves are use to
calculate the steady state i.e. initial conditions for water hammer and they are
kept on the same value for the water hammer analysis.

Control Rules
Simple Control Rules and Rule Based Controls are not allowed during the
water hammer analysis. Valve opening and pump scheduling is handled
directly by the specific valve and pump data.

Specific Pump Data

Input of pipes is the same as in the case of steady state analysis. Then you
have to specify rated rotational pump speed and its schedule - time series of
the rotational pump speed versus time.

Pumps may be located inside pipeline systems (booster pumps) or they may
be connected to a suction well. Pumps are frequently used for various pipe-
line systems, and may operate during hydraulic transients with constant
pump speed. Alternatively, the pump speed can decrease and/or increase
depending on pump shut-down and/or pump start-up. The greatest difficulties

185
 MIKE URBAN Water Hammer

come from hydraulic transient flows caused by turbopumps, since they may
work in four quadrants. Four quadrants pumps are currently not supported.

There are in principle four dependent variables describing any state of a

pump, namely:

 discharge - Q (m3/s)
 total dynamic head (tdh) - H (m)
 rotational speed - N (rpm)
 shaft torque - T (N.m)

The total dynamic head is defined as follows:

 V 2 p   V 2 
tdH  H   d  d  z  s  p s  z 
 2g  g d   2 g  g s 
   (4.41)

where the subscripts, d and s denote the discharge and suction flanges,
respectively. Power input P (kW) is defined as:

 gQ H T 2 N
P   T  
 60
(4.42)

where h is the pump efficiency and T (N.m) is the torque which may be calcu-
lated from this equation.

Manufacturers may provide pump performance characteristics using other

variables, e.g., {H, Q, N, P}, {H, Q, N, h}. If the pump operates only in the first
quadrant, the typical pump characteristics {H, Q, N, h} for a given rotational
speed of a centrifugal pump are shown in Fig. 3.2. The H - Q curve should be
a monotonously decreasing function and then it is called a stable pump
curve. The H - Q performance curve for a pump operating at constant rated
speed may be approximated as:

H=b + aQ2

where b is the shut-off head and a is determined for maximum efficiency of

the pump.

If the pump characteristics does not satisfy parabolic relation large errors may
be produced in GGM method and in all computation modules if the pump dis-
charge is out of the Q-H curve

186 MIKE URBAN - © DHI

Water Hammer Calculations

Figure 4.4 Q - H Curve

Another performance characteristic curve which should be specified by the

manufacturer is the net positive suction head (NPSH). The absolute pressure
at the inlet flange of the pump should be above NPSH in order to avoid cavi-
tation.

By applying the principles of dimensional analysis, the following relationships

can be written for a pump operating at two different speeds N1, N2

2 3
Q N H  N  P  N 
1  1 1   1  1   1 
Q N H  N  P  N 
2 2 2  2  2  2  (4.43)

Subscripts 1 and 2 are only for corresponding points on an affinity law parab-
ola. The affinity laws for discharge and head are accurate for all types of cen-
trifugal pumps. However, large errors may be produced using the affinity law
for a power requirement. It is recommended to compute P from head, dis-
charge and efficiency and not from affinity laws.

Many of the important transient analyses situations are caused by start-up

and shutdown of pumps. For a pump power failure the change in rotational
speed of the pump depends upon the unbalanced torque applied

 gQH T 2 N
P  T 
 60
(4.44)

where Iw (N.m.s) is combined moment of inertia and Dt (s) is time step used
for the calculation.

187
 MIKE URBAN Water Hammer

Pump start-up can be described by similar equation

N 
 Tm  T t30
I 
 (4.45)

where, Tm (N.m) is the motor torque.

The relation between the pump speed and the total pump dynamic head is
described by the following equation:

tdH100% 2
tdH t  2
Nt
N100%
(4.46)

where, index (100%) represents the 100% of the pump rated speed and the
time index t represents the actual value of tdH and N during the analysis.

Three different modes can be used in the transient flow analysis:

1. pump is controlled by a pump operation schedule (N-time) curve

2. pump is controlled by a pump operation schedule until time of the simula-
tion is equal time of the power failure, then pump shutdown is applied
and pump remains stopped till the end of the computation run.
3. pump is primarily stopped (N equals zero) until time of pump start-up is
reached, then pump start-up equation is applied.

Moment of Inertia, resistance of a rotating body to the change of its rotational

speed, sometimes called rotational inertia. In linear motion, inertial mass is
the measure of the resistance of a body to a change in its state of rest or uni-
form motion in a straight line. In rotational motion, moment of inertia is the
measure of the resistance of a body to a change in its rate of rotation.The
laws of motion of rotating objects are equivalent to the laws of motion for
objects moving in a line, with moment of inertia replacing mass, angular
acceleration replacing linear acceleration, and so on.

Force = mass x acceleration (F = ma) (linear motion)

Torque = moment of inertia x angular acceleration (T = Ia) (rotational motion)

The moment of inertia of a body can be calculated by dividing the object up

into many small elements each with mass, m. If each element is a distance,
ri, from the axis of rotation, the moment of inertia of the body is given by:

n
I  mrj
i 1
(4.47)

188 MIKE URBAN - © DHI

Water Hammer Calculations

The moment of inertia of a body depends on the axis about which the body is
rotated. If two axes of rotation have different distributions of mass around
them, then the body will have different moments of inertia for each of these
axes.

Torque, a twisting effort applied to an object that tends to make the object turn
about its axis of rotation. The magnitude of a torque is equal to the magnitude
of the applied force multiplied by the distance between the object's axis of
rotation and the point where the force is applied. In many ways, torque is the
rotational analogue to force. Just as a force applied to an object tends to
change the linear rate of motion of the object, a torque applied to an object
tends to change the object's rate of rotational motion.

Figure 4.5 Pump torque curve

Figure 4.6 Motor torque curve

189
 MIKE URBAN Water Hammer

Pump
A Pump can operate in 3 quadrant with constant and/or variable rotational
speed. Parameters which describe a Normal Pump are as follows:

[ID]

Pump ID

[N1]

Starting node ID

[N2]

Ending node ID

[rpm], [N]

Rated speed in r.p.m

[flow direction]

Direction of pumping, from beginning node to ending node, from end-

ing node to beginning node

[operation type]

Define pump operation. Each pump is operated by the predefined

operation schedule (always required) and pump start-up or trip-off can
be modelled by specifying the time level in seconds.

[operational schedule]

N-time curve

[pump characteristics for 100% pump speed]

Q-H curve in 3 quadrants.

[pump characteristics for zero pump speed]

Q-H curve in 3 quadrants.

[Iw]

combined moment of inertia

[pump torque]

T(pump)-Q curve

190 MIKE URBAN - © DHI

Water Hammer Calculations

[motor torque]

T(motor)-N curve

Figure 4.7 Pump Q-H curve defined in 3 quadrants for 100% rated speed

Specific Valve Data

Input of valves is the same as in the case of steady state analysis. Then you
have to specify valve characteristic curve (in case of TCV valves) and valve
schedule - relation between valve opening versus time.

The relationship between the flow Q and the head drop DH is expressed
using a discharge coefficient Cd for:

In-line valve

1
Q  Cd Av 2 gH Cd 

(4.48)

where Av is the valve area and x is the valve minor loss coefficient.

Free-discharge valve

1
Q  Cd Av 2 gH Cd 
 1
(4.49)

where xf is the valve minor loss coefficient for a free-discharge valve.

191
 MIKE URBAN Water Hammer

Values of the discharge coefficients as functions of the relative valve opening

(which is the ratio of valve and pipe area) have to be specified in the in Curve
Editor. Typical representative data is of the following form

Figure 4.8 Discharge coefficient for a typical in-line valve

Remarks:

TCV Throttle Control Valves can also be used as Isolation Valves for example
for isolation of a pipe section in case of repair, isolation of a pump, etc.
Parameters describing Control Valves are as follows:

[ID]

Valve ID

[N1]

Starting node ID

[N2]

Ending node ID

[Dd]

Valve diameter

[Valve characteristic]

discharge coefficient Cd as a function of the relative valve opening A

[%] is represented by table values [A-Cd table]. Cd units (-)

or,

discharge coefficient Kv as a function of the relative valve opening A

[%] is represented by table values [A-Kv table]. Kv coefficient (m3/hod)
represents discharge (m3) though the valve corresponding to the spe-
cific valve opening (%) under the pressure of 1 bar.

192 MIKE URBAN - © DHI

Water Hammer Calculations

[Valve schedule]

tabulated values of the relative valve opening A during the opening

and/or closing procedure

Figure 4.9 Valve Cd Characteristics

Figure 4.10 Valve Kv Characteristics (Example)

The relation between Cd and Kv valve coefficients is given by the following

equation:

Kv
Cd 
3600A 2 g
(4.50)

193
 MIKE URBAN Water Hammer

The relation between Cd valve coefficient and x minor loss coefficient is given
by the following equation:

1
Cd 
 1
(4.51)

or, for an in-line valve:

1
Cd 

(4.52)

Note, that the valve minor loss coefficient used for the steady state analysis
must correspond the initial valve opening used for the water hammer analy-
sis.

Specific Project Options Settings

Analysis type included fast transient flow analysis. Currently, only SI units
with LPS are allowed for the transient flow analysis along with Darcy-Weis-
bach friction expression. Specific numeric parameters, such as theta - used to
centre the high order finite difference scheme in time, default value of 0.5,
and others can be defined.

Specific Time Settings

Running the fast transient analysis requires entering specific time setting,
namely hydraulic time step and duration of the analysis. Pressure waves trav-
els with a high speed in the pressurized pipe networks; wave speed in steel
pipes is app. 1,200 m/s. In order to maintain Courant number criterion, dt -
time step has to be very small number such as dt = 0.1s.

t
Cr  1  a
 tx (4.53)

in which a - wave speed, dt - time step, dx - grid step, Cr - Courant number, a

non dimensional parameter.

194 MIKE URBAN - © DHI

Water Hammer Calculations

Specific Curves Data

Curve editor is extended by several new curve types, which are used in the
fast transient analysis.The following new curve types are available in the
Water Hammer simulations:

Table 4.3 Curve Data

Curve type Description

HGL transient boundary Define how HGL changes in time

Q transient boundary Define how flow changes in time (positive

value-outflow, negative value-inflow)

Valve schedule Define valve opening and closing as a func-

tion of time

Valve characteristic Flow coefficient versus valve opening

Pump schedule Define pump starting and closing as a func-

tion of time

Pump torque Pump torque versus flow

Motor torque Motor torque versus pump rotational speed

Dual-acting valve characteristic Air discharge versus gauge pressure

195
 MIKE URBAN Water Hammer

List of Supported Components

Table 4.4 List of supported components

Component Remark

Tank Available

Pump Available

Pressure reducing valve PRV Not available

Pressure sustaining valve PSV Not available

Pressure breaker valve PBV Not available

Flow control valve FCV Not available

Throttle control valve TCV Available

Closed pipes Available

Pipes with check valves CV Available

Node demands Multiple junction demands including their

patterns are kept constant during water
hammer analysis.

List of Unsupported Component

The following components are not supported by the Water Hammer simula-
tions.

Table 4.5 List of unsupported components

Component Remark

General purpose valve GPV Not available

Emitter Not available

Simple control rules Not available

Rule base controls Not available

Patterns Demand and Reservoir patterns need to be

entered as Transient Boundary Conditions

4.4.3 List of Additional Components

Several additional network components are used in Water Hamm er simula-

tions comparing to EPANET based simulation. These components (struc-

196 MIKE URBAN - © DHI

Water Hammer Calculations

tures) are classified according to their location either in nodes or on

branches.

Table 4.6 List of new components

Component Remark

Air Chamber Available

Vented air-chamber Available

Air Valve Available

4.4.4 Components Located in Nodes

One of the most frequently used components of water distribution networks

are tanks. Depending on their geometry, the tanks are classified as rectangu-
lar tanks, circular tanks, or tanks with the Depth-Volume curves. Tanks are
entered in the same way as in the case of steady state or extended period
analysis.

Tanks
Surge Tanks have been widely used for hydroelectric systems in order to pro-
tect the low-pressure supply tunnel. They may also sometimes be suitable for
water supply schemes. There are various types of Surge Tanks. The sche-
matic presentation of common Surge Tanks is the same as mentioned above
for Tanks.

The governing equations describing their hydraulic behavior are the dynamic
equation and the continuity equation. Losses are disregarded at the junction,
but are taken into account for pipes. Parameters characterizing the Surge
Tank are:

Parameters:

 Node ID.
 Maximum water depth above datum.
 Starting water depth for computation.
 Tank bottom level.
 Tank Type: Rectangular tank: [a] [b] right prism rectangular tank, the
base with sides a, b; Circular tank: vertical cylinder with diameter D; Var-
iable: depth versus volume curve.

Air-Chambers
Air Chambers contain compressed air which prevent very low minimum pres-
sures in the pipeline and hence column separation. They are frequently used

197
 MIKE URBAN Water Hammer

behind the pumps in water supply pipelines. Mostly they are cylindrical with a
vertical and/or horizontal axis. A horizontal cylinder may be preferred for a
very long pipeline when a large volume of air is required. The analysis is sim-
ilar for both cases, but the computation of the volume of air in a horizontal cyl-
inder is more difficult. Figure 4.11. illustrates an Air Chamber with a vertical
cylindrical tank.

The hydraulic behavior of an Air Chamber is described by the relation

between air pressure, its volume and continuity equation. It is assumed that
the enclosed air follows the polytropic relation for a perfect gas

C H * 
air air (4.54)

in which Hair and "air are the absolute pressure head and the volume of the
enclosed air, k is the exponent in the polytropic gas equation (k = 1.0 for an
isothermal expansion, k = 1.4 for adiabatic expansion). The orifice losses are
different for the inflow and outflow from the chamber.

Parameters:

 Node ID. Air chamber can be located only in the "dead-end" node, i.e.
node with only one adjacent pipe.
 Maximum water depth above datum.
 Starting water depth for computation.
 Tank bottom level.
 Tank Type: Rectangular tank: [a] [b] right prism rectangular tank, the
base with sides a, b; Circular tank: vertical cylinder with diameter D; Var-
iable: depth versus volume curve.
 Kappa: k = the exponent in the polytropic gas equation (default value
k=1.2).

198 MIKE URBAN - © DHI

Water Hammer Calculations

Figure 4.11 Air chamber

Vented Air Chambers

Vented Air Chambers contain air which prevent very low minimum pressures
in the pipeline and hence column separation. Vented Air Chambers are
equipped by dual-acting valves that allow air to be sucked into its chamber
and to escape there from, while preventing the outflow of liquid. The outflow
of air is usually restricted and the escaping air is compressed by the liquid in
the chamber. They are frequently used behind the pumps in water supply
pipelines. Mostly they are cylindrical with a vertical and/or horizontal axis.

The hydraulic behavior of an Air Chamber is described by the relation

between air pressure, its volume and continuity equation. It is assumed that
the enclosed air follows the izotermic relation for a perfect gas

C  H * 
air air (4.55)

in which Hair and "air are the absolute pressure head and the volume of the
enclosed air, k is the exponent in the polytropic gas equation (k = 1.0 for an
isothermal expansion, k = 1.4 for adiabatic expansion). m is the entrapped air
volume, pair is the air pressure inside a chamber for which the Qair (air dis-
charge) is taken from the dual-acting valve characteristics. Next chart shows
characteristics of Pont&Mousson, Ventex dual-acting valve, diameter of
150mm.

199
 MIKE URBAN Water Hammer

Figure 4.12 Dual-acting Valve Characteristics

When the pressure inside the Vented Air Chamber drops below the atmos-
pheric pressure, dual-acting valve opens and the air flow into a chamber. The
proper valve characteristics are required to set by a user. As soon as the liq-
uid starts flowing back into the dual-acting valve, valve closes.

Parameters:

 Node ID. Vented-air chamber can be located only in the "dead-end"

node, i.e. node with only one adjacent pipe. Vented air-chambers are
defined as junction nodes with an air-valve.
 Dual-acting Valve characteristics: Relation between gauge pressure and
the air flow through the dual-acting valve.

Figure 4.13 Vented Air-Chamber

200 MIKE URBAN - © DHI

Water Hammer Calculations

Air-Valve
Air valves, similar to Vented Air Chamber contain air which prevent very low
minimum pressures in the pipeline and hence column separation. Air valves
are modelled as small Vented Air Chamber equipped by dual-acting valves
that allow air to be sucked into its chamber and to escape therefrom, while
preventing the outflow of liquid. When the pressure inside the surrounding
pipes drops below the atmospheric pressure, air-valve opens and the air flow
into a system. The proper valve characteristics are required to set by a user.
As soon as the liquid starts flowing back into the dual-acting valve, valve
closes.

Parameters:

 Node ID. Air chamber can be located only in the "dead-end" node, i.e.
node with only one adjacent pipe. Air-valves are defined as junction
nodes with an air-valve type.
 Valve diameter.
 Dual-acting Valve characteristics: Relation between gauge pressure and
the air flow through the dual-acting valve.

201
 MIKE URBAN Water Hammer

Figure 4.14 Air Valve

4.4.5 Tutorial

This section contains brief summary describing how to use Water Hammer
when creating a new project.

1. Create new project as it is described for steady state analysis.

2. Change the project type to Transient Flow Analysis in the Project
Options dialog. Use Edit | Project Options to open Project Options dialog.
3. Modify time settings by selecting Extended | Time Editor. Use seconds to
define time settings. Typical time step for running the water hammer
analysis would be 0.1 second. Short section pipes are automatically
treated in a special way in order to avoid numerical problems associated
with high values of Courant numbers Cr.
4. Define Q = f (t) and H = f (t) boundary conditions for the transient flow
analysis. Use Edit | Water hammer | Boundary Conditions to assign the
appropriate boundary conditions to nodes.

202 MIKE URBAN - © DHI

Water Hammer Calculations

5. Add wave speed for each pipe in the model. Use the Field Calculator
from within the Pipe Editor to assign the wave speed globally. Example
of the command assigning the wave speed of 1200m/s to each pipe:
wavespeed=1200.
6. Edit specific data to pumps, valves, air-chambers, vented air-chambers,
or air valves where appropriate.
7. Run the water hammer analysis and load the analysis results.
8. Use standard way of reviewing the analysis results such as time series,
animated plans, or animated profile plots.

203
 MIKE URBAN Water Hammer

204 MIKE URBAN - © DHI

5 EPANET
MIKE URBAN consists of EPANET numerical engines (public domain
EPANET engine developed by EPA, US and EPANET engine modified by
DHI Water&Environment by adding variable speed drive pump, extended
controls, automated fire flow analysis, PID real-time control and other
ednanced modeling features) used for computing steady state, extended
period, and water quality analysis and of WATER HAMMER numerical engine
used for computing fast transient analysis.

The EPANET computer model used for water distribution network analysis is
composed of two parts: (1) the input data file and (2) the EPANET computer
program. The data file defines the characteristics of the pipes, the nodes
(ends of the pipe), and the control components (such as pumps and valves)
in the pipe network. The computer program solves the nonlinear energy
equations and linear mass equations for pressures at nodes and flowrates in
pipes.

EPANET Input Data File

The EPANET input data file, created automatically by MIKE URBAN WATER,
includes descriptions of the physical characteristics of pipes and nodes, and
the connectivity of the pipes in a pipe network system. The user can graphi-
cally layout the water distribution network, if desired. Values for the pipe net-
work parameters are entered through easy-to-use dialog boxes. MIKE
URBAN WATER then creates the EPANET input data file in the format
required to run the analysis. The pipe parameters include the length, inside
diameter, minor loss coefficient, and roughness coefficient of the pipe. Each
pipe has a defined positive flow direction and two nodes. The parameters of
nodes consist of the water demand or supply, elevation, and pressure or
hydraulic grade line. The hydraulic grade line (HGL) is the summation of node
elevation and pressure head at the node. The control components, which
usually are installed on pipes, include control valves and booster pumps.
They are also part of the input data file.

EPANET Computer Program

The EPANET computer program was developed by the U.S. EPA (Environ-
mental Protection Agency). The program computes the flowrates in the pipes
and then HGL at the nodes. The calculation of flowrates involves several iter-
ations because the mass and energy equations are nonlinear. The number of
iterations depends on the system of network equations and the user-specified
accuracy. A satisfactory solution of the flowrates must meet the specified
accuracy, the law of conservation of mass and energy in the water distribution
system, and any other requirements imposed by the user. The calculation of
HGL requires no iteration because the network equations are linear. Once the
flowrate analysis is complete, the water quality computations are then per-
formed.

205
 EPANET

5.1 History
Pipe network analysis of water distribution systems has evolved from a time-
consuming process done infrequently to a quick and easy process done reg-
ularly on systems of all sizes.

Pipe network analysis initially started early in 1940. Years later, two network
analysis programs were introduced by Shamir and Howard (1968) and Epp
and Fowler (1970). Both programs used the Newton-Raphson method to line-
arize the nonlinear mass and energy equations. The major differences
between these two programs are:

1. The Shamir-Howard program is based on node-oriented equations, while

the Epp-Fowler program is based on loop-oriented equations.
2. The Shamir-Howard program solves for pressure, demand, and the
parameters of pipes and nodes, while the Epp-Fowler program solves
only for pressures and flowrates.

Since then, several programs have been developed, based on improved

computing techniques as well as advances in computer hardware. Recently,
several computer programs running on personal computers, such as
EPANET, UNWB-LOOP, WADISO, U of K KYPIPE, and WATER have been
created and made available. Of the four programs, only EPANET and U of K
KYPIPE can perform dynamic simulation over a extended period of time and
only EPANET can perform water quality analysis. In addition, WADISO can
perform optimization analysis.

5.1.1 Analysis Methods

Three types of analysis may be conducted using MIKE URBAN WATER:

steady state (static) analysis, extended period (dynamic) analysis, and water
quality analysis. Steady state analysis is used to compute the pipe flowrates
and the node HGL in a steady state pipe network system. Extended period
analysis simulates the continuous flowrate and pressure changes over a
period of time. Water quality analysis is used to compute the age of water,
perform source tracking, calculate the fate of a dissolved substance, or deter-
mine the growth or decay of a substance. Each of these analysis types are
discussed in detail in the following sections.

Steady State Hydraulics

The calculation of flowrates and pressures for a steady state pipe network
system is called a steady state analysis. This analysis computes the pipe
flowrates and the node hydraulic grade line elevations (HGL) so that the con-
servation of energy and mass are satisfied. The pipe network system can
include pumps, check valves, and various types of control valves.

206 MIKE URBAN - © DHI

History

Extended Period Hydraulics

An extended period (dynamic) analysis is used to analyze a pipe network for
an extended period of time. The total simulation time is usually divided into
several time steps. At each time step an analysis is conducted for the pipe
network based on the current network parameters and the pipe flowrates cal-
culated from the previous time step.

In an extended period simulation, storage tanks and hydraulic switches are

often present as part of the water distribution system. The system operating
parameters at each time step depend on external conditions and the pipe
flowrates from the previous time step. External conditions are operating
parameters controlled by factors outside the system, such as external
demand or pump power. The previous time step flowrates are also used to
predict the storage tank HGL for the current time step.

Water Quality Analysis

A water quality analysis can be used by operations, planning, and engineer-
ing departments to study the flow and distribution of water. Source tracking,
travel time determination, water age, and concentration levels of chemical
constituents and contaminants are the primary concerns addressed by water
quality models. In addition, tracking paths of flow and distribution provide the
engineer insight to the origin and amount of water supplied to a particular
location, as well as concentration levels of chemical constituents and contam-
inants.

Water quality models can also be used to study water retention time for reser-
voir operations, pipeline travel times, and the percentage of water supplied to
a location from multiple sources (i.e., treatment plants, wells, and reservoirs).
Additionally, water quality models can be used to develop a hydrant flushing
program to reduce water stagnation at dead ends within the pipe network.
And, site sampling locations, future rechlorination facility locations, cross-
connection locations, and reservoir operating strategies can be designed and
analyzed.

5.1.2 Applications of MIKE URBAN WATER

Water distribution analysis software, such as MIKE URBAN WATER, is typi-

cally used for three broad areas of analysis. These areas of analysis are gen-
erally referred to as planning, design, and operation applications (AWWA
Manual M32, 1989).

Some examples of these applications include (1) analysis and design of

booster pumps and storage tanks for municipal or rural water distribution sys-
tems, and (2) analysis and design of chlorination satellite stations.

207
 EPANET

Planning
MIKE URBAN WATER can be used in the planning of pipe network systems
to meet forecasted demands of the next 10 years or 20 years. For example,
the program can be used to develop long term capital-improvement plans for
the existing pipe network system. These plans can include staging, sizing,
and locating future pipe network and water chlorination facilities. The soft-
ware can also be used in the development of a main rehabilitation plan or a
system-improvement plan. And, a network analysis can provide suggestions
and recommendations to prepare for the occurrence of any unusual events.

Design
MIKE URBAN WATER can be used to design a new pipe network system or
improve on the existing pipe network system. For example, the analysis con-
ducted using MIKE URBAN WATER could help users in selecting and sizing
pipe network components, such as pipes, booster pumps, and pressure regu-
lating valves. As a part of the analysis, the performance of the pipe network
system can be analyzed to verify that the system satisfies fire-flow demand
requirements.

Operation
The operating status of a water distribution system (e.g., the pipe flowrates
and junction node pressures) can be determined by MIKE URBAN WATER.
The analysis can then be used to develop operational strategies based on the
guidelines for maximum use of available water and efficient management of
electrical energy. MIKE URBAN WATER can also be used for system trouble-
shooting, such as finding the location of a pipe break.

5.1.3 Skeletonization

In the past, water distribution models have not included all of the pipes con-
tained in the network system due to the fact that the numerical modeling
schemes used and the memory requirements required could handle only a
limited number of pipes and nodes. These limitations required skeletonization
of the pipe network system, where only a subset of all the pipes contained
within the network system was defined. However, these limitations have
eroded over the past years as enhanced programming methods and
increased computer hardware capabilities have come into being.

Skeleton network models typically include only those pipes that are consid-
ered significant to the flow and distribution of water. For example, a skeleton
model might only consider 12 inch and larger diameter mains. Smaller diame-
ter mains might also be included if they supply water to a significant area or
complete a loop in the network. Examples of a complete network model and
an equivalent skeleton network model are shown in Figure 5.1 and
Figure 5.2.

208 MIKE URBAN - © DHI

History

Pipe Diameter
6" and smaller
8"
12"
16"

Figure 5.1 A complete water distribution network, showing all pipes contained
within the network

Pipe Diameter
6" and smaller
8"
12"
16"

Figure 5.2 An equivalent skeleton water distribution network, showing those pipes
that are considered significant to the flow and distribution of water for a
steady state simulation

A major advantage with working with a skeleton model is that they are much
easier to define since there is less data involved, and the simulation time is
shorter since not as many pipes are involved in computing a solution. Also,

209
 EPANET

the displaying of results is quicker and more readable since less pipes are
involved. In addition, the number of iterations required to converge to a solu-
tion is typically less as well.

Skeletonization, however, can adversely affect model accuracy. For example,

skeletonization should not be used when performing water quality modeling
since the flow rates, paths, and velocities for all pipes are critical components
to a water quality simulation. For steady state water distribution simulations,
though, skeletonization can produce results that have sufficient accuracy if
the water consumption has been properly assigned to the defined nodes. For
example, a skeleton model node might not only represent the water demand
within the immediate area, but also demand for a smaller pipe that services
an area much farther away. Also, if an existing calibrated model has been
converted into a skeleton model, it may be necessary to recalibrate the new
skeleton model since the nodal demands would be represented differently.

As was stated before, a skeleton model should not be used when performing
water quality modeling. However, in other situations, the modeler must decide
when it is appropriate to skeletonize a network model to produce results with
sufficient accuracy to meet the modeling requirements. Therefore, when and
where to skeletonize a network must be decided upon a case by case basis.

5.2 The Water Distribution Network

A water distribution system is a pipe network which delivers water from single
or multiple supply sources to consumers. Typical water supply sources
include reservoirs, storage tanks, and external water supply at junction nodes
such as groundwater wells. Consumers include both municipal and industrial
users. The pipe network consists of pipes, nodes, pumps, control valves,
storage tanks, and reservoirs.

EPANET views the water distribution system as a network containing nodes

and links, where the nodes are connected by links. Figure 5.3 illustrates a
node-link representation of a simple water distribution network.

210 MIKE URBAN - © DHI

The Water Distribution Network

Figure 5.3 Node-link representation of a water distribution network

As shown in Figure 5.3, links can represent the following components in a

network:

 Pipes
 Pumps
 Valves

Nodes, besides representing the connection point between pipes, can repre-
sent the following components in a network:

 Points of water consumption (demand nodes).

 Points of water input (source nodes).
 Locations of tanks or reservoirs (storage nodes).

How the EPANET program models the hydraulic behavior of each of these
components is described in the following sections. All flow rates in this dis-
cussion will be assumed as cubic feet per second (cfs), although the program
can also accept flow rates in gallons per minute (gpm), million gallons per day
(mgd), and litres per second (L/s).

5.2.1 Pipes

Every pipe is connected to two nodes at its ends. In a pipe network system,
pipes are the channels used to convey water from one location to another.
The physical characteristics of a pipe include the length, inside diameter,
roughness coefficient, and minor loss coefficient. The pipe roughness coeffi-
cient is associated with the pipe material and age. The minor loss coefficient
is due to the fittings along the pipe.

When water is conveyed through the pipe, hydraulic energy is lost due to the
friction between the moving water and the stationary pipe surface. This fric-

211
 EPANET

tion loss is a major energy loss in pipe flow and is a function of flowrate, pipe
length, diameter, and roughness coefficient.

The head lost to friction associated with flow through a pipe can be
expressed in a general fashion as:

b
h L = aq (5.1)

where

– hL = head loss, ft
– q = flow, cfs
– a = a resistance coefficient
– b = a flow exponent

EPANET can use any one of three popular forms of the headloss formula
shown in Equation 5.1: the Hazen-Williams formula, the Darcy-Weisbach for-
mula, or the Chezy-Manning formula. MIKE URBAN WATER allows the user
to choose the formulation to use.

The Hazen-Williams formula is probably the most popular head loss equation
for water distribution systems, the Darcy-Weisbach formula is more applica-
ble to laminar flow and to fluids other than water, and the Chezy-Manning for-
mula is more commonly used for open channel flow. Table 5.1 lists resistance
coefficients and flow exponents for each formula. Note that each formula
uses a different pipe roughness coefficient, which must be determined empir-
ically. Table 5.2 lists general ranges of these coefficients for different types of
new pipe materials. Be aware that a pipe's roughness coefficient can change
considerably with age.

While the Darcy-Weisbach relationship for closed-conduit flows is generally

recognized as a more accurate mathematical formulation over a wider range
of flow than the Hazen-Williams formulation, the field data on e values
(required for the Darcy-Weisbach formulation) are not as readily available as
are the C values for the pipe wall roughness coefficient (used in the Hazen-
Williams formulation).

Table 5.1 Pipe head loss formulas

Formula Resistance Coefficient (a) Flow Exponent (b)

Hazen-Williams 4.72 C-1.85 d-4.87 L 1.85

Darcy-Weisbach 0.0252 f(,d,q) d-5 L 2

Chezy-Manning 4.66 n2 d-5.33 L 2

(full pipe flow)

212 MIKE URBAN - © DHI

The Water Distribution Network

Notes:

– C = Hazen-Williams roughness coefficient

– e = Darcy-Weisbach roughness coefficient, ft
– f = friction factor (dependent on e, d, and q)
– d = pipe diameter, ft
– L = pipe length, ft

Table 5.2 Roughness coefficients for new pipe

Material Hazen-Williams C Darcy-Weisbach , Manning’s n

millifeet

Cast Iron 130 - 140 0.85 0.012 - 0.015

Concrete or Con- 120 - 140 1.0 - 10 0.012 - 0.017

crete Lined

Galvanized Iron 120 0.5 0.015 - 0.017

Plastic 140 - 150 0.005 0.011 - 0.015

Steel 140 - 150 0.15 0.015 - 0.017

Vitrified Clay 110 ---- 0.013 - 0.015

Pipes can contain check valves in them that restrict flow to a specific direc-
tion. They can also be made to open or close at pre-set times, when tank lev-
els fall below or above certain set-points, or when nodal pressures fall below
or above certain set-points. The normal initial condition for a pipe containing a
check valve or a pump is to be in open mode. The pipe will then switch to
closed mode only when flow is reversed.

In addition to the energy loss caused by friction between the fluid and the
pipe wall, energy losses also are caused by obstructions in the pipeline,
changes in flow direction, and changes in flow area. These losses are called
minor losses because their contribution to the reduction in energy is usually
much smaller than frictional losses. Head loss, which is the sum of friction
loss and minor losses, reduces the flowrate through the pipe.

Modeling Parallel Pipes

MIKE URBAN WATER can model parallel pipes. To define a set of parallel
pipes, simply draw in pipes with the same starting and ending nodes. How-
ever, to prevent the pipes from being drawn exactly on top of one another, it is
suggested that one of the pipes have at least one vertex (or bend point)
inserted in it so that the pipes are displayed slightly separated (see
Figure 5.3). For information on defining curved pipes, see the section titled
Pipe Editor on -18.

213
 EPANET

Figure 5.4 Vertices

Bend points are inserted to separate two parallel pipes to allow them to be
selectable.

5.2.2 Pumps

A pump is a device that raises the hydraulic head of water. EPANET repre-
sents pumps as links of negligible length with specified upstream and down-
stream junction nodes.

Pumps are described with a pump characteristic curve. The pump curve
describes the additional head imparted to a fluid as a function of its flow rate
through the pump. EPANET is capable of modeling several types of user
defined pumps, constant energy pumps, single-point pump curves, three-
point pump curves, multiple point pump curves, and variable speed pump.

Constant energy pumps operate at the same horsepower or kilowatt rating

over all combinations of head and flow. In this case the equation of the pump
curve would be:

h G = 8,81Hp
------------------ (5.2)
q

where

– hG = head gain, ft
– Hp = pump horsepower
– q = flow, cfs

A single point pump curve is defined by a single head-flow combination that

represents a pump’s desired operating point. EPANET fills in the rest of the
curve by assuming:

1. A shutoff head at zero flow equal to 133% of the design head.

2. A maximum flow at zero head equal to twice the design flow.

214 MIKE URBAN - © DHI

The Water Distribution Network

Figure 5.5 Single Point Pump Curve

A three-point pump curve is defined by three operating points:

1. Low Flow (flow and head at low or zero flow conditions).

2. Design Flow (flow and head at desired operating point).
3. Maximum Flow (flow and head at maximum flow).

Figure 5.6 Three Point Pump Curve

215
 EPANET

A multi-point pump curve is defined by providing either a pair of head-flow

points or four or more such points. EPANET creates the complete curve by
connecting the points with straight line segments.

Figure 5.7 Multi Point Pump Curve

For variable speed pumps, the pump curve shifts as the speed of the pump
changes. The relationships between flow (Q) and head (H) at speed N1 and
N2 are:

Q1 N1
-------- = ------- (5.3)
Q2 N2

2
H1 N
------- = ------1- (5.4)
H2 N2

EPANET will shut a pump down if the system demands a head higher than
the first point on a curve (i.e., the shutoff head.) A pump curve must be sup-
plied for each pump the system unless the pump is a constant energy pump.

Pumping Rate Limits

Flow through a pump is unidirectional (i.e., one flow direction only) and
pumps must operate within the head and flow limits imposed by their charac-
teristic curves. If the system conditions require that the pump produce more
than its shutoff head, EPANET will attempt to close off the pump and will
issue a warning message.

216 MIKE URBAN - © DHI

The Water Distribution Network

Controlling Pumps
EPANET allows you to turn pumps on or off at pre-set times, when tank levels
fall below or rise above certain set-points, or when nodal pressures fall below
or rise above certain set-points. Variable speed pumps can also be consid-
ered by specifying using the Control Editor (see the section titled Control Edi-
tor on -74 for further information) that their speed setting be changed under
these same types of conditions. By definition, the original pump curve sup-
plied to the program has a relative speed setting of 1. If the pump speed dou-
bles, then the relative setting would be 2; if run at half speed, the relative
setting is 0.5 and so on. Figure 5.5 illustrates how changing a pump's speed
setting affects its characteristic curve.

Figure 5.8 Effect of relative speed (n) on a pump curve

Modeling Pumps in Parallel and Series

As shown in Figure Figure 5.9, MIKE URBAN WATER can model parallel
pumps and pumps in series. To model parallel pumps, it is necessary to insert
the pumps on the same from and to nodes. To model pumps in series (where
the outlet of the first pump directly discharges into the inlet of the second
pump), it is necessary to insert the pumps one after the other on the same
pipe.

217
 EPANET

Figure 5.9 MIKE URBAN WATER can model both parallel pumps and pumps in
series

If desired, the two or more pumps can be modeled as an equivalent compos-

ite single pump that has a characteristic curve equal to the sum of the individ-
ual pump curves. For pumps that are in parallel, the discharge values for the
individual pump curves are added together to end up with the equivalent sin-
gle pump curve. If the pumps are connected together in series, then the head
values are for the individual pump curves are added together to end up with
the equivalent single pump curve.

For example, as shown in Figure 5.10, if two pumps are connected in parallel
and one pump operates at a flow rate of 50 gpm with a head of 75 ft and the
other pump operates at a 60 gpm with a head of 75 ft, then the resultant sin-
gle equivalent pump curve flow rate of 110 gpm would be available with a
head of 75 ft.

218 MIKE URBAN - © DHI

The Water Distribution Network

200

150

Head (ft)
100 Equiv
ilent P
ump

Pu
50

mp
Pum

2
p1
0
0 25 50 75 100 125 150
Discharge (gpm)

Figure 5.10 When modeling pumps in parallel, an equivalent single pump curve can
be determined by adding together the characteristic pump curve dis-
charge values from the other individual pump curves

Similarly, as shown in Figure 5.11 if two pumps are connected in series and
one pump provides a head of 75 ft at a flow rate of 50 gpm and the other
pump provides a head of 65 feet at a flow rate of 50 gpm, then the resultant
single equivalent pump curve head of 140 ft would be available at a flow rate
of 50 gpm.

219
 EPANET

200 Equiv
ilent P
ump

150

Head (ft)
Pump 1
100
Pump 2

50

0
0 25 50 75 100 125 150
Discharge (gpm)

Figure 5.11 When modeling pumps in series, an equivalent single pump curve can
be determined by adding together the characteristic pump curve head
values from the other individual pump curves

When modeling pumps in series, it is preferable to use an equivalent com-

posite pump to represent the multiple pumps in series. This is because it is
much easier to control a single pump than multiple pumps simultaneously
when the pumps turn on and off. In addition, multiple pumps in series can
cause numerical disconnections in the pipe network when EPANET checks to
see if the upstream grade is greater than the downstream grade plus the
available pump head.

5.2.3 Valves

Aside from the valves in pipes that are either fully opened or closed (such as
check valves), EPANET can also represent valves that control either the
pressure or flow at specific points in a network. Such valves are considered
as links of negligible length with specified upstream and downstream junction
nodes. The types of valves that can be modeled are described below.

Pressure Reducing Valves (PRV)

Pressure reducing valves (PRV) limit the pressure on their downstream end
to not exceed a pre-set value when the upstream pressure is above the set-
ting. If the upstream pressure is below the setting, then flow through the valve
is unrestricted. Should the pressure on the downstream end exceed that on
the upstream end, the valve closes to prevent reversal of flow.

220 MIKE URBAN - © DHI

The Water Distribution Network

Pressure Sustaining Valves (PSV)

Pressure sustaining valves (PSV) try to maintain a minimum pressure on their
upstream end when the downstream pressure is below that value. If the
downstream pressure is above the setting, then flow through the valve is
unrestricted. Should the downstream pressure exceed the upstream pressure
then the valve closes to prevent reverse flow.

Pressure Breaker Valves (PBV)

Pressure breaker valves (PBV) force a specified pressure loss to occur
across the valve. Flow can be in either direction through the valve.

Flow Control Valves (FCV)

Flow control valves (FCV) limit the flow through a valve to a specified
amount. The program produces a warning message if this flow cannot be
maintained without having to add additional head at the valve.

Throttle Control Valves (TCV)

Throttle control valves (TCV) simulate a partially closed valve by adjusting the
minor head loss coefficient of the valve. A relationship between the degree to
which the valve is closed and the resulting head loss coefficient is usually
available from the valve manufacturer.

General Purpose Valves (GPV)

A General Purpose Valve (GPV) provides the capability to model devices and
situations with unique headloss - flow relationships, such as reduced pres-
sure backflow prevention valves, turbines, and well drawdown behavior. The
valve setting is the ID of a Headloss Curve.

A Headloss Curve is used to described the headloss (Y in feet or meters)

through a General Purpose Valve (GPV) as a function of flow rate (X in flow
units). It provides the capability to model devices and situations with unique
headloss - flow relationships, such as reduced flow backflow prevention
valves, turbines, and well drawdown behavior.

5.2.4 Minor Losses

Minor head losses (also called local losses) can be associated with the added
turbulence that occurs at bends, junctions, meters, and valves. The impor-
tance of such losses will depend on the layout of the pipe network and the
degree of accuracy required. EPANET allows each pipe and valve to have a

221
 EPANET

minor loss coefficient associated with it. It computes the resulting head loss
from the following formula:

2
0,0252Kq
h L = -------------------------- (5.5)
4
d

where

– hL = head loss, ft
– d = diameter in, ft
– K = loss coefficients
– q = flow rate, cfs

Table 5.3 gives values of K for several different kinds of pipe network compo-
nents.

Table 5.3 Loss coefficients for common pipe network components

Component Loss Coefficient

Globe valve, fully open 10.0

Angle valve, fully open 5.0

Swing check valve, fully open 2.5

Gate valve, fully open 0.2

Short-radius elbow 0.9

Medium-radius elbow 0.8

Long-radius elbow 0.6

45 elbow 0.4

Closed return bend 2.2

Standard tee - flow through run 0.6

Standard tee - flow through branch 1.8

Square entrance 0.5

Exit 1.0

5.2.5 Nodes

Nodes are the locations where pipes connect. Two types of nodes exist in a
pipe network system, (1) fixed nodes and (2) junction nodes. Fixed nodes are
nodes whose HGL are defined. For example, reservoirs and storage tanks
are considered fixed nodes, because their HGL are initially defined. Junction

222 MIKE URBAN - © DHI

The Water Distribution Network

nodes are nodes whose HGL are not yet determined and must be computed
in the pipe network analysis.

Degree of freedom, elevation, and water demand are the three important
input parameters for a node (see Figure 5.3). A node's degree of freedom is
the number of pipes that connect to that node. In EPANET, a junction node
may be connected to more than one pipe, but a fixed node (i.e., storage tank
or reservoir) must be connected to exactly one pipe. Therefore, a fixed node's
degree of freedom is always one, and a junction node's degree of freedom
may be greater than one. The elevation of a node can sometimes be obtained
from system maps or drawings. More often, it is approximated using topo-
graphic maps. Water demand at a junction node is the summation of all water
drawn from or added to the system at that node.

All nodes should have their elevation specified above sea level (i.e., greater
than zero) so that the contribution to hydraulic head due to elevation can be
computed. Any water consumption or supply rates at nodes that are not stor-
age nodes must be known for the duration of time the network is being ana-
lyzed. Storage nodes (i.e., tanks and reservoirs) are special types of nodes
where a free water surface exists and the hydraulic head is simply the eleva-
tion of water above sea level. Tanks are distinguished from reservoirs by hav-
ing their water surface level change as water flows into or out of them—
reservoirs remain at a constant water level no matter what the flow is.

EPANET models the change in water level of a storage tank with the following
equation:

q
y = ---- t (5.6)
A

where

– Dy = change in water level, ft

– q = flow rate into (+) or out of (-) tank, cfs
– A = cross-sectional area of the tank, ft2
– Dt = time interval, sec

For each storage tank, the program must know the cross-sectional area as
well as the minimum and maximum permissible water level. Reservoir-type
storage nodes are usually used to represent external water sources, such as
lakes, rivers, or groundwater wells. Storage nodes should not have an exter-
nal water consumption or supply rate associated with them.

Modeling Pumped Groundwater Wells

The water supply to the water distribution network is furnished by reservoirs
and storage tanks. Reservoirs are treated as inexhaustible sources of water,
and as such their water level never varies. However, as a storage tank emp-
ties, its water level lowers and it has to be refilled by pumping from either a

223
 EPANET

reservoir or a groundwater well. In EPANET, a pumped groundwater well is

modeled the same as a pumped reservoir, as shown in Figure 5.12.

Storage Tank
Reservoir

Pump 11 12 13

11 12

111 112 113

Equivalent Pumped 21 22
Groundwater Well 21 23
22
121 122

31

31 32

Figure 5.12 A pumped groundwater well is modeled as a reservoir with an attached

pump in EPANET

In order to model a pumped groundwater well, an equivalent pumped reser-

voir must be defined (as shown in Figure 5.12) since a pumped groundwater
well cannot be explicitly defined with EPANET. Note that a pump cannot be
directly attached to reservoir or storage tank, therefore junction nodes are
automatically inserted by MIKE URBAN WATER when assigning a pump to
the link attached to a reservoir or storage tank.

As the storage tank empties and its water level falls to a certain amount, con-
trol rules turn on the groundwater (reservoir) pump to start refilling of the stor-
age tank. The control rules used to regulate the starting and stopping of the
groundwater pump are defined within the Control Editor (see the section titled
Control Editor on - 74 for further information).

As pumping of the groundwater occurs, drawdown of the water table eleva-

tion at the groundwater well can occur, as shown in Figure 5.13. At higher
pumping rates, the groundwater well may not be able to recharge fast enough
to maintain the pumping rate specified by the defined groundwater well pump
curve. To account for this effect, the groundwater well pump curve may need
to be adjusted downward, as shown in Figure Figure 5.13, if this effect is a
significant factor in modeling the refilling of the storage tank.

224 MIKE URBAN - © DHI

Time Patterns

Head
Static
ped
Pum
Discharge

Figure 5.13 The groundwater well pump curve may need to be adjusted downward
to account for groundwater well drawdown and recharge effects

Modeling Hydraulically Adjacent Storage Tanks

When performing an extended period simulation, if two or more storage tanks
are hydraulically adjacent to each other it is possible that oscillations can
occur between the tanks as the water bounces back and forth between them.
This fluctuation is caused by small differences in flow rates as the tanks refill
individually, causing the water level in the tanks to differ over time thereby
causing the oscillation between the tanks. To prevent this effect from occur-
ring, it is recommended that hydraulically adjacent tanks be modeled as a
single composite tank with an equivalent total surface area and storage vol-
ume equal to the sum of the individual tanks.

5.3 Time Patterns

EPANET assumes that water usage rates, external water supply rates, and
constituent source concentrations at nodes remain constant over a fixed
period of time, but that these quantities can change from one time period to
another. The default time period interval is one hour, but this can be set to
any desired value using the Time Editor (see the section titled Time Editor on
- 87 for further information). The value of any of these quantities in a time
period equals a baseline value multiplied by a time pattern factor for that
period. Figure 5.14.1illustrates a pattern of factors that might apply to daily
water demands, where each demand period is of two hours duration. Differ-
ent patterns can be assigned to individual nodes or groups of nodes.

225
 EPANET

1.6

1.4 1.4

Usage Factor
1.2 1.2
1.1
1.0
1 0.9
0.8
0.7
0.6

;;;;;
0.5

0
0 1 2 3 4 5 6 7 8 9 10 11 12

Demand Period

Figure 5.14 Time pattern for water usage

The Pattern Editor is used to define the time patterns to be used in an

extended period simulation. See the section titled Pattern Editor on -83 for
information on how to define time patterns.

5.4 Hydraulic Simulation Model

The method used in EPANET to solve the flow continuity and headloss equa-
tions that characterize the hydraulic state of the pipe network at a given point
in time can be termed a hybrid node-loop approach. Todini and Pilati (1987)
and later Salgado et al. (1988) chose to call it the "Gradient Method". Similar
approaches have been described by Hamam and Brameller (1971) (the
"Hybrid Method) and by Osiadacz (1987) (the "Newton Loop-Node Method").
The only difference between these methods is the way in which link flows are
updated after a new trial solution for nodal heads has been found. Because
Todini's approach is simpler, it was chosen for use in EPANET.

Assume we have a pipe network with N junction nodes and NF fixed grade
nodes (tanks and reservoirs). Let the flow-headloss relation in a pipe between
nodes i and j be given as:

H i – H j = h ij = r Q n ij + mQ 2 ij (5.7)

where H = nodal head, h = headloss, i = resistance coefficient, Q = flow rate,

n = flow exponent, and m = minor loss coefficient. The value of the resistance
coefficient will depend on which friction headloss formula is being used (see

226 MIKE URBAN - © DHI

Hydraulic Simulation Model

below). For pumps, the headloss (negative of the head gain) can be repre-
sented by a power law of the form

n
hij   2 (h0  r (Qij /  ) )
(5.8)

where h0 is the shutoff head for the pump, w is a relative speed setting, and r
and n are the pump curve coefficients. The second set of equations that must
be satisfied is flow continuity around all nodes:

 Qij – Di = 0 (5.9)
j

for i = 1,... N.

where Di is the flow demand at node i and by convention, flow into a node is
positive. For a set of known heads at the fixed grade nodes, we seek a solu-
tion for all heads Hi and flows Qij that satisfy Eqs. (5.8) and (5.9).

The Gradient solution method begins with an initial estimate of flows in each
pipe that may not necessarily satisfy flow continuity. At each iteration of the
method, new nodal heads are found by solving the matrix equation.

AH = F (5.10)

where A = an (NxN) Jacobian matrix, H = an (Nx1) vector of unknown nodal

heads, and F = an (Nx1) vector of right hand side terms.

The diagonal elements of the Jacobian matrix are:

Aii   pij
j
(5.11)

while the non-zero, off-diagonal terms are:

Aij   pij
(5.12)

where pij is the inverse derivative of the headloss in the link between nodes i
and j with respect to flow. For pipes,

1
pij  n 1
nr Qij  2m Qij
(5.13)

227
 EPANET

while for pumps

1
P ij = ----------------------------------
- (5.14)
 Q ij n – 1
n r -------
2
 

Each right hand side term consists of the net flow imbalance at a node plus a
flow correction factor:

1
pij  n 1
nr Qij  2m Qij
(5.15)

where the last term applies to any links connecting node i to a fixed grade
node f and the flow correction factor yij is: P

n 2
Y ij = p  r Q ij + m Q ij  sgn  Q ij  (5.16)

for pipes and


yij   pij 2 h0  r (Qij /  ) n  (5.17)

for pumps, where sgn (x) is 1 if x > 0 and -1 otherwise. (Qij is always positive
for pumps.)

After new heads are computed by solving Eq. (5.10), new flows are found
from:

Q ij = Qij –  y ij – p ij  H i – H j   (5.18)

If the sum of absolute flow changes relative to the total flow in all links is
larger than some tolerance (e.g., 0.001), then Eqs. (5.9) and (5.10) are
solved once again. The flow update formula (5.11) always results in flow con-
tinuity around each node after the first iteration.

EPANET implements this method using the following steps:

1. The linear system of equations (5.10) is solved using a sparse matrix

method based on node re-ordering (George and Liu, 1981). After re-
ordering the nodes to minimize the amount of fill-in for matrix A, a sym-
bolic factorization is carried out so that only the non-zero elements of A
need be stored and operated on in memory. For extended period simula-
tion this re-ordering and factorization is only carried out once at the start

228 MIKE URBAN - © DHI

Hydraulic Simulation Model

of the analysis.

2. For the very first iteration at time 0, the flow in a pipe is chosen equal to
the flow corresponding to a velocity of 1 ft/sec. The flow through a con-
ventional pump equals the design flow specified for the pump curve. An
initial flow of 1 cfs is assumed for other types of pumps. (All computa-
tions are made with head in feet and flow in cfs).

3. The resistance coefficient for a pipe (r) is computed as described in

Table 5.2. For the Darcy-Weisbach headloss equation, the friction factor f
is computed by different equations depending on the flow's Reynolds
Number (Re):

Hagen - Poiseuille formula for Re < 2,000 (Bhave, 1991):

64-
f = ------- (5.19)
Re

Swamee and Jain approximation to the Colebrook - White equation for

Re > 4,000 (Bhave, 1991):

0.25
f 2
   5.74 
 Ln  0.9 
  3.7d Re  (5.20)

Cubic Interpolation From Moody Diagram for 2,000 < Re < 4,000 (Dun-
lop, 1991):

f =  X1 + R  X2 + R  X3 + X4    (5.21)

Re
R = -------- (5.22)
64

X1 = 7FA – FB (5.23)

X3 = – 0,128 + 13FA – 2FB (5.24)

X4 = R  0,032 – 3FA + 0,5FB  (5.25)

X3 = – 0,128 + 13FA – 2FB (5.26)

229
 EPANET

X4 = R  0,032 – 3FA + 0,5FB  (5.27)

FA =  Y3  –2 (5.28)

0,00514215
FB = FA  2 – ---------------------------- (5.29)
  Y2   Y3  

 5,74
Y2 = ---------- + ------------- (5.30)
3,7d Re 0,9

 + ----------------
Y3 = – 0,86859Ln  ---------- 5,74 -
(5.31)
 3,7d 0,9
4000

where e = pipe roughness and d = pipe diameter.

4. The minor loss coefficient based on velocity head (K) is converted to one
based on flow (m) with the following relation:

0,02517K
m = ----------------------- (5.32)
4
d

5. Emitters at junctions are modeled as a fictitious pipe between the junc-

tion and a fictitious reservoir. The pipe's headloss parameters are n =
(1/g), r = (1/C) n, and m = 0 where C is the emitter's discharge coefficient
and g is its pressure exponent. The head at the fictitious reservoir is the
elevation of the junction. The computed flow through the fictitious pipe
becomes the flow associated with the emitter.

6. Open valves are assigned an r-value by assuming the open valve acts
as a smooth pipe (f = 0.02) whose length is twice the valve diameter.
Closed links are assumed to obey a linear headloss relation with a large
resistance factor, i.e., h = 108Q, so that p = 10-8 and y = Q. For links
where (r+m) Q < 10-7, p = 107 and y = Q/n.

7. Status checks on pumps, check valves (CVs), flow control valves, and
pipes connected to full/empty tanks are made after every other iteration,
up until the 10th iteration. After this, status checks are made only after
convergence is achieved. Status checks on pressure control valves
(PRVs and PSVs) are made after each iteration.

8. During status checks, pumps are closed if the head gain is greater than
the shutoff head (to prevent reverse flow). Similarly, check valves are

230 MIKE URBAN - © DHI

Hydraulic Simulation Model

closed if the headloss through them is negative (see below). When these
conditions are not present, the link is re-opened. A similar status check is
made for links connected to empty/full tanks. Such links are closed if the
difference in head across the link would cause an empty tank to drain or
a full tank to fill. They are re-opened at the next status check if such con-
ditions no longer hold.

9. Simply checking if h < 0 to determine if a check valve should be closed or

open was found to cause cycling between these two states in some net-
works due to limits on numerical precision. The following procedure was
devised to provide a more robust test of the status of a check valve (CV):
– if |h| > Htol then
– if h > -Htol then status = CLOSED
– if Q < -Qtol then status = CLOSED
– else status = OPEN
– else
– if Q < -Qtol then status = CLOSED
– else status = unchanged

where Htol = 0.0005 ft and Qtol = 0.0001 cfs.

10. If the status check closes an open pump, pipe, or CV, its flow is set to 10-
6 cfs. If a pump is re-opened, its flow is computed by applying the current
head gain to its characteristic curve. If a pipe or CV is re-opened, its flow
is determined by solving Eq. (5.9) for Q under the current headloss h,
ignoring any minor losses.

11. Matrix coefficients for pressure breaker valves (PBVs) are set to the fol-
lowing: p = 108 and y = 108Hset, where Hset is the pressure drop setting
for the valve (in feet). Throttle control valves (TCVs) are treated as pipes
with r as described in item 6 above and m taken as the converted value
of the valve setting (see item 4 above).

12. Matrix coefficients for pressure reducing, pressure sustaining, and flow
control valves (PRVs, PSVs, and FCVs) are computed after all other
links have been analyzed. Status checks on PRVs and PSVs are made
as described in item 7 above. These valves can either be completely
open, completely closed, or active at their pressure or flow setting.

13. The logic used to test the status of a PRV is as follows:

If current status = ACTIVE then

– if Q < -Qtol then new status = CLOSED

– if Hi < Hset + Hml - Htol then new status = OPEN
– else new status = ACTIVE

If current status = OPEN then

231
 EPANET

– if Q < -Qtol then new status = CLOSED

– if Hi > Hset + Hml + Htol then new status = ACTIVE
– else new status = OPEN

If current status = CLOSED then

– if Hi > Hj + Htol
– and Hi < Hset - Htol then new status = OPEN
– if Hi > Hj + Htol
– and Hj < Hset - Htol then new status = ACTIVE
else new status = CLOSED

where Q is the current flow through the valve, Hi is its upstream head, Hj
is its downstream head, Hset is its pressure setting converted to head,
Hml is the minor loss when the valve is open (= mQ2), and Htol and Qtol
are the same values used for check valves in item 9 above. A similar set
of tests is used for PSVs, except that when testing against Hset, the i
and j subscripts are switched as are the > and < operators.

14. Flow through an active PRV is maintained to force continuity at its down-
stream node while flow through a PSV does the same at its upstream
node. For an active PRV from node i to j:
– Pij = 0
– Fj = Fj + 108Hset
– Ajj = Ajj + 108

This forces the head at the downstream node to be at the valve setting
Hset. An equivalent assignment of coefficients is made for an active PSV
except the subscript for F and A is the upstream node i. Coefficients for
open/closed PRVs and PSVs are handled in the same way as for pipes.

15. For an active FCV from node i to j with flow setting Qset, Qset is added
to the flow leaving node i and entering node j, and is subtracted from Fi
and added to Fj. If the head at node i is less than that at node j, then the
valve cannot deliver the flow and it is treated as an open pipe.

16. After initial convergence is achieved (flow convergence plus no change

in status for PRVs and PSVs), another status check on pumps, CVs,
FCVs, and links to tanks is made. Also, the status of links controlled by
pressure switches (e.g., a pump controlled by the pressure at a junction
node) is checked. If any status change occurs, the iterations must con-
tinue for at least two more iterations (i.e., a convergence check is
skipped on the very next iteration). Otherwise, a final solution has been
obtained.

17. For extended period simulation (EPS), the following procedure is imple-
mented:

232 MIKE URBAN - © DHI

Hydraulic Simulation Model

After a solution is found for the current time period, the time step for the
next solution is the minimum of:
– The time until a new demand period begins.
– The shortest time for a tank to fill or drain.
– The shortest time until a tank level reaches a point that triggers a
change in status for some link (e.g., opens or closes a pump) as
stipulated in a simple control.
– The next time until a simple timer control on a link kicks in.
– The next time at which a rule-based control causes a status change
somewhere in the network.

In computing the times based on tank levels, the latter are assumed to
change in a linear fashion based on the current flow solution. The activa-
tion time of rule-based controls is computed as follows:

– Starting at the current time, rules are evaluated at a rule time step.
Its default value is 1/10 of the normal hydraulic time step (e.g., if
hydraulics are updated every hour, then rules are evaluated every 6
minutes).
– Over this rule time step, clock time is updated, as are the water lev-
els in storage tanks (based on the last set of pipe flows computed).
– If a rule's conditions are satisfied, then its actions are added to a list.
If an action conflicts with one for the same link already on the list
then the action from the rule with the higher priority stays on the list
and the other is removed. If the priorities are the same then the orig-
inal action stays on the list.
– After all rules are evaluated, if the list is not empty then the new
actions are taken. If this causes the status of one or more links to
change then a new hydraulic solution is computed and the process
begins anew.
– If no status changes were called for, the action list is cleared and the
next rule time step is taken unless the normal hydraulic time step
has elapsed.

Time is advanced by the computed time step, new demands are found,
tank levels are adjusted based on the current flow solution, and link con-
trol rules are checked to determine which links change status.

A new set of iterations with equations are begun at the current set of flows.

5.4.1 Water Age and Source Tracing

Water age is the time spent by a parcel of water in the network. It provides a
simple, non-specific measure of the overall quality of delivered drinking water.
New water entering the network from reservoirs or source nodes enters with
age of zero. As this water moves through the pipe network it splits apart and
blends together with parcels of varying age at pipe junctions and storage
facilities. EPANET provides automatic modeling of water age. Internally, it

233
 EPANET

treats age as a reactive constituent whose growth follows zero-order kinetics

with a rate constant equal to 1 (i.e., each second the water becomes a sec-
ond older).

Source tracing tracks over time what percent of water reaching any node in
the network had its origin at a particular node. The source node can be any
node in the network, including storage nodes. Source tracing is a useful tool
for analyzing distribution systems drawing water from two or more different
raw water supplies. It can show to what degree water from a given source
blends with that from other sources, and how the spatial pattern of this blend-
ing changes over time. EPANET provides an automatic facility for performing
source tracing. The user need only specify which node is the source node.
Internally, EPANET treats this node as a constant source of a non-reacting
constituent that enters the network with a concentration of 100.

5.5 Water Quality Simulation Model

The governing equations for EPANET’s water quality module are based on
the principles of conservation of mass coupled with reaction kinetics. The fol-
lowing phenomenon are represented (Rossman et al., 1993; Rossman and
Boulos, 1996):

5.5.1 Advective Transport in Pipes

A dissolved substance will travel down the length of a pipe with the same
average velocity as the carrier fluid while at the same time reacting (either
growing or decaying) at some given rate. Longitudinal dispersion is usually
not an important transport mechanism under most operating conditions. This
means there is no intermixing of mass between adjacent parcels of water
traveling down a pipe. Advective transport within a pipe is represented with
the following equation:

C i C
= – ui i + r  C i  (5.33)
t x

where Ci = concentration (mass/volume) in pipe i as a function of distance x

and time t, ui = flow velocity (length/time) in pipe i, and r = rate of reaction
(mass/volume/time) as a function of concentration.

5.5.2 Mixing at Pipe Junctions

At junctions receiving inflow from two or more pipes, the mixing of fluid is
taken to be complete and instantaneous. Thus the concentration of a sub-

234 MIKE URBAN - © DHI

Water Quality Simulation Model

stance in water leaving the junction is simply the flow-weighted sum of the
concentrations from the inflow pipes. For a specific node k one can write:

 j  IkQCj x = Lj + Qk ext Ck ext

C i x = 0 = -----------------------------------------------------------------------
- (5.34)
 jIkQj + Qk ext
where i = link with flow leaving node k, Ik = set of links with flow into k, Lj =
length of link j, Qj = flow (volume/time) in link j, Qk,ext = external source flow
entering the network at node k, and Ck,ext = concentration of the external
flow entering at node k. The notation Ci|x=0 represents the concentration at
the start of link i, while Ci|x=L is the concentration at the end of the link

5.5.3 Mixing in Storage Facilities

It is convenient to assume that the contents of storage facilities (tanks and

reservoirs) are completely mixed. This is a reasonable assumption for many
tanks operating under fill-and-draw conditions providing that sufficient
momentum flux is imparted to the inflow (Rossman and Grayman, 1999).
Under completely mixed conditions the concentration throughout the tank is a
blend of the current contents and that of any entering water. At the same
time, the internal concentration could be changing due to reactions. The fol-
lowing equation expresses these phenomena:

 V C  =
t s s  iIsQi C i x = Li –  jOsQj C s + r  Cs  (5.35)

where Vs = volume in storage at time t, Cs = concentration within the storage

facility, Is = set of links providing flow into the facility, and Os = set of links
withdrawing flow from the facility.

5.5.4 Bulk Flow Reactions

While a substance moves down a pipe or resides in storage it can undergo

reaction with constituents in the water column. The rate of reaction can gen-
erally be described as a power function of concentration:

r  kC n (5.36)

where k = a reaction constant and n = the reaction order. When a limiting con-
centration exists on the ultimate growth or loss of a substance then the rate
expression becomes

235
 EPANET

for n > 0, Kb > 0

R  K b (C L  C )C ( n 1)
(5.37)

for n > 0, Kb < 0

R  K b (C  C L )C ( n 1)
(5.38)

where CL = the limiting concentration.

Some examples of different reaction rate expressions are:

Simple First-Order Decay (CL = 0, Kb < 0, n = 1)

R  K bC
(5.39)

The decay of many substances, such as chlorine, can be modeled ade-

quately as a simple first-order reaction.

First-Order Saturation Growth (CL > 0, Kb > 0, n = 1):

R  K b (C L  C )
(5.40)

This model can be applied to the growth of disinfection by-products, such as

trihalomethanes, where the ultimate formation of by-product (CL) is limited by
the amount of reactive precursor present.

Two-Component, Second Order Decay (CL ¹ 0, Kb < 0, n = 2):

R  K bC (C  C L )
(5.41)

This model assumes that substance A reacts with substance B in some

unknown ratio to produce a product P. The rate of disappearance of A is pro-
portional to the product of A and B remaining. CL can be either positive or
negative, depending on whether either component A or B is in excess,
respectively. Clark (1998) has had success in applying this model to chlorine
decay data that did not conform to the simple first-order model.

Michaelis-Menton Decay Kinetics (CL > 0, Kb < 0, n < 0):

Kb C
R = ----------------
- (5.42)
CL – C

236 MIKE URBAN - © DHI

Water Quality Simulation Model

As a special case, when a negative reaction order n is specified, EPANET will

utilize the Michaelis-Menton rate equation, shown above for a decay reaction.
(For growth reactions the denominator becomes CL + C.) This rate equation
is often used to describe enzyme-catalyzed reactions and microbial growth. It
produces first-order behavior at low concentrations and zero-order behavior
at higher concentrations. Note that for decay reactions, CL must be set higher
than the initial concentration present.

Koechling (1998) has applied Michaelis-Menton kinetics to model chlorine

decay in a number of different waters and found that both Kb and CL could be
related to the water's organic content and its ultraviolet absorbance as fol-
lows:

1,365  100UVA 
K b = – 0,32UVA – -------------------------- (5.43)
DOC

CL  4.98UVA  1.91DOC
(5.44)

where UVA = ultraviolet absorbance at 254 nm (1/cm) and DOC = dissolved

organic carbon concentration (mg/L).

Note: These expressions apply only for values of Kb and CL used with
Michaelis-Menton kinetics.

Zero-Order growth (CL = 0, Kb = 1, n = 0)

R = 1.0

This special case can be used to model water age, where with each unit of
time the "concentration" (i.e., age) increases by one unit.

The relationship between the bulk rate constant seen at one temperature (T1)
to that at another temperature (T2) is often expressed using a van't Hoff -
Arrehnius equation of the form:

K b 2  K b1 T 2T 1
(5.45)

where q is a constant. In one investigation for chlorine, q was estimated to be

1.1 when T1 was 20 deg. C (Koechling, 1998).

5.5.5 Pipe Wall Reactions

While flowing through pipes, dissolved substances can be transported to the

pipe wall and react with material such as corrosion products or biofilm that
are on or close to the wall. The amount of wall area available for reaction and
the rate of mass transfer between the bulk fluid and the wall will also influ-

237
 EPANET

ence the overall rate of this reaction. The surface area per unit volume, which
for a pipe equals 2 divided by the radius, determines the former factor. The
latter factor can be represented by a mass transfer coefficient whose value
depends on the molecular diffusivity of the reactive species and on the Reyn-
olds number of the flow (Rossman et. al, 1994). For first-order kinetics, the
rate of a pipe wall reaction can be expressed as:

2k w k f C
r = -------------------------
- (5.46)
R  kw + kf 

where kw = wall reaction rate constant (length/time), kf = mass transfer coef-

ficient (length/time), and R = pipe radius. For zero-order kinetics the reaction
rate cannot be any higher than the rate of mass transfer, so

r  MIN (k w , k f C )(2 / R)
(5.47)

where kw now has units of mass/area/time.

Mass transfer coefficients are usually expressed in terms of a dimensionless

Sherwood number (Sh):

D
k f  Sh
d (5.48)

in which D = the molecular diffusivity of the species being transported

(length2/time) and d = pipe diameter. In fully developed laminar flow, the
average Sherwood number along the length of a pipe can be expressed as

0.0668(d / L) Re Sc
Sh  3.65 
1  0.04(d / L) Re Sc 
2/3
(5.49)

in which Re = Reynolds number and Sc = Schmidt number (kinematic viscos-

ity of water divided by the diffusivity of the chemical) (Edwards et.al, 1976).
For turbulent flow the empirical correlation of Notter and Sleicher (1971) can
be used:

Sh  0.0149 Re 0.88 Sc1/ 3 (5.50)

5.5.6 Lagrangian Transport Algorithm

EPANET's water quality simulator uses a Lagrangian time-based approach to

track the fate of discrete parcels of water as they move along pipes and mix

238 MIKE URBAN - © DHI

Water Quality Simulation Model

together at junctions between fixed-length time steps (Liou and Kroon, 1987).
These water quality time steps are typically much shorter than the hydraulic
time step (e.g., minutes rather than hours) to accommodate the short times of
travel that can occur within pipes. As time progresses, the size of the most
upstream segment in a pipe increases as water enters the pipe while an
equal loss in size of the most downstream segment occurs as water leaves
the link. The size of the segments in between these remains unchanged.

The following steps occur at the end of each such time step:

1. The water quality in each segment is updated to reflect any reaction that
may have occurred over the time step.

2. The water from the leading segments of pipes with flow into each junc-
tion is blended together to compute a new water quality value at the junc-
tion. The volume contributed from each segment equals the product of its
pipe's flow rate and the time step. If this volume exceeds that of the seg-
ment then the segment is destroyed and the next one in line behind it
begins to contribute its volume.

3. Contributions from outside sources are added to the quality values at the
junctions. The quality in storage tanks is updated depending on the
method used to model mixing in the tank (see below).

4. New segments are created in pipes with flow out of each junction, reser-
voir, and tank. The segment volume equals the product of the pipe flow
and the time step. The segment's water quality equals the new quality
value computed for the node.

To cut down on the number of segments, Step 4 is only carried out if the new
node quality differs by a user-specified tolerance from that of the last seg-
ment in the outflow pipe. If the difference in quality is below the tolerance then
the size of the current last segment in the outflow pipe is simply increased by
the volume flowing into the pipe over the time step.

This process is then repeated for the next water-quality time step. At the start
of the next hydraulic time step the order of segments in any links that experi-
ence a flow reversal is switched. Initially each pipe in the network consists of
a single segment whose quality equals the initial quality assigned to the
downstream node.

239
 EPANET

Figure 5.15 Behavior of Segments in the Lagrangian Solution Method

5.5.7 Tank Mixing Models

EPANET can use four different types of models to characterize mixing within
storage tanks: Complete Mixing, Two-Compartment Mixing, First In First Out
(FIFO) Plug Flow and Last In Fist Out (LIFO) Plug Flow. Different models can
be used with different tanks within a network.

Complete Mixing assumes that all water that enters the tank is instantane-
ously and completely mixed with the water already in the tank.

240 MIKE URBAN - © DHI

Water Quality Simulation Model

Figure 5.16 Complete mixed tank mixing model schematic

It is the simplest mixing behavior to assume, requires no extra parameters to

describe it, and seems to apply well to a large number of facilities that oper-
ate in a fill and draw fashion.

Two-Compartment Mixing divides the available storage volume into two com-
partments.

Figure 5.17 Two compartment tank mixing model schematic

The inlet/outlet pipes of the tank are assumed to be in the first compartment.
New water that enters the tank is mixed with the first compartment. If this
compartment is full, then it sends the overflow water to the second compart-
ment where it completely mixes with the water stored there. When water
leaves the tank, it exits from the first compartment, which if full, receives an

241
 EPANET

equal volume of water from the second compartment to make up the differ-
ence. The first compartment is capable of simulating short circulating
between inflow and outflow, while the second compartment represents dead
zones. The user must define a single parameter, which is the fraction of the
volume of the tank to be dedicated to the first compartment.

FIFO Plug Flow assumes that there is no mixing during the residence time of
the water in the tank.

Figure 5.18 FIFO tank mixing model schematic

Water parcels move through the tank in a segregated fashion, where the first
parcel to enter the tank is also the first to leave. Physically speaking, this
model best represents baffled tanks with constant inflow and outflow. There
are no additional parameters to describe in this model.

LIFO Plug Flow assumes that there is no mixing between parcels of water
that enter the tank.

242 MIKE URBAN - © DHI

Model Calibration

Figure 5.19 LIFO tank mixing model schematic

However, in contrast to the FIFO Plug Flow model, the parcels stack up one
on top of another, where water enters and leaves from the bottom. In this
fashion, the first parcel to enter the tank is also the last to leave. Physically
speaking this type of model might apply to a tall narrow standpipe with an
inflow/outflow pipe at the bottom and a low momentum inflow. It requires no
additional parameters be provided

5.6 Model Calibration

Once a water distribution model has been developed, it must be calibrated so
that it accurately represents the actual working real-life water distribution net-
work under a variety of conditions. This involves making minor adjustments to
the input data so that the model accurately simulates both the pressure and
flow rates in the system. Note that both the pressure and flow rates need to
be matched together, since pressure and flow rates are dependent upon
each other. Therefore, matching only pressures or flow rates is not sufficient
enough. Also, ideally the model should be calibrated over an extended period
of time, such as a time range for the maximum day—not just the maximum
and minimum hour for the maximum day.

This section discusses the concepts and steps involved in calibrating a water
distribution model.

5.6.1 Accuracy Concerns

The computed pressure and measured field (actual) pressure will not exactly
match for every node contained within the network system—there will be dif-
ferences within the system. However, the maximum amount of these differ-
ences needs to be considered when performing a model calibration.

243
 EPANET

Generally there are three types of measurements used to determine the

degree of accuracy reached when calibrating a model.

Pressure Differential Method

The simplest and easiest method in determining how accurate a simulated
network model is to the actual network is to look at the maximum pressure
differential between simulated and actual node pressures. For example, in a
large system of several hundred or thousands of nodes, the model is gener-
ally considered calibrated if the pressure difference between the simulated
and actual node pressures is within 5 to 10 psi of each other. For smaller net-
work systems, with maybe a 100 or less nodes, this pressure differential
would likely be considered too high.

Percent Pressure Differential Method

A more precise determination of accuracy is to look at the pressure differen-
tial as a percentage. For example, at a base pressure of 100 psi a 4 psi pres-
sure differential between simulated and actual pressures represents a 4
percent differential, but at a base pressure of 40 psi this represents a 10 per-
cent differential. Therefore, looking at the pressure differential as a percent-
age is a more precise measure of accuracy.

Head Loss Differential Method

The most precise determination of accuracy is to examine the difference in
pipe head loss between simulated and actual. Head loss is more sensitive to
calibration errors than pressure. As was discussed previously, it is better to
look at the percent differential in head loss rather than the actual head loss
values.

Although the head loss method is better than the other methods in determin-
ing accuracy, from a practical standpoint the pressure differential method and
the percent pressure differential method are far easier to use and understand.
Therefore, since the main point is to establish a common unit of measure to
determine the degree of accuracy reached when calibrating a model, any of
these methods can be used.

During the calibration process there should be a point where the modeler has
decided that enough time has been spent on calibrating the model to the
actual network and he should move on to analyzing the pipe network system.

5.6.2 Reasons for Calibrating a Model

Calibration is an important process that should be performed. A calibrated

model establishes model credibility, sets a benchmark, can be used to predict
potential problems, establish understanding of the system operation, and
uncover errors with the system.

244 MIKE URBAN - © DHI

Model Calibration

Model Credibility
Through the process of calibrating a model, credibility of the model is estab-
lished. The data and modeling assumptions of an uncalibrated model are
unlikely to match the actual system. On the other hand, a calibrated model is
known to simulate a network system for a range of operating conditions. Its
input data has been examined and adjusted to insure that the model can be
used as an accurate, predictive tool. Hence, use of an uncalibrated model is
not good engineering practice, since it will most likely lead to inaccurate
model results and poor engineering decisions based upon these results.

Benchmark
Once a network model has been calibrated to a known range of operating
conditions, it can be used as a benchmark. Pressure and flow rates com-
puted by this model become the benchmark from which pressure and flow
rates computed by subsequent, modified models can be compared. The dif-
ferences between the two models can then be used to analyze the changes
brought about in the modified system.

Predict Potential Problems

A calibrated model can be used to predict any potential problems due to
changes in the system operation. A modified model, for example, might
include additional valves or pumps to increase the number of pressure zones
in a system. These changes can then be compared with the benchmark cali-
brated model to see what changes in pressure and flow rates have occurred.

Understand System Operations

In the process of calibrating a model, by collecting and analyzing data used to
define the model and studying the existing network system, the engineer
gains valuable insight and knowledge of the network system. The engineer
needs to simulate many of the same system settings and operations that a
network operator makes in order to calibrate the model to the actual opera-
tion of the network system. In addition, by analyzing the system operation,
possible improvements to system operation may be determined.

Uncover Errors
Calibration requires collecting data about the system. Many times during the
process of calibrating a model, questionable model results are investigated.
Inconsistencies between the modeled results and the actual field conditions
are examined, with additional field data being collected and analyzed. Incor-
rect pipe diameters may be determined, or even incorrectly closed valves dis-
covered.

245
 EPANET

5.6.3 Calibration Model Data Requirements

The data requirements for performing a model calibration is made up with the
fact that some data is fixed and unchanging (e.g., pipe diameter, length, etc.),
some data is variable with time (e.g., demand patterns, pump rates, dis-
charge pressures, reservoir elevations, etc.), and some data is assumed
(e.g., pipe roughness values, consumption rates, etc.). Some data, such as
consumption, is measured but sometimes also assumed.

Fixed data is easily obtained or measured. Variable data is more difficult to

obtain and measure, but generally can be gathered from SCADA (Supervi-
sory Control And Data Acquisition) system information or manual chart and
graph systems. And, although it would be ideal to gather this information in
order to perform an extended period simulation, it is extremely difficult to do
for a large network system. This task is easily performed on a small system.
But, for a large system, it may only be possible to gather this data for a single
time period (i.e., maximum hour).

Simulation Type Considerations

Before collecting the data, it is important to decide whether a steady state or
an extended period calibration simulation is to be performed. A steady state
simulation is used to calibrate the maximum (peak) hour demand, whereas
an extended period simulation is needed to calibrate a maximum day
demand or when performing water quality modeling.

Complexity of the network system determines the degree of difficulty in cali-

brating the model. The more pipes, pumps, tanks, reservoirs, valves, pres-
sure zones, and demand patterns there are, the more difficult the calibration
process is and the less accurate the calibrated model will be.

The availability of data is also a limiting factor. For example, calibrating to the
maximum hour requires data for only one time period. Calibration should also
be performed for the minimum hour of the maximum day. Therefore, data
would be required for two time periods. Extending the calibration process to
an extended period simulation for the maximum day increases the data
requirements to 24 time periods.

What can be done to simplify the calibration process is to approach the cali-
bration process incrementally. To elaborate, first calibrate the model for the
maximum hour. This will provide a good understanding of the network sys-
tem, and can usually be accomplished with a reasonable degree of accu-
racy—even for large network systems. Then, the minimum hour can be
calibrated. Knowledge learned in calibrating the maximum hour can be used
to speed-up the calibration process for the minimum hour. Once both of these
simulations have been calibrated, calibration of the maximum day extended
period simulation can be performed if time permits and the data is readily
available.

246 MIKE URBAN - © DHI

Model Calibration

Data Acquisition
Calibration data should be compiled and organized so that it can be used in
an efficient manner. In addition, a map of the network system is essential to
properly calibrate the system. All pipes, pumps, valves, tanks, and reservoirs
should be identified, including nodal elevations, pressure zone boundaries,
and other important information.

SCADA data and graphs are useful since they contain data that will be refer-
enced when developing the calibrated model. Even though the maximum
hour is typically of primary interest, values for the entire day are also impor-
tant since they provide an overall understanding of the operation of the net-
work system. This information, for example, allows the user to compare the
discharge pressure value at a particular location on the system during the day
with the maximum hour value.

Operational Review
Operational rules must be determined for the network system for all major
system components. Discussions with system operators should be performed
to determine the operational rules and strategy behind how specific opera-
tions are performed. For example, under what conditions do system opera-
tors turn on a pump, close a control valve, or adjust a pressure-regulating
valve? Is water pressure at particular key locations in the network act as a
flag to turn on a particular pump? Are pumping schedules ever changed to try
to minimize power costs? Are all facilities currently available for use, or are
some facilities off-line for maintenance repair?

An operational review is essential in order to accurately calibrate the model

and to make recommendations to improve system operation. In addition, by
discussing and reviewing this information with system operators, a good
working relationship is generally established between the modeler and the
operators. And, the modeler and operators can teach each other about the
system. The modeler can then use this information toward planning and
improving the water distribution system so that the operators can then more
easily manage the system.

Recalibration Frequency
Whenever changes occur in the water distribution network, such as opera-
tional changes, network configuration, or increases in water consumption, the
degree of accuracy for the calibrated model is reduced. If these changes are
severe enough, the model will need to be recalibrated. In practice this means
that the model should be recalibrated whenever major new facilities are
added to the network system, a new record for maximum hour is set, or oper-
ational procedures change significantly. Therefore, recalibration may be nec-
essary every few years.

247
 EPANET

Many experts feel that calibration should be performed on an annual basis,

regardless of what changes have occurred to the network system. Many
times the primary reason to recalibrate the system is so that knowledge
learned about the network can be used to improve the system operation.
And, if the system changes are rather insignificant, then the recalibration pro-
cess can be quickly and easily performed. By updating the model calibration
annually, changes to the model are less extensive and are easier to imple-
ment in order to get the model back in tune with the actual network system.

5.6.4 Calibration Simulations

Once all of the necessary data has been gathered, the network model can be
constructed. It is important that the network model accurately represents the
physical layout of the system. Pipes and nodes must be accurately located in
the model. Pipe roughness values should be estimated—based upon the age
of the pipe.

Next, consumption values must be defined at the nodes. To save time initially,
MIKE URBAN WATER provides a method of globally applying the total net-
work demand to each of the individual network nodes using the Distributed
Demand dialog box. See the section titled Distributed Demands on - 59 for
more information on computing distributed demands. Also, operational data
must be defined for all pumps, valves, storage tanks, and reservoirs. Finally,
a network simulation can be performed with EPANET.

Initial Simulation
An initial simulation is performed to simply determine what the resulting pres-
sures and flow rates are in the pipe network. This simulation may be simpli-
fied by using single operational values rather than complete operational data.
At this stage of the network calibration, exact operational data is not required.
For example, a reservoir elevation could be entered as rounded to the near-
est whole number elevation, rather than to a tenth of a foot. Similarly, a single
value pump definition could be used rather than a pump curve for each of the
pumps defined in the network. As long as the simplified data is generally
accurate, a balanced run can be produced so that the pipe network and oper-
ational input data can be checked.

Comparing Model Output

After the initial analysis output has been checked and verified, the modeled
results can then be compared with actual measured field values. A table of
actual flow rates and pressures should be prepared at key locations in the
system so that comparisons with the computed analysis results can be
quickly performed. Table 5.4 and Table 5.5 illustrate a comparison of flow
rates and pressures between actual field measurements and computed val-
ues.

248 MIKE URBAN - © DHI

Model Calibration

Table 5.4 Comparison of actual field flow rates and computed flow rates

Location Actual Actual Com- Com- Actual Percent

Pipe Flow puted puted Differ- Difference
ID Rate Pipe Flow ence (%)
(mgd) ID Rate (mgd)
(mgd)

University 346 32 597 34 +6 +2

Avenue @
Hilldale
Shopping
Center

East John- 1067 42 234 40 -2 -5

son at
State Capi-
tol Building

Middleton 2734 11 112 12 +1 +8

Heights
Pump Sta-
tion

249
 EPANET

Table 5.5 Comparison of actual field pressures and computed pressures

Locations Actual Actual Computed Computed Actual Percent

Node Pressure Node Pressure Differ- Differ-
ID (psi) ID (psi) ence ence
(psi) (%)

Pump sta- 2934 55 132 60 +5 +8

tion dis-
charge
pressure
at
6300 Uni-
versity
Avenue

Pressure 1367 59 253 56 -3 -5

regulating
valve
down-
stream
pressure
at
6612 Min-
eral Point
Road

Control 4589 87 306 90 +3 +3

valve
down-
stream
pressure
at
State Capi-
tol Building

Note that the comparison process is best handled by comparing values on a

broad scale, and then working downward to a more detailed examination of
values on a more localized level. For example, examine the flow rates at the
network supply points first. The computed and actual flow rates from these
supply points should be compared. Next, flow rates and discharge pressures
at pump stations should be examined. Continue this narrowing of scope as
the calibration process progresses.

5.6.5 Model Adjustments

Once the differences between the computed and actual measured values can
be determined, adjustments to the model data can be performed to make the

250 MIKE URBAN - © DHI

Model Calibration

computed results match more closely to that of the actual data. When per-
forming comparisons, differences in flow rates and pressures should be
examined.

The computed flow rates and pressures may be lower or higher from the
measured values. Table 5.6 allows you to quickly determine the cause for
these differences—allowing you to adjust the model input to more closely
match the measured values.

Table 5.6 Causes for differences between computed and mesured flow rates and
pressures

Network Parameter Computed Flow Computed Flow

Rate and Rate and Pressure
Pressure Too Low Too High

Total System Demand High Low

Pipe Roughness Low High

Pump Lifts Low High

Pressure Regulating Valve Settings Low High

Reservoir Elevation Settings Low High

Flow Control Valves Not opened enough Opened too much

Individual Nodal Demand High Low

In adjusting the model input data, there are four areas where adjustments can
be considered. These input data adjustments are discussed in the following
sections.

Operational Data Adjustments

Possible operational data adjustments to increase or decrease flow might
include raising or lowering of reservoir elevations, or increasing or decreasing
pump lifts. However, if the original operational data has been collected to a
reasonable degree of accuracy, then major adjustments may not be valid.
However, minor adjustments can be made to this data since there is always
some degree of uncertainty.

Consumption Data Adjustments

Generally, a system-wide consumption value is known to reasonable accu-
racy due to pumping records. However, the distribution of this consumption
may not be known to the same degree of accuracy. Therefore, a redistribution
of nodal consumption may need to be performed.

251
 EPANET

Network Data Adjustments

There always exists the chance of input errors when defining a network
model. For example, pipe diameters or lengths could be defined in error. Or,
perhaps a pressure regulating valve was missed in the initial model definition.
Therefore, it is important to check over the defined network input data to
make certain errors were not made.

Pipe Roughness Adjustments

Pipe roughness values are typically estimated based upon material type and
age. Sometimes pump tests are conducted to more accurately estimate pipe
roughness, but this is typically for a single pipe. Therefore there exists a great
deal of uncertainty in the assigned pipe roughness values. Therefore, pipe
roughness values should be adjusted after the previously identified adjust-
ments have been performed.

Number of Simulation Iterations

After performing these input data adjustments, a new computational analysis
must be performed and the computed results compared again with the actual
measured results. Numerous iterations may be necessary until the desired
degree of accuracy is achieved.

For a small network system with fairly reliable input data, a calibrated model
might be achieved within 10 or fewer iterations. For more complicated net-
work systems, a calibrated model may require more than 100 iterations.

252 MIKE URBAN - © DHI

 APPENDIX A
Literature
Abbott, M.B. and Basco, DE.R., "Computational Fluid Dynamics".

253
 EPANET

Abbott, M.B., "Hydroinformatics - Information Technology and Aquatic Envi-

ronment".

Altman, T. and Boulos, P. F., Convergence of Newton Method in Nonlinear

Network Analysis, Journal Mathematical and Computer Modeling, Vol. 21,
No. 4, 1995, pp. 35-41.

Altman, T. and Boulos, P. F., On Solving Flow Constrained Networks: The

Inverse Problem, Journal Hydraulic Engineering ASCE, Vol. 121, No. 5,
1995, pp. 427-431.

AWWA Manual M32, Distribution Network Analysis for Water Utilities, First
Edition, American Water Works Association, Denver, CO, 1989.

Babbitt, H. E., Plumbing, Third Edition, McGraw-Hill, Inc., New York, NY,
1960.

Barlow, J. F., and Markland, E., Converting the Hazen-Williams Equation to

the Colebrook Function, Water Power & Dam Construction, September,
1975, pp. 331-333.

Bhave, P. R., Analysis of Flow in Water Distribution Networks, Technomic

Publishing Co., New York, NY, 1991.

Bhave, P. R., Calibrating Water Distribution Network Models, Journal of Envi-

ronmental Engineering, Vol. 114, No. 1, February, 1988, pp. 120-136.

Bhave, P. R., Rules for Solvability of Pipe Networks, Journal of Indian Water
Works Association, January-March, 1990, pp. 7-10.

Bhave, P. R., Unknown Pipe Characteristics in Hardy Cross Method of Net-

work Analysis, Journal of Indian Water Works Association, April-June, 1990.
pp. 133-135.

Bishop, M. M., Morgan, J. M., Cornwell, B., and Jamison, D. K., Improving the
Disinfection Detention Time of a Water Plant Clearwell, Journal of American
Water Works Association, Vol. 85, No. 3, 1993, pp. 68-75.

Boulos, P. F., Altman, T., and Liou, J. C. P., On the Solvability of Water Distri-
bution Networks with Unknown Pipe Characteristics, Journal Applied Mathe-
matical Modelling, Vol. 17, July, 1993, pp. 380-387.

Boulos, P. F., Altman, T., and Sadhal, K. S., Computer Modelling of Water
Quality in Large Multiple Source Networks, Journal Applied Mathematical
Modelling, Vol. 16, 1992, pp. 439-445.

Boulos, P. F., Altman, T., Bowcock, R. W., Dhingra, A. K., and Collevati, F., An
Explicit Algorithm for Modeling Distribution System Water Quality with Appli-
cations, 2nd BHR International Conference on Water Pipeline Systems, D. S.
Miller ed., Edinburgh, Scotland, 1994, pp. 405-423.

254 MIKE URBAN - © DHI

Model Calibration

Boulos, P. F., Altman, T., Jarrige, P. A., and Collevati, F., An Event-Driven
Method for Modeling Contaminant Propagation in Water Networks, Journal
Applied Mathematical Modelling, Vol. 18, 1994, pp. 84-92.

Boulos, P. F., Altman, T., Jarrige, P. A., and Collevati, F., Discrete Simulation
Approach for Network Water Quality Models, Journal Water Resources Plan-
ning and Management ASCE, Vol.121 No. 1, 1995, pp. 49-60.

Boulos, P. F., and Altman, T, A Graph-Theoretic Approach to Explicit Nonlin-

ear Pipe Network Optimization, Journal Applied Mathematical Modelling, Vol.
15, September, 1991, pp. 459-486.

Boulos, P. F., and Altman, T., An Explicit Approach for Modelling Closed
Pipes in Water Networks, Journal Applied Mathematical Modelling, Vol. 17,
August, 1993, pp. 437-443.

Boulos, P. F., and Altman, T., Explicit Calculation of Water Quality Prameters
in Pipe Distribution Networks, Journal Civil Engineering Systems, Vol. 10,
1993, pp. 187-206.

Boulos, P. F., and Ormsbee, L. E., Explicit Network Calibration for Multiple
Loading Conditions, Journal Civil Engineering Systems., Vol. 8, 1991, pp.
153-160.

Boulos, P. F., and Wood, D. J., An Explicit Algorithm for Calculating Operating
Parameters for Water Networks, Journal Civil Engineering System, Vol. 8,
1991, pp. 115-122.

Boulos, P. F., and Wood, D. J., Explicit Calculation of Pipe-Network Parame-

ters, Journal Hydraulic Engineering ASCE, Vol. 116, No. 11, 1990, pp. 1329-
1344.

Boulos, P. F., Rossman, L. A., and Vasconcelos, J. J., A Comparison of Meth-

ods for Modeling Water Quality in Distribution Systems, Proceedings of the
AWWA 1994 Annual Conference and Exposition, New York, NY, pp. 523-541.

Cesario, A. L., Modeling, Analysis, and Design of Water Distribution Systems,

American Water Works Association, Denver, Colorado, 1995.

Clark, R. M. and Colye, J. A., Measuring and Modeling Variations in Distribu-

tion System Water Quality, Journal of American Water Works Association,
Vol. 82, No. 8, 1990, pp. 46-53.

Clark, R. M., Grayman, W. M., Goodrich, J. A., Deininger, R. A., and Hess, A.
F., Field-Testing Distribution Water Quality Models, Journal of American
Water Works Association, Vol. 83, No. 7, 1991, pp. 67-75.

Clark, R. M., Smalley, G., Goodrich, J., Tull, R., Rossman, L. A., Vasconcelos,
J. J., and Boulos, P. F., Managing Water Quality in Distribution Systems: Sim-
ulating TTHM and Chlorine Residual Propagation, Journal Water SRT-AQUA,
Vol. 43, No. 4, 1994, pp. 182-191.

255
 EPANET

Cole, G. B., Leak Detection -- Two Methods That Work, AWWA Seminar on
Operational Techniques for Distribution Systems, Los Angeles, California,
1980.

Cormen, T. H., Leiserson, C. E., and Rivest, R. L., Introduction to Algorithms,

The MIT Press, Cambridge, MA, 1990.

Cross, H., Analysis of Flow in Networks of Conduits or Conductors, Bulletin

No. 286, University of Illinois Engineering Experimental Station, Urbana, Illi-
nois, 1936.

D. & L. George Ltd., Herts, England, 1991.

Daugherty, R.L. and Franzini, J.B. and Finnemore, E.J., "Fluid Mechanics
with Engineering Applications".

Dennis, J. E. and Schnabel, R. B., Numerical Methods for Unconstrained

Optimization and NonLinear Equations, Englewood Cliffs, New Jersey, Pren-
tice-Hall, 1983.

Duff, I. S., Erisman, A. M., and Reid, J. K., Direct Methods for Sparse Matri-
ces, Oxford University Press, Inc., New York, NY, 1992.

E. Cabrera and F. Martínez, "Water Supply Systems - State of the art and
future trends".

Edwards, D. K., Denny, V. E., and Mills, A. F., Transfer Processes, McGraw-
Hill, New York, NY, 1976.

Epp, R., and Fowler, A. G., Efficient Code for steady state Flows in Networks,
Journal of the Hydraulics Division, Proceedings of the American Society of
Civil Engineers, Vol. 96, No. HY1, January, 1970, pp. 43-56.

Fawell, J. K., and Miller, D. G., UK Drinking Water: A European Comparison,

Journal of the International Water and Environmental Management, Vol. 8,
February, 1994, pp. 76-83.

FEB Press, Ann Arbor, MI, U.S.A., 1967, 1983.

Fletcher, R. and Reeves, C. M., Function Minimization by Conjugate Gradi-

ents, The Computer Journal, Vol. 7., July, 1964, pp. 149-154.

Fletcher, R., Function Minimization without Evaluating Derivatives -A Review,

The Computer Journal, Vol. 8., April, 1965, pp. 33-41.

Fok, A.T.K, "Design Charts for Air Chamber on Pump Pipe Lines".

George, A., and Liu, J., Computer Solution of Large Sparse Positive Definite
Systems, Prentice-Hall, Inc., Englewood Cliffs, New Jersey, 1981.

256 MIKE URBAN - © DHI

Model Calibration

Gofman, E. and Rodeh, M., Loop Equations with Unknown Pipe Characteris-
tics, Journal of the Hydraulics Division, ASCE, Vol. 107, No. HY9, September,
1981, pp. 1047-1060.

Grayman, W. M., and Clark, R. M., Using Computer Models to Determine the
Effect of Storage on Water Quality, Journal of American Water Works Associ-
ation, Vol. 85, No. 7, 1993, pp. 67-77.

Grayman, W. M., Clark, R. M., and Males, R. M., Modeling Distribution-Sys-

tem water Quality Dynamic Approach, Journal of Water Resources Planning
and Management, Vol. 114, No. 3, May, 1988, pp. 295-311.

Gupta, R. S., Hydrology and Hydraulic Systems, Prentice Hall, 1989.

Hamam, Y. M., and Brameller, A., Hybrid Method for the Solution of Piping
Networks, Proceedings IEE, Vol. 118, No. 11, 1971, pp. 1607-1612.

Housing and Home Finance Agency, The Uniform Plumbing Code for Hous-
ing, Section 8.3.5, February, 1948.

Chaudhry, M.H., "Applied Hydraulics Transients".

Chow, V. T., Maidmant, D. R., and Mays, L. W., Applied Hydrology, McGraw-
Hill, Inc., New York, NY, 1988.

International Conference on Computer Application for Water Supply and Dis-

tribution, Leicester Polytechnic, UK, September 8-10, 1987.

Jeppson, R. W., Analysis of Flow in Pipe Networks, 5th Printing, Ann Arbor
Science, Ann Arbor, Michigan, 1981.

John Wiley and Sons, New York, 1989.

Journal of the Hydraulic Division, Vol. 104, No. HY9, September, 1978, pp.
1289-1303.

Journal of the Hydraulic Division, Vol. 107, No. HY10, October, 1981, pp.
1145-1161.

Lansey, K. E. and Mays, L. W., Optimization Model for Water Distribution Sys-
tem Design, Journal of the Hydraulic Engineering, Vol. 115, No. 10, October,
1989, pp. 1401-1418.

Lemieux, P. F., Efficient Algorithms for Distribution Networks, Journal of the

Hydraulics Division, ASCE, Vol. 98, No. HY11, November, 1972, pp. 43-56.

Lindeburg, M. R., Civil Engineering Reference Manual, Sixth Edition, Profes-

sional Publications, Inc., Belmont, CA, 1992.

Longman, Essex, 1989.

257
 EPANET

Male, J. W., and Walski, T. M., Water Distribution Systems: A Troubleshooting

Manual, Lewis Publishers, Inc., 1991.

Nayyar, M. L., Piping Handbook, Sixth Edition, McGraw-Hill, Inc., New York,
NY, 1992.

Nielsen, H. B., Methods for Analyzing Pipe Networks, Journal of the Hydrau-
lic Engineering, Vol. 115, February, 1989, pp. 139-157.

Nielsen, L. S., Standard Plumping Engineering Design, McGraw-Hill, Inc.,

New York, NY, 1963.

Ohtmer, O., NonLinear Flow Analysis in Pipe Networks, International Journal

for Numerical Methods in Engineering, Vol. 19, 1983, pp. 373-392.

Ormsbee, L. E. and Wood, D. J., Hydraulic Design Algorithms for Pipe Net-
works, Journal of Hydraulic Engineering, Vol. 112, No. 12, December, 1986,
pp. 1195-1207.

Ortega, J. M. and Rheinbolt, W. C., Iterative Solution of Non-Linear Equations

in Several Variables, Academic Press, New York, NY, 1970.

Osiadacz, A. J., Simulation and Analysis of Gas Networks, E & F Spon Ltd.,
London, UK, 1991.

Osiadacz, A. J., Simulation and Analysis of Gas Networks, Gulf Publishing

Company, Book Division, Houston, TX, 1987.

Pierre, D. A., Optimization Theory with Application, Wiley, New York, NY,
1969.

Pitman, London, 1991.

Powell, M. J. D., An Efficient Method for Finding the Minimum of a Function of

Several Variables without Calculating Derivatives, The Computer Journal,
Vol. 7, July, 1964, pp. 155-162.

Press, W. H., Teukolsky, S. A., Vetterling, W. T., and Flannery, B. P., Numeri-
cal Recipes in C, Second Edition, Cambridge University, New York, NY, 1992.

Proceedings of the 1st International Conference on Hydroinformatics, Delft,

The Netherlands, September 1994.

Proceedings of the Conference on Hydraulics in Civil Engineering, Brisbane,

1994.

Proceedings of the XXV th IAHR Conference, Tokyo, Japan, August 1993.

Richardson, A. K., Estimating Water Storage Requirements, Lectures pre-

sented at Water Distribution System Analysis and Design Workshops at the
University of Wisconsin- Extension, University of Wisconsin-Madison CEE

258 MIKE URBAN - © DHI

Model Calibration

325 class handout, summer semester, 1991, Alan K. Richardson is with Black
& Veatch Consulting Engineers at Kansas City, Missouri.

Rossman, L. A., and Boulos, P. F., Numerical Methods for Modeling Water
Quality in Distribution Systems: A Comparison, Journal Water Resources
Planning and Management ASCE, Vol. 122, No. 2, 1996, pp 137-146.

Rossman, L. A., Boulos, P. F., and Altman, T., Discrete Volume-Element

Method for Network Water-Quality Models, Journal Water Resources Plan-
ning and Management ASCE, Vol. 119, No. 5, September/October, 1993, pp.
505-517.

Rossman, L. A., Clark, R. M., and Grayman, W. M., Modeling Chlorine Resid-
uals in Drinking-Water Distribution Systems, Journal Environmental Engi-
neering ASCE, Vol. 120, No. 4, 1994, pp. 803-820.

Rossman, L. A., EPANET -An Advanced Water Quality Modeling Package for
Distribution System, Proceedings of American Water Works Association 1993
Annual Conference.

Rossman, L. A., EPANET Users Manual, EPA-600/R-94/057, Environmental

Protection Agency, Risk Reduction Engineering Laboratory, Cincinnati, Ohio,
1994.

Second ed., Van Nostrand Reinhold, New York, NY, 1987.

Shames, I. H., Mechanics of Fluids, Third Edition, McGraw-Hill, Inc., New

York, NY, 1992.

Shamir, U., and Howard, D. D., Water Distribution Systems Analysis, Journal
of the Hydraulic Division, ASCE, Vol. 94, No. HY1, January, 1968, pp. 219-
234.

Schweitzer, P. A., Handbook of Valves, Industrial Press, Inc., New York, NY,
1972.

SI Metric Edition, Mc-Graw Hill Book Company, Singapore, 1989.

Streeter, V. L., and Wylie, E. B., Fluid Mechanics, McGraw-Hill, Inc., Seventh
Edition, 1979.

Thorley, A.R.D., "Fluid Transients in Pipeline Systems".

Todini, E. and Pilati, S., "A Gradient Method for the Analysis of Pipe Net-
works".

Todini, E., and Pilati, S., A Gradient Method for the Analysis of Pipe Net-
works, Proceeding of the International Conference on Computer Applications
for Water Supply and Distribution, Leicester Polytechnic, UK, September 8-
10, 1987.

259
 EPANET

Travers, K., The Mesh Method in Gas Network Analysis, Gas Journal,
November, 1967, pp. 167-177.

Tullis, J.P., "Hydraulics of Pipelines".

University of Kentucky, Lexington, KY, 1980.

Valencia, 19-23 October 1992, Computational Mechanics Publications,

Southampton, ISBN 1-85312-253-X.

Vanicek, S. and Verwey, A. and Abbott, M., "HYPRESS: An Exercise in

Object-Orientation for Water Hammer and Water Distribution Simulation in
Pipe Networks",

Verwey, A. and Yu, J.H., "A Space-Compact High-Order Scheme for Water
Hammer Simulations".

Verwey, A., "Linkage of Physical and Numerical Aspects of Models Applied in

Environmental Studies",

Walski, T. M., Analysis of Water Distribution Systems, Van Nostrand Rhein-

hold Company, Inc., New York, NY, 1984.

Walski, T. M., Gessler, J., and Sjostrom, J. W., Water Distribution Systems:
Simulation and Sizing, Lewis Publishers, Inc., 1990.

Warga, J., Determination of steady state Flows and Currents in a Network,

Proceedings of Instrument Society of America, Vol. 9., Pt. 5, September,
1954, Paper No. 54-43-4.

Warring, R. H., Handbook of Valves, Piping and Pipelines, First Edition, Trade
& Technical Press Ltd., England, 1982.

Watters, G.Z., "Analysis and Control of Unsteady Flow in Pipe Networks

Including Extended Period Simulations".

Webler, T., and Brown, H. S., Exposure to Tetrachloroethylene via Contami-

nated Drinking Water Pipes in Massachusetts: A Predictive Model, Vol. 48,
No. 5, September/October, 1993, pp. 293-297.

Wood, D. J., 1991, Comprehensive Computer Modeling of Pipe Distribution

Networks, KYPIPE Manual, Civil Engineering Software Center, College of
Engineering, University of Kentucky, Lexington, KY, 1991.

260 MIKE URBAN - © DHI

INDEX

261
 Index

Numerics Comparing Alternative Solutions . . . . 128

1xQ-H Point Pump Curve . . . . . . . . 27 Completely Mixed . . . . . . . . . . . .41
3xQ-H Point Pump Curve . . . . . . . . 27 Computational Grid . . . . . . . . . . . 185
Computational Parameters . . . . . . . 182
A Concentration . . . . . . . . . . . . . .93
Accumulated Street Demands-Polygons . Concentration Source . . . . . . . . . .94
123 Condition . . . . . . . . . . . . . . .54, 75
Accuracy . . . . . . . . . . . . . . . . 50 Constant Power . . . . . . . . . . . . .27
Accuracy Concerns . . . . . . . . . . . 243 Construction Year . . . . . . . . . . . .20
Additional Components . . . . . . . . . 196 Consumption Data . . . . . . . . 115, 251
Advective Transport in Pipes . . . . . . 234 Control Editor . . . . . . . . . . . . . . .74
Aggregate Demands . . . . . . . . . . 68 Control Level . . . . . . . . . . . . . . .76
Aggregation . . . . . . . . . . . . . . . 68 Control Node . . . . . . . . . . . . . . .32
Air-Chamber . . . . . . . . . . . . . . 197 Control Pressure . . . . . . . . . . . . .32
Air-Chamber Editor . . . . . . . . . . . 44 Control Rules . . . . . . . . . . . . . . 185
Air-CHamber ID . . . . . . . . . . . . . 45 Control Settings . . . . . . . . . . . . 151
Air-Valve . . . . . . . . . . . . . . 16, 201 Control Type . . . . . . . . . . . . . . .32
Analysis . . . . . . . . . . . . . . . . . 125 Control Valve . . . . . . . . . . . . . . 145
Analysis Duration . . . . . . . . . . . . 87 Controlling Pumps . . . . . . . . . . . 217
Analysis Methods . . . . . . . . . . . . 206 Controls . . . . . . . . . . . . . . . . .73
Animation . . . . . . . . . . . . . . . . 132 Convergence . . . . . . . . . . . . . . .50
Architecture of RTC Systems . . . . . . 145 Counted Lots . . . . . . . . . . . . . . 123
Asset ID . . . . . . . . . 12, 19, 26, 35, 40 Critical Node Pressure . . . . . . . . . 137
Atmospheric Pressure . . . . . . . . . 51 Curve and Relations Editor . . . . . . . .56
Available Operational Controls . . . . . 77 Cyclic Patterns . . . . . . . . . . . . . .86

B D
Base Elevation . . . . . . . . . . . 42, 47 Darcy-Weisbach . . . . . . . . . . . . 139
Benchmark . . . . . . . . . . . . . . . 245 Data Acquisition . . . . . . . . . . . . 247
Boundary Conditions . . . . . . . . 58, 182 Data Source . . . . . . 13, 19, 26, 35, 41
Boundary Curve . . . . . . . . . . . . 59 Defining Curved Pipes . . . . . . . . . .24
Boundary Type . . . . . . . . . . . . . 59 Defining the Model . . . . . . . . . . . 111
Branches . . . . . . . . . . . . . . . . 184 Definition of Network Layout . . . . . . 183
Browser . . . . . . . . . . . . . . . . . 130 Demand . . . . . .52, 65, 70, 71, 72, 129
Browsing Results . . . . . . . . . . . . 138 Demand Allocation . . . . . . . . . 64, 124
Bulk Flow Reactions . . . . . . . . . . 235 Demand Category . . . . . . . 15, 66, 70
Bulk Reaction Order . . . . . . . . . . 96 Demand Coefficient . . . . . . 14, 21, 70
Demand Description . . . . . . . . . . .15
C Demand Editing . . . . . . . . . . . . 124
Calculation . . . . . . . . . . . . . . . 171 Demand Pattern ID . . . . . . . . . .15, 70
Calibration Model Data Requirements . 246 Demand Processing . . . . . . . . . . 119
Calibration Simulations . . . . . . . . . 248 Demand Scenarios . . . . . . . . . . . 124
Category . . . . . . . . . . . . . . 71, 84 Demand Statistics . . . . . . . . . . . .71
Category Type . . . . . . . . . . . . . 72 Design . . . . . . . . . . . . . . . . . 208
CHeck Valve . . . . . . . . . . . . . . 22 Design Fire Flow . . . . . . . . . . . . 137
Check Valves . . . . . . . . . . . . . . 38 Design Fire Pressure . . . . . . . . . . 136
CHemical Concentrations . . . . . . . . 90 Design Flow . . . . . . . . . . . . . . .27
Closed . . . . . . . . . . . . . 23, 29, 37 Design Head . . . . . . . . . . . . . . .27
Developing Pipe Demand Coefficients . 122

262 MIKE URBAN - © DHI

Index

Diagrams . . . . . . . . . . . . . . . 175 G
Diameter . . . . . . . . . . . 20, 37, 43, 47 General Purpose Valve . . . . 36, 115, 221
Distributed Demands . . . . . . . 59, 119 Genetic Algorithms Calibration . . . . . 143
Diurnal Curves and Patterns . . . . . . 88 Geocoding . . . . . . . . . . . . . . . 66
Diurnal Pattern . . . . . . . . . . . . . 86 Getting Started . . . . . . . . . . . . . 111
Diurnal Profiles . . . . . . . . . . . . . 83 Global Bulk Reaction Rate Coefficient . 95
Dual-Acting Valve CHaracteistics Curve 17 Global Editing . . . . . . . . . . . . . . 125
Dual-Acting Valve Characteristics . . . . 56 Global Pipe Wall Reaction Rate Coefficient
95
E GPV . . . . . . . . . . . . . . . . . . . 221
Effect of Pressure Reducing Valves . . 129 Graph . . . . . . . . . . . . . . . . . . 85
Efficiency . . . . . . . . . . . . . . . . 53 Graphical Input . . . . . . . . . . . . . 117
Efficiency Curve . . . . . . . . . . . . . 30 Graphical Placement 17, 23, 33, 38, 44, 47
Elevation . . . . . . . . . . . . . . . . 14 Gravity . . . . . . . . . . . . . . . . . 51
Emitter . . . . . . . . . . . . . . . 16, 114 Grid Points . . . . . . . . . . . . . . . 185
Emitter Exponent . . . . . . . . . . . . 50 Groundwater Wells . . . . . . . . . . . 223
Energy . . . . . . . . . . . . . . . . 53, 54 Group Definition . . . . . . . . . . . . . 142
Energy Editor . . . . . . . . . . . . . . 53
Energy Price . . . . . . . . . . . . . . . 30 H
Energy Price Pattern . . . . . . . . . . 30 Hazen-Williams . . . . . . . . . . . . . 139
Engineering Tables . . . . . . . . . . . 98 Head Loss Differential Method . . . . . 244
Entering Data . . . . . . . . . . . . . 116 Headlosses . . . . . . . . . . . . . . . 49
EPANET . . . . . . . . . . . . . . 108, 205 Hiding and Showing Graphical Data . . 119
EPANET Analysis Results . . . . . . 129 High End Flow . . . . . . . . . . . . . 28
EPANET Computer Program . . . . . 205 High End Head . . . . . . . . . . . . . 28
EPANET Input Data File . . . . . . . 205 History . . . . . . . . . . . . . . . . . 206
Equivalent Person EP . . . . . . . . . . 66 Hydraulic Simulation Model . . . . . . . 226
Equivalent Tenement ET . . . . . . . . 66 Hydraulic Structures . . . . . . . 180, 185
Estate Height . . . . . . . . . . . . . . 14 Hydraulic Time Step . . . . . . . . . . 87
Evolutionary Algorithms . . . . . . . . 139
Executing the Analysis . . . . . . . . 125 I
Extended Period Data . . . . . . . . . 115 Importing Graphical Data . . . . . . . . 118
Extended Period Hydraulics . . . . 49, 207 Inactive Volume . . . . . . . . . . . . . 43
Extended Period Simulations . . . . . . 73 Initial Conditions . . . . . . . . . . . . 182
Extended Period Water Quality . . . . . 49 Initial Level . . . . . . . . . . . . . 43, 47
Extended Rule-Based Controls . . 110, 157 Initial Simulation . . . . . . . . . . . . 248
Initial Water Quality . . . . . . . . . . . 91
F Input Data Requirements . . . . . . . . 113
FCV . . . . . . . . . . . . . . . . . . 221 Interactive Data Entry . . . . . . . . . . 116
FIFO . . . . . . . . . . . . . . . . . 41, 42 Introduction . . . . . . . . . . . . . . . 141
Fire Flow Analysis . . . . . . . . . 109, 135
Fire Flow Data . . . . . . . . . . . . . 116 J
Fixed HGL . . . . . . . . . . . . . . . . 42 Junction . . . . . . . . . . . . . . . . . 12
Flow Coefficient . . . . . . . . . . . . . 16 Junction Node Data . . . . . . . . . . . 114
Flow Control Valve . . . 36, 115, 145, 221 Junction Node Demands . . . . . . . . 185
Flow Direction . . . . . . . . . . 23, 33, 38 Junction Node Pressure Based Controls 77
Flow Paced Booster Source . . . . . . . 94
Forward Tracking . . . . . . . . . . . 100

263
 Index

L Operational Review . . . . . . . . . . 247

Lagrangian Transport Algorithm . . . . 238 Operational Schedule . . . . . . . .31, 37
Length . . . . . . . . . . . . . 20, 43, 47 Operational Schedule Type . . . . . . .30
LIFO . . . . . . . . . . . . . . . . 41, 42 Owner . . . . . . . . . . . . . . . . . .66
Link Type . . . . . . . . . . . . . . . . 75
Links . . . . . . . . . . . . . . . . . . 55 P
Looped Network Solution Algorithm . . 178 Parallel Pipes . . . . . . . . . . . . . . 213
Loss Coefficient . . . . . . . . . . . 21, 37 Pattern . . . . . . . . . . . . . . . . . .53
Pattern Editor . . . . . . . . . . . . . . .83
M Pattern ID . . . . . . . . . . . . . . . .84
Manning . . . . . . . . . . . . . . . . 139 Pattern Time Step . . . . . . . . . . . .87
Mass Booster Source . . . . . . . . . . 94 Patterns . . . . . . . . . . . . . . . . .73
Material . . . . . . . . . . . . . . . . . 20 PBV . . . . . . . . . . . . . . . . . . 221
Maximum Flow . . . . . . . . . . . . . 28 PID . . . . . . . . . . . . . . . . . 56, 147
Maximum Level . . . . . . . . . . . 43, 47 PID Control . . . . . . . . . . . . 109, 145
Messages . . . . . . . . . . . . . . . . 54 Pipe Data . . . . . . . . . . . . . . . . 114
Method of Pipe Lengths 21, 22, 59, 61, 120 Pipe Demand Coefficients . . . . . 59, 120
Method of Reduced Pipe Lengths . 61, 120 Pipe Diameter . . . . . . . . . . . . . .68
Method of Two Coefficients . 21, 22, 60, 61, Pipe Editor . . . . . . . . . . . . . . . .18
120 Pipe Flow Module . . . . . . . . . . . 108
MIKE NET . . . . . . . . . . . . . . . 107 Pipe Groupings . . . . . . . . . . . . . 142
Minimum Level . . . . . . . . . . . 43, 47 Pipe Reaction Rate Assignments . . . .96
Minor Losses . . . . . . . . . . . . . . 221 Pipe Roughness . . . . . . . . . 129, 252
Mixing . . . . . . . . . . . . . . . 234, 235 Pipe Roughness Calibration . . . 109, 139
Molecular Diffusivity . . . . . . . . 50, 91 Pipe Wall Reactions . . . . . . . . . . 237
Moment of Inertia . . . . . . . . . . . . 31 Pipes . . . . . . . . . . . . . . . . . . 211
Motor Torque . . . . . . . . . . . . 31, 56 Pipes with Check Valves . . . . . . . . .77
Multiple Controls . . . . . . . . . . . . 76 Planning . . . . . . . . . . . . . . . . 208
Multiple Demand . . . . . . . . . . 15, 69 Point Source Data . . . . . . . . . . . 115
Multiple Demand Editor . . . . . . . . . 48 Polytropic Expansion . . . . . . . . . . .46
Multipliers . . . . . . . . . . . . . . . . 85 Pressure and Flow Measurements . . . 143
Pressure Breaker Valve 36, 115, 145, 221
N Pressure Differential Method . . . . . . 244
Network . . . . . . . . . . . . . . . . . 210 Pressure Normal . . . . . . . . . . . . .21
Network Component Editor . . . . . . . 11 Pressure Reducing Valve 36, 115, 145, 220
Network Data . . . . . . . . . . . 113, 252 Pressure Sustaining Valve . . 36, 115, 145,
Network Demand . . . . . . . . . . . . 59 221
Node Demand Coefficients . . . . . . . 62 Pressure Zone . . . . . . . . . . . . . .71
Node Demand Multiplier . . . . . . . . 137 Pressure Zone Editor . . . . . . . . . . .52
Node Pressure Zone ID . . . . . . 60, 62 Pressure Zone ID 13, 20, 27, 36, 41, 46, 52,
Node Type . . . . . . . . . . . 58, 93, 98 . . . . . . . . . . . . . . . . . . . . . .65
Nodes . . . . . . . . . . . . . . . 184, 222 Price . . . . . . . . . . . . . . . . . . .53
Number of Simulation Iterations . . . . 252 Profile Plots . . . . . . . . . . . . . . 131
Numerical Scheme and Algorithm . . . 176 Project Data . . . . . . . . . . . . . . 113
Project Options . . . . . . . . . . . . . .48
O Project Specification Data . . . . . . . 113
Operation . . . . . . . . . . . . . . . . 208 Project Type . . . . . . . . . . . . . . .49
Operational Data Adjustments . . . . . 251 Properties . . . . . . . . . . . . . . . .49
PRV . . . . . . . . . . . . . . . . . . 220

264 MIKE URBAN - © DHI

Index

PSV . . . . . . . . . . . . . . . . . . 221 Scenario Manager . . . . . . . . . . . 98

Pump . . . . . . . . . . . . 114, 145, 190 Select Time Level . . . . . . . . . . . . 138
Pump Editor . . . . . . . . . . . . . . . 25 Send Data . . . . . . . . . . . . . . . . 98
Pump Efficiency . . . . . . . . . . . . . 56 Set Point Settings . . . . . . . . . . . . 152
Pump Energy Settings . . . . . . . . . 29 Setpoint Booster Source . . . . . . . . 94
Pump ID . . . . . . . . . . . . . . . . . 26 Setting . . . . . . . . . . . . . . . 36, 75
Pump Operational Schedule . . . . . . 56 Shutoff Head . . . . . . . . . . . . . . 27
Pump Q-H Curve . . . . . . . . . . . . 56 Simple Control . . . . . . . . . . . . . 74
Pump Speed . . . . . . . . . . . . . . 32 Simulate Fire Hydrant . . . . . . . . . . 137
Pump Start-Up Time . . . . . . . . . . 31 Simulation Menu . . . . . . . . . . . . 111
Pump State . . . . . . . . . . . . . . . 29 Simulation Type Considerations . . . . 246
Pump Torque . . . . . . . . . . . . 31, 56 Skeletonization . . . . . . . . . . . . . 208
Pump Trip-Off Time . . . . . . . . . . . 31 Snap Tolerance . . . . . . . . . . . . . 24
Pump Type . . . . . . . . . . . . . . . 27 Source Tracing . . . . . . . . .90, 97, 233
Pump Variable Speed . . . . . . . . . . 31 Source Type . . . . . . . . . . . . . . 94
Pumping Rate Limits . . . . . . . . . 216 Specific Curves Data . . . . . . . . . . 195
Pumping Scenarios . . . . . . . . . . 129 Specific Gravity . . . . . . . . . . . 49, 91
Pumps . . . . . . . . . . . . . . . . . 214 Specific Pipe Data . . . . . . . . . . . 185
Specific Project Options Settings . . . . 194
Q Specific Pump Data . . . . . . . . . . . 185
Q-H Pump Curve . . . . . . . . . . . . 28 Specific Time Settings . . . . . . . . . 194
Quality . . . . . . . . . . . . . . . . . . 92 Start Clock Time . . . . . . . . . . . . 88
Quality Menu . . . . . . . . . . . . . 111 State . . . . . . . . . . . . . . . . 14, 22
Quality Time Step . . . . . . . . . . . . 88 Statistics . . . . . . . . . . . . . . . . 88
Status . . . . 13, 20, 26, 35, 41, 46, 54, 73
R Steady State Hydraulics . . . . . .49, 206
Reaction Rate Editor . . . . . . . . . . 95 Storage Tank . . . . . . . . . 77, 114, 225
Recalibration Frequency . . . . . . . 247 Street Name . . . . . . . . . . . . . . 22
Redistribute . . . . . . . . . . . . . 72, 85 Summary . . . . . . . . . . . . . . . . 54
Reference ID . . . . . . . . . . . . . . 64 Supported Components . . . . . . . . . 196
Refresh . . . . . . . . . . . . . . . . . 72 Surface Elevation . . . . . . . . . . . . 14
Reinforced Concrete Pipe . . . . . . . 174
Relative Speed . . . . . . . . . . . . . 29 T
Report Critical Node . . . . . . . . . . 137 Tabular Results . . . . . . . . . . . . . 138
Report Editor . . . . . . . . . . . . . . 53 Tank Depth-Volume Curve . . . . . . . 56
Report Start Time . . . . . . . . . . . . 88 Tank Editor . . . . . . . . . . . . . . . 39
Report Time Step . . . . . . . . . . . . 88 Tank Mixing . . . . . . . . . . . . . . . 41
Reservoir Data . . . . . . . . . . . . 114 Tank Mixing Models . . . . . . . . . . . 240
Rigid Conduit . . . . . . . . . . . . . 173 Tank Size . . . . . . . . . . . . . . 43, 47
Rotational Pump Speed . . . . . . . . . 31 Tank Type . . . . . . . . . . . . . . . . 41
Roughness . . . . . . . . . . . . . . . 21 Tanks . . . . . . . . . . . . . . . . . . 197
Roughness Coefficient . . . . . . . . 113 Target . . . . . . . . . . . . . . . . . . 69
Roughness Correlation . . . . . . . . . 96 Target Demand . . . . . . . . . . . 62, 63
RTC . . . . . . . . . . . . . . . . 145, 147 TCV . . . . . . . . . . . . . . . . . . . 221
Rule Based Controls . . . . . . . . . . 78 Temperature . . . . . . . . . . . . . . 52
Running Water Hammer Module . . . 182 Thematic Mapping . . . . . . . . . . . 130
Thematic Maps . . . . . . . . . . . . . 138
S Theoretical Background . . . . . . . . . 172
Save . . . . . . . . . . . . . . . . . . . 76 Theta . . . . . . . . . . . . . . . . . . 51

265
 Index

Thick-Walled Elastic Conduit . . . . . . 173 Variable Speed Drive . . . . . . . . . . 110

Thin-Walled Elastic Conduit . . . . . . 173 Variable Speed Drive Pump . . . . . . 157
Throttle Control Valve . 36, 115, 145, 221 Variable Valves . . . . . . . . . . . . . .38
Time . . . . . . . . . . . . . . . . . . 87 Vented Air Chambers . . . . . . . . . 199
Time Based Controls . . . . . . . . . . 77 View Menu . . . . . . . . . . . . . . . 111
Time Editor . . . . . . . . . . . . . . . 87 Viewing Analysis Results . . . . . . . . 127
Time Patterns . . . . . . . . . . . . 94, 225 Viscosity . . . . . . . . . . . . . . .49, 91
Time Series Plot . . . . . . . . . . . . 131 Volume Curve . . . . . . . . . . . .43, 47
Time Step . . . . . . . . . . . . . . . . 84 VSD pump . . . . . . . . . . . . . . . 157
Time Units . . . . . . . . . . . . . . . 88
Times . . . . . . . . . . . . . . . . . . 73 W
Title . . . . . . . . . . . . . . . . . . . 49 Wall Reaction Order . . . . . . . . . . .96
Tools . . . . . . . . . . . . . . . . . . 98 Wall Speed . . . . . . . . . . . . . . . .20
Tools Menu . . . . . . . . . . . . . . . 111 Wall Thickness . . . . . . . . . . . . . .20
Total Network Water Demand . . . 60, 62 Warning . . . . . . . . . . . . . . . . .23
Tracking . . . . . . . . . . . . . . . . 100 Water Age . . . . . . . . . . . . . 90, 233
Transient Flow Analysis . . . . . . . . 49 Water Hammer . . . . . . . . 51, 109, 171
Transient H-Boundary . . . . . . . . . 56 Water Hammer Calculation . . . 171, 181
Transient Q-Boundary . . . . . . . . . 56 Water Hammer Data . . . . . . . . . . 116
Tunnels Through Solid Rock . . . . . . 174 Water Hammer Data Preparation . . . 181
Tutorial . . . . . . . . . . . . . . . . . 202 Water Hammer Model . . . . . . . . . 172
Two-Compartment mixing . . . . . . . 42 Water Hammer Result Presentation . . 181
Type . . . . . . . . . . . . . . . . 14, 71 Water Quality Analysis . . . . . 89, 90, 207
Water Quality Data . . . . . . . . . . . 115
U Water Quality Module . . . . . . . . . 108
Unbalanced System . . . . . . . . . . 50 Water Quality Parameters . . . . . . . 129
Uncover Errors . . . . . . . . . . . . . 245 Water Quality Simulation . . . . . . . . .89
Understand System Operations . . . . 245 Water Quality Simulation Model . . . . 234
Units . . . . . . . . . . . . . . . . . . 49 Water Quality Tolerance . . . . . . . . .50
Unsupported Component . . . . . . . . 196 Water Source Price . . . . . . . . . . .56
Wave Speed . . . . . . . . . . . . . . 172
V WD-Tools . . . . . . . . . . 1, 81, 109, 135
Valve Characteristics Cd . . . . . . . . 56 Width . . . . . . . . . . . . . . . . .43, 47
Valve Characteristics Kv . . . . . . . . 56
Valve Curve . . . . . . . . . . . . . . . 37 X
Valve Data . . . . . . . . . . . . 115, 191 X and Y Coordinate . . . . 13, 41, 46, 65
Valve Diameter . . . . . . . . . . . . . 17
Valve Editor . . . . . . . . . . . . . . . 34
Valve Head Loss . . . . . . . . . . . . 56
Valve ID . . . . . . . . . . . . . . . . . 35
Valve Operation Schedule . . . . . . . 56
Valve State . . . . . . . . . . . . . . . 37
Valve Type . . . . . . . . . . . . . . . 36
Valve Water Hammer . . . . . . . . . . 37
Valve Water Hammer Settings . . . . . 37
Valves . . . . . . . . . . . . . . . . . 220
Vapour Cavity Pressure . . . . . . . . 52
Variable . . . . . . . . . . . . . . . . . 55
Variable Pumps . . . . . . . . . . . . . 32

266 MIKE URBAN - © DHI