ICPTT 2012 © ASCE 2013 453

Building Water Distribution Network Hydraulic Model by Using WaterGEMS

Baiyi Jiang1, Fabei Zhang2, Jian Gao3 and Hongbin Zhao4

1 Professor, Municipal And Environmental Engineering Institute, Shenyang Jian Zhu

University; No.9, Hunnan Road, Shenyang 110000; Tel：13840267090; Email：
Downloaded from ascelibrary.org by New York University on 08/02/15. Copyright ASCE. For personal use only; all rights reserved.

842365314@qq.com
2 Master, Municipal And Environmental Engineering Institute, Shenyang Jian Zhu
University; No.9, Hunnan Road, Shenyang 110000; Email: 493112993@qq.com;
18722094467
3 Project Manager, Tianjin Sambo Water Technology Co., Ltd, Tianjin 300041; Tel:
13502634972; Email: gaojian@8mwater.com
4 Professor, Municipal Faculty, Harbin Institute of Technology; No.73, Huanghe
Road, Harbin 150090; Email: zhb169@126.com Tel: 13871509445

ABSTRACT

Dynamic pipe network hydraulic modeling is the realization of the municipal water
supply network to optimize an effective means of scheduling and science operations.
Pipe network modeling processes; the basic data collection, including flow and
pressure points SCADA data; acquisition and importing of model work, is the key to
establishing a precise model. GIS data in an urban pipeline network, for instance,
creates a dynamic water supply network model. A detailed analysis of the operation,
using WaterGEMS provides practical examples to establish a water supply network
model based on SCADA data.

KEYWORDS

WaterGEMS; Water Demand Patterns; SCADA; Optimal Operation; Calibration of

Pressure

Municipal Water Supply Pipe Network Hydraulic Model of Data Collection

Water supply systems strive to achieve automation and informatization, to improve

water quality, ensure water supply, enhance service and increase efficiency is an
important way to create a new and efficient enterprises. However, a municipal water
supply system is a complicated structure, which is composed by a large scale water
supply system with strong randomicity and multiple operation control targets.
Therefore, to establish a water supply network system of scientific management, a
water supply network micro-dynamic hydraulic model needs to be built. This can be

ICPTT 2012
 ICPTT 2012 © ASCE 2013 454

done using computer technology simulation of the actual network operation

condition, which is the need of water utilities for the development.
By computer simulation, the establishment of the pipe network hydraulic and water
quality models, and simulation of various system scenarios, a reliable water supply
plan for the water supply can be developed. The water supply network model for the
water supply network establishes the computer simulation, including the following
parts: junctions, pipes, valves, pumps, reservoirs, hydrants.
Downloaded from ascelibrary.org by New York University on 08/02/15. Copyright ASCE. For personal use only; all rights reserved.

Network data, water data and operation data are necessary for the water supply
network model. A GIS system is currently used by the city’s water company to
represent the water supply network.

Pipe network object properties. The simulated pipe network consists of many
components, stored in vector form. It contains a lot of information about pipes,
junctions, pumps, valves, etc.; namely, the properties of the components.
WaterGEMS can effectively simulate and manage the pipeline network properties,
the formation of pipes, junctions, pumps, valves, and other databases.
Junctions: ID / unique ID, position of X and Y, elevation and demand collection;
Pipes: ID / unique ID, pipes of start node / stop node, diameter, number of breaks,
status (closed or open), material, Hazen-Williams C, length (user defined);
Hydrants: ID / unique ID, position of X and Y, demand collection, elevation, hydrant
status (closed or open);
Pumps: ID / unique ID, downstream pipe, position of X and Y, elevation and pump
definition;
Reservoirs: ID / unique ID, position of X and Y, elevation, hydraulic grade pattern;
Valves: ID / unique ID, downstream pipe, elevation, installation year and valve type
(ball, butterfly, circular gate, globe, needle, user defined);
Other elements of the pipeline: ID / unique ID, position of X and Y, elevation, the
relevant parameters and so on.

The properties of the relevant parameters. Constant speed pump flow - lift curve:
The pump is put into use, due to the long-term wear and tear of running as well as
technological innovation and other reasons, resulting in actual pump characteristic
curve deviation from the sample curve. The pump characteristic curve can be defined
in constant power, design point, multiple point and other convenient ways.
Open degree of valve: Valve open state, the hydraulic characteristics of the valve
opening degree of the simulation to calculate the coefficient of resistance.
Water demand variation graph: In order to master the water law to better simulate the
situation of water, it is necessary to draw a water characteristic curve of big water
user.
Flow simulation node: The node demand consists of a few large users and leakage.
The demand of big water users is a constant, based on location and the measured

ICPTT 2012
 ICPTT 2012 © ASCE 2013 455

curve; small water uses are assigned to the entire pipe network by weights. The
missing part is calculated by the demand, considering the diameter of laying age,
stress and other factors according to different weights.
Bentley of WaterGEMS has universal data interface, can be directly connected with
other GIS systems, and the user can import data including ArcView, ArcInfo, the
latest ArcGIS, AutoCAD DWG, DXF files, Shapfile files and so on. It can also be
connected with any relational database, such as the Microsoft Access, dBase, Excel,
Downloaded from ascelibrary.org by New York University on 08/02/15. Copyright ASCE. For personal use only; all rights reserved.

Oracle, and SQL Server. In addition, it has built-in ODBC support, through which is
can be connected to any virtual database.
Here is the project example for the city of the ArcGIS data, obtained through sorting
the GIS documents which contain information shown in Table 1:

Table 1. GIS system information table

ID File Name Component Types
1 FITING.shp Junctions
2 HYDRANT.shp Hydrants
3 VALVE.shp Throttle Control Valves
4 PIPESECTIONMAIN.shp Trunk Mains Of Water Distribution
5 PIPESECTIONUSER.shp Branch Pipe Of Water Distribution

Model Building Process

WaterGEMS can leverage virtually any data source, just as been show in Figure 1.

Data CAD Databases & Geospatial

drawings spreadsheets data

Model

Figure 1. The type of data source

Preliminary data processing and checking. In order to ensure the accuracy of the

ICPTT 2012
 ICPTT 2012 © ASCE 2013 456

water supply pipe network model, WaterGEMS software and GIS data import
function (ModelBuilder) must be correctly applied; and examination of the
continuity of the network model must be verified. The next step in the
implementation of the project is checking and correction of existing GIS data,
primarily using the technology by checking of the orphaned notes. The method to
find disconnected nodes and the pipes in WaterGEMS is using the Network
Navigator menu, which can found the existence of nodes, pipe-connection errors, and
Downloaded from ascelibrary.org by New York University on 08/02/15. Copyright ASCE. For personal use only; all rights reserved.

the data connection status at the visualization level. This stage of work is emphasis
on the connection of the pipes and the nodes, checking the topology of the GIS
system, and checking the connectivity of the pipe network.
This phase of work focuses on the collation of pipe and node connection to complete
the GIS system data topology check; then to complete the inspection and data import
of the pipe network connectivity.

GIS data partition imported. According to the marketing division, divide the pipe
networks. This stage is to determine the pipe network partition boundaries work,
when combined with the consideration of water statistics and outlined division of
water. Complete the focus of the observation region and the layout of the observation
points.
City meter to regional partition is the basis of water division, provided by the
Division of Marketing zoning map, the pipe network is divided into 10 partitions
model. After the partition model is exported, the exported file contents include
pipeline information and node information in an MDB file.

Loaded water demand based on the XML interface. Water consumption variation
simulation is one of the microscope models of the water supply network, and is the
basis of dynamic simulation. Urban water consumption is the dynamic, random,
multi-variable combined effect of many factors. It is difficult to find explicit
expressions to describe. In this paper, the week of hourly water consumption data
when submitted to cluster analysis, was divided into 13 categories of major water
variation, as shown in Figure 2. The DBF database files, Excel files, can be written in
the XML files by using the software developed in this subject. And the XML files
can be re-used by the Engineering Libraries module in WaterGEMS.

ICPTT 2012
 ICPTT 2012 © ASCE 2013 457
Downloaded from ascelibrary.org by New York University on 08/02/15. Copyright ASCE. For personal use only; all rights reserved.

Figure 2. The XML interface of WaterGEMS in Engineering Library module

The data of the water-using charges can be imported by the WaterGEMS of the
LoadBuilder module. And then the average of the water demand data can be loaded.
This part of the work is dependent upon completion of the static pipe network
hydraulics calculations, and integrity of the testing and calibration model.
Dynamic simulation process according to the user node 168 hours the amount of
water changes, as well as pump station pumps 168 hours continuously to the water
changes the hydraulic calculation, node pressure, pipe flow, velocity, and changes in
circumstances within 168 hours, and water towers, reservoirs, water level change
curve. In the city modeling project implementation, the pipe network node traffic
data based on the measured data using a separate assignment of large users, general
users of the global assignment method, combined with the urban pipe network data
to arrange the measured work, and the measured data into the model.

The pipe network is not connected to the inspection and handling of the pipe
sections and isolated points. Disconnect the pipe network is divided into mainly
two kinds: one for the isolated node; for the isolated pipe sections.
Isolated nodes junction of pipe sections are not any tube segment connected to which
the connection point of the pipe sections upstream and downstream nodes, valves,
pumps, water meters, fire hydrants, etc., of WaterGEMS pipe isolated node
(Orphaned the Nodes) query tool, able to quickly find the these isolated nodes.

The collection of the field data and the determination of pump characteristic
curve. Week observation period determined by the partition and observation points
in the third step, the implementation of the observations of the scene flow and
pressure in the project implementation period for the different seasons and water
conditions, to observe the flow and pressure of the pipe sections. There are special
conditions, such as fire or burst pipes, to increase the observation frequency and

ICPTT 2012
 ICPTT 2012 © ASCE 2013 458

density. At the same time, to carry out site pipe resistance coefficient determination.
The stage is to complete the field data, surveying, completion of the calibration data
checking.
Because of the pump operated for many years, the appearance of characteristic curve
can’t be used for model, the need for field data collection. The data collected at the
scene, combined with the waterworks recent test data, application of analysis of
covariance, the differences between the comparisons of these data, to get more
Downloaded from ascelibrary.org by New York University on 08/02/15. Copyright ASCE. For personal use only; all rights reserved.

information. There is a relationship H  H 0  S pQn between the actual flow rate Q

and head H, the head of the differences in the comparison of different data sources,
net of the impact of flow rate Q, and we should use the analysis of covariance.

Application of SCADA data pipe network hydraulic model checking.

WaterGEMS 'Darwin validator and SCADA data and field observation data, revised
pipe network hydraulic model parameters and application of observational data pipe
network hydraulic model checking and accuracy evaluation. The simulation will
achieve 80% accuracy of data after the completion of the measurement point
pressure data, such as the figure 3 in this program.

Figure 3. The simulated pressure and the field data

Arrange of key observation points. Facilitate the application of model checking the
results, the key observation points in the pipe network, to submit a list of key
observation points, and key observations. This section aims to examine the model
calculation error, the first calibration error analysis report.

Re-check of the model. Application key observation points, observational data and
the observational data of the pre-observation points, the second check of the pipe

ICPTT 2012
 ICPTT 2012 © ASCE 2013 459

network hydraulic model. To complete network model secondary check the network
model pressure accuracy of the technical agreement.

B/S structure of the network model of secondary development and application

of

Bentley's Haestad Methods provides the WaterObject secondary development tools.

Downloaded from ascelibrary.org by New York University on 08/02/15. Copyright ASCE. For personal use only; all rights reserved.

WaterObject is a set of COM components, and provide a range of methods to operate

the WaterGEMS the data. Secondary development, including the Microsoft Office
users can use different software tools, ArcMap and AutoCAD VBA description. You
can also use advanced programming language, such as .NET, C++, in Java, the
Visual Basic, C, or FORTRAN, etc.

B / S structure of the proposed. B / S, which is the abbreviation of Brower / Server,

the client, as long as the installation of a browser (Browser), such as Netscape
Navigator or Internet Explorer, the server install the Oracle, the Sybase, Informix, or
SQL Server and other databases. Browser severs through the Web Server for data
interaction with the database.
C / S, also known as Client / Server or client / server model, need to install special
client software. The servers are typically high-performance PC, workstation or
minicomputer, and large database systems such as Oracle, Sybase, for Informix, or of
SQL Server.
The biggest advantages of the B / S is that we can operate anywhere without having
to install any special software, the whole general framework show in figure 4. As
long as there is a computer with Internet access can use, zero client maintenance. The
expansion of the system is very easy, as long as the Internet, and then by the system
administrator to assign a user name and password, you can use. Can even apply
online through the company's internal safety certification (such as the CA certificate),
does not require the participation of people, the system can automatically assigned to
an account of the user into the system.

ICPTT 2012
 ICPTT 2012 © ASCE 2013 460
Downloaded from ascelibrary.org by New York University on 08/02/15. Copyright ASCE. For personal use only; all rights reserved.

Figure 4. The whole general framework

Model data in the B / S architecture real-time call. Simulation and calculation of

the pipe network in the WaterGEMS model has saved two files, which are a database
file DBF files and an MDB file, and DBF file MDB file records information on
almost all models, including node, pipeline physical information, such as the node x,
y coordinates, the pipeline start and end position, including all the attributes of the
element, such as the node's water demand, pipe C values.
The results are saved in the MDB file after the network analysis by WaterGEMS,
then real-time call information in these MDB files, released on the B / S architecture.
Of course, in the B / S architecture there can be further pipe network analysis, as
long as the preparation of the corresponding function, such as pipe location, contour
simulation, open the pump stops program analysis.

REFERENCES

Bentley Systems Incorporated (2006) “WaterGEMS v8Users Manual”, Haestad

Methods Solution Centre, 27 Siemon Co Dr. Suite200W, Watertown, CT06795,
USA.
Shamir, U. and Howard, C.D.D. (1968) “Water Distribution System Analysis,”

ICPTT 2012
 ICPTT 2012 © ASCE 2013 461

Journal of Hydraulics Division, ASCE, 94(1), 129.

Walski, T.M., Chase, D.V., Savic, D.,A., Grayman, W., Beckwith, S. and Koelle, E.,
(2003) Advanced Water Distribution Modeling and Management,” Haestad Methods,
Inc. Waterbury, CT.
Wu, Z. Y. Wang, R. H., Diezo, D., Walski, T. (2003) “Mining Water Consumption
and GIS-based Data for Loading Water Distribution Models”, Proceeding of 2003
World Water and Environmental Resource Congress, June 22-26, Philadelphia,
Downloaded from ascelibrary.org by New York University on 08/02/15. Copyright ASCE. For personal use only; all rights reserved.

Pennsylvania.

ICPTT 2012