Smart Water Networks need Smart Data Presentation - IT&Water Feb 10 2015

Smart Water Networks need Smart Data Presentation
L. Stakenborg*, E. Trietsch** and B. de Graaf**
* Quasset, Seinstraat 20, 1223 DA, Hilversum, the Netherlands (E-mail: luc.stakenborg@quasset.com)
** Vitens, PO Box 1205, 8001 BE, Zwolle, the Netherlands (E-mail: eelco.trietsch@vitens.nl and
bendert.degraaf@vitens.nl)
Abstract
The increasing use of sensors in water distribution networks is generating vast amounts of
data. This data will only be valuable if presented to users (operators, analysts and
management) in an effective and efficient way. This paper describes the approach and results
achieved in Vitens using a web-based, GIS oriented presentation layer, integrating real-time
sensor data and network and customer data from multiple data sources.
Keywords
Smart water networks, sensors, real-time data, data integration, presentation layer, Vitens,
Quasset, IMQS, GIS, browser based.
Smart Water Networks
Water utilities create and maintain an ever growing amount of data on their business operations and
infrastructure. The increasing use of sensors that monitor the networks in real-time is even
accelerating the rate at which the amount of data is increasing. However, all this data is only of real
business value if it is used to support analysis and decision making. Ensuring that users (e.g.
operators, analysts and management) are presented with network information in an effective
(meaningful/relevant) and efficient (easy-to-use, responsive) way is therefore a key success factor for
Smart Water Networks.
In this paper we share the approach and results achieved in the Vitens Innovation Playground (VIP).
Quasset[1]
is working with Vitens to provide a web-based, GIS oriented information presentation layer
on top of existing data sources containing network, customer and real-time sensor data. The work is
part of SmartWater4Europe[2]
, an EU FP7 project that shows water innovation in the demo sites of
Vitens, Acciona, Thames Water and the Université de Lille.
Vitens[3]
, the largest water utility in the Netherlands, is testing innovative solutions in the Vitens
Innovation Playground (VIP), a dedicated area comprising 5% of the total network. The goal of the
VIP is to test new solutions and tools that help build the business case before being rolled out to the
whole of Vitens. Currently information about the network is contained in multiple data sources, e.g. a
network database, hydraulic modelling tools, customer databases and sensor data. Many types of
sensors are deployed in the VIP, measuring physical properties like flow, pressure, temperature,
conductivity and water quality (e.g. with Optiqua sensors). In addition, soft sensors represent
computational outcomes, e.g. net flow to DMAs (District Metered Areas) and leak detection
algorithms.

Software architecture
The figure below shows the software architecture at the Vitens demo site:
Figure 1: Software architecture at Vitens demo site
Real-time data platform
All sensor data is collected in a real-time data platform, based on the OSIsoft[4]
PI system. At the
heart of PI is a data historian: a data collection and retrieval system optimized for time series of data.
The real-time data platform provides a central location to interface with and collect data from a wide
variety of sensors. In addition, soft sensors (algorithms using sensor data as input) can be
implemented in PI.
IMQS presentation layer
The presentation layer is provided by IMQS[5]
, a web-based, GIS oriented system, integrating real-
time data and network and customer data from multiple sources. The IMQS presentation layer
presents network and sensor information in a geospatial context, for example the location of sensors
in DMA zones:
Figure 2: DMAs in the VIP area (blue areas) and sensor locations (purple dots)

A user can select several map themes to quickly view and compare pipe details such as year of
construction:
Figure 3: Map theme 'Year of construction’'
Other themes include pipe material, pressure, flow direction and system type. Detailed data about
pipes, DMAs, hydraulic regions, production sites and sensors are displayed by selecting the objects
on the map.
Figure 4: Display properties of a selected pipe

Real-time data and historic trends
Users have immediate access to current values through the property pane on the right. Clicking on
values provides historic trends of the sensor data:
Figure 5: Historic trends of a selected sensor measurement
Feedback from Vitens’ users shows that the user interface is both responsive (quick to navigate and
display data) and easy to use (avoiding deeply nested menu structures). This is especially important
for e.g. operators.
Flow map
Not only does the geospatial presentation provide an intuitive way to access the information at
various aggregation levels (e.g. zooming in from region, via a city, all the way to street level), it also
offers the opportunity for novel information tools. A good example is the Flow map, which displays
flow arrows proportionate to the actual flow measured at sensor locations. The flow map is a tool to
quickly identify the location area of a large break (by just following the flow).
The time slider at the bottom can be used to go back in time which is useful in a post-event analysis
or when comparing night/day patterns.

Figure 6: Flow map with time slider
Sensor measurements, states and event notifications
Sensors provide a time series of measurement values of (continuous) physical properties, e.g. water
temperature in degrees Celsius. It is often useful to categorize these values in value bands, e.g. too
low, low, normal, high and too high. Thresholds or other decision logic can be used to categorize the
measurement value into these so called ‘states’. State transitions signal events.
This relationship between these concepts is illustrated in the figure below.
Figure 7: Measurement values and states
Sensor states also have associated alarm levels (normal, warning, alarm and error). Warning and
alarm events are pushed immediately from PI to IMQS and to the user’s browser. This provides a
map of events:

Figure 8: Notifications for Water Quality, Pressure, Leak detection and Temperature events
The warning/alarm events can be used to trigger appropriate response procedures.
Sensor metadata repository
The network of sensors deployed in the Smart Water Network is quite dynamic: sensors get
(re)placed or relocated, sensors may temporarily be out-of-service, new sensor types from other
vendors may be added, etc. To accommodate these changes, the presentation layer needs to have
access to a central, up-to-date repository of data about the sensors, i.e. sensor metadata (e.g. their
location, vendor, model, operational status, type etc.).
We designed the following logical data model to structure the sensor data in this repository:
Sensor
location
Sensor
parameter
Sensor device
State EventMeasurement
Info layer
Info object
Sensor
network
(Dynamic)
property
Figure 9: Logical data model for sensor data and metadata

A sensor network is essentially a collection of sensor devices. It is useful to allow for multiple sensor
networks (e.g. representing water quality sensors, pressure sensors, etc.). Each sensor device
measures one or more physical properties, referred to as sensor parameters. Each of these sensor
parameters has associated measurements (a time series of values), states and events. Furthermore
we have identified the minimum required set of attributes and verified that the model caters for a
range of physical and soft sensors being used in the Vitens demo site.
The right-hand side illustrates how the sensor data is presented by IMQS: sensor networks are
represented as (GIS) layers, sensors as selectable objects inside these layers and sensor parameters
are considered as the dynamic (or real-time) properties of sensor objects.
Based on the model, the sensor metadata repository has been implemented in Vitens PI real-time
data platform, using PI-AF (PI Asset Framework). IMQS and PI are connected through a JDBC
interface and Web services are used for event notifications.
Figure 10: Sensor repository in PI-AF
RESULTS
In conclusion, the following results have been achieved in the Vitens Innovation Playground:
 A wide variety of sensors have been installed in the Vitens demo site. Both physical sensors
measuring Pressure, Temperature, Flow, Water Quality, as well as soft sensors (algorithms) for
leak detection, energy efficiency, net flows, etc.
 Sensor data is centrally collected and stored in a real-time data platform based on the OSIsoft PI
system.
 A sensor metadata repository based on OSIsoft PI-AF has been designed and implemented,
keeping track of all the data about the sensors that are available in the network.
 A web-based, GIS oriented presentation layer based on IMQS provides a single, integrated and
user-friendly interface to multiple data sources.

 The geospatial presentation and access to information offers novel means to interpret and act on
real-time data, e.g. the localization of pipe bursts. It also demonstrates the added value of
offering sensor data in the context of other relevant asset and customer information.
 It also has been very effective in presenting the outcomes of various activities and themes in the
VIP to the senior management and other stakeholders.
Within the SmartWater4Europe program we will continue to build on these results and provide
opportunities for other water utilities to visit and learn more from the Smart Water Network demo
sites.
REFERENCES
[1] Quasset: http://www.quasset.com
[2] SmartWater4Europe: http://www.smartwater4europe.com
[3] Vitens: http://www.vitens.nl
[4] OSIsoft: http://www.osisoft.com
[5] IMQS: http://www.imqs.co.za

Smart Water Networks need Smart Data Presentation - IT&Water Feb 10 2015

More Related Content

Viewers also liked (10)

Similar to Smart Water Networks need Smart Data Presentation - IT&Water Feb 10 2015 (20)

Smart Water Networks need Smart Data Presentation - IT&Water Feb 10 2015