A Zigbee Based Wireless Sensor Network for Sewerage Monitoring

#1 #1 #1 #1 #1
C.H. See , K.V. Horoshenkov , S.J. Tait , R.A. Abd-Alhameed , Y.F. Hu , E.A.Elkhazmi#2 and J.G.Gardiner#1
#
School of Engineering, Design and Technology
University of Bradford
Richmond Road,
Bradford, West Yorkshire, UK, BD7 1DP
1
c.h.see2@bradford.ac.uk
#
The Higher Institute Of Electronics, Bani Walid-Libya
2
eaelkhazmi@hotmail.com

Abstract — Blockages in sewers are major causes of both sewer
flooding and pollution. Water companies which fail to tackle this
problem face hefty fines and high operational costs if they
unsuccessful to provide a practical solution to prevent flooding.
As a result, the detection of sewer condition is routinely required
to inform on the best course of action to eliminate this critical
problem. This paper presents a novel low cost wireless sensor
technology to detect blockages proactively, and feed these event
data back to a central control room. The practical deployment
of the proposed WSN in an urban area will be demonstrated. In
addition, the challenges of this technology in a field trial and the
recorded data in terms of the sensor and communication
reliability will be addressed.
Index Terms — WSN, Wireless sensor
This paper describes a practical implementation of a
low cost wireless WSN using Zigbee communication and
acoustic sensor technologies to monitor the water level of the
gullies in a residential urban area. The purpose of this field
trial was to evaluate the preliminary design of the proposed
system in terms of durability of sensors, sensor nodes and
gateways and reliability of communication under real
operational conditions and within a typical urban
environment.

I. INTRODUCTION
Sewer flooding (DG5 Other Causes) and pollution
incidents are significant issues in the wastewater business
process in the water industry. One of the most pressing issues
for prevention of sewer flooding and pollution is sewer and
gully blockages [1]. Monitoring and maintenance are
important part of the many water companies business to
prevent catastrophic failures that can shut down a facility
which may cost several millions pounds per day. On top of
that, water industries regulated by several UK’s government
agencies, such as the OFWAT and EA, may be required to
pay hefty fines for not meeting the basic standard
Fig.1: Wireless Sensor Network- System Architecture
performance of their services. Currently, many water
companies have deployed telemetry systems to replace some
of the manual operations, running costs remain expensive. II. SYSTEM ARCHITECTURE
Low cost wireless sensors may be the only cost-efficient
option to replace traditional visual inspection which is In this section, the architecture of the proposed low power
extremely inefficient and costly. Moreover, existing telemetry mesh network wireless sensor system will be discussed.
systems require extensive cabling for (Public Switch Inherent power limitations of radio communication devices
Telephone Network) PSTN and power and cannot be might introduce hard restrictions on the coverage of the WSN
deployed over a large catchment area because of the cost. [2]. Due to this limitation, direct communication between
Low cost and low power sensors could be deployed over an sensor and base station is not always possible especially over
extensive footprint network and provide early warning of a difficult radio environment with strong attenuation. In order
impending failure offering time for maintenance teams to to overcome this difficulty and to extend the communication
prevent service or regulatory compliance failure. distance, an obvious way forward is to use multi-distributed
nodes that use multi-hops to send the data over these nodes

978-1-4244-2802-1/09/$25.00 ©2009 IEEE 731

 Fig.3: Wireless Sensors Distribution on field trial

(sensors) on the way to the base station. As can be illustrated
in Fig.1, by implementing a mesh network communication
configuration, this WSN allows for continuous connections
and reconfigurations around blocked paths. This might
results in hopping from sensor node to node until a
connection can be established with the base station [3-5]. It
should be noted that the mesh networks posses the self-
healing capability that will operate even when a node breaks
down or a connection fails. As a result, it forms a very
reliable network. Once the data is received by the data
gatherer/hub, it will be stored and published on a webpage
via connection to the internet. By accessing the internet,
remote mobile devices can request the recorded data with the
right user name and password.
A Zigbee based short range WSN was selected for this
application due to its attractive features such as low data rate,
low power consumption, simple communication
infrastructure, low latency and capability to support one
master and up to 65000 slave control units. In general, the
proposed system has led to the development of knowledge
and expertise in four areas of research: a) Embedded Antenna
design; b) Sensors and Instrumentation for use in water
industry assets; c) Wireless communications and distributed
wireless sensor networking; d) Remote monitoring of water
related assets.
There are two ways of using a sonic transmission to
detect the water level, i.e. (i) By measuring the time of flight
from transmitter to receiver – it will be faster in water, (ii) By
using the level of the received signal – the received signal
will be louder if the transmitter and receiver are both under
water. Both of these methods of detection will require a
driver for the acoustic transmitter and an amplifier for the
received signal. Method (i) suffers from the leaves or debris
and most likely to cause a false alarm. Therefore, method (ii)
was adopted in this project. A novel acoustic probe [7] and
data acquisition circuit (DAC) board were developed to
identify three status of water level, i.e., Low, Normal and
High. The DAC board was designed to operate in low power
consumption mode and considerably extend the life time of
the sensor.

A. Crossbow Mica2 Sensor Node

The Crossbow Mica2 [6] node is an advanced tiny wireless
platform for smart sensors, which is constituted by CC1000
radio, Atmega 128L processor, 128kB Flash, 4kB RAM and
10 bits ADC. By enabling the Low Power Listening (LPL)
mode of this platform, it will perform the operation in the
lowest duty cycle mode. Hence, this wireless module can
Fig.2: Deployment of wireless sensors in gullies and data gatherer
prolong the battery life significantly. An important part of on lamppost.
this system is the water level sensor element. There are a
number of water level sensors available in the market.
However, many commercial low cost water level sensors are B. Antenna design
unreliable and too delicate to be used in the hostile gully
environment. In order to reliably and effectively measure the Unforeseen RF transmission disruption can occur when
water level of a gully, a novel and low cost acoustic sensor deployment of wireless sensors is implemented over a large
was designed and developed. area. Multipath reflections occur when an RF signal takes
different paths on the propagation channel between the source

732
nodes (e.g., a radio NIC) to a destination node (e.g., access
point). In this scenario the received signal can be routed
along different paths, which causes the signal to bounce in
different directions. As a result, some of the radio signals
arrive with a delay as they travel longer paths to the receiver.
This is likely to corrupt the information contained in the
broadcast message and cause significant delay to the whole
communication system. To alleviate this problem in mesh
radio network, hop-by-hop algorithm can be adopted.
Conventional low power mesh network of wireless
communications sensors suffers from limited communication
range. In order to achieve an optimum reasonable
communication distance with minimum power consumption,
the antenna plays an important role in implementation of the
wireless sensor network (WSN).
The objective of this project is to design, develop and
implement a gully pots wireless monitoring system.
Therefore, the proposed wireless sensors have to be able to
operate in a harsh environment, with high radio signal
attenuation (lossy and watery) surroundings and invisible to
third party. As a result, the commercial off the shelf (COTS)
type aerial which is a monopole, is not suitable for this
application. This requires redesigning an aerial to satisfy the
demands of the proposed WSN. Ideally, the proposed antenna
has to provide the following characteristics: (1) able to
operate from 902MHz to 926 MHz frequency band, (2) have
omni- directional radiation patterns, (3) high gain, (4)
robustness, (5) low cost and profile and (6) ease of
installation and maintenance. For the proposed system, an
embedded microstrip antenna was designed, tested and
implemented within a waterproof enclosure. It is believed to
have a high potential and feasibility to be adopted in
underground infrastructure monitoring. The description of
this antenna is provided in [8].
C. Stargate Platform
Stargate platform is a linux operating system based on a
mini-computer [9-10]. In this present application, it is acted
as a data gatherer/hub which collects the data from all the
wireless sensors. As can be seen in Fig. 2, it was mounted on
a lamppost for optimum coverage of the monitored area.
Stargate is programmed to operate into sleep mode when it is
idle to conserve the battery power. In addition, GPRS was
used to transfer the received data from Stargate to the remote
server. Constantly turning on the GPRS connection will
flatten the battery within days so in order to maintain the
battery life for months; Stargate sends the recorded data back
to the remote server once a day based on GPRS connection.
III. FIELD TRIALS
A dense residential area in Bradford, UK, was chosen to carry
out the field trial. There were two main reasons for the selection of
this site: (i) In this area, every single house only has one gully
which collected all the wastewater, such as hot water, cold
water, bath water full of hair, rain water, kitchen sink water
733
full of fat. ie a high risk wastewater disposal system, and (ii)
the topology of the site allowed the proposed Zigbee mesh
network wireless sensor system to be tested easily. The site
was a row of residential houses, sensors could relay
information between its neighbour to reach the data gatherer.
Fig.3 shows the distribution of the Zigbee-based
communication system which currently includes eight gully
monitoring units and a stargate. As can be observed, the
shortest and longest distances between the sensor to sensor
(STS) and sensor to hub(STH) are 5.5m and 38.5m, and
12.3m and 66.5m respectively. As can be noticed, the red
triangle spot and the white ring symbols in Fig.3, indicate the
location of the data gatherer and sensors respectively.
For the purpose of extending the battery life of the
Zigbee flooding monitor up to a long as possible (ideally two
years), the proposed Zigbee monitor was programmed to
operate in the following modes:
1. A two way low power mesh network communication
was designed and implemented into the proposed
wireless system. In order to keep the minimal power
consumption in the sensor nodes, a low power listening
(LPL) mode was enable on the radio module (CC1000)
of the Crossbow transceiver. The LPL mode enabled the
radio module to go into sleep i.e. extreme LPL mode,
instead of turn off the radio completely. By implementing
this mode, the sensor and base station were in a 1% duty
cycle, maximum 0.89 packet transmission/reception per
second and 0.258kbps effective throughput operation
mode [11].
2.The sensor was programmed to wake up from sleep mode
to measure the data in every five minutes. If it detected
five consecutive flooding/leaking alarms, then it
transmitted the radio packet to either nearby sensor node
or directly to data gatherer/hub depend on the detected
receive signal strength indication (RSSI) level, otherwise,
it operated in sleep mode to keep the current
consumption to a few microamperes. In general, there
were three operation modes in this application, which
were sleep mode, sensing mode and radio broadcast
mode, these modes were corresponding to average 20uA,
9mA and 18mA current consumption, respectively.
3.Once it detected a low/high alarm, it sensed the RSSI,
from the hub as well as its neighbour sensor’s node. Then,
it compared the RSSI and find out the best route to relay
the data back to the GPRS server. It should be noted that
the open source Xmesh reliable route protocol [12] was
employed for this application. However, for the health
check purpose, the sensor broadcasted a health condition
packet back to the hub to indicate its battery level and
water level condition on a daily basis.
 IV. RESULTS AND DISCUSSION
From the field trial, according the daily data received from
the server, it was observed that the reliability of the proposed
Zigbee system can reach up to 80%. Fig.4 interprets the
reliability of the proposed wireless communication system
over each stage of the field trial. In general, this field trial has
been divided into four stages. In the first stage, eight sensors
and one data gatherer (i.e. stargate) were installed. It should
be noted that after the installation there are 2 sensors (ID
4520 and 4521) working reliably which operate under LOS
condition.
By adding relay points (i.e. repeaters) which was about 40m
to 70m away from the hub and changing the orientation of the
aerial of the hub, in addition of installation a high gain aerial
of the hub, which is corresponding to stage 2,3 and 4 in Fig.4,
the efficiency of the proposed system can be improved
considerably. As a result, 5 out the 8 of the sensors were
working reliably over trial period. However, this
communication reliability can be improved in future and
applied to large scale deployment by increasing the frequency
of the transmission of the Zigbee sensor, installing additional
low cost relay points and improve the aerial design on the
hub. This can be done by introducing multiple aerials on the
hub to mitigate the fading effects of a multipath environment.
Moreover, adding amplifier on the reception side of the
Zigbee radio module will improve the receive sensitivity of
the hub. As for the sensors that did not work with a high
reliability, this was due to unforeseen man-made issues, such
as covering the sensors with rubbish and concrete blocks.
These activities impaired the radio communication
significantly and invalidated the communication reliability of
the proposed system. An investigation has been carried out to
pinpoint the reasons on what possibly blocked the radio
communication on those gully monitoring units (ID: 4510,
4512 and 4513). According to findings from the investigation,
it seemed that failure of Sensors ID 4510, 4512 and 4513 can
be attributed to covering the gully.
Fig.4: Communication Reliability of the proposed system
734
V. CONCLUSIONS AND RECOMMENDATIONS
This paper presents a practical implementation of the Zigbee
based wireless communication system on an urban residential
area. These proposed technologies will enable to transfer
effectively and process data from large numbers of
potentially diverse sensors distributed within a sewer
infrastructure network to achieve potentially zero pollution
and DG5 other causes. Outcomes from the field trial enabled
researcher in this research area to gain expertise in the issues
associated with the practical monitoring of the performance
of elements of the urban sewerage infrastructure managed in
the UK water industry.
ACKNOWLEDGMENT
The authors acknowledge the support provided by funding
from Yorkshire Water Services and the Technology Strategy
Board, via the KTP Project “The development of a new
sewerage telemetry system”
REFERENCES
[1] OFWAT, “Service and Delivery- performance of the Water
Companies in England and Wales 2007-08
Report”,http://www.ofwat.gov.uk/regulating/reporting/rpt_los_
2007-08.pdf
[2] Cauligi S. Raghavendra, Krishna M. Sivalingam, and Taieb
Znati, “Wireless Sensor Networks”, Springer-Verlag New
York Inc.; Sep 2006.
[3] D. Geer, "User make a Beeline for ZigBee sensor Technology,"
Computer, vol. 33, pp. 16-19, Dec.2005.
[4] H.-C. Huang, Y.-M. Huang, and J.-W. Ding, "An
Implementation of Battery-aware Wireless Sensor Network
Using ZigBee for Multimedia Service," International
Conference on Consumer Electronics (ICCE '06) Digest of
Technical Papers, pp. 369-370, Jan. 2006.
[5] C.Evans-Pughe, "Bzzzz zzz [ZigBee Wireless Standard]," IEE
Review, vol. 49, pp. 28-31, March 2003.
[6] Crossbow, Inc., Mica2 wireless module,
http://www.xbow.com/Products/productdetails.aspx?sid=174
[7] A.Tolstoy, K.V.Horoshenkov and MT Bin Ali, “Detecting pipe
changes via acoustic matched field processing”, Applied
Acoustics, Vol.70, May 2009, pp.695-702
[8] C.H.See, R.A.A.Alhameed, D.Zhou, Y.F.Hu and
K.V.Horoshenkov, “Measure the Range of Sensor Networks”,
Microwave and RF, vol.47, Oct. 2008, pp. 69-76.
[9] Crossbow, Inc., Stargate gateway,
http://www.xbow.com/Products/Product_pdf_files/Wireless_pd
f/Stargate_Datasheet.pdf
[10] I.Stoianov, L.Nacham, S.Madde and T.Tokmouline,
“PIPENET: A Wireless Sensor Network for Pipeline
Monitoring”, IPSN’07, pp.264-273
[11] TinyOS, http://www.tinyos.net/tinyos-1.x/doc/mica2radio/
CC1000.html
[12] A.Woo, T.Tong and D.Culler, “Taming the underlying
challenges of reliable multhop routing in sensor networks, In
proceedings of the 1st International conference on Embedded
networked sensor systems (SenSys 2003), pp. 14-27. Los
Angeles, California:ACM Press.