FMGM 2015 – PM Dight (ed.

)
© 2015 Australian Centre for Geomechanics, Perth, ISBN 978-0-9924810-2-5
https://papers.acg.uwa.edu.au/p/1508_14_Malcolm/

Wireless data collection systems in the real world

PC Scott itmsoil Australia, Australia

J Paretas-Martinez Worldsensing, Spain
J Pérez-Arcas Worldsensing, Spain
KR Malcolm itmsoil Australia, Australia

Abstract
In the past, the instrumentation and monitoring component of any project was carried out manually. With
the introduction of dataloggers, the transformation of how the data was collected, delivered and presented
improved a quantum leap with near real time readings becoming possible and alarms being triggered. But
in so doing, it did present an issue with massive cable runs being required to link all sensors to the system.
These took a long time to install and were always difficult for the contractor to avoid damaging.
This paper will present the next generation of data collection through two different wireless systems each
with their own unique capabilities. Firstly the Rippa wireless system where case studies in the Perth City Rail
Link Project as well as the Gateway Project will be presented. Secondly, the Loadsensing mesh networked
wireless system will be presented which has been successfully utilised in the Perth City Bus and Roy Hill
Projects.
The details of each system will be examined in detail along with the applications and case studies.
In conclusion, the future of the wireless data collection systems will be discussed.

1 Introduction
Since Terzaghi, the observational technique of construction has been the best way to confirm that the
predicted parameters of construction are being maintained. To enable this to be conducted in a more
efficient manner the engineer has desired to have realtime data presented to their desktop computer in a
timely and efficient manner. As technology improves the capability of instrumentation specialists to better
provide this service has continued to be a work in progress. The development of dataloggers in the late
20th century saw major breakthrough in the capacity to achieve the goal. With telephone modem the data
was presented with ease and in real time but on the construction site a problem developed in how to
handle and protect the vast quantity of cable that was required to link all the sensors to a central
datalogger.
Determining the best data delivery system for your instrumentation and monitoring program will have a
large effect on the quality and reliability of the data, as well as initial and ongoing costs to your project. In
selecting the most suitable data collection method, Sections 1.1 to 1.4 are taken into account.

1.1 Purpose of the data

If the monitoring is serving to form a baseline against which to compare behaviours during construction,
then regular recordings will be required, but collection on a monthly or less basis is sufficient.
Measurements used to control the construction process are often reviewed daily so daily delivery is
sufficient. For monitoring critical structures or providing an alert function, much shorter intervals are
required.

FMGM 2015, Sydney, Australia 249

Wireless data collection systems in the real world PC Scott et al.

1.2 Type and number of sensors

Nested or profile sensors will need a multichannel datalogger to monitor them. Some systems not originally
designed for geotechnical instrumentation will not be compatible with vibrating wire (VW) sensors.
Likewise multiplexed or serial sensors will require specific controllers in their respective dataloggers.

1.3 Site layout and topography

Distribution of monitoring locations, structures and landscape as well as cellular network availability will
determine viable telemetry options within your site. Suitable cable routing options may be restricted,
dictating the position and number of dataloggers required. Line of sight between monitoring locations will
determine necessity for repeaters or additional gateways. Public access may mean the logger will need to
be positioned in a safe positon away from the risk of vandalism. Hillocks or vegetation may necessitate
elevating your transmitter on a pole.

1.4 Construction process

Running cables to a centralised multichannel logger will often be initially the most cost efficient method,
especially on sites with a large number of sensors within a small area. However, changes to the site
conditions must be considered. Removing and rerouting cables can be a costly exercise and lead to gaps in
data, often at critical points of the construction process. Likewise erecting structures — permanent or
temporary — between a radio transmitter and its Gateway will mean a loss of communication.

2 SI- based datalogger

The SIM-based datalogger (Rippa) is a self-contained datalogger unit with up to two channels. The unit
contains the recording hardware, long-life battery and 3G modem in an IP68 and UV stable housing. The
logger will measure the instrument at a preset interval, for example, every 15 minutes and transmits all
recorded data via the cellular network at a pre-set interval, typically 6-hourly. The ‘store and forward’
method ensures firstly data integrity; even in the event of a communications failure, all data will eventually
be transmitted. It also minimises communication up time, meaning lower power consumption and lower
cellular data usage. Transmitting at 24-hour intervals, the internal battery on the datalogger will support
the unit for 10+ years. At 6-hourly uploads the unit is self-sufficient for up to three years. The device uses
the standards based licenced radio, allowing 2-way communication via 3G (UMTS/HSDPA) allowing data to
be sent, and configuration instructions, clock synchronising and firmware updates to be received.
Data is made available through the cloud based software — web accessible software, which is used for both
hardware configuration and data analysis and reporting. The software contains powerful graphing tools for
live and interactive charting, as well as an export function for further processing in Excel. It allows
establishment of virtual sensors for automatic compensation or comparison. It integrates Bureau of
Meteorology data for barometric compensation, as well as the ability to analyse data against rainfall,
temperature changes and wind speeds.
The software also contains the hardware management tools. This allows adjustment of the reading and
upload interval, setting of the sensor calibration factors, establishment of virtual sensors, and displays of
hardware statistics such as battery life, signal strength, last upload and next scheduled upload. Since all
hardware configurations are done online no field connections are required and all commissioning can be
done remotely or via the mobile specific website. The mobile software sets GPS location of the device and
utilises your phone’s camera for recording installation photographs. Configuration is also done on a ‘store
and forward’ basis, compiling the logger’s instructions and sending them as a packet at the next scheduled
connection. For immediate configuration changes, data confirmation or commissioning, a forced
connection can be made by swiping the logger with a magnet.
Using a cloud-based software offers global access, requires no IT infrastructure costs or software licensing;
all software updates are done remotely and in real-time, and any firmware updates for the loggers can be

250 FMGM 2015, Sydney, Australia

 Civil tunnelling

carried out automatically. It also allows remote technical support for both data analysis tools and hardware
configuration.
New low-power cellular modems, coupled with advances in long-life battery technology have allowed the
practical development of self-contained cellular logging devices, with low-cost SIM card options making
them extremely economical. Self-contained cellular dataloggers — simple, low-cost, internally-powered
dataloggers with cellular connectivity — offer a number of advantages over short haul/unlicensed radio
loggers. By utilising the cellular network rather than a local network, small projects with isolated sensors
can quickly and easily establish a remote monitoring system. For example, a number of landslide projects in
southeast Queensland containing only one or two piezometers can have their data telemetered and
accessible by web cost effectively. For larger projects, the advantages come when sensors are likely to be
relocated, or are spread over a long distance, such as highway projects or large tailings dams.

2.1 Gateway WA project

The Gateway WA project is a billion dollar road initiative designed to improve access to Perth Airport and
surrounding industrial areas, including road and bridge improvements. The project allowed for monitoring
wells spread over 6 km to be monitored online. When work has progressed and the boreholes are no longer
of critical importance, the sensor and logger is relocated to a new position with no Gateway or network
configuration required. Likewise a number of tilt sensors were used to monitor a sheet pile wall during
excavation. Installation takes only as long as mounting the sensor and the logger, and commissioning only
requires powering the unit up.
With regards to network stability, licensed radio — as utilised by cellular network providers — are closely
monitored for spectrum crowding and quickly upgraded where spectrum crowding is a risk.
These devices do have their limitations, most obviously the requirement for and reliance upon, a cellular
network. Where cellular networks are limited, short haul radio devices allow you to channel the data to a
location where communications are available — external via cellular network, or internal/local via local
area network (LAN), WiFi or direct connection.

2.2 Perth City Link Project

The Perth City Link — Rail project presented a number of challenges with regard to data delivery:
The AUD 360 million rail project, completed in 2014, is the first stage of the Perth City Link. The scope of
works includes:
 Sinking the Fremantle Line between William Street and Lake/King Street. This has created almost
600 m of a new covered section of the Fremantle Line.
 Upgrading Perth Station with new tiling, lighting and services and convert the current Platform 9
into an island Platform 8/9.
 Created a new tunnel under the northern end of the Barrack Street Bridge to service the new
Platform 9.
 Extended the existing Joondalup line tunnel roof from Lake to Milligan Streets to be in line with
the new Fremantle line tunnel.
 Created a new pedestrian underpass connecting Perth Underground Station to all the platforms at
Perth Station.
Existing structures were being monitored by tilt sensors and crackmeters. While the majority of these were
contained within the Perth Train Station and Horseshoe Bridge (Figure 1), an area of roughly 300 × 300 m,
many were on bridge columns in between rail tracks, or on heritage listed structures. These constraints
meant running cables to central logging points was either time consuming and impractical, or outright

FMGM 2015, Sydney, Australia 251

Wireless data collection systems in the real world PC Scott et al.

impossible. This was overcome with single channel radio devices transmitting short distances to a central
Gateway.

Figure 1 Overview of the Perth City Link Project

2.3 Lowering the water table

As the site is located above what was once Lake Kingsford, a major step in the construction process was to
lower the water table, or dewater, using a number of wells installed between the tunnel walls. The water
was treated to remove the high level of iron and adjust pH levels before being pumped back into the
ground by nearby recharge wells to maintain water level around the site.
This required close monitoring of both groundwater levels and water flow from the dewatering pumps.
The purpose of the data was to ensure the water levels were matching the model with regard to the water
being pumped out of the ground. This meant a clear trend plot was required, but would only be assessed
once to twice per day. The sensors being monitored were pressure transducers suspended in observation
wells with a 4-20 mA output, and a combination of flow meters with either a pulse output or 4-20 mA.
Sensors were positioned separately and widely across the site.
We utilised the Rippa 2G datalogger/transmitter for recording and transmitting the data (Figure 3).
Over the life of the project approximately a gigalitre (one billion litres) of water was pumped out of the
construction area, treated and pumped back into the ground. To monitor this massive undertaking the
consultants opted for an innovative and successful installation of the Rippas. These units record the data
and regularly upload to a dedicated web page (Figure 2) providing the engineers with near real time
knowledge of how the water table is behaving. This was completely new technology and began with small
numbers being installed, but as confidence grew the quantity was finally well over 100 units.

252 FMGM 2015, Sydney, Australia

 Civil tunnelling

Figure 2 Software graphic of the various installations

Figure 3 Rippa connected to water meter

In a site as congested as the Perth City Link Rail Alliance (PCLRA) project having cables running from the
water monitoring bores to centralised dataloggers would have been impractical.

3 Mesh networked system — technical explanation

Wireless communications are sometimes perceived as less secure, because there is no physical connection
(wires) between network devices. More specifically, there is the belief that the network can be sensitive to
being hacked or provide interference problems. However, the configuration of the Loadsensing’s mesh
network is one of the safest solutions on the market. Loadsensing’s protocol, based on frequency-hopping
and time synchronisation, allows the operation of hundreds of sensors in the same network without any
interference. All information broadcast by the network is encrypted, with mechanisms for verifying
message integrity and authentication to access the network.
The difference and advantage facing a classical wired system is that thanks to a redundant mesh network, a
malfunctioning device will be isolated, minimising the impact of his disappearance on the network, because
this is dynamically adjusted to find another path to bring the data. In a classical cabling system, the
break/malfunctioning of a cable most times cause a complete system loss.
The mesh technology allows each node to act as sender of its own data but also as a repeater of its
neighbours (Figure 4). A network is a group of nodes with the same ID, security key, and synchronised
between them. The network gateway serves as the time base for nodes and serves its configuration. The
routing information is dynamic and always looking for the best path to bring data from the nodes to the
gateway.

FMGM 2015, Sydney, Australia 253

Wireless data collection systems in the real world PC Scott et al.

Figure 4 Diagram of a load sensing (LS) wireless network

When a new node joins the network, it is assigned an ID and receives information from time and frequency
synchronisation of its neighbours. This ensures that the network is working properly synchronised with the
new device. This allows the operation of different networks in one place without interaction between
them.
The wireless nodes are activated periodically, following its programming. Additionally, to improve the
robustness of communications different communication channels are used that exist at 2.4 GHz frequency
band, using a technique called channel-hoping which avoids interference: each data packet sent from one
node to its neighbour is sent by a different channel, choosing the best available channel depending on the
level of interference observed in the different channels.
All devices in a LS network share the same time base; this means that each node transmits, hears or sleeps
during time slots assigned. Each time slot has a length of a few milliseconds, this keeps the devices asleep
most of the time and only be active in the sending window data. This allows you to create not only highly
secure and robust networks, but also a very low power performance that allows years of autonomy running
on standard internal batteries.
The wireless communication is based on long used industrial standards known as highway addressable
remote transducer (HART). The HART standard is one of the first implementations of the standard Fieldbus
also well known in industrial environments. The current specification on which the communication system
is based is HART 7 September 2007: in this specification an included section is defined as Wireless HART,
which has been implemented in the LS nodes. The protocol offers a fully synchronous communication Mesh
type dynamic network: this ensures redundancy in communications, resilience and resistance to failures,
and thus ensures the reliability of the communication system even in the most complex environments.

254 FMGM 2015, Sydney, Australia

 Civil tunnelling

Additionally, the network synchronisation allows:

 In each window, submission seeks the sending node that is best to send the data packet.
 In case of specific high demand network reset communication times for all data packets to be
delivered on time.
 The possibility of negotiating the frequency in each window of communication, offering a highly
robust system against radio frequency interference. The system used is called frequency hopping
spread spectrum (FHSS) combined with a method called direct sequence spread spectrum (DSSS),
offering significant advantages in terms of scope and protection against fading.
Furthermore, we implement the following security systems:
 Confidentiality: there is an encryption system data using the algorithm (Hardware) 128-bit AES.
This prevents interception of packets.
 Data integrity: each message sent is protected by a code that looks the integrity packet sent. If
this is not correct the packet is forwarded.
 Protection against replay: replay attacks are protected by individual meters on each node to
ensure its invulnerability network level.
 Access control: each packet sent by a node carries an ID generated by the node itself. On receiving
each packet is checked with a unique key of the Gateway.
 Compartmentalised security: the loss of a portion of a data packet does not compromise the
network. Part of lost data is recovered in successive retransmissions.
 Mechanically the LS nodes are also very robust. These are protected with an IP67 encapsulated to
protect them from rain, shock, lightning, dust etc.

4 Case studies
4.1 Slope stability
Monitoring is fundamental for the prediction and analysis of landslide triggering factors and dynamic
behaviour, major issues in the hazard assessment and risk mitigation. Wired monitoring systems have
traditionally been used in landslide monitoring. However, wireless technologies are escalating in this field
as a consequence of their multiple advantages against standard wired systems, such as their versatility or
their lower power consumption. Wireless monitoring is the perfect solution for the acquisition of data on
geological processes placed in remote areas where power availability is scarce, and the position of the
sensors is often a critical issue due to the landscape conditions. In this context, a complete landslide
wireless monitoring system was successfully installed in the Rebaixader catchment (Central Pyrenees,
Spain) (Hürlimann et al. 2014).
The Rebaixader constitutes a typical high mountain catchment where landslides and torrential processes
occur with a sub-annual frequency. Rainfall is the principal triggering factor of the type of landslides
occurring at the Rebaixader known as debris flows, but the specific details of the geotechnical mechanisms
that originate the events are still not clear. In order to increase the knowledge on the processes occurring
at the catchment, this was equipped with a LS network for the monitoring of triggering factors and a
Worldsensing’s Spidernano Seismic Remote Unit for the acquisition of the ground vibration generated by
the moving mass (Figure 5).
The sensor network includes seven low-power wireless dataloggers. Digital, pulse and voltage sensors are
connected to these dataloggers in order to monitor soil water content, soil water potential, snow height,
5-minute rainfall intensity and air temperature and humidity. The dataloggers communicate in a multi-hop
fashion to deliver the information into the gateway, placed 1 km far down the hill but in line-of-sight, with

FMGM 2015, Sydney, Australia 255

Wireless data collection systems in the real world PC Scott et al.

two repeaters placed along the line. The gateway offers enhanced computational and storage capabilities
as well as 3G modem communication to the data centre.
The Spidernano seismic datalogger was connected to three 1D geophones recording simultaneously the
ground vibration generated by the pass of the debris-flow mass. The Seismic Remote Unit (SRU) equipped
with GPS clock discipline, and is placed nearby the gateway and connected to it via Ethernet. The SRU has a
low power consumption (0.5 W), specially adapted for field campaigns or permanent monitoring.
The recordings of SRU revealed that debris-flow occurrence can be perfectly detected by ground vibration
signal, and the approach of the flowing mass can even be detected before the arrival. Data acquired by the
wireless sensor network provided valuable results on the understanding of the failure and post-failure
mechanisms such as the 15 mm in one hour as a preliminary rainfall threshold for debris-flow initiation in
summer season. All these achievements are promising results for the application of low power wireless
technologies not only for standard landslide monitoring but also for landslide early warning systems.

Figure 5 Rebaixader catchment, Central Pyrenees (Catalonia)

Figure 5 shows LS wireless dataloggers are placed on top of the catchment to monitor triggering conditions,
one seismic datalogger is located at the base of the catchment to monitor ground vibration after debris
flow and mass movement. One repeater was located at the middle of the catchment to repeat data from
the dataloggers to reach the gateway, which gathers data from LS dataloggers and the seismic unit.

4.2 Rockfall
Railways crossing mountainous regions are often threatened by rockfalls. On one side, trains circulating
along threatened areas are exposed to relevant hazards. On the other side, damages caused by rockfalls
produce delays in railway traffic due to unexpected events, which usually generates important economic
losses. Supervision of these areas results in costly operations of maintenance for the railway operators.

256 FMGM 2015, Sydney, Australia

 Civil tunnelling

Aiming to provide an optimal solution to improve operations for ÖBB (Austrian Railways) two solutions
using LS wireless sensor networks have been implemented.
First, an automatic low-power warning system implementing wireless dataloggers is being used in the
railway crossing the Hieflau tunnel, in central Austria. An unstable rock mass, in the middle of a rock wall at
the tunnel entry, is directly threatening the railway (Figure 6). Two crackmeters connected to a datalogger
have been installed to monitor a critical rock aperture of the unstable mass, in the middle of the rock wall.
The crack aperture is being monitored every hour and data is processed in order to determine whether it
exceeds a predefined threshold. In case that the aperture threshold is exceeded, an e-mail is sent to ÖBB.
Thanks to the LS wireless system the occurrence of rockfall events can be predicted at a real low cost of
installation and maintenance. Thanks to this information, authorities are able to ask maintenance operators
to be there only when necessary (LS alarm notification) and without any need of long cable revisions,
avoiding periodical expensive on-site operations.

Figure 6 Hieflau tunnel, central Austria

Second, the LS system has been implemented to monitor rockfall barriers to prevent rockfall events in
Austrian railways. This project is based on multiple sensors, which are combined to form a wireless radio
network of the nets. Strain gauge load cells measure the tension of the steel cable of the barrier, while
tiltmeters (+-180º range) measure the barrier inclination, and wire extensometers measure the barrier
movement (Figure 7). Each type of sensor is read by a wireless datalogger, which at the same time creates a
wireless network that sends all data to the gateway. The gateway (solar-powered) delivers all data via
global system for mobile communication to the customer’s servers. This low-maintenance and
self-sufficient solution provides a 24 hour monitoring of the rockfall barriers in inaccessible areas in all
weather conditions, delivering data every hour, and in case of particular events, the sampling rate can be
increased remotely thanks to the bi-directional capacities of the system. Thanks to this, the rockfall barriers
can be monitored in real time without any need for human maintenance in a difficult-to-access remote site.

FMGM 2015, Sydney, Australia 257

Wireless data collection systems in the real world PC Scott et al.

Figure 7 Rockfall barrier, central Austria

4.3 Roy Hill Project

Roy Hill is an iron ore mining project in the Chichester Range in the Pilbara region of Western Australia,
located 115 kilometres (71 mi) north of Newman and 277 kilometres (172 mi) south of Port Hedland.
The project requires a new railway from the mine to port to move the ore to market in the quickest time
possible. A major section of this construction is the rail loop at the port which required a significant
embankment to be constructed on soft marine material. Standard instrumentation would typically be
settlement cells. In an effort to expedite construction the contractor wanted to automate the readings and
remove the settlement plates from being an obstruction to the truck movements. To achieve this
requirement settlement cells were installed. This then raised the next issue of collecting the data in the
most appropriate method. Many scenarios were contemplated until it was decided to use the mesh
network wireless system.
With an active construction site approximately 1,500 by 1,000 m incorporating 45 settlement cells and
14 piezometers it was a problematic installation. The wireless system relayed all the data to three
Gateways (Figures 8 and 9), incorporating a SIM card that transmitted the results to LS’s internet based
software for easy access and alarms (Figure 8).

Figure 8 Gateway installations

258 FMGM 2015, Sydney, Australia

 Civil tunnelling

Figure 9 Second Gateway with node in close proximity

5 Future
Wireless sensor networks (WSN) have mostly been based on the concept of tiny devices, with a set of
sensors and a radio to send the results to a dedicated server or the Internet. This radio infrastructure has
been typically designed keeping in mind the state-of-the-art on radio transceivers with low power
consumption, but implying a short range (100 m max). The solution to large projects that almost everybody
has found has led to develop and build a mesh network of devices, with a medium to low range between
the nodes of the mesh. In these designs, data is sent from one device to another until the data reaches the
central node (gateway) which itself is connected to the Internet in a more traditional way (3G, GPRS,
Ethernet).
However, depending on the applications, mesh networks might present several problems. One of them is
the synchronisation between devices; because each device has a different clock drift, it is often mandatory
to synchronise the whole network, leading to raising energy consumption. Another problem is the
asymmetry on the devices consumption, meaning that not all devices drain power at same rate. This is
because nodes closer to the gateway will act as repeaters to other devices. This phenomenon complicates
the operational costs of the network maintenance, even limiting the use of energy harvesting techniques in
some of the devices in the same network. Other operational costs of mesh networks are in the use of
repeaters to enhance coverage of some parts of the deployment, and many times it is not easy to detect
these dark zones. The last problem arises on networks where the density of nodes is not high (due to costs
of the nodes depending on the application). In these kinds of networks, the entire network sometimes
relies on a single node to repeat the signal form others, and the malfunctioning of this node can break
down the communication of all data.
Recently, a new concept of radio systems has arisen. This new idea is focused on sending few bits of data
(enough for the sampling rates and data used in geotechnical instrumentation) at a very slow transmission
rate (less than 1 kbps) but with a long range (up to kilometres). These solutions are based on techniques
like spread spectrum and specifically designed modulation schemes for high sensitivity. With this new
concept of radio in mind, the landscape of WSNs is changing. The use of a mesh network is no longer
needed, because with these long-range radios all devices can communicate directly with the gateway.
Hence, from a mesh network topology we are changing to a star network, and a single gateway receives
data directly from all devices. This implies that the loss of one device does not affect any other device.

FMGM 2015, Sydney, Australia 259

Wireless data collection systems in the real world PC Scott et al.

This change simplifies enormous operational procedures and costs, because it avoids the use of repeaters.
It also simplifies coverage tests and, hence, usually needs fewer trained people to do the installation. Once
these new long-range radios are implemented in the geotechnical/structural monitoring world, deploying a
WSN will be as easy as placing the sensor+datalogger wherever fits best, checking that they are covering
your deployment zone and just dropping your devices and starting to receive data into your servers. There
is no need of complex planning to deploy a mesh network. That is, concentrating on what your true
business is: providing services to specific industries (and not spending most of your efforts in dealing with
coverage problems).

Reference
Hürlimann, M, Abancó, C, Moya, J & Vilajosana, I 2014, ‘Results and experiences gathered at the Rebaixader debris-flow monitoring
site, Central Pyrenees, Spain’, Landslides, vol. 11, no. 6, pp. 939-953.

260 FMGM 2015, Sydney, Australia