CONTROL ENGINEERING LABORATORY

WIRELESS TECHNOLOGIES IN
PROCESS AUTOMATION
- REVIEW AND AN APPLICATION
EXAMPLE

Marko Paavola

Report A No 33, December 2007

University of Oulu
Control Engineering Laboratory
Report A No 33, December 2007

Wireless Technologies in Process Automation

- A Review and an Application Example

Marko Paavola

University of Oulu, Control Engineering Laboratory

Abstract: This literature review studies the wireless technologies, their applications, characteristics,
design criteria and requirements for process automation applications, focusing especially on wireless sensor
networks (WSN). Additionally, implementation of an application example, wireless monitoring of
temperature and pressure of a steam boiler, is also presented.

The review indicates that the wireless automation systems are often hybrid in nature, combining legacy
fieldbus systems and heterogeneous networks. The design criteria and requirements for industrial
applications are stricter than compared to some other domains, for example environmental monitoring,
home and building automation. According to review, wireless technologies provide potential for industrial
applications, but additional research may be required on interoperability, reliability, data processing
techniques and energy consumption, the last especially from whole system point-of-view.

Keywords: wireless sensor networks, industrial automation

ISBN 978-951-42-8705-3 University of Oulu

ISSN 1238-9390 Control Engineering Laboratory
PL 4300
FIN-90014 University of Oulu
1 INTRODUCTION ...................................................................................................... 1
2 WIRELESS TECHNOLOGIES IN PROCESS AUTOMATION.............................. 2
2.1 Overview............................................................................................................. 2
2.2 Wireless Local Area Networks (WLAN, 802.11) .............................................. 3
2.3 Wireless Metropolitan Area Networks (IEEE 802.16, IEEE 802.20) ................ 4
2.4 Wireless Personal Area Networks (WPAN) ....................................................... 6
2.4.1 Bluetooth and Wibree (IEEE 802.15.1)...................................................... 6
2.4.2 Ultra-wideband (IEEE 802.15.3) ................................................................ 6
2.4.3 ZigBee, Wireless Hart, 6loWPAN, (IEEE 802.15.4) ................................. 7
2.4.4 Radio Frequency Identification (RFID)...................................................... 8
2.5 Satellite ............................................................................................................... 9
2.6 Cellular Networks ............................................................................................... 9
2.7 Additional Standards......................................................................................... 10
2.8 Technology Roadmap ....................................................................................... 11
3 WIRELESS SENSOR NETWORKS IN INDUSTRIAL ENVIRONMENT........... 13
3.1 Introduction to Wireless Sensor Networks ....................................................... 13
3.2 Industrial Environment Characteristics............................................................. 15
3.3 Industrial Applications...................................................................................... 16
3.3.1 Introduction............................................................................................... 16
3.3.2 Classification of Applications................................................................... 19
3.4 Design Considerations ...................................................................................... 21
3.4.1 Design Challenge ...................................................................................... 21
3.4.2 Performance and Quality of Service ......................................................... 21
3.4.3 Operational Power .................................................................................... 24
3.4.4 Scalability and Modularity........................................................................ 25
3.4.5 Topology ................................................................................................... 25
3.4.6 Radio ......................................................................................................... 26
3.4.7 System and Software................................................................................. 28
4 CASE EXAMPLE – WIRELESS MONITORING OF A STEAM BOILER .......... 31
4.1 Application........................................................................................................ 31
4.2 Use Case Scenario............................................................................................. 32
4.3 Requirements .................................................................................................... 33
5 DISCUSSION ........................................................................................................... 36
5.1 General Considerations..................................................................................... 36
5.2 Monitoring of the Steam Boiler: Experiences and Conclusions....................... 37
5.3 Future Work ...................................................................................................... 38
6 CONCLUSION......................................................................................................... 39
REFERENCES ................................................................................................................. 40
 1

1 INTRODUCTION

The field of wireless communications is diverse and it may be difficult to get a picture of
the different technologies utilised in a certain field. In industry, the proposed and already
employed technologies vary from short-range personal area networks to cellular
networks, and in some cases even global communications via satellite are applied.

Recently, the use of wireless sensor networks (WSN) in industrial automation has gained
attention. WSNs are technically a challenging system, requiring expertise from several
different disciplines. Therefore, the information about important design criteria is often
scattered. Additionally, characteristics for the industrial automation applications are often
stricter than the other domains, since the failure of the communication system may lead
to loss of production or even lives.

This report attempts first to give a short review about the emerging and already employed
wireless technologies in process automation, excluding proprietary solutions provided by
different manufacturers. Additionally, some standards and regulations, which may be
relevant for wireless applications in the industrial automation, are shortly introduced.

After the technology review, WSN are introduced briefly. Next, industrial applications of
WSN and categories of wireless communication are discussed. Finally, some important
design criteria in the context of industrial automation applications are presented.

The importance of gaining experience of applying wireless sensor networks to process

automation environments has been addressed in the literature. The experience is
important, as it can be used to show, as well as to get rid of, the problems in the current
technology, and to enable larger variety of applications.

To study the applicability of the wireless sensor network in process monitoring

applications, an example application to steam boiler was implemented. This report
describes the implementation of wireless monitoring of temperature and pressure in the
steam boiler. Finally, some conclusions and future research suggestions are presented.
 2

2 WIRELESS TECHNOLOGIES IN PROCESS

AUTOMATION

2.1 Overview
A survey investigated wireless technologies used in industrial automation. The most
exploited were Wi-Fi (34%), Radio frequency identification (RFID) (21%), Bluetooth
(14%), and ZigBee (6%) (Hoske, 2006). Similar results were presented in (Welander,
2007), though ZigBee had been strengthening its position. Namely, RFID and ZigBee
were reported to have highest number of new uses (RFID +17%, ZigBee +16%)
(Welander, 2007). The landscape of wireless technologies is presented in Figure 1.

Figure 1. Wireless technologies landscape described in terms of throughput (y-axis) and

range (x-axis) (Koumpis, 2006).

Several of the technologies in Figure 1 are applied (or planned to be applied) in process
industry (see Section 2). The choice between different (or combination of) technologies is
dependent on application: different approach is applicable e.g. to monitor remote tanks
and monitoring robotic installations (See APPENDIX 1 for examples). Moreover,
adoption of several network technologies in a single environment is becoming more
common nowadays (Low et al., 2005). Figure 2 illustrates an example about wireless
sensor/actuator network.
 3

Figure 2. An example about industrial sensor/actuator network combining several

network technologies (Low et al., 2005).

Next, the wireless technologies applied in process industry are shortly introduced.
Proprietary technologies are excluded, and therefore, Z-Wave and Wavenis mentioned in
Figure 2 are not discussed (it should be noted, however, that proprietary technologies
may have applications in some cases (Sallinen et al., 2007). After introduction, some of
the related standards in industry are presented and finally, a technology roadmap is
discussed.

2.2 Wireless Local Area Networks (WLAN, 802.11)

Wireless local area networks (WLAN), or commercially Wireless Fidelity (Wi-Fi),
operate at 2.4 GHz or 5 GHz ISM bands and offer speeds up to 54 Mbps. They support
two modes: ad-hoc and infrastructure. The ad-hoc mode allows stations to spontaneously
form a wireless LAN, in which all stations communicate with each other in peer-to-peer
manner. The infrastructure mode, the network has an access point (AP), through which
each client station communicates. A typical Wi-Fi AP may have range of 45 m indoors
and 90 m outdoors. The properties of the most important 802.11 standards (802.11a,
802.11b, and 802.11g) can be found for example from (Cisco).

In process automation, Wi-Fi Networks can serve as the backbone for data concentration
and networking. For example, it can used in conjunction with short-range, low-power
devices in a wireless field network to collect data from the gateway further sending it to
the control room or other data collection point (Emerson, 2006b).

The advantages of WLAN in process automation include high range, open standards,
robustness, cost-effectiveness, easy accessibility to process control by mobile workers
and high data transmit rate (Emerson, 2006b). Recently, also the security has increased
with two important security enhancements, Wi-Fi Protected Access (WPA) and WPA2,
which replace older and less secure technology, Wireless Equivalent Policy (WEP).
 4

Additionally, used in conjunction with for example wireless sensor networks, they
provide sufficient bandwidth for supporting multiple gateways. However, WLAN may
not be applicable for wireless device-to-device communication because of high power
consumption, requiring line power or power over Ethernet (PoE). The expense of
providing the power lines may limit the size of the network (Emerson, 2006b).
Additionally, WLAN is more designed to high throughput applications for small number
of terminals (Dzung et al., 2005), which is quite opposite to some WSN characteristics
(see Section 3.1). Figure 3 presents an example of Wi-Fi backbone implementation by
Honeywell.

Figure 3. Wi-Fi backbone architecture concept developed by Honeywell (Gonia and

Budampati, 2006). Information is transmitted between field equipment (wireless sensors,
Programmable Logic Controller (PLC), mobile station, IntelaTrac, mobile phone) and
Wi-Fi backbone wirelessly and connected to the whole automation system via gateway
(IML Group plc, 2007a).

2.3 Wireless Metropolitan Area Networks (IEEE 802.16, IEEE

802.20)
WiMAX (Worldwide Interoperability for Microwave Access), synonym for IEEE 802.16
standard, is a metropolitan access technique, and its characteristics are long distance
transmissions and high data capabilities (Sun et al., 2007). WiMAX coverage can reach
up to 50 kilometres and the capacity of network backbone can be increased to 75 Mbps
(Cungor and Lambert, 2006). WiMAX not only provides wireless access but may also be
used to expand the network access coverage e.g. to remote or suburban areas (Sun et al.,
2007).
 5

In industry, WiMAX could serve as the backbone in a hierarchical network strategy, e.g.
for ZigBee (see Section 2.4.3) (Cleaveland). In (Cungor and Lambert, 2006), a WiMAX
backbone for electric system automation is presented (see Figure 4).

Figure 4. Illustration of hybrid architecture of WiMAX and wireless mesh networks

(Cungor and Lambert, 2006).

(Cungor and Lambert, 2006) also discusses several advantages of utilising WiMAX as a
part of hybrid architecture in electric system automation, namely: increased reliability,
low installation costs, large coverage area and automatic network connectivity. As a
challenge for the proposed solution (Cungor and Lambert, 2006) mentions harsh
monitoring environments, optimal placement of WiMAX towers, mobility support,
integration of heterogeneous networks, scalability, coordinated resource management and
security issues.

An emerging standard in WMAN is Mobile Broadband Wireless Access (MBWA), IEEE

802.20. However, according to (VAN, 2005), the standardisation is ongoing and products
are not yet available.
 6

2.4 Wireless Personal Area Networks (WPAN)

2.4.1 Bluetooth and Wibree (IEEE 802.15.1)

Bluetooth operates at 2.4 GHz ISM band and employs frequency hopping spread
spectrum (FHSS) modulation technique (VAN, 2005). Bluetooth has been considered as
one possible alternative for WSN implementation (Aakvaag et al., 2005). However, due
to its high complexity and inadequate power characteristics for sensors, the interest
towards Bluetooth for WSN applications has decreased (Aakvaag et al., 2005).
Additionally, Bluetooth is designed for high throughput applications between small
numbers of terminals (Dzung et al., 2005). IEEE 802.15.1 specification covers physical
and link layers from Bluetooth 1.1 -specification (VAN, 2005).

Wibree specification is a part of Bluetooth specification as an ultra-low power technology

addressing devices with very low battery capacity. This extension to Bluetooth allows for
data rates of up to 1 Mbit/second over distances of 5 to 10 meters in the 2.45 GHz band.
Though Wibree is similar to Bluetooth and can employ the same chips and antennas, it
has some important differences. Wibree does not use frequency hopping to avoid
interference, as Bluetooth does. It has a variable-length packet structure, compared to
Bluetooth's fixed length. It also employs a different-unspecified-modulation scheme.
Additionally, the implementation of the security algorithm (AES) has been taken into
account from the start. It has been discussed, that Wibree also may not be applicable to
industrial environments due to short range (<10 meters), lack of mesh networking
capabilities and limited number of radios in network. (Mannion, 2006)

2.4.2 Ultra-wideband (IEEE 802.15.3)

Ultra-wideband (UWB) is a short-range technology based on transmission of impulses

lasting only fractions of nanoseconds emitted in periodic sequences (Hancke and Allen,
2006; Zeng et al., 2006). IEEE 802.15.3 specification defines the physical layer for
UWB.

Formerly, the applications of UWB mainly considered multimedia and personal area
networking (Porcino and Hirt, 2003), but recently also the industrial applications have
gained attention as can be seen from (Hancke and Allen, 2006; van de Kar et al. 2006;
Zeng et al., 2006). In (van de Kar et al., 2006), the results of the brainstorming session
about UWB use in process industry are presented. UWB was estimated to be very
applicable e.g. for automated registration and localization of equipment and persons; the
rapid commissioning of the process control equipment (especially in pilot plants); remote
control; plant wide communications not related to process control; and secondary sensor
& control networks like gauging in tanks (van de Kar et al., 2006). UWB is not seen
applicable for communication over longer distances or measuring data from unsafe zone
(latter because high peak energy of pulses). Additionally, it should be noted that
maximum amount of devices per network, 2 (Kinnunen, 2007), may provide restrictions
for utilisation of UWB in some applications.
 7

The advantages of UWB are good localization capabilities (Zeng et al., 2006), possibility
to share previously allocated radio frequency bands by hiding signals under noise floor
(Hancke and Allen, 2006), ability to transmit high data rates with low power (Hancke and
Allen, 2006), good security characteristics due to the unique mode of operation, and
ability to cope with multipath environments (Hancke and Allen, 2006). The existing
challenges lie e.g. in hardware development, dealing with medium access control (MAC)
and multipath interference, and understanding propagation characteristics (Hancke and
Allen, 2006).

2.4.3 ZigBee, Wireless Hart, 6loWPAN, (IEEE 802.15.4)

ZigBee

ZigBee aims for cost sensitive, very low-power and low-speed wireless networking,
targeting to industrial, home and building automation (Aakvaag et al., 2005). The good
characteristics of the ZigBee are extremely low energy consumption and support for
several different topologies, which makes it a good candidate for several sensor network
applications (Aakvaag et al., 2005). However, it is reported in (Dzung et al., 2005) that
ZigBee cannot meet all the requirements for at least some industrial applications. For
example, it cannot serve the high number of nodes within the specified cycle time (Dzung
et al., 2005). IEEE 802.15.4 standard defines physical and data link layers for ZigBee
(VAN, 2005).

Wireless HART

In (IML Group, 2007b) release of the wireless HART specification has been reported.
The technology employs 802.15.4-based radio, frequency hopping, redundant data paths,
and retry mechanisms. Devices that comply with the wireless HART specification are
interoperable and support tools that work with both wired and wireless HART equipment.
(IML Group, 2007b)

Wireless HART networks utilize mesh networking, in which each device is able to
transmit its own data as well as relay information from other devices in the network. Each
field device has two routes to send data to the network gateway; the alternative route is
used when the primary route is blocked either physically or by interference. Additionally,
a new route is established if a route is seen permanently blocked. The range between
devices is approximately 200 meters. (IML Group, 2007b)

The standard uses time division multiple access (TDMA) to schedule transmissions over
the network (IML Group, 2007b). In TDMA, a series of timeslots (10 ms for Wireless
HART) are used to coordinate transmissions. The media access control (MAC) header is
designed to support the co-existence of other networks, such as ZigBee. Wireless HART
is aimed for new applications (e.g. for asset monitoring) rather than replacing the wired
solutions. (IML Group, 2007b)
 8

6loWPAN

6loWPAN aims for standard Internet Protocol (IP) communication over low-power,
wireless 802.15.4 networks utilising Internet Protocol version 6 (IPv6). The advantages
of 6loWPAN from the industrial point of view are ability to communicate directly with
other IP devices locally or via IP network (e.g. Internet, Ethernet) (Culler and Hui, 2007),
existing architecture and security (Culler, 2006), established application level data model
and services (e.g. HTTP, HTML, XML) (Culler, 2006), established network management
tools (Culler, 2006), transport protocols and existing support for IP option in most
industrial wireless standards (Culler, 2006).

2.4.4 Radio Frequency Identification (RFID)

RFID uses a small radio-frequency transponder called an RF tag (Emerson, 2006a). The
tag is electronically programmed with unique information, which can be read from a
distance (Emerson, 2006a). The RFID operates from frequencies ranging from several
hundred kilohertz to tens of gigahertz (Yeager et al., 2006). Typically, the RFID reader
transmits the required power and data to the tag (Yeager et al., 2006). There are two types
of RFID tags, active and passive (Yeager et al., 2006). Active tags are battery-powered,
more expensive, and use higher frequencies, whereas the passive tags use lower
frequencies, and do not have the internal power source (Yeager et al., 2006). The reading
range is up to 100 meters for active and about 20 meters for passive tags (Yeager et al.,
2006).

RFID is used for example in inventory-tracking applications (industrial examples can be

found e.g. from (IDG Communications, 2005; King, 2005)). It is also possible to update
some types of tags, e.g. as the item progresses down an assembly line (Emerson, 2006a).
However, because RFID information is static and must be programmed into the tag it
cannot be used directly for the measurement or diagnostic data that is constantly subject
to change (Emerson, 2006a).

Future research seeks to further expand the capabilities of RFID technology. One
promising area of development is the integration of sensors and dynamic computing
power into RFID chips. Advances in integrated circuit (IC) fabrication have reduced the
power requirements of microcontrollers to the point that they can operate on the
scavenged power of passive RFID tags. Coupled with the breakthroughs in micro-electro-
mechanical-systems (MEMS) and IC sensor technology, this may enable wireless,
battery-free sensors. (Yeager et al., 2006)

Near field communication (NFC) is a very short-range (max. 20 cm), wireless point-to-
point interconnection technology, which has evolved from a combination of earlier RFID
contactless identification and interconnection technologies. It is based on inductive
coupling and uses 13.56 MHz carrier frequency and the available data rates are 106, 212
and 424 kbps. (Ailisto et al., 2007).
 9

NFC technology is becoming into large-scale use in mobile phones. The expansion of the
use of NFC is driven by an industrial alliance, which opens new possibilities for mobile
devices also in industrial applications. The advantages of the NFC compared to RFID are
its diverse and flexible interfaces. Additionally, NFC offers more potential to implement
battery-operated sensors (or even sensors without battery) with battery replacement
cycles of months or years. (Sallinen et al., 2007)

2.5 Satellite
Satellite communication can offer innovative solutions for remote control and monitoring
of geographically dispersed assets (Emerson, 2006a; Cungor and Lambert, 2006). Due to
extensive geographical coverage, it may be a good alternative, when communications
infrastructure, such as telephone or cellular network, does not exist (Emerson, 2006a;
Cungor and Lambert, 2006). However, satellite communication has some disadvantages,
e.g. long delays (Cungor and Lambert, 2006), varying channel characteristics (depending
on fading and weather conditions) (Hu, 2001), cost (Emerson, 2006a; Cungor and
Lambert, 2006) and lack of existing standards (Emerson, 2006a).

Example of utilising satellite as a part of the monitoring system, in conjunction with

WSN, can be found in (King, 2005). London-based company, BP, used battery-operated
ultrasonic sensors to remotely monitor the fill levels of industrial customers’ gas tanks.
The radio signal from sensors was transmitted to low-earth orbit satellite and then further
to BP to enable timely deliveries. The number of tanks was about 200 and the reported
delivery efficiency improvement was over 33%.

2.6 Cellular Networks

In process automation, cellular technologies can be used to collect measurement data
from remote locations via public cell-phone networks (Emerson, 2006a; Hoske, 2006).
This enables more cost-effective and up-to-date information about asset status than
sending a person to the site (Emerson, 2006a). In addition to providing primary network
connectivity, the cellular connections can serve as backup for industrial automation
devices, such as SCADA devices, programmable logic controllers (PLCs) and remote
terminal units (RTUs) (Hoske, 2006).

Compared to satellite connection, cellular networks have several advantages, namely

lower costs, existing standards and economics-of-scale (Emerson, 2006a). As a
disadvantage, it is reported that cellular technologies are not able to meet strict Quality of
Service (QoS) requirements needed for some real-time applications (Cungor and
Lambert, 2006). However, opposite opinions exist. (Liu and Jäntti, 2006) proposes a
hybrid communication scheme utilizing general packet radio service (GPRS) and circuit-
switched data call (CSD) to improve reliability performance requirement of the system.
The results implied that with the improvement scheme the reliability requirements could
be met using commercial cellular networks (Liu and Jäntti, 2006).
 10

2.7 Additional Standards

In addition to standards mentioned above, some other standards may be of concern when
developing wireless industrial applications. IEEE 1451, a family of Smart Transducer
Interface Standards, describes a set of open, common, network-independent
communication interfaces for connecting transducers (sensors or actuators) to
microprocessors, instrumentation systems, and control/field networks (IEEE 1451). For
example, wireless communication protocol standards (e.g. 802.11, 802.15.1, 802.15.4)
are being considered as some of the physical interfaces for IEEE P1451.5 (IEEE 1451).
The aim for IEEE P1451.5 is to make possible to get the same sensor data from the
wireless sensor implementing any of these three wireless protocols (IEEE 1451).

The International Electrotechnical Commission (IEC) 61508 standard for functional

safety in programmable electronic systems defines the functionality of safety networks.
The standard defines the requirements of a safety system to comply with the appropriate
safety integrity level (SIL). (VAN, 2005)

Table 1 presents the IEC61508 safety integrity levels. The top-most is used for critical
equipment, for example at nuclear power plants. SIL 3 is the highest level found in the
traditional manufacturing and process applications. (VAN, 2005)

Table 1. IEC 61508 safety integrity levels (VAN, 2005).

European standard EN 954-1 was developed for safety machinery and is titled “Safety of
machinery, Safety related parts of control systems” (VAN, 2005). For some wireless
applications, especially related to control, it may be relevant.

Since 2003, ATEX directive has become mandatory for all electrical and mechanical
equipment used in facilities with danger of exploitation. Any networked embedded
devices needs to comply with the directive (Koumpis, 2005).

ISA 100 committee is working on establishing standards (ISA100.11a), recommended

practices, technical reports, and related information that will define procedures for
implementing wireless systems in the automation and control environment with a focus
on the field level. The standard will likely adopt 6loWPAN (ABIresearch, 2007) and
Wireless HART technologies (Pinto, 2007).
 11

2.8 Technology Roadmap

Figure 5 presents a technology roadmap developed in RUNES project (see (RUNES)).
The roadmap was discussed in (Koumpis, 2005). As a conclusion, it was stated that:

• Adoption of wireless systems to industrial automation will take longer period than
in other sectors (e.g. building automation and control, medical care, disaster
response, automatic meter reading)
• Companies dealing with automotive, food process, petrochemical and asset
tracking will be the early adopters
• Adoption of wireless system to control applications will take considerably longer
than adoption to monitoring applications
• Monitoring of hazardous / inaccessible areas was given priority in short-term by
adopters of wireless technologies
• Many wireless systems in market today do not meet the local/national regulations,
either because they transmit too much power or operate on non-licence free
frequency area
• Any new networked embedded components should comply with the ATEX
directive

Figure 5. RUNES technology roadmap for industrial control and automation (Koumpis,
2005).
 12

In the following Section, a brief introduction to wireless sensor networks is presented.

Then, some characteristics for industrial environment with respect to wireless systems are
introduced followed by industrial applications. Finally, some technical considerations are
discussed.
 13

3 WIRELESS SENSOR NETWORKS IN INDUSTRIAL

ENVIRONMENT

3.1 Introduction to Wireless Sensor Networks

Wireless sensor networks (WSN) have been enabled by advancements in silicon chip
technology, low-power micro-controllers, chip-based radios, ad-hoc networking
protocols, and operating systems and languages for embedded systems. The WSN have a
wide range of applications. These can be divided to: (Culler et al., 2004)

• Monitoring space
• Monitoring things
• Monitoring interactions with things with each other and the encompassing
environment (monitoring complex interactions)

More detailed, the categories include, respectively: (Culler et al., 2004)

• Environmental and habitat monitoring, precision agriculture, indoor climate

control, surveillance, treaty verification and intelligent alarms
• Structural monitoring, ecophysiology, condition-based equipment maintenance,
medical diagnostics, and urban terrain mapping
• Wildlife habitats, disaster management, emergency response, ubiquitous
computing environments, asset tracking, healthcare, and manufacturing process
flow

The sensor nodes are physically small and they are typically provided with radio
transceivers, microcontrollers, and batteries (Low et al., 2005). The nodes include
wireless communications capability as well as sufficient computing resources for signal
processing and data transmission (Aakvaag et al., 2005).

A node in sensor network consists of a microcontroller, data storage, sensor, analogue-to-

digital converters (ADCs), a data transreceiver, controllers that tie the pieces together,
and an energy source. A simple microcontroller may operate at 1 mW/ 10 MHz. When
most of the circuits are turned off (in standby / sleep mode) the microcontroller power
consumption is typically about 1 μW. These low-power microcontrollers have limited
data storage, typically 10 kbytes of RAM for data and 100 kbytes of ROM for data
storage. However, designers typically incorporate larger amounts of flash storage, for
example a megabyte on a separate chip. (Culler et al., 2004)

Transreceivers are based on CMOS technology. The amount of energy needed to

communicate wirelessly increases rapidly with distance and obstructions further attenuate
the signal. WSN radios’ energy consumption is about 20 mW and their range is typically
measured in tens of meters. (Culler et al., 2004)
 14

Depending on the surrounding environment, communication in WSN can be

implemented in several ways (Aakvaag et al., 2005). Therefore, several network
topologies are possible: star, cluster-tree and mesh. In different topologies, sensor nodes
can act as a simple data transmitters and receivers or routers working in a multi-hop
fashion (Aakvaag et al., 2005). IEEE 802.15.4 standard considers star and mesh
topologies (Baronti et al., 2007). Figure 6 illustrates star, cluster-tree and mesh network
topologies applied in ZigBee network.

Figure 6. Different topologies in ZigBee network. Network topologies are general, but
personal area network (PAN) coordinator, full function devices (FFDs) and reduced
function devices (RFDs) are characteristic for ZigBee. FFD can act as a network
coordinator or an end-device. Therefore, FFD is capable of relaying data, providing
synchronisation or network join. The FFD can also take the role of personal area network
(PAN) coordinator, administering network services. RFD can act only as end device.
(Baronti et al., 2007)

The network tries to minimize the energy consumption either by eliminating

communications or turning off the radio, when communications do not occur. This can be
done either by: (Culler et al., 2004)

• Local processing of data in nodes

• Communicating, only if something of interest occurs
• Data aggregation (for example measuring average temperature for certain area
• Compression and scheduling
• Assigning certain tasks for special nodes
• Turning off radio, when uninteresting packet is received
 15

In addition to limited energy resources, several other issues must be addressed in order to
fully gain the advantage of WSN. These issues include: (Culler et al., 2004)

• Limited processing speed

• Limited storage capacity
• Limited communication bandwidth
• Wide variety in connectivity between nodes
• Harsh operating environments
• High degree of interaction between nodes (positive and negative)

The network must also be able to allocate limited hardware to multiple concurrent
activities, for example sampling sensors, processing, and streaming data. Also,
discovering the interconnections between the devices as well as routing data effectively
from where it is produced to where it is used, is required. (Culler et al., 2004)

Each of the aforementioned factors complicates networking protocols. However, despite

of the challenges, deploying and maintaining of the nodes must be inexpensive (Culler et
al., 2004). Especially, because of the large number of nodes expected in some
applications, the unit price of a node is an important parameter in the overall network cost
(Aakvaag et al., 2005): according to (Rabaey et al., 2000), the unit price of a node should
be less than 1 USD to achieve cost-effectiveness

3.2 Industrial Environment Characteristics

In industrial environments, the coverage area of WSN as well as the reliability of the data
may suffer from noise, co-channel interferences, multipath propagation and other
interferers (Low et al., 2005). For example, the signal strength may be severely affected
by the reflections from the walls (multipath propagation) (Werb and Sexton, 2005),
interferences from other devices using ISM bands (Werb and Sexton, 2005), and by the
noise generated from the equipments or heavy machinery (Low et al., 2005). In these
conditions, it is important that data integrity is maintained for operation-critical data, for
example alarms (Low et al., 2005).

Interference signals can be classified in two different categories, broadband and

narrowband. Broadband interference signals have constant energy spectrum over all
frequencies and high energy. They are usually emitted unintentionally from radiating
sources where as narrowband interference signals are intentional and have less energy.
Both interferences have varying type of degradation effect on wireless link reliability.
(Low et al., 2005)

Table 2 presents the general sources found in industry. Both broadband and narrowband
interferences are considered.
 16

Table 2. Interference sources. (Low et al., 2005)

Broadband interferences Narrowband interferences

Motors Cellular telephone
Inverters, Silicon-controlled
rectifier
(SCR) Circuits Radio and TV transmitters
Electric switch contacts Power-line hum
Computers, ESD Signal generators
Ignition systems Local oscillator, UPS system
Voltage regulators Test equipment
Lightning electromagnetic Microwave&Ultrasonic
pulses equipment
Arc/vapor lams Electronic ballasts
Pulse generators Medical equipment
Thermostats Microprocessor systems
Welding equipment Pager transmitters
Frequency converters High-frequency generators

Other types of interference sources are for example wide operating temperatures, strong
vibrations, and airborne contaminants. It is important to study and understand the radio
channel characteristics in order to predict the communications performance in these
operating conditions (Low et al., 2005). Moreover, radio modulation techniques can be
applied to reduce the interferences (Low et al., 2005).

Spread spectrum radio modulation techniques are applicable to reduce the interferences
because of their multiple access, anti-multipath fading and anti-jamming capabilities. The
two main spread spectrum techniques employed are direct sequence spread spectrum
(DSSS) and frequency hopping spread spectrum (FHSSS). They have different physical
mechanisms and therefore, react differently in industrial settings. The choice between
radio techniques is dependent on application (Low et al., 2005) and discussed more
detailed in Section 3.4.6.

3.3 Industrial Applications

3.3.1 Introduction

According to survey results presented in (Hoske, 2006), the leading application for
industrial networks (both wired and wireless) is supervisory control and data acquisition
(SCADA). It is followed by diagnostics, testing, maintenance; both continuous and batch
processing; motion control, robotic equipment; and machine control, CNC equipment
(graphic). Furthermore, the applications include pump, fan, blower applications;
continuous processing; packaging machines; materials handling equipment (elevators,
cranes, hoists); and discrete product manufacturing. The most used means of
communication were Ethernet TCP/IP, RS232 and 4-20 mA. Ten most used networks,
communications and protocols did not include wireless alternatives.
 17

However, as mentioned in Section 2.1, wireless technologies are reported to grow. A

categorization for wireless sensor network applications is presented in (Low et al., 2005):

• Rare event detection

• Periodic data collection
• Real-time data acquisition
• Control
• Industrial mobile robots
• Real-time inventory management

WSN can be deployed for example to bearings of motors, oil pumps, engines, vibration
sensors on packing crates, or many inaccessible or hazardous environments. For these
environments, the wired solution may be impractical due to e.g. isolation required for
cables running near to high humidity, magnetic field or high vibration environment. Also,
for applications that have frequent re-location, wireless solution may be more feasible.
(Low et al., 2005). APPENDIX 1 contains some application examples collected from the
literature.

Characteristic for industrial applications is the existence of legacy communication

systems (Shen et al., 2004). For example, fieldbuses are used to transport information
across manufacturing or process equipment (Aakvaag et al., 2005). Figure 7 presents an
example of such a communication scenario.

Figure 7. An example architecture of a plant automation system with wireless interface to

sensors and actuators (WISA) provided by ABB. Complexity of architecture may pose
challenges for managing the entity. (Virrankoski, 2007)
 18

Compared to wired solutions in industrial applications, the wireless systems and WSN
have several advantages. These include, for example:

• Flexibility in installing/upgrading network (Aakvaag et al., 2005)

• Reduced deployment and maintenance costs (Shen et al., 2004; Aakvaag et al.,
2005)
• Decentralization of automation functions (Aakvaag et al., 2005)
• Better coping with regulatory and safety obstacles in running cables in constricted
or dangerous areas (Shen et al., 2004
• Applicable for moving and rotating equipment (Low et al., 2005)
• Improved fault localization and isolation: for example critical tasks are often
ensured with redundant wires, which may pose difficulties for fault location and
isolation (Low et al., 2005).
• Incorporating short-range technologies to automation system (which has possible
interfaces to wide area networks) forms a heterogeneous network, which may
improve automation system efficiency (Low et al., 2005). Figure 8 presents a
heterogeneous system in industrial automation.
• Exploitation of micro-electromechanical systems (MEMS): integrated wireless
sensors with built-in communication capabilities offer a more robust design than
attaching wires to small-sized devices. (US Department of Energy, 2002)

Figure 8. Heterogeneous system in an automation application. Different communication

protocols are needed for based on wireless application scenario. Interworking of different
technologies must be taken into account. (Virrankoski, 2007)
 19

3.3.2 Classification of Applications

From industrial point of view, ISA SP100 workgroup introduced six classes for wireless
communications, which include applications from each of the aforementioned categories.
The classes were based on analysis of industrial, inter-device wireless communication
applications (ISA-SP100.11, 2006). The classes are presented in Figure 9.

Class 5 defines items related to monitoring without immediate operational consequences.

This class includes applications without strong timeliness requirements. The reliability
requirement may, however vary. Some, like sequence-of-events logs, require high
reliability; others, like reports of slowly-changing information of low economic value,
need not to be so reliable since loss of a few consecutive samples may not be important.
(ISA-SP100.11, 2006).

Class 4 defines monitoring with short-term operational consequences, which includes

high-limit and low-limit alarms and other information that may require further checking
or involvement of a maintenance technician. Timeliness of information in this class is
typically low (slow), measured in minutes or even hours. (ISA-SP100.11, 2006).

Class 3 covers open loop control applications, in which an operator, rather than a
machine, “closes the loop” between input and output. For example, an operator could
take a unit offline, if required. Timeliness for this class is human scale, measured in
seconds to minutes. (ISA-SP100.11, 2006)

Class 2 consists of closed loop supervisory control, in which applications usually have
long time constants, with timeliness of communications measured in seconds to minutes.
Examples in this category are batch unit and equipment selection. (ISA-SP100.11, 2006)

Class 1, closed loop regulatory control, includes motor and axis control as well as
primary flow and pressure control. The timeliness of information in this class is often
critical. (ISA-SP100.11, 2006)

Class 0 defines emergency actions related to safety, which are always critical to both
personnel and plant. Most safety functions are, and will be, carried out by dedicated
wired networks in order to limit both failure modes and vulnerability to external events or
attacks. Examples in this category are safety interlock, emergency shutdown, and fire
control. (ISA-SP100.11, 2006)
 20

Figure 9. Classification of industrial automation applications. (ISA-SP100.11, 2006)

Another way of classifying the communications in automation domain in general is

presented in (VAN, 2005). The classification is based on guaranteed response times of an
automation system and is as follows (additionally there is a “non real-time class”, which
has not been considered here):

• Class 1, soft real-time. This category has scalable cycle time. Systems of this class
are used in factory floor and process automation.
• Class 2, hard real-time: Cycle time in this category varies between 1…10 ms.
Systems of this class are used for control applications.
• Class 3, isochronous real-time. This category has time/clock synchronisation and
routing with time schedule. Cycle time varies from 250 µs…1 ms; jitter less than
1 µs. Systems of this class are used for motion control.

Either one of these classifications does not take into account all the real-time aspects
presented in Section 3.4.2.
 21

3.4 Design Considerations

3.4.1 Design Challenge

Developing a wireless sensor network for industrial applications requires a combination

of expertise from several different disciplines (VAN, 2005; Bonivento et al., 2006;
Howitt et al., 2006). According to (Howitt et al., 2006):

• Industrial expertise is essential for application domain specific knowledge.

• Sensor expertise is needed to understand issues associated with calibration, drift,
digitalization, sampling and transducers.
• Wireless device and RF environment expertise is needed to address issues for
technology suitability, electromagnetic compatibility, and electromagnetic
interference in industrial environments.
• Network expertise is required for understanding the complex hierarchical network
architectures which are required for industrial WSN.

3.4.2 Performance and Quality of Service

For industrial WSN, performance can be measured in terms of reliability, security,

throughput, latency, adaptability and affordability. In aggregate, these measures can be
used to define quality of service (QoS) (Howitt et al., 2006).

Reliability

Reliability addresses the ability of the network to carry out its functional requirement of
carrying industrial application information over the network over a broad range of
operational conditions (Howitt et al., 2006). Networks ability to deal with communication
disruptions, unanticipated variations in traffic and variations in the operational
environment should be considered (for issues causing disturbances refer to Section 3.2)
(Howitt et al., 2006). The system should be able to react to these disturbances using a
predictable”fail-safe” mechanism (Howitt et al., 2006), which is sometimes referred as
“robustness of design”.

As mentioned before, WSN are often energy constrained and therefore reliability includes
the ability of the network to maintain operational integrity under energy constraints such
as dirty power, battery operation, or energy scavenging (Howitt et al., 2006). Operational
power issues are discussed more detailed below.

Security

According to (Hamalainen et al., 2006), WSN share the common networking security
threats presented in (Stallings, 2005), namely message interception, message
modification, message fabrication, and interruption of communication and operation.
Furthermore, attacks can be carried out by malicious outsiders or insiders (Shi and Perrig,
2004; Karlof and Wagner, 2003). According to (Shi and Perrig, 2004), it is desired that
 22

outsider attacks are blocked and in case of insider attack, the security gracefully degrades
(is resilient). Different types of attacks are presented in Figure 10.

Figure 10. Different types of attacks in WSN: a) interruption b) modification c)

fabrication d) interruption (Hamalainen et al., 2006).

WSN have special characteristics that enable new ways of security attacks (Hamalainen
et al., 2006). Passive attacks are carried out by eavesdropping on transmissions and may
include for example traffic analysis or disclosure of message contents (Stallings, 2005).
Active attacks consist of modification, fabrication and interruption, which in WSN cases
may include node capturing, routing attacks, or flooding (Hamalainen et al., 2006). It
should also be noted, that even though WSN themselves are resource-limited, an attacker
may have more powerful equipment available, such as laptop or PC, for making attacking
more effective (Karlof and Wagner, 2003).

When designing the security of the wireless sensor network, both low-level (key
establishment and trust control, secrecy and authentication, privacy, robustness to
communication denial of service, secure routing, resilience to node capture) and high-
level (secure group management, intrusion detection, secure data aggregation) security
primitives should be addressed. It should also be noted, that due to resource limitations
some of the previously proposed solutions may not work due to resource limitations.
(Perrig et al., 2004)

As discussed in Section 3.3.1, WSN in industrial environments may be hybrid solutions.

This, in turn, may provide another challenge. (Roman et al., 2005) separates the security
of the wireless sensor network to two different categories. First, there is the security of
the data acquisition network, which is a collection of sensor nodes with the task of
measuring the physical data of its surroundings, and one or more base stations in charge
of collecting data from the nodes and forwarding control information from the users.
Second, there is the security of the data dissemination network which is a combination of
wired and wireless networks providing an interface of the data acquisition network to any
user.

A short review about some security technologies employed in wireless communications

can be found in (VAN, 2005). A good review about ZigBee security issues (and others) is
presented in (Baronti et al., 2007).

Throughput

Throughput defines the network’s ability to carry out the offered traffic by industrial
applications. It is impacted by the communication environment (Howitt et al., 2006).
 23

When considering data rate required for application, it should be noted that although
required bandwidth for sensor data may normally be low, dealing with multiple sensor
data in small time interval may cause bandwidth requirement to increase (Pellegrini et al.,
2006; Körber et al. 2007). Section 3.4.6 discusses the effect of different radio modulation
techniques and their effect to throughput.

Latency, Timeliness, Jitter

Latency defines the time delay from sensor node to base station. Latency requirements
are dependent on certain communication traffic, for example machine’s servo control
sensors have stricter requirements than sensors in supporting supply chain management
may have. Latency is influenced by sensor node density, interrogation rates, network
topology and number of simultaneous actions. (Howitt et al., 2006)

In addition to latency, other two concepts for defining real-time performance are
timeliness and jitter. Timeliness defines the absolute deadline for the actions to be
executed or completed. Jitter defines the time window in which action or co-ordinated
actions have to be executed or completed. In the latter case it is also a fault if an action is
completed earlier. Jitter can therefore be used to define requirement of synchronization.
Concepts of timeliness and jitter are illustrated in Figure 11. (VAN, 2005)

Figure 11. Concepts of timeliness and jitter. Time/utility functions are used to describe
the utility of executing and completing a certain action as a function of the point of time
when the action is executed and completed. The utility values describe the relative
importance of an action. (VAN, 2005)

Adaptability

A major advantage of WSN is its ability to adapt to new configurations or tasks (Howitt
et al., 2006). In industrial environment, adaptations may be required due to mobility of
some sensor elements and scaling and/or reconfiguring the WSN to handle new process
 24

or modification to an existing process (Howitt et al., 2006). Configuration of WSNs may

change also because of losses of key components, such as router nodes and gateways
(ISA-SP100.11, 2006).

Also, physical industrial environment may change over time, which in turn causes
changes to RF environment (Howitt et al., 2006). Additionally, there may be changes in
traffic flow due to changes in sensor location or process utilization (Howitt et al., 2006).
The WSN should be able to adapt to all these changes while maintaining the required
levels of throughput, latency, reliability and security (Howitt et al., 2006). Proposals to
improve WSN reliability especially in presence of changing RF environment can be
found e.g. from (Jiang et al., 2006; Howitt et al., 2006)
Affordability

Affordability includes all the costs of ownership. These incorporate packaging

requirements, modifications, maintainability, implementation costs, replacement and
logistics costs, and training and servicing costs as well as the per-unit costs. (Howitt et
al., 2006)

Coexistence

Interferences to the mission-critical data can have costly consequences in terms of

money, manpower, and even lives of employees and public (Low et al., 2005). Therefore,
the coexistence of WSN with the other systems operating in the same band should be
examined (ISA-SP100.11, 2006). The coexistence should be considered both from
system and to system –point of views, i.e. how does the system reduce its effects to other
systems or reduce effects from other systems and interference sources to itself (ISA-
SP100.11, 2006).

An example about coexistence impact analysis between 802.15.4 and 802.11b is

presented in (Howitt and Gutierrez, 2003). As a conclusion, it is proposed that 802.15.4
should have little or now impact to 802.11b .

3.4.3 Operational Power

Energy consumption is an important design constraint in battery-powered wireless

sensors. In order to save power, network nodes are kept in sleep mode for a significant
fraction of time. A synchronization algorithm is required to ensure simultaneous sleep
and awake times for nodes. The development environment should provide a MAC layer
access to node sleep/awake states in order to reduce additional stack delays and enable
precise synchronization. (Aakvaag et al., 2005)

Communication is the most energy intensive operation that the node performs (Culler et
al., 2004): packet transmitting, receiving and listening consume energy (ISA-SP100.11,
2006). For example, in (Körber et al., 2007), data rate (Mbit/s) and energy per bit -ratio
(µJ/bit) were used as selection criteria.
 25

However, the use of batteries can be troublesome due to their limited lifetime, making
periodic replacements unavoidable (Anton and Sodano, 2007). Energy scavenging
devices are used to capture the ambient energy surrounding the electronics and convert it
into usable electrical energy (Anton and Sodano, 2007). Some energy scavenging
techniques’ characteristics are presented in Table 3. A survey about power scavenging
systems is presented for example in (Paradiso and Starner, 2005).

Table 3. Characteristics of some energy scavenging techniques.(Körber et al., 2007)

3.4.4 Scalability and Modularity

WSN should be both scalable and modular, i.e., support of both small and large numbers
of sensors and actuators without affecting system performance (Körber et al., 2007). The
characteristics of different radio communication components of the system should be
carefully analysed (Ramamurthy et al., 2007). For example, Bluetooth can support only
up to seven connections per device in a piconet setup (Ramamurthy et al., 2007).
Additionally, when developing a security approach for wireless sensor networks, the
scalability of the solution is important (Shi and Perrig, 2004).

3.4.5 Topology

According to (Horton and Suh, 2005), mesh networks are more scalable, offer better
QoS, and it is shown analytically that they are more energy efficient than equivalent
single-hop networks (Horton and Suh, 2005). They also introduce redundancy in
topology, which makes them more robust against the failure of the single node, since
messages may be routed through alternative paths (Shen et al., 2004; Horton and Suh,
2005). In star topology, if a signal is blocked by physical or RF interference, the network
cannot recover, until the source of interference is removed (Shen et al., 2004).

However, (Körber et al., 2007) points out that energy awareness achieved by low or very
low duty cycles causes a trade-off between quality of service and node lifetime and with
respect to real-time applications this is not acceptable. Hence, (Körber et al., 2007)
proposes star topology for industrial real-time wireless sensor/actuator applications.
 26

A hybrid topology can combine the benefits of the both aforementioned networks by
combining efficiency and flexibility of star and mesh topologies (Shen et al., 2004). In
hybrid networks, star nodes are controlled by mesh nodes, therefore providing efficiency
and flexibility (Shen et al., 2004).

3.4.6 Radio

As mentioned in Section 3.2, radio technologies can be used to reduce interferences. For
this, spread spectrum radio modulation techniques are applicable because of their
multiple access, anti-multipath fading and anti-jamming capabilities (Low et al., 2005).
The two main spread spectrum techniques employed are direct sequence spread spectrum
(DSSS) and frequency hopping spread spectrum (FHSS) (Low et al., 2005). They have
different physical mechanisms and therefore, react differently in industrial settings (Low
et al., 2005). The choice between radio techniques is dependent on application and
environment (Low et al., 2005).

Due to the nature of frequency hopping in FHSS, other data packets still may reach the
receiver even if one packet fails because of strong narrowband interference. Additionally,
users may choose not to use certain frequencies, if there are known sources of
interference. Therefore, FHSS is suitable for harsh environments, because only part of the
spectrum is affected if packets jam at some hops. In general, FHSS scheme is less likely
to collide with other transmissions since changing of transmitting frequency makes
colliding in certain time frame more difficult than in DSSS scheme. (Low et al., 2005)

DSSS, for its part, can remove the interference completely, if the interfering signal power
is within the jamming margin, especially in case of low two medium narrowband
interference (Low et al., 2005). In case of heavy narrowband interference, though, FHSS
can still perform, but DSSS fails (Low et al., 2005). To summarize, DSSS is ideal for
systems that require high speed and are applied to environments with low or medium
narrowband interference, whereas FHSS is suitable for applications that require low cost,
lower data rate and applicability to high narrowband interference environment (Low et
al., 2005).

Radio modulation can be incorporated with other strategies to improve link quality in
industrial environments. These strategies include channel diversity, path diversity,
temporal diversity an increased transmit power. (Low et al., 2005) Discussion about
reduced radio link quality and countermeasures can be found in (Werb and Sexton, 2005)

The frequencies used for wireless communication in industrial automation are presented
in Table 4. The frequencies can be used also by other wireless applications, e.g. RFID
applications, WLAN applications or ISM applications. There is no guaranty for minimum
radio transmission quality. (VAN, 2005)
 27

Table 4. Frequencies used for wireless communication in industrial automation (VAN,

2005).

In (Körber et al., 2007), a selection of radio technology for complementary or substitional

system for a wired sensor/actuator bus is presented. A selected technology was a
compromise of design flexibility, availability, data rate, supportability for large group of
network nodes (at least 64), real-time capability, energy consumption (short wake-up
time) and radio characteristics. Interestingly, 802.15.4-based radio was not selected
because of it failed to provide support for guaranteed time slots for more than 7
nodes.(Körber et al., 2007)

Other issues to be considered are maximum transmission range and possible regulatory
issues concerning radio technology in question (ISA-SP100.11, 2006). Additionally,
antenna type and possible restrictions to it may be analyzed (ISA-SP100.11, 2006). An
open-field test for different antenna types was carried out in (Buckley, et al., 2006). It
was shown that the antenna performance is related directly to the size, type and
orientation of the antenna under test. The results for different antenna types are presented
in Table 5.

Table 5. Performance comparison of different antenna types. Good link threshold

distance (GLTD), is defined as a distance, where 65% of transmitted packets is received
successfully. (Buckley et al., 2006)
 28

3.4.7 System and Software

General Software and System Considerations

According to (Dzung et al., 2005), the system must require minimum operator
intervention. For WSN, this would mean a system that is self-organizing, self-configuring
and self-healing (Shen et al., 2004). (Culler et al., 2004) addresses this issue also: manual
configuration and managing of large WSN is impractical, and therefore, nodes must be
able to organize themselves. Additionally, it should be possible to administrate and
program the nodes as an ensemble rather than individual devices alone (Culler et al.,
2004). Additionally, the architecture should support existing (legacy) fixed-installation
systems and standards (in (Dzung et al., 2005), the smart sensors (IEEE1451) and
Ethernet were mentioned as example).

Application software should be accessible through a simple API and be easily customized
for both standards-based and customer-specific requirements enabling rapid development
and network deployment (Shen et al., 2004). Moreover, the WSN should integrate
seamlessly with the host PLC (Dzung et al., 2005) and legacy fieldbus systems (Shen et
al., 2004; Willig et al., 2005).

An application utilising existing wireless fieldbus architecture was presented in

(Pellegrini et al., 2006). In the proposed solution, the physical and data link layers were
taken directly from existing WLAN/WPAN protocols and the application layer was
adequately modified from the existing wired fieldbus system. Additionally, an adaptation
layer between the application and radio data link layers was required. Figure 12 presents
the architecture of the system bridged with wired fieldbus system. Wireless technologies
in industrial automation and fieldbus systems are discussed also in (Willig et al., 2005.
 29

Figure 12. Wireless fieldbus bridged with wired fieldbus. The physical and data link
layers of the wireless system are taken directly from existing WLANs / WPANs. The
wired fieldbus-like application layer is modified from existing fieldbus systems. The
adaptation layer adapts the application and data link layers.

After deployment in the field, network management and monitoring tools are essential
(Shen et al., 2004). A graphical user display could, for example, display network
connectivity and set node parameters (Shen et al., 2004). The management software
could provide whole network performance analysis and other management features, such
as, detecting failed nodes (e.g. for replacement), assigning sensing tasks, monitoring
network health, upgrading software and providing QoS provisioning(Shen et al., 2004).
Possibility of firmware updates should also be considered (ISA-SP100.11, 2006).

Operating System Considerations

According to (Horton and Suh, 2005), embedded software in nodes must have precise
control over their hardware and OS should allow high level applications of direct and
efficient control over low level HW, when necessary. An important property in this sense
is the access to medium access control (MAC) layer in order to enable proper time
synchronization, as mentioned in (Aakvaag et al., 2005). Without access to the MAC
layer, the synchronization had to be implemented to the network (NTW) layer, which
introduced an additional stack delay (Aakvaag et al., 2005). Taking into account the
application classes presented in Section 3.3.2 and real-time requirements presented in
Section 3.4.2, the additional stack delay may not be acceptable for all applications. The
other alternative would have been to keep nodes awake all the time, which would not
have been feasible in terms of energy consumption (Aakvaag et al., 2005).

A survey about MAC protocols can be found for example in (Demirkol et al., 2006).
Note, that it is mentioned there that latency, throughput, and bandwidth utilization may
be secondary in sensor networks, in industrial applications these maybe essential
properties, depending on application (see for example (Körber et al., 2007) for reference).
 30

An another example of where hardware control is needed is the power-saving feature.

There may be a need to control a sensor (such as a turn vibration sensor on/off), turn the
memory system off immediately after sending data or application may need direct access
to radio RSSI to make routing decisions. (Horton and Suh, 2005) suggests that traditional
layered abstraction for both networks and network stack leads to inefficiencies. Instead,
one should seek for the balance between the precise hardware control and the ease of
general purpose programming and efficiency (Horton and Suh, 2005).
 31

4 CASE EXAMPLE – WIRELESS MONITORING OF A

STEAM BOILER

4.1 Application
Section 4.1 is freely quoted from Juha Kemppainen: Master’s Thesis

A steam boiler produces steam for a laboratory-scale chemical pulp process. The steam
production process uses fuel oil and includes water storage tank, boiler, and pipelines.
Feed water temperature is approximately 20 °C and after the boiler, steam temperature is
approximately 200 °C. Four measurements are implemented to the steam boiler: three for
temperature and one for the steam pressure. The measurements are presented in Figure
13.

Figure 13. Steam boiler environment with measurements. T1, T2 and T3 are the
temperature measurements and P1 is the pressure measurement.

Temperature is measured from the flame in the combustion chamber (required measuring
range from the room temperature to approximately 1500 °C), from the combustion gas
pipe (from the room temperature to over 300 °C) and from the surface of the steam pipe
(from the room temperature to approximately 300 °C). The pressure, which normally is
approximately 13 bar, is measured from the bypass manifold. During shutdowns and
maintenance operations, however, the pressure may vary between 0-40 bars.

Lower temperatures from the combustion gas pipe and from the surface of the steam pipe
are measured by Pt100-sensors. Since measurements are located closely to each other, the
sensors are attached into one two-channel wireless transmitter node.
 32

The higher temperature from the flame in the combustion chamber is measured with the
S-type thermocouple and it has its own wireless transmitter node. For the thermocouple,
mains power is required for the transmitter head.

Additionally, the pressure sensor has its own wireless transmitter node and requires
mains power. Altogether, the equipment has three nodes and one gateway. The gateway
uses OPC interface, which passes the measurement information to the LabVIEWTM
development system. Test environment has potential sources of disturbances such as
thick cement walls, metal pipes, humidity and varying temperatures. The wireless sensor
network is presented in Figure 14.

Figure 14. The WSN and sensors applied for monitoring the temperatures and the
pressure of the steam boiler.

4.2 Use Case Scenario

A use case scenario was developed about the application example. The scenario is
presented in Table 6. The template for the use case scenario was obtained from Tapio
Frantti / VTT Technical Research Centre of Finland.

Table 6. Developed use case scenario.

 33

Narrative Description:
An operator needs to monitor steam pressure and temperature, flame temperature and combustion gas temperature of a
steam boiler. The monitoring is carried out both during operation and shutdown mode. During operation, the
measurements can be utilised to control process operating temperature and pressure as well as to detect abnormal
process conditions. Additionally, the data can be analysed for predictive maintenance purposes. During shutdown, the
ambient temperatures are monitored to detect fire and prevent equipment from freezing.

Operator uses wireless sensor network to monitor process parameters. The WSN can be connected to the host PC via
Ethernet/hub connection or over Local Area Network (LAN). The measured data is collected via OPC interface and
displayed as trends in PC’s monitoring software. Additionally, the data is stored for further analysis.

Use case:
Wireless_Monitoring_of_Temperature_And_Pressure_Of_A_Steam_Boiler
Description:
An operator monitors process parameters using wireless sensor network.
Actors:
-Operator (Primary)
-Wireless Sensor Network
-Local Area Network
Preconditions/Assumptions: -
Steps:
1. Configure the Wireless Sensor Network
1.1. Find out network structure
1.2.Configure connected measurements
1.3a)Configure Ethernet/hub connection between host PC and Bridge node
1.3b)Configure LAN connection between host PC and Bridge node
1.4 Start sending data via OPC interface

2. Display and store measured data in host PC

2.1 Start reading data via OPC interface
2.2 Display trends in monitoring application
2.3 Continuously store data for further use
2.4. Indicate abnormal situations for operator
Variations (optional):-
Non-functional requirements:
1. A/D-conversion should have accurate enough resolution
2. Node should have support for different types of sensor connections
3. Long battery life (high sampling rate may consume power significantly
4. Easy installation
Issues:-
Comments from the use case: External power was required for some sensors. Therefore, power wires were
needed.

4.3 Requirements
A requirement analysis was performed for the application example. The original set of
requirements was obtained from ESNA project. The requirements are presented in Table
7. Some requirements for the industrial real-time application were proposed to be added.
 34

The added requirements are presented in Table 8. A set of requirements can also be found
from (Willig, 2002).

Table 7. The original set of requirements.

Numerical
Requirement Verbal value Remarks
value

Type of data VALUES: event indication (on/off, something Slowly

1 meas / 2 min Shutdown
happened); slowly changing phenomenon such as temperature; Changing
6 meas/ 1 min Running
rapidly changing phenomenon; streaming data (vibration, sound) Phenomenon

Sensing and control VALUES: only need to sense data; both

Sensing
sensing and control

Data rate VALUES: very low (bytes / hour); low (bytes / minute);
Low to Medium
medium (bytes / second); high (kilobytes / second)

Data delivery rate VALUES: very low (1 / day) low (1 / hour) Medium to 0.5/minute to
medium (1 / minute) high (1 / second) High 1/10s

Sensing rate VALUES: very low (1 / day) low (1 / hour) medium

Medium
(1 / minute) high (1 / second) very high (1 / millisecond)

Acceptable latency(how long OK to wait for data) VALUES:

very high (day); high (hour); medium (minute), low (second); low 1-2 seconds
very low (millisecond)
Network scale (number of nodes) VALUES: 1-10; 10-100; 100-
1-10
1000; 1000-
Density (nodes / square meter) VALUES: 10 / m2 1 / m2 1 / 10
1/m2
m2 1 / 100 m2
Price sensitivity VALUES: very high (0.1€) high (1€) medium
low 100 €
(10€) low (100€) very low (1000€)
Localization precision VALUES: 0.01 m 0.1 m 1 m 10 m
Not applicable
100m

Harshness of environment(likeliness of hardware failure) Nodes might get

VALUES: rough (nodes might get destroyed) medium low rough destroyed by high
(controlled laboratory experiment) operating temperatures.

Expected lifetime VALUES: years; one season; months; days;

years
hours
Distance to monitoring entity VALUES: other side of the world;
meters
kilometers; meters
Mobility VALUES: local; global local
No process control loop,
Security VALUES: none; light; normal; heavy normal therefore no need for
high security.

self
connection,
Automatic system VALUES: self connection; self routing; self self routing,
adaptation; automatic fault recovering self adaptation,
automatic fault
recovering
 35

Table 8. Added proposals for requirements. When source is not marked in remarks,
requirement is written by author.

Numerical
Requirement Verbal value Remarks
value

Network setupe: need to

configure e.g. used ports
in nodes, sensor types
and update rates.
Management /Configuration SW needed VALUES: Yes; No Yes -
During operation: all
these may be altered.
Source: Shen et al., 2004

In example, firmware
Application development SW needed VALUES: Yes; No Yes -
updates: nice-to-have.

OPC compliance was

Compliance with existing process automation standards and essential for data
Yes -
interfaces required VALUES: Yes;No gathering.
Source: Shen et al., 2004

Includes Performance,
QoS and adaptability
QoS VALUES: Low / Medium / High Low - issues (Section XX.)
Source: Howitt et al.,
2006
Support for heterogeneous network architecture needed
Yes - TCP / IP
VALUES: Yes / No
Hardware Redundancy Needed VALUES: Yes / No No -

No e.g. Wi-Fi links near.

Coexistence Values: Yes / No No - Source: Körber et al.
2007

Interference mainly time-

variant (e.g. frequency
Harshness of RF environment VALUES: Low Interference / Medium to
- converters)
Medium Interference / High Interference High
Source: Werb and
Sexton, 2005
Source: ISA-SP100.11,
Support for Multiple Gateways Required? VALUES Yes;No No -
2006

Devices from one

Interoperability between devices (Nodes, Gateways) VALUES:
No - manufacturer.
Yes;No
Source: Low et al., 2005

Safety Integrity Level (IEC 61508) VALUES: SIL 1 / SIL 2 / SIL 3 /

SIL 1 - Source: VAN, 2005
SIL 4
Source: Virrankoski,
Data processing needed VALUES: Yes /No Yes -
2007
 36

5 DISCUSSION

5.1 General Considerations

Based on this review, it seems that there is a diverse, evolving field of wireless
technologies suitable for process automation communications. As indicated in Sections 2
and 3; and in APPENDIX 1, in process automation, the resulting systems could often be
hybrid by nature, incorporating existing automation systems and possibly heterogeneous
wireless communication schemes. Therefore, interoperability of current and future
wireless and wired applications may be a challenge, especially in the presence of rapidly
evolving new technologies, industrial standards and proprietary solutions available.

Considering the wireless technologies presented in Section 2, the existing industrial

automation wireless applications could be divided to two categories. First, a category, in
which wireless communications is used as a backbone in communication architecture. In
second category, wireless technology is applied to device level communication,
delivering data from sensor to higher level in network hierarchy.

Employing wireless technologies in several levels of plant network hierarchy – from

device level to remote connection – may introduce increased uncertainty in QoS
indicators as the whole system perspective must be considered. This would, in turn,
suggest distributed algorithms for some applications. For example, control applications
are sensitive for varying delay in the feedback loop. Applying such a loop over wireless
sensor network, wireless fieldbus and cellular link would provide increased uncertainties
in terms of timeliness, jitter and latency, making the delay in feedback loop possibly
more variable. Distributing the control to lower levels of network hierarchy might reduce
the uncertainty.

At device level, based on use cases presented in APPENDIX 1, it seems that most of the
current applications of wireless sensor networks in industrial automation have been
focused on condition-based maintenance. Control applications are rare; the focus is
mainly theoretical, although commercial application by ABB (wireless proximity
switches) could be found. Additionally, remote monitoring approaches with complex
network architectures, incorporating satellite and internet communications, could be
found. Taking into account the manufacturing process example at aluminium plant and
the event-based monitoring of relief valves, the applications support the proposed in the
literature. Additionally, as presented in (Koumpis et al., 2006), applications related to
monitoring are easier to find than those related in control.

Most of the wireless sensor network applications found during survey were based on
802.15.4-radio, with proprietary upper layer protocol stacks. For example, industrial
applications using ZigBee are hard to find, but this may simply be because of the novelty
of the technology. Quite recently, though, Mitsubishi announced a ZigBee-based WSN
application for plant monitoring. Interestingly, in some cases it was mentioned that
ZigBee is unable to cope with the requirements of industrial automation. (Dzung et al.,
 37

2005). and some discussion is ongoing about the ZigBee’s applicability to industrial
automation in internet publications. This suggests that the utilisation of ZigBee may also
be application-specific, and its characteristics should be carefully considered based on
requirements of application.

In general, it appears that most of the research in wireless sensor networks is based on
communications technology and some less QoS intensive and harsh domains, such as
environmental monitoring and building / home automation. Compared to that, relatively
few articles could be found about industrial applications. Additionally, it seems that quite
often the design of the applications starts from the network communications and wireless
devices; instead, in process automation, the design should start from the process-point-of-
view. One should consider e.g. process dynamics, required sampling rate (defined by the
Nyquist criteria) and required accuracy for measurements. Similar approach is suggested
also in (Bonivento et al., 2006).

5.2 Monitoring of the Steam Boiler: Experiences and Conclusions

Setting up the wireless sensor network for steam boiler monitoring was relatively
straightforward and required very little previous knowledge about WSN, or networking in
general. The selected network was self-organizing, so network was immediately
functional when management software was installed and devices were turned on. Clearly,
self-organization was considered as a beneficial and possibly mandatory requirement,
which probably eases the technology adoption for industrial customers. Also the
management software was considered to be very useful, at least in the setup phase and in
problem solving (later, own application was developed for monitoring and data
collection).

The only problems encountered in setup phase were related to Windows low-level
application interfaces. The language version mismatch caused errors, preventing OPC-
interface from working correctly. Later, another problem was detected: the host PC was
unable to receive the messages from WSN gateway, when extended to monitor over
LAN. The fault had not been previously experienced by the manufacturer’s customers, so
it might have been characteristic / configuration problem in the test environment. The
solution was to separate both the host PC and WSN to their own network segment.
However, this solution prevents the full usage of LAN, which may not be acceptable for
industrial clients. These two issues in relatively simple environment lend support for
importance of interoperability considerations.

Despite the wireless communication channel, lots of power wiring was needed. This was
because of the amplifying and signal processing required by the S-type thermocouple and
excitation power required by the pressure sensor. In order for the system to be truly
wireless, these power wires should be replaced by wireless alternatives. However, the
power need is beyond the capacity of current energy scavenging technologies. For
improved energy efficiency, also advances in sensor technology (e.g. adaptation of
MEMS to industrial measurements) to save power may be required. Indeed, when
 38

discussing the power usage and energy efficiency of the wireless system, one should
consider whole system, rather than just a piece of it.

As suggested in the literature, it was noticed, that when designing a wireless sensor
network for an application, the specification of the measurements (measuring ranges,
operating temperatures, etc.) and knowledge about process and sensors is essential. Based
on the requirements and knowledge, it is possible to draw requirements for the wireless
sensor network. For example, when selecting the system provider, it could be noticed,
that all the manufacturer’s did not have accurate enough (12-bit was required) A/D-
conversion for the pressure and temperature measurements needed to be implemented.

5.3 Future Work

Due to diversity of wireless technologies and hybrid nature of industrial implementations,
interoperability may be an important issue to be investigated. A possible solution to
tackle some interoperability problems could be the use of some design principles from
software defined radio (SDR) and reconfigurable systems. Traditionally, SDR is
considered as flexible end device reconfiguration by replacing radios implemented
completely in hardware by those that are reconfigurable or even reprogrammable in
software to a large extent. Reconfigurable systems extent the concept to take into account
required network support and reconfiguration of application and services. The use of
SDR in WSN has been proposed in (Ali et al., 2006) and in (Silverstrim et al., 2006).

It may also be useful to further deepen the knowledge about the different categories of
industrial applications (see Section 3.3.2). The classification of applications could be
supplemented for example by QoS issues and other relevant design criteria (presented in
Section 3.4). This classification might be useful in e.g. deriving requirements, designing
WSN architecture as well as selecting technology for an industrial automation
application.

In order to gain advantage of the additional information the WSN provide, data
processing methods may need to be addressed. Especially, data stream mining might be
applicable in dealing with large amount of sensor data. The applications used in data
processing should take into account the characteristics of the wireless sensor networks.

Based on the review, there seems to be some opposing opinions about interferences in
industrial environments. According to some authors, the interferences may cause
problems with wireless channel, while the others claim that these interferences are
negligible. Therefore, systematic field experiments, taking into account different
interference sources and proposed countermeasures might be needed.
 39

6 CONCLUSION

This review firstly presented shortly the emerging and employed wireless technologies in
process automation. Some standards, which may be relevant to industrial wireless
systems, were then introduced. Wireless sensor networks and interferences in industrial
environments, application classification and design criteria, were discussed a bit more
detailed. Finally, a case example utilised in studying WSN applicability for process
monitoring, was presented.

The review indicates that industrial implementations of wireless systems are often hybrid
by nature, incorporating existing (legacy systems) and heterogeneous networks. Wireless
technologies were used as a backbone in the network, or at device level, delivering data
from sensor to upper level in network hierarchy. Compared to other domains, wireless
industrial automation applications have stricter requirements for certain design criteria,
e.g. QoS. Additionally, process automation applications should be designed from the
process point-of-view, rather than the communication network.

Wireless sensor network could be utilised in monitoring of a steam boiler, although the
solution was not truly wireless, due to required power cabling. In order to enable truly
wireless sensor networks, the energy efficiency needs to be improved significantly, which
may require advancements in several technology areas, for example in sensor technology
and wireless communications. Interoperability showed to be an important design criteria
also in case example.

Based on review, it was proposed, that the in future work, interoperability in hybrid
networks and more detailed application classification taking into account the design
criteria of industrial applications should be addressed. Also a systematic understanding of
interferences in industrial environment should be acquired. Finally, the use of data
processing techniques, e.g. data stream mining, to take advantage of the information
provided by the WSN should be studied.
 40

REFERENCES

Aakvaag, N., M. Mathiesen and G. Thonet (2005). Timing and Power Issues in Wireless
Sensor Networks – an Industrial Test Case. In: Proceedings of the 2005 International
Conference on Parallel Processing Workshops, 419-426.

ABIresearch® (2007). Key Developing Industrial Wireless Sensor Networking Standard

Looks to IP. http://www.abiresearch.com/abiprdisplay.jsp?pressid=934 (30.10.2007).

Ailisto H., T. Matinmikko, J. Häikiö, A. Ylisaukko-oja, E. Strömmer, M. Hillukkala, A.

Wallin, E. Siira, A. Pöyry, V. Törmänen, T. Huomo, T. Tuikka, S. Leskinen and J.
Salonen (2007). Physical Browsing with NFC Technology.
http://www.vtt.fi/vtt_show_record.jsp?target=julk&form=sdefe&search=57624
(30.10.2007).

Akan, Ö.B. and I.F. Akyildiz (2005). Event-to-Sink Reliable Transport in Wireless
Sensor Networks. IEEE/ACM Transactions on Networking, 13, 1003-1016.

Akyildiz, I.F., W. Su, Y. Sankarasubramaniam and E. Cayirci (2002). A Survey on

Sensor Networks. IEEE Communications Magazine, 40, 102-114.

Ali, M., S. Umar, A. Dunkels, T. Voigt, K. Römer, K. Langendoen, J. Polastre and Z.A.
Uzmi (2006). Medium Access Control Issues in Sensor Networks. ACM SIGCOMM
Computer Communication Review, 36, 33-36.

Anton, S.R. and H.A. Sodano (2007). A Review of Power Harvesting Using Piezoelectric
Materials (2003-2006). Smart Materials and Structures, 16, 1-21.

Baronti, P., P. Pillai, V.W.C. Chook, S. Chessa, A. Gotta and Y.F Hu (2007). Wireless
Sensor Networks: A Survey on the State of the art and the 802.15.4 and ZigBee
Standards. Computer Communications, 30, 1655-1695.

Bonivento, A., L.P. Carloni and A. Sangiovanni-Vincentelli (2006). Platform-Based

Design of Wireless Sensor Networks for Industrial Applications. In: Proceedings of
Design, Automation and Test in Europe, 2006, 1, 1-6.

Buckley J., K. Aherne, B. O’Flynn, J.Barton, A. Murphy and C. O’Mathuana (2006).

Antenna Performance Measurements Using Wireless Sensor Networks. In: Proceedings
of the 56th Electronic Components and Technology Conference, 2006, 1652-1657.

Cisco. Capacity Coverage & Deployment Considerations for IEEE 802.11g.

http://www.cisco.com/en/US/products/hw/wireless/ps4570/products_white_paper09186a
00801d61a3.shtml (30.10.2007).
 41

Cleaveland, P. The Explosive Growth of Wireless Communications Requires Well-

Informed Decisions to Avoid Confusion. IMPO.
http://www.impomag.com/scripts/ShowPR~PUBCODE~032~ACCT~0000100~ISSUE~
0604~RELTYPE~PR~PRODCODE~0000~PRODLETT~A.asp (30.10.2007).

Culler, D., D. Estrin, and M. Srivastava (2004). Overview of Sensor Networks.

Computer, 37, 41-49.

Culler, D.E. and J. Hui. 6LoWPAN Tutorial.

http://www.archrock.com/downloads/resources/6LoWPAN-tutorial.pdf (30.10.2007).

Cungor, V.C. and F.C. Lambert (2006). A Survey on Communication Networks for
Electric System Automation. Computer Networks, 50, 877-897.

Demirkol, I., C. Ersoy and F. Alagoz (2006). MAC Protocols for Wireless Sensor
Networks: A Survey. IEEE Communications Magazine, 44, 115-121.

Dzung, D., C. Apneseth, J. Endersen and J.-E. Frey (2005). Design and Implementation
of a Real-Time Wireless Sensor/Actuator Communication System. In: 10th IEEE
Conference on, Emerging Technologies and Factory Automation, 2005, 443-442.

Emerson (2006a). Technology Selection.

http://plantweb.emersonprocess.com/university/Library_Downloadable_Courses.asp#wir
(30.10.2007).

Emerson (2006b). Wi-Fi Networks.

http://plantweb.emersonprocess.com/university/Library_Downloadable_Courses.asp#wir
(30.10.2007).

Eriksson L., and H.N. Koivo (2005). Tuning of Discrete-Time PID Controllers in Sensor
Network Based Control Systems. In: Proceedings 2005 IEEE International Symposium
on Computational Intelligence in Robotics and Automation, 359-364.

Gonia, P. and R. Budampati (2006). Wireless Network for Secure Industrial Applications.
http://www1.eere.energy.gov/industry/sensors_automation/pdfs/meetings/0606/hw-
wireless_gonia_0606.pdf (30.10.2007).

Hancke, G.P. and B. Allen (2006). Ultrawideband as an Industrial Wireless Solution.

Pervasive Computing, IEEE, 5, 78-85.

Horton, M. and J. Suh (2005). A Vision for Wireless Sensor Networks. In: 2005 IEEE
MTT-S International Microwave Symposium Digest, 361-364.

Hoske, M.T. (2006). Industrial Networks. Control Engineering,

http://www.controleng.com/article/CA6347555.html (19.10.2007).
 42

Howitt, I. and J.A. Gutierrez (2003). IEEE 802.15.4 Low Rate – Wireless Personal Area
Network Coexistence Issues. In: 2003 IEEE Wireless Communications and Networking,
3, 1481-1486.

Howitt, I., W.W. Manges, P.T. Kurugati, G. Allgood, J.A. Gutierrez and J.M. Conrad
(2006). Wireless Industrial Sensor Networks: Framework for QoS Assessment and QoS
Management. ISA Transactions, 45, 347-359.

Hu Y. (2001).: Satellite-Based Internet: A Tutorial. IEEE Communications Magazine, 39,

154-162.

Hämäläinen P., M. Kuorilehto, T. Alho, M. Hännikäinen and T.D. Hämäläinen (2006).

Security in Wireless Sensor Networks: Considerations and Experiments. In: Embedded
Computer Systems: Architectures, Modeling and Simulation, 167-177.

IEEE 1451. Introduction to IEEE P1451. http://ieee1451.nist.gov/ (30.10.2007)

IDG Communications (2005). GM Cuts the Cords to Cut Costs. Techworld.

http://www.techworld.com/mobility/features/index.cfm?featureid=1530 (30.10.2007).

IML Group plc (2007a). Wireless Networking in Process Automation: Building New
Applications in 2007. Control Engineering Europe, February 2007, 36-39.

IML Group plc (2007b). Wireless Hart Specification Released. Control Engineering
Europe, June/July 2007, 8.

Intel. Sensor Network Operation.

http://www.intel.com/research/exploratory/sensornetwork_operation.htm# (31.10.2007)

ISA-SP100.11 (2006). Call for Proposal: Wireless for Industrial Process Measurement
and Control. http://www.isa.org/filestore/ISASP100_11_CFP_14Jul06_Final.pdf
(30.10.2007).

Jiang, P., H. Ren, L. Zhang, Z. Wang and A. Xue (2006). Reliable Application of
Wireless Sensor Networks in Industrial Process Control. In: Proceedings of the 6th World
Congress on Intelligent Control and Automation, 99-103.

van de Kar, E., Z. Lukszo and G. Leus (2006). Wireless Networks in the Process
Industry: Opportunities for Ultrawideband Applications. In: Proceedings of the 2006
IEEE International Conference on Networking, Sensing and Control, 2006, 607-612.

Karlof, C. and D. Wagner (2003). Secure Routing in Wireless Sensor Networks: Attacs
and Countermeasures. In: Proceedings of the 2003 IEEE International Workshop on
Sensor Network Protocols and Applications, 113-127.
 43

Kinnunen, L. (2007). Wireless Automation Provides both Opportunities and Challenges.

Automaatioväylä, 23, 26-28.

King, J. (2005). BP Pioneers Large Scale Use of Wireless Sensor Networks.

Computerworld, http://www.computerworld.com/newsletter/0,4902,99962,00.html
(19.10.2007).

Koumpis, C. (2006). Wireless in Industrial Applications.

http://host.quid5.net/~koumpis/talks/2006-10-26_IET_Reading.pdf (30.10.2007).

Koumpis, K., L. Hanna, M. Andersson and M. Johansson (2005). Wireless Industrial

Control and Monitoring Beyond Cable Replacement, Profibus International Conference,
June 2005, Coombe Abbey, Warwickshire, 1-7.

Körber H.-J., H. Wattar and G. Scholl (2007). Modular Wireless Real-Time

Sensor/Actuator Network for Factory Automation Applications. IEEE Transactions on
Industrial Informatics, 3, 111-119.

Liu, Y. and R. Jäntti (2006). A Study Towards Enhanced Reliability Performance of

Remote Control and Monitoring Application Over Commercial Wireless Communication
Networks. In: International Conference on Wireless Communications, Networking and
Mobile Computing, 2006, 1-4.

Low, K.S., W.N.N Win, and J.E. Meng (2005). Wireless Sensor Networks for Industrial
Environments. In: International Conference on Computational Modelling, Control and
Automation, 2005 and International Conference on Intelligent Agents, Web Technologies,
and Internet Commerce, 2, 271-276.

Mannion, P. (2006). Handset Forms Sensor Gateway. Electronic Engineering Times,

http://www.eetonline.com/issue/fp/showArticle.jhtml;jsessionid=WUC44CIYYW0N4QS
NDLRCKH0CJUNN2JVN?articleID=193105229&_requestid=912996 (30.10.2007).

Millenial Net. Wireless Networking for Remote Monitoring.

http://www.millennial.net/industries/ia_remotemonitoring.php (31.10.2007)

Paradiso, J.A. and T. Starner (2005). Energy Scavenging for Mobile and Wireless
Electronics. Pervasive Computing, 4, 18-27.

de Pellegrini, F., D. Miorandi, S. Vitturi and A. Zanella (2006). On the Use of Wireless
Networks at Low Level of Factory Automation Systems. IEEE Transactions on
Industrial Informatics, 2, 129-143.

Perrig, A., J. Stankovich and D. Wagner (2004). Security in Wireless Sensor Networks.
Communications of the ACM, 47, 53-57.
 44

Pinto, J. (2007). Industrial Wireless Standards.

http://www.isa.org/Content/ContentGroups/News/20071/August37/Industrial_wireless_st
andards.htm (30.10.2007).

Porcino, D. and W. Hirt (2003). Ultra-Wideband Radio Technology: Potential and

Challenges Ahead. IEEE Communications Magazine, 41, 66-74.

Rabaey J., J. Ammer, J.L. da Silva Jr. and D. Patel (2000). PicoRadio: Ad-hoc Wireless
Networking of Ubiquitous Low-Energy Sensor/Monitor Nodes. In: Proceedings of the
IEEE Computer Society Workshop on VLSI, 2000, 9-12.

Ramamurthy H., B.S. Prabhu, R. R. Gadh and A.M. Madni (2007). Wireless Industrial
Monitoring and Control Using a Smart Sensor Platform. IEEE Sensors Journal, 7, 611-
618.

Roman, R., J. Zhou and J. Lopez (2005). On the Security of Wireless Sensor Networks.
In: Computational Science and Its Applications – ICCSA 2005, 681-690.

RUNES. http://www.ist-runes.org/introduction.html (29.10.2007).

Sallinen, M., E. Strömmer and K. Känsälä (2007). Lyhyen kantaman tiedonsiirto

mullistaa teollisuuden tiedonsiirron. Automaatioväylä, 23, 32-35.

Sensicast (a). Montoring Giant Furnaces.

http://www.sensicast.com/uploadedFiles/Resource_Center/Monitoring_Giant_Furnaces.p
df (31.10.2007).

Sensicast (b). Sensinet® Improves Efficiency in Compressed Air Systems While

Reducing Operating Costs in Manufacturing Facilities.
http://www.sensicast.com/uploadedFiles/hv.case.study.ver.6bb_casestudy.pdf
(30.10.2007).

Sensicast (c). Sensinet® Wireless Temperature Monitoring Solutions Increases

Throughput in Aluminium Plant. http://www.sensicast.com/uploadedFiles/CS-
Alum.10.06_casestudy.pdf (31.10.2007).

Sereiko, P. (2007). Building the Perfect Beast.

http://www.isa.org/InTechTemplate.cfm?Section=Article_Index1&template=/ContentMa
nagement/ContentDisplay.cfm&ContentID=59875m (31.10.2007)

Sifakis E. (2006). Routing Protocols for Wireless Networked Sensing and Control
Systems. http://etd.nd.edu/ETD-db/theses/available/etd-05252006-
114853/unrestricted/SifakisE052006.pdf (31.10.2007)
 45

Shen, X., Z. Wang and Y. Sun (2004). Wireless Sensor Networks for Industrial
Applications. In: Fifth World Congress on Intelligent Control and Automation, 2004, 4,
3636-3640.

Shi, E. and A. Perrig (2004). Designing Secure Sensor Networks. IEEE Wireless
Communications, 11, 38-43.

Silverstrim, J.E., W.E. Holland and K. Colling (2006). Method and Apparatus for Multi-
Waveform Wireless Sensor Network. http://www.patentstorm.us/patents/7119676-
claims.html (30.10.2007).

Sun, Y., Y.Q. Bian, P. Strauch and A.R. Nix (2007). Study of Radio Resource Sharing
for Future Mobile WiMAX Applications with Relays. In: IEEE Mobile WiMAX
Symposium, 2007, 91-97.

Stallings, W. (1995). Network and Internetwork Security: Principles and Practice.

Prentice-Hall, USA (1995).

US Department of Energy (2002). Industrial Wireless Technology for the 21st century.
http://www.energetics.com/rep_products.asp?Product=90 (30.10.2007).

VAN (2005). State of the Art and Trends in Safety, Security, Wireless Technologies and
Real-Time Properties. http://www.van-eu.eu/deliverables (29.10.2007).

Virrankoski R. (2007). Langaton tiedonsiirto tehdasautomaatiossa.

http://www.smartsystems.fi/konferenssimateriaalit.html (30.10.2007)

Welander, P. (2007). Industrial Networking. Control Engineering, 54, 59.

http://proquest.umi.com/pqdlink?Ver=1&Exp=07-09-
2012&FMT=7&DID=1255353221&RQT=309&cfc=1 (30.10.2007).

Werb, J. and D. Sexton (2005). Improved Quality of Service in IEEE 802.15.4 Mesh
Networks. International Workshop on Wireless and Industrial Automation, San
Fransisco, California, March 7, 2005, 1-6.

Wiberg, P.-A. and U. Bilstrup (2001). Wireless Technology in Industry – Applications

and User Scenarios. In: Proceedings of the 2001 8th IEEE International Conference on
Emerging Technologies and Factory Automation (2001), 1, 123-131.

Willig, A. (2002). Investigations on MAC and Link Layer for a Wireless PROFIBUS
over IEEE 802.11. http://edocs.tu-berlin.de/diss/2002/willig_andreas.pdf (29.10.2007).

Willig, A., K. Matheus and A. Wolisz (2005). Wireless Technology in Industrial

Networks. In: Proceedings of the IEEE, 93, 1130-1151.
 46

Yeager, D.J., A.P. Sample, J.R. Smith, P.S. Powledge and A.V. Mamishev (2006).
Sensor Applications in RFID Technology. In: International Conference on Actual
Problems of Electron Devices Engineering, 449-452.

Zeng, W., H. Wang , H. Yu and A. Xu (2006). The Research and Application of UWB
Based Industrial Network. In: The Third Industrial Conference on Ultrawideband and
Ultrashort Impulse Signals, 153-155.
 APPENDIX 1(1)

Condition-Based Maintenance, General Use Case

Narrative Description:
Physical structure, such as machine, motor, airplane wing, bridge or building has typical modes of vibration, acoustic
emissions and response to stimuli. Variations in these behaviours can indicate wear, fatigue, or other mechanical
challenges. For example, a bearing will often squeak and shudder before it seizes up. (Culler et al., 2004)

Often, an operator is required to gather data manually/with a computing device from the sensors attached to production
equipment. The logs are then carried out to central computer, which analyses the data and e.g. for the signs of wear.
The time between visits at a particular machine may be months. WSN offers alternative approach to condition
monitoring: performing local processing at each device and transporting the data continuously to operations staff.
(Culler et al., 2004)

Some example sampling rates for condition monitoring: about 100 Hz for vibration analysis and several kHz for
acoustic analysis, which are considerably higher than in environmental monitoring. Higher sampling rate requires more
energetic sampling activity and greater care in how network performs sampling. Buffering data needs greater amounts
of storage and the system may perform more extensive local processing on data chunks sampled intermittently. (Culler
et al., 2004)

Instead of transmitting large amount of raw data, the sensor nodes can perform signal analysis, communicating only the
modes of vibration and anomalies. Sensor nodes could also monitor control networks to find out what is active when
the sample is taken or even to determine when to sample. Because reducing costs of obtaining and processing data
reduces overall costs, improving timeliness of analysis can improve plant performance (Culler et al., 2004).
Use case:
Condition_Based_Maintenance_General
Description:
Preventive maintenance is carried out to reduce unwanted downtime by condition-based maintenance.
Sources: (Culler et al., 2004)
Actors:
Operator (Primary), Wireless Sensor Network, Local Area Network, Production Equipment
Preconditions/Assumptions:
Steps:
a)
1. Measure defined physical stimuli from production equipment
2. Send data continuously to sink
3. Transfer data e.g. to control room PC over LAN
4. Display data in monitoring application.
5. Analyze data for anomalies.
b)
1. Measure defined physical stimuli from production equipment
2. Analyze data locally
3. If an anomaly is detected, send alert (and data) to sink
4. Transfer data e.g. to control room PC over LAN
5. Display alert (and data) in monitoring
Variations (optional):
Non-functional requirements:
1. External power may be required for some sensors
2. A/D-conversion should have accurate enough resolution
3. Node should have support for different types of sensor connections
4. Long battery life (high sampling rate may consume power significantly)
Issues:
Comments from the use case:
 APPENDIX 1(2)

Condition-Based Maintenance, Monitoring Compressed Air System

Narrative Description
Compressed air systems degrade over time and tend to come leaky and inefficient, which is compensated by increasing
line pressure. If the inspections are run seldom (according to (Sensicast, (b)), typically once a year), the system works
inefficiently and use excessive electricity for long periods of time. (Sensicast, b)

Wireless Sensor Network has been implemented to monitor compressed air system in a paper plant /Sensicast/.
Monitored parameters were for example temperature, line pressure, airflow, and compressor energy usage. Monitoring
software was used to display data and detect trends leading to system degradation. The monitoring and reporting
interval was set to one minute. Hence, knowledge about small changes in system was available accurately. The system
was reported to increase compressed air systems efficiency by 50%. (Sensicast, b)

Use case:
Monitoring_Compressed_Air_System
Description:
Increase compressed air system’s efficiency by monitoring operating parameters and detecting degradation.
Source: (Sensicast, b)
Actors:
Operator, Wireless Sensor Network, Compressed Air System

Preconditions/Assumptions:
Steps:
1. Measure defined physical stimuli from production equipment
2. Send data continuously to sink
3. Display data in monitoring application.
4. Analyze data for anomalies.

Variations (optional):
Non-functional requirements:
1. External power may be required for some sensors
2. A/D-conversion should have accurate enough resolution
3. Node should have support for different types of sensor connections
4. Long battery life (high sampling rate may consume power significantly)
Issues:
Comments from the use case:
 APPENDIX 1(3)

Condition-Based Maintenance, Vibration Monitoring of a Robot in a Manufacturing Cell

Narrative Description
A manufacturing cell is a stage of automation line in an industrial plant. In this scenario, robots equipped with drilling
tips work with metal surfaces. Monitoring the state of the drilling tip is considered to be important because they wear
out and need to be replaced before damaging both the piece under construction and the robot itself. (Bonivento et al.,
2006)

Typically, the monitoring is carried out by different vibration sensors placed on the robot. If a certain threshold of
vibration intensity is exceeded, the equipment is stopped in order to safely perform maintenance. The monitoring task is
carried out by a PLC unit, which queries periodically each cluster for information on the vibration intensity for the
robots. (Bonivento et al., 2006)
Use case:
Vibration_Monitoring_Manufacturing_Cell
Description:
Detect tool wear by monitoring vibration in production equipment.
Sources: (Bonivento et al., 2006)
Actors:
Operator (Primary), Wireless Sensor Network, Production Equipment
Preconditions/Assumptions:
Steps:
1. PLC queries WSN for vibration patterns
2. WSN measures vibration patterns
3. WSN sends information to PLC
4. PLC: Anomaly detected?
4.1 No Î Continue production
4.2 Yes Î Stop production equipment
Variations (optional):
Non-functional requirements:
1. External power may be required for some sensors
2. A/D-conversion should have accurate enough resolution
3. Node should have support for different types of sensor connections
4. Long battery life (high sampling rate may consume power significantly)
Issues:
Comments from the use case:
 APPENDIX 1(4)

Condition-Based Maintenance, Detecting Vibration in Water Purication Equipment

Narrative Description
A trial deployment of a wireless sensor network to monitor the health of semiconductor fabrication
equipment is presented in (Intel). Specifically, the network monitors the vibration signature of water purification
equipment, providing data for preventive maintenance operations.

Currently manual monitoring is used to predict failures and schedule maintenance or replacement to avoid costly
manufacturing downtime. Deploying wireless sensor networks would provide more frequent and more reliable data,
reduce equipment downtime and eliminate costly manual equipment monitoring. (Intel)
Use case:
Vibration_Monitoring_Water_Purification
Description:
Monitor vibration signature of water purification equipment for preventive maintenance operations.
Sources: (Intel)
Actors:
Operator (Primary), Wireless Sensor Network, Production Equipment
Preconditions/Assumptions:
Steps:
1. Query for vibration patterns to WSN
2. WSN measures vibration patterns
3. WSN sends information to sink
4. Anomaly detected?
4.1 No Î Continue purification
4.2 Yes Î Perform maintenance activities
Variations (optional):
Non-functional requirements:
1. External power may be required for some sensors
2. A/D-conversion should have accurate enough resolution
3. Node should have support for different types of sensor connections
4. Long battery life (high sampling rate may consume power significantly)
Issues:
Comments from the use case:
 APPENDIX 1(5)

Condition-Based Maintenance, Monitoring Steel Furnace

Narrative Description:
In /Sensifur/, a wireless sensor network was used to monitor temperature fluctuations in water jackets of steel furnaces.
In the monitoring environment, there are extreme temperatures and harsh chemical, metallurgical, and mechanical
processes. At some point of the steel making process, electrical current often arcs to the sides of the furnaces, heating
its walls and sometimes overloading its temperature control mechanism. If the cooling system would fail, the
production equipment as well as the lives of the operating personnel would be endangered. (Sensicast, a)

Wireless solution was selected, because the installation of the sensors had to be carried out quickly, and high heat and
difficult positioning made the use of wired sensors impossible. Temperature sensing nodes were installed between the
inner and outer walls of the furnace. They were reported to be able to withstand the heat from melting steel, wide
temperature variations, powerful magnetic fields, water spray, and vibration. (Sensicast, a)

The wireless technology routes data around any temporary obstacles that might occur. Frequency hopping was
employed to select the best channel for clear communication. OPC interface was utilised in delivering data directly to
the control system. (Sensicast, a)
Use case:
Monitoring_Steel_Furnace
Description:
A wireless sensor network is used to monitor the cooling system of the steel furnace.
Source: (Sensicast, a)
Actors:
Operator (Primary), Wireless Sensor Network, Plant’s control system
Preconditions/Assumptions:
Steps:
1. Measure temperature data
2. Try to send temperature data to sink
2.1 a) If sending fails, try another route / channel
2.1 b) If sending was successful, continue monitoring task
Variations (optional):
Non-functional requirements:
1. External power may be required for some sensors
2. A/D-conversion should have accurate enough resolution
3. Node should have support for different types of sensor connections
4. Long battery life (high sampling rate may consume power significantly)
5. Perform reliably in extremely harsh conditions (strong magnetic and electrical fields, large masses of steel
interrupting communications, wide temperature variations, water sprays, vibrations)
Issues:
Comments from the use case:
 APPENDIX 1(6)

Condition-Based Maintenance and Diagnostics, Pulp and Paper Mills

Narrative Description
An example about using wireless sensor networks for diagnostic and preventative maintenance purposes at pulp and
paper mills is presented in (Shen et al., 2004). The rolling machines at the mills are massive, complex mechanisms.
Even small variations in the speed, temperature or alignment of the rollers can have serious effects on quality or
operation. (Shen et al., 2004)

WSN could be utilised in investigating and resolving unanticipated circumstances such as variations in output quality,
unusual vibration or noise, or other signs for potential problems. Examples about measured quantities are temperature,
speed, pressure, and vibration. Technicians could attach the required sensor nodes to the areas of interest in the machine
or process line. The nodes could use wireless RF communications to relay data to portable or handheld device. The
wireless measurement system should be quickly installed and rapidly removed once problems are identified and
resolved. (Shen et al., 2004)
Use case:
Pulp_And_Paper_Mills

Description:
WSN is utilized to solve quality problems and for preventive maintenance operations at pulp and paper mills.
Source: (Shen et al., 2004)
Actors:
Operator (Primary), Wireless Sensor Network, Portable / Handheld Device, Production Equipment
Preconditions/Assumptions:
Steps:
1. Attach nodes to area of interest
2. Form wireless sensor network
3. Measure specified quantity
4. Transmit data to portable / handheld device
Variations (optional):
Non-functional requirements:
1. External power may be required for some sensors
2. A/D-conversion should have accurate enough resolution
3. Node should have support for different types of sensor connections
4. Long battery life (high sampling rate may consume power significantly)
5. Rapid installation and removal of WSN
Issues:
Comments from the use case:
 APPENDIX 1(7)

Condition Based Maintenance, Monitoring Pump Wear

Narrative Description
In (Aakvaag et al., 2005), an experimental network developed by ABB and Metso is presented. The WSN was
implemented at a mining company, Metso supplying automation equipment and ABB delivering the automation
system. At the facility, there are 200 pumps, which were reported to have uneven and unpredictable wear of rubber
parts. On average, a wear part had to be replaced three times a year. Metso was working on “smart wear parts”, which
would detect tear and wear in rubber and then deliver acquired information from pumps to central location for
integration. (Aakvaag et al., 2005)

The wear sensing unit consisted of two parts: RFID tag and a radio-equipped unit. The RFID tag would transmit pump
identification and wear data across to radio unit. (Aakvaag et al., 2005)

The, developed solution, EmberNet, was a multi-hop network using an ad-hoc on-demand distance routing (AODV)
algorithm. The network consisted of seven nodes, which would exchange data once per minute, and a gateway node.
The application programming interface API supported two types of packets, broadcast and unicast, from which
broadcast was used. The broadcast method sent packet three times (time between retransmissions was 100 ms) in order
to ensure successful transmission (the number of retransmissions could not be altered, which was a limitation for the
system). Maximum radius (maximum number of radio hops that a packet can propagate from originator) was utilised to
prevent network from flooding. (Aakvaag et al., 2005)
Use case:
Monitoring_Pump_Wear
Description:
Wireless Sensor Network is used to detect wear and tear in rubber parts of pumps.
Sources: (Aakvaag et al., 2005)
Actors:
Operator (Primary), WSN, Production Equipment
Preconditions/Assumptions:
Steps:
1. Measure wear data
2. Transmit data to gateway node 3 times (broadcast)
3. Go to sleep mode, wait for the next measurement period.
Variations (optional):
Non-functional requirements:
1. A/D-conversion should have accurate enough resolution
2. Node should have support for different types of sensor connections
3. Long battery life (high sampling rate may consume power significantly)
4. Rapid installation and removal of WSN
Issues:
Comments from the use case:
 APPENDIX 1(8)

Control Applications, Wireless Networked Control Systems

Narrative Description
Wireless Networked Control Systems, WNCS, consists of sensors, actuators and controllers, which are connected over
a wireless communication channel. The sensing nodes take measurements related to a phenomenon and route these
measurements to the actuator/controller; second, actuators and controllers in charge of receiving sensed data and
sending a control command to the plant. The system must satisfy timeliness requirements in order to avoid instabilities.
Compared to wireless sensor networks, they usually require sending more data packets of smaller size. However, in
some cases, the time frame for sending packets may be small, which may increase periodic data rate. Additionally,
wireless sensor networks usually have one base station or data sink while wireless networked control systems have
many actuators/controllers which behave like data sinks. (Sifakis, 2006)

Wireless communication channel for control applications brings new challenges in the form of higher bit error rates and
communication latencies. These characteristics may lead to decreased phase margins, which can destabilize the system.
(Sifakis, 2006)
Use case:
Wireless_Network_Control_Systems
Description:
Wireless Networked Control Systems perform sensing and control over wireless communication channel.
Sources: (Sifakis, 2006)
Actors:
Controller (Primary), WNCS, Plant Equipment
Preconditions/Assumptions:
Steps:

Variations (optional):
Non-functional requirements:
Issues:
Comments from the use case:
Steps not included due to general nature of use case.
 APPENDIX 1(10)

Control Applications, A PID Controller in a Feedback Loop

Narrative Description
In (Erikkson and Koivo, 2005), a new tuning method is developed for discrete-time PID controllers in networked
control systems where time-varying delays are present. The control system is comprised of a process, a discrete-time
PID controller, an actuator and a wireless sensor network.

Although many processes have multiple inputs and multiple outputs (MIMO), they are often controlled with single
input-single output (SISO) controllers. Therefore, data fusion needs to be applied over multiple sensor readings.
(Erikkson and Koivo, 2005)

The use of sensor networks in process control is based on fact that there is no need for expensive cabling of sensors,
and free mobility. There is also more freedom if the monitored process is far away from the controller or if the process
is very wide, and it is difficult or impossible to use wired coupling. However, multiple sensors are required for
redundancy and for guaranteeing reliable operation of the whole measurement system. (Erikkson and Koivo, 2005)
Use case:
PID_Control_System_In_Wireless_Feedback_Loop

Description:
PID control system is used in wireless feedback loop
Source: (Erikkson and Koivo, 2005)
Actors:
Controller(Primary), Wireless Networked Control System, Process, Actuator
Preconditions/Assumptions:
Steps:
1. Data is acquired from multiple sensors.
2. Data is transmitted over wireless communication channel to e.g. sink/host PC.
3. Data fusion is carried out to fuse multiple sensor readings.
4. Fused data is transmitted to controller over wireless channel.
4. Controller calculates control signal and transmits the signal to actuator using wireless communication channel.
Variations (optional):
Only one potential case of multiple sensors described in steps.
Non-functional requirements:
1. Low battery consumption
2. QoS (strict with timeliness, jitter Î variable delays cause instability in control loop)
3. Sensor nodes should have support for different sensor connections

Issues:
Comments from the use case:
 APPENDIX 1(10)

Control Applications, Wireless Sensor/Actuator Network in Robotic Installation

Narrative Description
(Dzung et al., 2005) describes design and implementation of a real-time wireless sensor/actuator communication
system. Requirements for industrial automation applications using production machines and robotic installations are
also presented (in addition to the general requirements of high reliability and low and predicable data transfer): (Dzung
et al., 2005)

• Fast response time (generally less than 10-15 ms for real-time applications)
• Serve large number (up to 120) of sensors/actuators located in a cell of several meters radius
• guarantee high reliability of data transmission even in unfriendly environment of factories, where radio
propagation may be affected by many obstacles and where various sources of interfering signals must be
expected
• have low power consumption

Fast response times were achieved. The system was tested in factory environment and products based on it are
commercially available (ABB, wireless proximity switches). According to authors, the design criteria related to
response times as well as number of sensors / actuators differ by orders of magnitude from the design goals of existing
wireless systems such as Bluetooth or ZigBee. (Dzung et al., 2005)
Use case:
WSAN_Production_Machines_Robotic_Installations
Description:
Wireless Sensor and actuator network (WSAN) is utilised in real-time wireless sensor/actuator communication system
(wireless proximity switches).
Source: (Dzung et al., 2005)
Actors:
Controller (primary), WSAN, Production Equipment
Preconditions/Assumptions:
Steps:
1. Node transmits proximity data from to base station (BS)
2. Base station maps the wireless links to fieldbus addresses and connects links to automation controller (PLC)
3. Base station sends the data to PLC
4. PLC sends control command to BS (e.g. based on data)
6. Control command is send to actuator from BS (wirelessly)
Variations (optional):
Non-functional requirements:
1. A/D-conversion should have accurate enough resolution
2. Node should have support for different types of sensor connections
3. Fast response times
4. High number of nodes to be served
5. High reliability
6. Low power consumption
Issues:
Comments from the use case:
 APPENDIX 1(11)

Remote Monitoring, Monitoring For Tank Levels of Natural Gas By-Products

Narrative Description
The process of extracting natural gas produces by-products, such as such as hydrocarbons, oil and water, which are
stored in tanks in the gas fields and trucked off site on a scheduled basis. Utilising wireless communication channels for
measurements (pressure, temperature, flow and level) was considered to enable the trucks to make pickup based on
need rather than timed schedule. (Millenial Net)

A self-sufficient, self-powered wireless sensor network was employed to collect and communicate vital operational data
to a remote monitoring centre via satellite to an Internet backbone. The solution was reported to enable moving away
from scheduled maintenance to just-in-time, as-needed basis. Additionally, gathered data could be utilized in improving
the gas extraction process. (Millenial Net)
Use case:
Remote_Monitoring_Natural_Gas
Description:
WSN is utilised to enable just-in-time basis maintenance pickups from remote tanks.
Source: (Millenial Net)
Actors:
Operator (Primary), Satellite Network, Internet, WSN, Production Equipment

Preconditions/Assumptions:
Steps:
1. Sensor node detects high level for some by-product measurement.
2. Data is sent via satellite to internet backbone
3. Data is displayed at the monitoring application and maintenance is delivered.
Variations (optional):
Non-functional requirements:
1. Ability to withstand changes in weather conditions
2. Long battery life
Issues:
Comments from the use case:
 APPENDIX 1(12)

Remote Monitoring, Monitoring LPG Level

Narrative Description
A London-based company is also remotely monitoring its industrial customers' LPG tank fill levels, using battery-
powered ultrasonic sensors that transmit information by radio signal to a low Earth orbit satellite, which relays the data
to the company information system. The technology is operational on about 200 tanks in England and is being deployed
across Europe. WSN was utilised to gauge how much LPG remained in large tanks to enable just-in-time deliveries.
(King, 2005)
Use case:
Remote_Monitoring_of_Tank_Levels

Description:
WSN is utilised in tank level monitoring to enable just-in-time deliveries.
Source: (King, 2005)
Actors:
Operator (Primary), Satellite, WSN, Refinery Equipment
Preconditions/Assumptions:
Steps:
1. Sensor measures the LPG level in the tank
2. Data is transmitted to sink via wireless communication channel
3. Data is transmitted to remote office via low-Earth satellite.
4. Tank level data is displayed at the monitoring software.
Variations (optional):
Non-functional requirements:
1. Ability to withstand changes in weather conditions
2. Long battery life
Issues:
Comments from the use case:
 APPENDIX 1(13)

Other Applications, Monitoring Aluminium Rolls in Cooling Phase

Narrative Description
A company’s uses a cold rolling process to produce rolled aluminium goods. As a result of the rolling process, metal is
heated to 225 degrees and is let cool to 100 degrees at the cooling stations, which cover an area about a football field.
An overhead crane manually monitors the temperatures of about 100 rolls to determine when they are cool enough to
move. The measurement process was considered to be time-consuming, because the operator had to climb down from
an overhead crane to perform the measurements. (Sensicast, c)

A wireless temperature monitoring system was deployed in the plant, which automatically monitors the temperature of
the rolls in the cooling stations every minute. The data is wirelessly transmitted to a computer in the cab of the crane.
The need for manual measurements could be eliminated. And an increase of 2.3 million kilograms in annual throughput
was observed due to the improvement. (Sensicast, c)

Use case:
Monitoring_Aluminium_Rolls_In_Cooling_Phase

Description:
WSN is utilised to improve temperature monitoring at a aluminium factory.
Source: (Sensicast, c)
Actors:
Operator (Primary), Wireless Sensor Network, Aluminium Rolls
Preconditions/Assumptions:
Steps:
1. Temperature monitoring sensor node is attached to aluminium roll.
2. Network accepts a new node to the network.
3. Node measures a temperature of an aluminium roll.
4. Temperature is transmitted over wireless communication channel to the gateway in the overhead crane.
5. Data is transferred to PC and displayed at the monitoring software.
6. Once the roll is cooled down, remove the sensor node and dismiss the node from the network.
Variations (optional):
Non-functional requirements:
1. Rapid installation and removal of WSN
2. Nodes should withstand high temperatures
Issues:
Comments from the use case:
 APPENDIX 1(14)

Other Applications, Relief Valve

Narrative Description:
Petroleum companies are among the largest worldwide consumers of control and automation products, and
therefore, a potential applier of wireless automation technology. In an oil refinery, a relief valve is a purely
mechanical device used to protect equipment from overpressure, when refining process is out of balance.
(Sereiko, 2007)

A typical refinery covers over 8 hectares and can have between seven and 10 different processing areas
where hydrocarbon emissions may occur. Each of these areas is about 100 square meters and can contain
over 200 relief valves. (Sereiko, 2007)

There may be several reasons for gas pressure to builds up in the refinery setting. When the pressure goes
above a predetermined, safe level of pressure the relief valve opens and gas vents to the atmosphere. When
the pressure drops back to acceptable level, the relief valve closes. Regulatory limitations provide
thresholds for the emissions, and if those are exceeded, the refinery must pay fine. Therefore, it is
economically feasible to monitor the emissions. (Sereiko, 2007)

Wireless sensor networks are applicable for monitoring the emissions, because they do not require cabling,
which would be costly due to large area covered by oil refinery. The measurements would be based on
acoustic emission, produced by the hissing sound, which gas makes as it escapes from the valve.
Furthermore, after capturing the information is solved, the emission information could be utilised in plant’s
control system to further enhance efficiency of the project. (Sereiko, 2007)
Use case:
WSN_Relief_Valve_Monitoring
Description:
WSN is utilised in monitoring of relief valves.
Source: (Sereiko, 2007)
Actors:
Operator (Primary), WSN, Relief Valve, Data Storing Equipment
Preconditions/Assumptions:
Steps:
1. Valve opens.
2. Sensor node detects hissing sound produced by gas escaping from the valve. Node starts timer.
3. Sensor node detects end of the sound and stops timer.
4. The duration of gas emission is routed e.g. to PC via mesh network and stored in database.
Variations (optional):
Non-functional requirements:
1. Ability to withstand changes in weather conditions
2. Long battery life
Issues:
Comments from the use case: