View
10
Download
0
Category
Preview:
Citation preview
AUTOMATION
AUTOMATIONP.O. Box 13001
33101 Tampere, Finlandwww.vtt.fi/aut
forename.surname@vtt.fi
AU
TO
MA
TIO
N T
EC
HN
OLO
GY
RE
VIE
W 2
00
1V
TT
AU
TO
MA
TIO
N
AAAAAUTUTUTUTUTOMAOMAOMAOMAOMATIONTIONTIONTIONTIONTECHNOLOGYTECHNOLOGYTECHNOLOGYTECHNOLOGYTECHNOLOGYREVIEW 2001REVIEW 2001REVIEW 2001REVIEW 2001REVIEW 2001
kansi aukeamana.p65 5.12.2001, 11:311
AU
TOM
ATIO
N T
EC
HN
OLO
GY
RE
VIE
W 2
001
2
AUTOMATIONTECHNOLOGYREVIEW 2001
Publisher: VTT Automation
Editor in chief: Väinö Kelhä
Editorial board: Hannu Lehtinen, Matti Lehtimäki, Raija Koivisto, Aarne Oja,Timo Varpula and Olli Ventä
Publication date: December, 2001
Design: Kaisa Kuisma
Cover: Kaisa Kuisma
Language: Barole Oy
Postal address: Automation Technology Review, VTT Automation,P.O. Box 13001, 33101 Tampere, Finland
Telephone: +358 3 316 3111
Fax: +358 3 316 3494
ISSN: 1238-8688
Printed in Finland by Erweko Painotuote Oy
Copyright: VTT Automation. If the text, figures or tables of thismagazine are cited, the source must be mentioned.
Photos: Ateljee Nygård/Eija Nygård, Auvo Ahvenjärvi, Foto Strömmer Oy, JoukoJärvinen, Kuvakulma/Harri Lundelin, KuvaKabinetti, Merja Tulokas, Nokian Tyres plc,Paroc Oy Ab, Ponsse Oyj, Rautaruukki, Suomen Kuvapalvelu Oy
ATR_2001.p65 5.12.2001, 11:322
3
AU
TOM
ATIO
N T
EC
HN
OLO
GY
RE
VIE
W 2
001
C O
N T
E N
T S
Wireless Communication in Work MachinesPentti Vähä, Jarmo Alanen, Klaus Känsälä andHannu Lehtinen
6
Low Cost Wireless RF SensorsTimo Varpula and Olli Jaakkola12
Seamless Mobile ServicesPasi Viitanen
12
28 Future Mobility Services in Urban Areas -Selected CasesTapani Mäkinen, Jari Kaikkonen and Henrik Huovila
Next Generation Industrial Automation -Needs and OpportunitiesTeemu Tommila, Olli Ventä and Kari Koskinen
34
ATR_2001.p65 5.12.2001, 11:323
AU
TOM
ATIO
N T
EC
HN
OLO
GY
RE
VIE
W 2
001
4
Customized Dynamic Simulator Supports Process &Control Engineering at Mill SiteSami Tuuri, Jari Lappalainen and Kaj Juslin
48
C O
N T
E N
T S
SCM is about integrationIlkka Seilonen, Juha Nurmilaakso, Jari Kettunen,Stefan Jakobsson, Petri Kalliokoski, Markku Mikkolaand Veli-Pekka Mattila
Implementing ERP Systems in SME EnterprisesMagnus Simons and Raimo Hyötyläinen
Managing Electrostatic Discharges by Protective ClothingSalme Nurmi, Terttu Peltoniemi, Markku Soini, Mika Tukiainen, TuijaLuoma, Inga Mattila and Raija Ilmén
Integrated Safety Management in Teamwork OrganisationJarmo Karlund
60
66
42
54
ATR_2001.p65 5.12.2001, 11:334
5
AU
TOM
ATIO
N T
EC
HN
OLO
GY
RE
VIE
W 2
001
E D
I T
O R
I A
LTowards Next GenerationAutomationIn our vision of the future, we see automation evolving through wireless technology, new sensors,
networking and system technology. New business strategies, management styles and operational
practices are linked to technological change. This progress facilitates activities at work and at
home. This vision forms the basis of our strategy.
Wireless technologies and the Internet were first developed for mobile users. We now see our
mission as continuing to develop these technologies for citizens while also transferring them to
meet industrial needs. Mobility is essential for mobile working machines, mobile parts of process
devices, plant operators and maintenance personnel, to mention just a few industrial examples.
New sensors, the Internet and wireless communication can all provide us with more accurate
information on process or operating conditions, production performance,
stresses and the development of failure mechanisms in critical process or
machine components and structures.
The cost of new investments in process industries such as the pulp and
paper industry, in the chemical industry, and in energy production is
becoming ever higher. One challenge to automation is to reduce the
investment in machines and process installation. When more accurate real
time information becomes available for plant models and simulators, we
can increase the efficiency and yield of processes, and enable simpler plant
or machine construction, by replacing some traditional parts in installations
with improved control system. Considerable economic benefits can also be
gained by expanding the economic lifetime of an investment by better
system-level management of dependability combined with on-line
monitoring of the failure mechanisms of critical components.
These challenges require the integration of several technologies.
Networking with experts is our way of carrying out R&D.
This Automation Technology Review throws light on the technology trends
in automation in general, and on our work in industrial and machine automation in particular, as well
as in the development of mobile services to citizens and in developing a safer world to live.
Jouko SuokasResearch Director
ATR_2001.p65 5.12.2001, 11:335
AU
TOM
ATIO
N T
EC
HN
OLO
GY
RE
VIE
W 2
001
6
Wireless Communication in Work MachinesPentti Vähä, Jarmo Alanen, Klaus Känsälä and Hannu Lehtinen
Mobility has become an integral part of our everyday life, whether for work or forleisure. Indeed, there are a number of ways work machines can take advantage ofthe developments in information technology that have taken place over the last fewyears. Mobility is in fact essential to work machines performing given tasks. Wirelesscommunication from control rooms to on-site working machines is becoming morecrucial as automation increases in these machines. Today, wireless communicationis increasingly becoming part of work machine applications, and increasingly offeringdifferent value added services as well.
IntroductionMachines developed for performing
work in non-manufacturing lines are
called work machines. The demand for
mobility is typified where material
cannot be transported to them, e.g. by
conveyors, so they must be capable of
moving to the material. Mobility is, in
fact, essential for work machines when
performing given tasks. Today they are
typically semiautonomous, possessing
several autonomous features. In the
future, they will be autonomous, and
the greater the increase in the level of
automation in work machines, the
greater the need for development in
wireless communication for the
purposes of monitoring, commanding
and controlling them. In the early
stages, in some cases for safety reasons,
the operator controlled the machine
remotely by giving commands via the
cable tethered to the machine.
Machines were manually operated and
the operator had to have eye contact
with the machine and with the work
being done. By replacing the tethered
cable with a radio link, the disturbing
cables could be removed, but the
operation of the machine would still
remain manual. In remote control, the
operator commands the machine’s
operations directly, making direct eye
contact or a video connection with the
performed work necessary. When the
automation level gradually increased,
and several machine operations could
ATR_2001.p65 5.12.2001, 11:336
7
AU
TOM
ATIO
N T
EC
HN
OLO
GY
RE
VIE
W 2
001
be performed automatically, it became
possible to give task level commands
to the machine. These task level
commands make the remote control
easier for the operator. Several
applications using remote control
methods have been reported [#1].
Basic WirelesscommunicationtechnologiesFor several years, the satellite based
positioning system (GPS) has been
used for vehicle positioning in open
space to make vehicle navigation
possible [#2]. Radio modems have
been typically used for giving task
commands to these automated
guided vehicles. Commonly, those
task commands start or stop a
sequence of automatic operations.
Radio modems are also used to ask
vehicle status information to monitor
the state of the vehicle. Baud rate of
the modem is typically 1200 - 4800
bps, which is enough for task
commands and status information.
The GSM phone became available
in 1993, and is nowadays an
affordable item and commonly used
in everyday life, whether for work or
for leisure [#3]. This technology also
offers the possibility of connecting a
PC to the Internet, although the
connection is rather slow compared
to the wired one. Nevertheless, it
gives the freedom to be mobile and
still be able to be connected to a
vehicle. In this case, a portable PC
including a GSM mobile phone
adapter is available for interfacing the
machine with the user [#4]. GSM
phones, and quite soon the first
representative of the 3G or the GPRS,
offer also the possibility of data
collection and transmission to and
from vehicles or work machines, e.g.
a forest harvester can transfer the file
describing the amount of chopped
trees to the sawmill, and then obtain
the pay check based on that
information.
Typically, the radio modem is used
as a command line connection
between the vehicle and the task
level controller. This connection has
to be reliable in order to guarantee
the transfer of the commands to the
vehicle. Usually, acknowledgement is
used to make sure the command data
transfer between vehicle and task
controller has been successful. This
of course slows the transfer of the
packets. In reference [#5], this is done
with a stream socket. The command
line is, however, not sufficient in
remote operations, since there is
usually no direct eye contact with the
machine. A video connection is
needed to see how the work is going
on and what is happening in the near
environment. This connection can be
an analog link or a digital one, and
the connection does not need to be
guaranteed all the time and does not
result in faulty operations in cases
when short breaks appear. However,
if remote control is performed
through the Internet, then the video
signal is grabbed as a sequence of
digital pictures and sent via the net
to the user.
Today wireless local area networks
(WLAN) [#6] are becoming more
common due to the fact that more
and more equipment is available for
that purpose. This brings new
possibilities for work machines to
take advantage of this technology to
develop the services, including value
added services, that the machines
and equipment can offer end users.
Typical examples of work
machines are mining machines (rock-
drilling and hauling machines), forest
harvesters and forwarders,
agricultural machines, road and earth-
moving machines as well as different
transportation machines in harbours,
goods terminals and warehouses.
Next, a short overview of the use and
demand of wireless communication
is given with respect to these
application areas.
Wirelesscommunication in workmachine applications
MinesWireless communication has been
used in the context of mining
machines for many years to control
machines remotely by visual contact.
With the emergence of autonomously
operated machines, the wireless
communication infrastructure gets
more complicated as there can be
several autonomous machines
roaming around in the tunnel
network. Hence, a wireless local area
network facilitating dependable
communications is needed. It should,
however, be noted that, compared to
those of traditional remote operation,
the dependability requirements of
wireless communications of
autonomous machines might even be
even alleviated. This is because the
autonomous machine is able to
continue the task given to it, even if a
short communication break occurs.
The safety functions, such as a longer
communication break stopping the
machine, can be implemented.
Nevertheless, in practice, the
machines are also operated
remotely, hence the dependability
requirements remain at the same
level as those of traditional systems.
An example of an autonomously
operated mine can be found at LKAB
mines in Kiruna in Sweden. The
wireless communications system of
the mine is called WUCS (Wireless
Underground Communication
System), and is delivered by a Finnish
company called Electrobit Oy. WUCS
is based on ATM technology and
provides three data channels, one
video channel and one audio channel
over a radio interface. Each
autonomous machine is equipped
with a mobile terminal (WUCS-MT)
that is able to roam the wireless
network built by a set of base stations
(WUCS-BS) attached to the tunnel
ATR_2001.p65 5.12.2001, 11:337
AU
TOM
ATIO
N T
EC
HN
OLO
GY
RE
VIE
W 2
001
8
walls. The base stations are connected
to the operator’s control station
through an optical ATM network (see
Fig. 1). Sandvik Tamrock Oy is also
using WUCS communication system
in their first prototype of AutoMine™
autonomous mine system.
HarboursAs more and more cargo is shipped
in containers, moving them optimally
������������������������ ������
���������
����������
���������
�������������
�� ��!���"
�� ��!���"
�� ��!���"
�������������
�������������
!�������
!������� !�������
��������������������� ��
��������� �� � ����������� �
!����� ���������������������� ������
�����
�����
in the intermediate storage area
called a port requires a large amount
of communication. As the number of
vehicle types moving containers in
ports - straddle carriers, rubber tired
gantry cranes (RTG), fork lift trucks
and trailers etc. - probably exceeds
10, it is not possible to give commands
to the drivers with a “voice radio”
connection. It is better to send the
transfer task information digitally to
the vehicle. It has been said that
Figure 1. Principle of the WUCS communication system (picture from Electrobit web-site).
misplaced containers correspond to
five percent of the operating costs of
typical ports.
The digitally received current task
can be checked on screen all the time
and need neither be written down nor
memorised. And digital information
can be used to automatically operate
mechanical subsystems of the
vehicles. Adopting the gripping
mechanism to the length of the
container to be picked up is a typical
ATR_2001.p65 5.12.2001, 11:338
9
AU
TOM
ATIO
N T
EC
HN
OLO
GY
RE
VIE
W 2
001
example.
As port vehicles should never stop,
the next task should be sent to the
vehicle while it is fulfilling the current
one. As an ongoing operation affects
the operation of the other vehicles,
since containers are typically stacked
on top of each other, the result of an
operation needs to be transferred to
the central computing system often
called the “port operation system” or
“port optimisation system”. The key
element in this is a positioning system
to measure and estimate where - in
which storage slot and tier (altitude) -
the container was stored. Typically
differential GPS systems are used to
confirm the slot. Transferring this
differential information from the static
GPS base station to all vehicles creates
another communication need.
As there is plenty of cost
optimisation potential in integrated
vehicle fleet and quay crane operation
GPS systems, digital radio systems are
rapidly being installed in ports. The
central radio station of the current
practical product checks the
transmission needs of a mobile station
at certain intervals. Intervals are
typically about one second. If needed,
a bi-directional radio connection is
established. A central station can
handle about 100 mobile stations.
Autonomous container transfer
vehicles have been in use [#7] and are
under development. They increase
communication demands, because
several things have to be checked
before autonomous motion is safe.
ForestsForest harvester includes an embedded
controller for controlling the cutting
process and for measuring the total
amount of logs and trees. The amount
of chopped trees and logs is calculated
and stored in the file according to the
spieces (pine, spruce etc.) and to the
type (saw timber, pulp logs) of logs. This
file is then transferred to the buyer of
the trees, e.g. to the sawmill, as a data
transfer via the GSM link. The harvester
contractor gets his/her pay-check
based on that information, and the saw
mill knows where, of what type and
how many logs are available. This
information can be used when
planning the transportation of logs to
the mill, according to the demands of
the sawn timber.
Another application, a radio
controlled remote control for a
lumberjack intended for the first
thinning phase, has also demonstrated
[#1]. The machine has no cabin, but is
remote-controlled by the operator
from the neighbourhood of the
machine via a radio link. This gives the
operator an opportunity to select a
suitable command place according to
the task. This radio link is especially
designed for outdoor applications in
rugged environments. The reliability
of the link is designed to be very high,
and is based on three factors: constant
monitoring of the radio signal, a
#���$���%��&�&'�(�%)*
!� )+)�����,
-)�.��)�����'� �/0 #�*�1�'1
�$))0�+�&2
����%/31�*)%)
Figure 2. Road construction site with wireless connections.
Wireless Communication in Work Machines
ATR_2001.p65 5.12.2001, 11:339
AU
TOM
ATIO
N T
EC
HN
OLO
GY
RE
VIE
W 2
001
10
special address function (fingerprint)
and digitally coded messages with
CRC polynomials. The components of
the radio are approved standard
components.
WarehousesAGVs (Automated Guided Vehicles)
have been used in the process and
machine industry since the 1970’s.
The basic structure of an AGV has not
changed much since then, and its
outlook is more or less the same as
well. The biggest changes are inside;
the amount of electronics has grown
and the control systems have become
more sophisticated. The AGV today is
laser guided and computer controlled,
and has graphical user interfaces.
Communication between production
control and the AGV is wireless.
(Typically, 450 MHz, point to point).
The next phase in the development
is to connect the AGV to the factory
WLAN network. Then transportation
will become a service available from
the transportation server. The AGV
fleet is acting as a service provider for
the in-factory logistics system. This
will cause certain demands on the
AGV control system. The system has
to be task based and able to respond
to the following questions: what tasks
are active right now when can a new
transportation task be carried out and
what routes are available in the
system? The capacity optimisation
takes place automatically when a new
transportation task is passed to the
transportation server. VTT Automation
has actively been involved in the
development together with some
Finnish companies [#8]. At the
moment, a new generation AGV is
ready for demonstration: it has got a
WLAN interface, the operation system
is based on the real-time RT-Linux and
the system hardware is built from
commercially available PC/104 cards.
The user interface is coded with JAVA
and the AGV can be remotely
controlled with a web browser
anywhere from the web (for more
details visit www.e-wfff.com).
Road constructionAt first glance, a road construction
site seems to be far removed from an
ideal wireless application case.
However, construction work is very
expensive and lots of different
materials have to be spread on the
road before it is ready for the final
layer of asphalt. The first goal of
automation is to increase the
accuracy of the construction work so
that all the layers will be inside
desired construction tolerances. This
will increase the quality of the work
and decrease the loss of material. If,
for instance, sand is spread on the
highway construction site and the
layer has one centimetre extra
thickness, it means that on a 30 km
long, 30 m wide road about one
million euro will be wasted due to
extra work and material. So there is a
huge potential for savings. VTT
Automation has been developing an
automatic blade control system for
road scrapers. The blade control is
based on the model of the road
scraper. The blade control system gets
the set values for road layers from 3D
CAD model of the road. The position
of the blade is measured with a robot
tachometer located on the site
nearby the scraper. The road scraper
control system has been on field tests,
and results have been very promising:
both the accuracy and the efficiency
have improved. The test drivers have
been enthusiastic about the new
working environment. The work is
more interesting and causes less
stress.
The next roadmap in the
development of road construction is
called “The Wireless Construction
Site”, where the main goal is to
improve the logistics inside the
construction site. Since lots of material
is transported, it is important to plan
and control the movements of the
fleet of trucks and loaders in order to
avoid traffic jams and queuing on the
site. Another important issue is to
reduce time-consuming paperwork -
as also in [#9]. The complete set of
drawings for a typical construction
site can consist of as many as several
thousand pages of drawings and
contracts. Typically this archive is very
difficult to keep up to date, especially
as far as all the changes are concerned.
Therefore, the goal is to get access to
the corresponding files through a
server located on the construction
site. This server keeps track of the
changes and has always got the latest
information about the situation. The
material transportation and progress
information is also stored in the same
database, which is then used by the
site manager as he/she is planning
future actions.
ConclusionsIt is fairly obvious that the mobile
information society is coming, and
that it will be mostly technology-
enabled. The EU TRD programmes
support this trend. This means that the
work machine sector has to follow
this trend and take advantage of the
technology in order to be competitive
and able to attract workers. When
work machines are autonomous, the
workers perform production control
by giving task commands to several
autonomous machines from the
control room, taking care of the fleet
control also. Production control
information is needed at management
level to run the business and give new
production plans. Today machine
contractors can have several
machines at different work sites; the
owner needs to follow up the working
hours of each machine, to keep them
running, and to keep getting paid for
the work they have done. Wireless
communication and the Internet will
make this possible in the future. In the
same way, the manufacturer of the
machine will be interested in
following up its use during the period
of guarantee. This information may be
used either for R&D in forthcoming
ATR_2001.p65 5.12.2001, 11:3310
11
AU
TOM
ATIO
N T
EC
HN
OLO
GY
RE
VIE
W 2
001
References1. Känsälä K., Vähä P., Kerva J., Saavalainen P., Workhorse with flexible remotecontrol unit for lumberjacks. Proceedings of the 2nd International Conferenceon Machine Automation, ICMA’98, 16 – 18 September 1998, Tampere, Finland,pp. 305 – 309.
2. Rintanen, K., Mäkelä, H., Koskinen, K., Puputti, J., Sampo, M., Ojala, M.,Development of an autonomous navigation system for an outdoor vehicle,Control Eng. Practice, Vol. 4, No 4, 1996, pp. 499 – 505.
3. Kaikkonen J., Viitanen P., Towards a Mobile Information Society - a QuickGlance at Tomorrow. Automation Technology Review 2000, VTT Automation,Espoo, Finland, December 1, 2000, ISSN 1238-8688, pp. 93 - 97.
4. Rehu J., Kaarmila P., Känsälä K., Vähä P., Merin P., Automated guidedvehicle with intelligent clamp for paper roll handling in warehouses. Proc.of the Scandinavian Symposium on Robotics ’99, October 14-15, 1999, Oulu,Finland, pp. 257 - 263.
5. Annala, M., Vähä, P., Matsushita T., Remote Control of an Intelligent Vehiclein an Electronics Manufacturing Facility via the Internet. Proc. of the 9thIEEE Int. Workshop on Robot and Human Interactive Communication,ROMAN2000, Sept. 27 - 29, 2000, Osaka, Japan, pp. 173 - 177.
6. http://www.ndclan.com/Wireless/wlanW1.htm, 6.10.2001
7. Gelderland, J., Case study: ECT Delta/Sea-land - First results, Proc. 8th Int.Conf. on Terminal operations, Amsterdam, 1994.
8. Känsälä K., Kaarmila P., Ruokonen K., Lassila K., Kaarlenkaski J., Joensuu P.,Mämmelä M., Production of electronics - Enhanced by flexibility. AutomationTechnology Review 1998, VTT Automation, Espoo, Finland, 1998, pp. 23 - 27.
9. Peyret, F., Jurasz, J., Carrel, A., Zekri, E., Gormam, B. The ComputerIntegrated Road Construction project. Automation in Construction, 9/2000,Elsevier Science, pp. 447 - 461.
versions of the machine, and also to
follow up how hard the machine is
used during the period of guarantee.
Definitely, today’s work machines
can’t get along without wireless
communications. In closed areas like
mines, harbours and warehouses
they typically have their own local
area network, and in mines, for
example, these have been used to
control machines remotely. LKAB
uses an ATM technology based WUCS
system that has three channels for
data, one for video and one for audio.
The base stations are attached in the
tunnel walls, and the vehicle’s mobile
terminal roams within the wireless
network. A large amount of material
enters to a road construction site in
lorries. Transportation commands are
typically given outside of the local
construction site network. Therefore
the vehicles have to be able to switch
and adapt between local and global
area networks.
Wireless Communication in Work Machines
Hannu LehtinenD. TechGroup ManagerVTT Automation
Klaus KänsäläM. Sc. (Tech)Group ManagerVTT Automation
Jarmo AlanenResearch ScientistVTT Automation
Pentti VähäResearch ProfessorVTT Automation
ATR_2001.p65 5.12.2001, 11:3311
AU
TOM
ATIO
N T
EC
HN
OLO
GY
RE
VIE
W 2
001
12
Low Cost Wireless RF SensorsTimo Varpula and Olli Jaakkola
VTT Automation, Atmel, Idesco, and Rafsec havedeveloped, as part of a joint EU project, a new radiofrequency identification (RFID) system known asPALOMAR (Passive Long Distance Multiple Access HighRadio Frequency Identification System). It consists ofbatteryless transponders whose 1 kbit memory can beread wirelessly by a base station operating in 869 MHzor 2.45 GHz ISM (Industrial Scientific Medical) bands.The integrated circuit (IC) of the transponder ismanufactured with standard CMOS technology thatcombines EEPROM and RF features. The PALOMARconsortium is seeking a reading distance of 4 m, whichsubstantially exceeds that of the competing RFIDsystems, a frequency-independent IC solution, thecontactless rewrite possibility of EEPROM, and an anti-collision capability of up to 100 transponders. ThePALOMAR offers a platform for developing newubiquitous sensor concepts that allow remote wirelessmeasurements.
IntroductionA sensor can be regarded as a device
that transforms a physical quantity to
be measured into information that can
be further processed. Traditionally, the
information from the sensor is
transferred as an analogue electric
voltage or current via a twisted pair
or a coaxial cable. Currently a number
of other media for carrying sensor
information is in use: mains supply
lines, optical cable, infrared, and cable
based field bus, e.g. LON, and CAN.
Besides the value of the measured
quantity, the identification of the
sensor is equally important in
multiple-sensor set-ups. When each
sensor is individually wired, the
identification is obvious. A wireless
sensor must transmit an identification
code along with the measured value.
In many applications it is
advantageous if a measurement can be
done without leads between the
sensing element and indicator, control
unit, data logger or equivalent device.
In recent years, wireless radio
frequency (RF) communication has
become increasingly popular. This is
based on the revolution in wireless
telecommunications, during which
new silicon based RF components
were developed. These new
technologies allow the RF wireless
concept to expand into new fields. The
wireless revolution continues in the
sensor business. Bluetooth and
Wireless Local Area Network (WLAN)
are entering the sensor market. Sensors
that use a modem and the GSM
network for their communications
have been available for some time.
Bluetooth, WLAN and GSM are,
however, too expensive for many
applications, but now technologies
developed for radio frequency
identification (RFID) are being adapted
to sensors to provide a cheaper option.
The number of applications of RFID
systems is expected to grow in the
coming years. RFID will accompany
and replace optically read bar codes.
Traditionally, a sensor is powered by
an external supply or a battery. The
Prototype of a passive wireless RF sensor (869 MHz)being held in front of the reader antenna.
ATR_2001.p65 5.12.2001, 11:3312
13
AU
TOM
ATIO
N T
EC
HN
OLO
GY
RE
VIE
W 2
001
RFID concept allows the sensor to be
operated without external power
supply. This will make it possible to
realize new measurements and to
reduce the installation costs of the
sensors. Requirements for different
measurements vary and there is no
single technology for all applications.
Besides costs, technical specifications
such as sensitivity, stability, speed,
reliability, operation distance needs,
maximum size or weight of the sensor
unit, operation time without battery
replacement, must be considered. A
conventional, externally powered and
wired sensor will remain a competitive
choice for most applications.
Wireless RF sensortypesWe divide wireless RF sensors into
three types:
• Active sensors that have a power
supply, active RF components, and
are able to send a radio signal.
• Semi passive sensors that have a
power supply, but use it only after
a wake-up signal from the reader
unit. These sensors use back
scattering, i.e. they modulate the
signal of a base station or reader
to send data instead of sending an
RF carrier.
• Passive sensors have no power
supply or a battery, but also use back
scattering for RF communication.
For performing measurements and
transferring the information, passive
sensors take energy from the RF
field emitted by the reader.
Wireless RF sensors are also divided
according to the type of
electromagnetic coupling to the
reader:
• Inductive uses magnetic field.
• Capacitive uses electric field.
• Radiating uses radiating field.
Low-frequency inductively coupled
systems are in widespread use in RFID.
Capacitive sensors require in practice
a very short distance, near contact, in
fact, between the sensor and the
reader head, and have therefore given
way to inductive and radiating
systems. When properly designed, the
production costs of capacitive sensors
might be very low. Sensors using
Bluetooth or GSM technology are,
according to these definitions,
radiating active sensors. In this article
we focus mainly on radiating passive
(batteryless) sensors.
Passive wirelesssensorThe simplest and cheapest wireless
sensor is based on an inductor-
capacitor resonator. One example of
such a sensor is the 8 MHz resonator
used in electrical article surveillance
systems. The manufacturing price of
this kind of sensor is in the order of
a few cents when manufactured in
large volumes. The functionality of
this sensor is, however, limited. Data
can only be coded into the resonant
frequency or the Q-value of the
resonance.
If an integrated circuit (IC) chip is
connected to an antenna, the wireless
sensor can have much more
complicated functions. The IC is like
a small micro-controller that
communicates in both directions
wirelessly. It also contains a circuit, a
voltage rectifier, that derives power for
the sensor from the field of the reader
or the base station. Because the power
available from this supply gets smaller
when the distance between the base
station and sensor increases, the
power consumption must be very
small if long operation distances are
required. This means that there can be
no active RF components or radio
transmitter on the chip.
Fig. 1 shows the concept of a
passive wireless RF sensor system. The
antenna of a wireless sensor is
Figure 1. Concept of a passive wireless RF sensor system.
� �3)���
�)*�0�+/%�
���)+)3% �&�3�
������
�0�+/%��&
�)&�� )+)3% �&�3�
)*� (�)&��
�&%)&&/��+%/') )3%���) /&0
0)%)3%�
� ��
!����������,���
�� �,���
/%31�&' ��'�3
ATR_2001.p65 5.12.2001, 11:3313
AU
TOM
ATIO
N T
EC
HN
OLO
GY
RE
VIE
W 2
001
14
fabricated on a laminate or a printed
circuit board. The chip, fabricated with
CMOS technology, consists of an
antenna matching circuit, a voltage
rectifier, detector, logic, sensor, sensor
electronics and memory. The reader
interrogates the wireless sensor by
sending a measurement command
and then continues sending a
constant RF signal for powering the
sensor. The value of the measured
quantity is either converted into
digital form and stored in the memory
or is immediately transmitted by
modulating the antenna impedance.
At each interrogation the sensor
transmits the content of its memory
that contains the identification code
of the sensor. When the impedance
of the sensor antenna is modulated,
the back scattering from the antenna
is also modulated. The back scattering
is then detected by the reader.
The schematic diagram of the
reader RF electronics is shown in Fig. 2.
Figure 2. Schematic diagram of the RF electronics of a reader unit of a wireless sensor.
In principle, this type of electronics is
suitable for reading all the RF sensors
using back scattering. The electronics
is, in effect, a sensitive impedance
measurement device. When the wireless
sensor modulates the impedance of its
antenna, these modulations are
reflected to the impedance of the
antenna of the reader. The back-
scattered signal from the sensor is
detected by a sensitive impedance
measurement of the reader antenna.
Standards andregulationsTable 1 gives the frequency bands
and power levels allocated for short-
range radio devices (SRD), mainly in
Europe, but also in America. No
license is needed if the device
operates within the given bands and
power. The regulations for wireless
sensor communication are not well
harmonised worldwide. This is a
major obstacle for the wide spread
acceptance of the passive RF sensors
at the moment.
In the VHF (30 - 300 MHz) and
UHF (300 - 3000 MHz) bands there
is only one band accepted
worldwide, the 2.45 GHz ISM
(Industrial Scientific Medical) band.
In this band, however, the allocated
radiated power is only 0.5 W in
Europe, while in the US the reader
can use a power of 4 W. A 0.5 W
power limits the maximum reading
distance of a passive sensor to well
below 1 m, which is too low for most
applications. At frequencies below 1
GHz there are no common frequency
bands. In Europe the most used band
will be 869 MHz, whereas in the US
the corresponding band is around
915 MHz. Again, a much higher
power is allowed in the US (4 W vs.
0.5 W). Because frequencies above
900 MHz are reserved for GSM in
�0�+/%��&
���������
� )4�)&3(3�&% �+ ���
�� ����5���% ��
����%
��
6�
76�
��
�,�
����
��
ATR_2001.p65 5.12.2001, 11:3314
15
AU
TOM
ATIO
N T
EC
HN
OLO
GY
RE
VIE
W 2
001
Europe, a discussion of allocating the
band of 865 - 868 MHz with a power of
4 W to SRD is underway.
Concerning older inductive passive
sensors, the worldwide harmoni-
sation is more mature, especially in the
13.56 MHz, 8.2 MHz and lower-
frequency bands. Due to shorter
reading distance, however, the
inductively coupled systems will give
way to radiating sensors except for
certain applications.
CommunicationPerformanceBesides communication speed, the
maximum reading distance is the
most important issue concerning the
passive RF sensor. Both are primarily
FREQUENCY BAND POWER, LIMITATIONS, REGION
<125 kHz Allowed in many countries for inductively coupled RF sensors
1.95 MHz, 3.25 MHz and 8.2 MHz Inductively coupled theft tags, world wide
13.56 MHz Inductively coupled RFID tags and sensors, worldwide
27 MHz and 40 MHz 0.1 W ERP, Europe
138 MHz 0.05 W ERP, Duty cycle <1%, Europe
402-405 MHz Medical implants, 25 µW ERP
433.05-434.79 MHz 25 mW ERP, Duty cycle <10 %, Europe
468.200 MHz 0.5 W ERP, Band width 25 kHz, Europe
869.40 - 869.65 MHz 0.5 W ERP, Duty cycle <10%, Europe
902-928 MHz 4 W EIRP, America
2400 - 2483.5 MHz ISM band, 0.5 W EIRP Europe, 4 W America, Bluetooth
5725 - 5875 MHz 25 mW EIRP
24.00 - 24.25 GHz 0.1 W EIRP (Police radars)
61.00 - 61.50 GHz 0.1 W EIRP
122 - 123 GHz 0.1 W EIRP
244 - 246 GHz 0.1 W EIRP
EIRP = Equivalent Isotropic Radiated PowerERP = Equivalent Radiated Power
limited by the bandwidth and power
allocated by the authorities. It is
expected that within the present
regulations the effective baud rate of
a passive radiating sensor will be in
the range of 10 - 50 kbits/s.
An inductively coupled sensor is
read via a coil of the reader unit. In
principle, the reading distance is
limited by the signal-to-noise ratio and
the maximum magnetic field strength
allowed to the reader. The magnetic
field due to a coil decays as 1/r3 for
distance r longer than the largest
dimension of the coil. It means that
the power available to the sensor
decays very strongly, as 1/r6. In
practice therefore, the size of the coil
of the reader determines the reading
distance which is typically well below
1 m for the inductively coupled
sensors. The reading distance may
exceed 1 m with only large and
expensive reader coils and/or large
sensor antennas.
In passive radiating systems, the
maximum reading distance depends
on the manufacturing technology of
the sensors. When manufactured
with the present CMOS technology,
the distance is determined by the
power the reader can supply to the
sensor. It means that if the power the
sensor receives is high enough for
operation, the reader can always read
the back-scattered signal. For
antennas with aligned polarisation,
the power transfer is obtained from
the transmission formula
Low Cost Wireless RF Sensors
Table 1. Frequency bands and power levels allocated for SRD in Europe. Some American bands are also given.
ATR_2001.p65 5.12.2001, 11:3315
AU
TOM
ATIO
N T
EC
HN
OLO
GY
RE
VIE
W 2
001
16
where Pr is the received power, λ the
wave length of the radiation, r the
distance between the sensor and
reader, GT the gain of the reader
antenna, GR the gain of the sensor
antenna, and PT the transmitted
power of the reader. The parameter
n has the following values: n = 2 in
free space, experimentally it has been
found that n = 2.5 inside soft-
partition buildings, and n = 3 inside
hard-partition buildings. In some
favourable cases, e.g. along a corridor,
it is found that n ≈ 1. Using Eq. 1, we
obtain for the maximum reading
distance
where η is the efficiency of the
rectenna (=antenna+voltage rectifier)
of the sensor when transforming RF
power into dc power, and PR is now
the power needed by the sensor.
In Fig. 3, the maximum reading
distance of Eq. 2 is plotted for a
typical case. We see that the
European 0.5 W at 2.45 GHz would
give a reading distance barely above
1 m while the US 4 W at about 900
MHz would allow a distance of even
11 m.
Limitations andProblems with PassiveWireless SensorsTraditional sensors will keep their
positions in many applications. If a
cable is drawn for some other reason,
e.g. for powering an actuator, or
because the cable already exists as
in households, the sensor can utilise
the cable for communication by
using a field bus protocol. In addition,
wireless RF sensors cannot be used
if they are embedded in a medium
where the RF field does not
penetrate, such as metal, thick layers
of water or some weakly conducting
materials. Lower frequencies
penetrate deeper into a conductive
medium. Microwave frequencies are
attenuated considerably even when
Figure 3. Maximum reading distance vs. power (EIRP) emitted by the reader PT at869 MHz (red) and 2.45 GHz (blue) in free space. The calculation is made for η =0.15 and a chip power consumption of 5 µW attainable with 0.5 µm CMOStechnology. The sensor has a dipole antenna (GR = 2 dB).
going through the human body.
Operation distance is in most
cases less than one meter for
inductive coupling. With radiating
coupling the operation distance is
longer in ideal conditions, but
interference and blocking-objects
can reduce the operation distance. In
addition to interference, radio
communication is impeded by multi-
path phenomena and fading. In
applications where security is an
issue, data encryption must be used
in wireless communication.
ApplicationsWireless measurements find many
applications in situations where
measurements should be made
through material. Medical implants or
sensors moving autonomously in the
human body have been developed.
Another application field is
measurements when the object is
moving so that the wires cannot be
installed. Because wireless means
also contactless, some applications
can be found in situations where
corrosion of contacts is a problem.
Also, cases where the number of
measurement points is large can
benefit from the properties of
wireless sensors. A large number of
sensors can be installed at low cost
and read fast. Because an IC chip of
the sensor contains memory that can
be read and written from a distance,
a wireless sensor also serves as a
RFID tag, and helps book keeping and
logistics problems in many
applications.
The application areas of the
passive wireless sensors will be very
wide. We list here only some
examples. Automotive industry:
continuous monitoring of tyre
pressure, fuel level, and moving or
rotating engine parts. Medical
industry: implantable sensors, body
temperature and heart rate through
clothes, moisture content of diaper.
Process industry: new measurements
that cannot be realised with wired
sensors, measurements in explosion
hazardous environments. Building
industry: temperature, humidity, and
CO2 concentration of air, moisture
˜
ATR_2001.p65 5.12.2001, 11:3316
17
AU
TOM
ATIO
N T
EC
HN
OLO
GY
RE
VIE
W 2
001
inside building structures. Food
industry: monitoring quality of food
in production chains and retail shop
packages, measuring usability of
frozen food.
References1. For telecommunications standards see web pages of TelecommunicationsAdministration Centre Finland: http://www.thk.fi/ or ETSI, EuropeanTelecommunications Standards Institute: http://www.etsi.org/.
2. J.D. Kraus, Antennas, McGraw-Hill Inc., 1988, ISBN 0-07-100482-3.
3. For description of a joint European PALOMAR RFID project see:http://dbs.cordis.lu/
Olli JaakkolaResearch AssistantVTT Automation
Timo VarpulaGroup ManagerVTT Automation
Low Cost Wireless RF Sensors
ATR_2001.p65 5.12.2001, 11:3317
AU
TOM
ATIO
N T
EC
HN
OLO
GY
RE
VIE
W 2
001
18
Wireless Flexible Factory Floor:Remote Control and Monitoring for WirelesslyConnected Production Devices
Leila Rannanjärvi,Tuomo Näyhä andKristiina Valtanen
Our vision is “Wireless technologies serve man, and control machines and production”.This vision, together with developing wireless technologies and experience gatheredin earlier projects, gives us a strong basis from which to work on the development ofservice architectures for the wireless flexible factory floor (WFFF). WFFF meansincreasing communication between production and transportation devices that areable to complete their tasks independently. Communication does not always needto be wireless, since it can be also hardwired if the device is not mobile or wires arealready existing. Adopting communication possibilities between production andtransportation devices will provide more flexible production and enable a remoteuser to control production tasks or to monitor devices.
IntroductionMost new wireless technologies are
designed to enable mobile users to
connect with the Internet and other
network services - like e-mail - on the
road. In future, the entertainment
business also will provide more
services in houses. The network
requirements are totally different when
considering an office worker (printing
services), a traveller (mobile phone, e-
mail etc.), a production device on the
factory f loor or a mobile work
machine. We propose classification into
four classes: Home networks, Office
networks, In-house production
(Factory Floor) networks and Outdoor
production (road/bridge construction)
networks. In the following paragraphs,
we’ll concentrate on the third one,
Factory Floor, and describe our pilot
network environment, which was
implemented using commercial
products, and which enables a user
outside the factory floor to control and
monitor production devices on the
factory floor.
This pilot network provides
development and test facilities for
industry as well as for ourselves. The
first application we developed
provides AGV’s user-interface for an
external user over the Internet. AGV
(Autonomous Guided Vehicle)
transports components over the
factory floor without any rails. This AGV
application was already demonstrated
at the Hanover Fair (April 2001) and
ATR_2001.p65 5.12.2001, 11:3318
19
AU
TOM
ATIO
N T
EC
HN
OLO
GY
RE
VIE
W 2
001
NextGen (Oulu, May 2001). In this pilot
network implementation, we decided
to use wireless technologies that use
license free radio frequencies using the
2,4 GHz radio band. In the very near
future, extranet applications dealing
with material flow over factory floors,
and even between several factories or
distributors, will also be developed.
In this article, we explain how
communication between different
devices on the factory floor can be
implemented, and draft the
possibilities for a remote user to
monitor a device, or control a task,
on the factory floor. On our modest
pilot factory floor, there is one AGV
(Autonomous Guided Vehicle,
transporting components from a
production device to a miniwarehouse
or vice versa), one robot (as a produc-
tion device) and two mini-warehouses
(containing components needed in
production).
Wireless technologieson a factory floor
Factory floor communicationSwitched Ethernet technology is
penetrating the factory f loor,
threatening even proprietary field bus
territory. It is already the dominant
network technology at the controller
supervisory level in new installations.
There are many discussions about the
ultimate extent of the Industrial
Ethernet dominance in the future.
W i r e l e s s , l i c e n c e - f r e e ,
communication products are
emerging in the factor y
communication domain in this very
sensitive phase, scaling up confusion.
Many questions arise. What are the
benefits of wireless technologies?
Will there be any new difficulties?
Where are the products? How will
this new technology amend the
network infrastructure? Today, we
can find a growing number of
answers to these questions.
Wireless technology for theindustryThe two primary problems for radio
transmission are interference and
multipath fading. Interference
reduces data throughput by
increasing retransmission rate and
thereby latency grade. Multipath
fading is a special problem in
industrial environments because it
reduces range, which is a critical
factor in most plants. There are many
testimonies that favour frequency
hopping radios over their direct
sequence counterparts in industrial
applications, especially for the better
latency and range grades. The spectral
robustness of frequency hopping
systems also tolerates better other
networks in the same geographic
area, retaining the margin for the new
radio networks in the future. The
drawback of frequency hopping is
high frequency synchronisation
overhead, which degrades data
capacity. Using parallel networks
could, however, relieve this
complication.
The hype around wireless
technologies has also invoked
negative side effects. One of them is
that the standardisation effort has
centred on the economically more
appealing office IT and
telecommunication business areas.
The characteristics of industrial
communication have not gained
much attention in wireless
standardisation work groups. The
result is that today nearly all wireless
industrial products are proprietary,
without the backing of established
international standards.
The greatest benefit of the wireless
factory floor communication will be
the reduction of the inflexible and
expensive control level Ethernet
wiring on the factory floor. Today, we
have not yet seen these kind of true
”Wireless Industrial Ethernet”
products. The radio technologies used
in some wireless industrial modems
are, however, not too far from the
required level. These are based on
frequency hopping radio technology
and have a range that surpasses
Ethernet cabling. Standard short-range
wireless products, e.g. 802.11b and
Bluetooth ® certainly also have
relevance to factory floor applications,
but perhaps they will not induce
strategic impacts on factory floor
communication system architecture.
Network architecture impactsWireless technology is not a major
threat to the wired field bus systems.
The real challenges for them are the
wired Ethernet and pervasive Internet
technology. Wireless technology is
merely a good ally to fight the
aggressive Switched Ethernet diffusion.
A wireless gateway between wired
Ethernet and field bus systems isolates
the rival technologies effectively, and
gives field bus systems extra time to
survive. This method is also applicable
for the majority of legacy automation
systems, which lengthens the pay-off
periods of earlier investments.
The most radical impact of the
wireless technology on the factory
floor will be on the controller level
networks. This territory has already
been considered as the future home
ground for the Industrial Ethernet.
The f lexibility demands for the
modern factory f loor production
systems nonetheless favours wireless
technologies. So there will be some
kind of back off for the wired
Ethernet, analogous to the retreat of
the wired office Ethernet under
pressure from office WLAN products.
Wireless factory sceneThe comprehensive vision of the
future of factory floor communications
system may as well be as follows: the
Switched Ethernet technology evolves
into the effective broadband layer of
the wired backbone system around
technologically diverse but coherent
automation islands, which build up a
flexible and effective production
machinery. The connecting technology
between these islands and the
backbone network will be a robust
ATR_2001.p65 5.12.2001, 11:3319
AU
TOM
ATIO
N T
EC
HN
OLO
GY
RE
VIE
W 2
001
20
Wireless Industrial Ethernet network that
runs over the factory floor.
Consequently, the wireless tech-
nology is not an element that pro-
pagates chaos to the factory floor
communications system. On the
contrary, it can substantially clarify the
network scenery with the new
abstraction layer. Network architecture
planning process benefits from this
new layer and the flexibility of the
network, and the overall production of
the system will increase. This new level
of f lexibility also creates new
challenges for the other system-
architecture disciplines.
Flexible factory floorOn the flexible factory floor there are
two co-operating sets of devices. The
first completes transportation tasks
around the factory f loor and the
second completes the actual pro-
duction. Both of these have independent
task managers: the transportation task
manager and the production task
manager, respectively. The former
communicates with production devices
and mini-warehouses to find out
• what has to be transported
• where to find it and
• where to deliver it.
When transportation tasks have been
completed, the manager asks
transportation devices (AGVs) to
complete the task. An AGV is an
autonomous guided vehicle that
transports goods/elements/parts/
components from machines to the
warehouse or mini-warehouse and
vice versa. On our modest factory
floor there exists only one AGV and
no decision problem (which of the
AGVs should complete the task).
AGVs, mini-warehouses and Robot
Cell Control also represent physical
components, which either complete
a task or contain some components
needed in production. Upside down,
the Virtual Warehouse does not
necessarily find itself on the factory
floor, because it is just a user interface
for a remote user, who wants to find out
what the situation in mini-warehouses
is. A mini-warehouse is a kind of local
small warehouse near the production
line, where parts, goods, elements
supposed to be needed shortly, can be
stored before (or after) production. Mini-
warehouses are located dispersed over
the factory floor.
Infrastructure for theWFFF - Pilot networkThe wireless communication is not the
main issue in WFFF, although it provides
a shortcut on the wireless flexible factory
floor. Wireless communication is useful
there, where one or both of the com-
munication nodes are moving or will be
moved from time to time. Wireless
connection is useful when, for example, a
distributor delivers components to
several mini-warehouses. At the factory
gate he connects to the virtual warehouse
Figure 1. Pilot network implemented using commercial products enabling a user outside the factory floor to control productionor monitor production devices.
ATR_2001.p65 5.12.2001, 11:3320
21
AU
TOM
ATIO
N T
EC
HN
OLO
GY
RE
VIE
W 2
001
and wises up where to deliver which
components. When he is loading new
components into a mini-warehouse, he
will update the data storage of the mini-
warehouse, and the system (physical
components/data) will stay balanced. The
user interface device for a miniwarehouse
can be located beside each mini-ware-
house (for example, a bar code reader
and touch screen) or carried around by
the distributor (Personal Digital Assistant
PDA, handheld).
We have implemented a pilot network
using commercial products enabling a
user outside the factory floor to control
production devices. This pilot network
contains (Fig. 1):
• An open external access into our
public network with a domain
name www.e-wfff.com
• A firewall server to keep out
unwanted visitors
• A PC for firewall management
• A Linux server responsible for
www services
• A camera for live video images
• Several application servers for
production devices
• Access points for Wireless LAN
(802.11b) communication
• Production devices capable of
communicating wirelessly
• A non 802.11b standard wireless
serial multi-point network, which
enables wireless communication
both indoors and outdoors.
Architecture for theremote control of adevice on WFFFAs an example of how to provide
remote access (i.e. extranet
application) into a mobile device on
a factory f loor, we describe the
architecture of the AGV at our pilot
network (Fig. 2). The AGV works as a
transportation agent on the factory
floor, delivering components to/from
mini-warehouses and to/from
production devices. The hard real-time
control of AGV is implemented using
RT-Linux. AGV receives transportation
orders from the transportation server,
which is connected to the AGV by
Wireless LAN, using unambiguous and
predefined AGV commands.
In our remote control architecture,
there are two Jabber servers [Jabber].
One is for the transportation tasks,
while the other is an intermediator
forwarding all Jabber messages to the
correct receiver. The remote user
connects with his/her common
browser to our public http-server and
(if he/she is authenticated) downloads
a Java applet, which allows him/her to
communicate (ask transportation
tasks) with the AGV via the graphical
user interface (GUI in Fig. 3). Live video
image feedback (Fig. 4) is available and
gives the current view over the factory
floor. When compression techniques for
images develop further, and trans-
mission delays over the Internet are
overcome, then even direct control of
devices might be possible.
ExperiencesTo date, we have implemented the
infrastructure for the wireless flexible
factory floor (WFFF) and developed
the first remote user interface for an
AGV. The infrastructure with a firewall
server and a separate www server
enable security against remote un-
wanted visitors. Production devices are
not in the public network; they may be
reached through the firewall and www
server if the password is known by the
visitor, and if his/her current IP address is
correct and approved.
Remote control of a device is possible,
if we accept the fact that no hard real
Figure 2. Architecture of the remote control of an AGV.
Wireless Flexible Factory Floor:Remote Control and Monitoring for Wirelessly Connected Production Devices
ATR_2001.p65 5.12.2001, 11:3321
AU
TOM
ATIO
N T
EC
HN
OLO
GY
RE
VIE
W 2
001
22
time feedback can be provided. For
example, now (current maturity of
network and image compressing) it
would be possible to drive a machine
remotely by joysticks connected to a PC
that could send the control information
to the machine. But, when a human
needs feedback (has to see where the
machine currently exists and in which
alignment), the transfer of live video
images is not fast enough. While a human
is waiting for the image to update, the
machine may collide or run away. That’s
why we keep the hard-real-time control
encapsulated in the machine, which is
able to complete tasks given from other
units.
The remote control of AGV over the
Internet has been introduced at
Hanover Fair (April 2001), at NextGen
(Oulu, May 2001) and in several smaller
occasions. Everywhere it has been
greeted with sincere interest and has
created an exciting atmosphere. In
September (Automaatiopäivät, Helsin-
ki) we demonstrated also co-operation
between transportation and produc-
tion devices. The developed extranet
applications follow definitions and
descriptions presented in this article.
Perspectives on theFactory Floor NetworkArchitecturesThe requirements set on factory floor
networks may differ in many ways from
the features of public network.
Networks on a factory floor are usually
closed in order to guarantee data security.
Figure 3. The graphicaluser interface GUI for theremote control of an AGVis a JAVA applet. You mayrun it on your PC having anInternet connection andcommon browser. Throughthe GUI you are able to askfor transportation tasks, seethe current location on themap (red spot) and see theposition of the AGV(coordinates).
Figure 4. Live video image of the factoryfloor. Viewing this window, the remoteuser can verify that the AGV is movingin the desired direction.
However, limited access to the factory
floor network may be provided for the
remote control of a production device
or for the presentation of some assorted
production data. In contrast to the public
network, the majority of users in the
factory floor network may be different
types of machines or devices whose
intelligence and needs for communi-
cation may vary considerably. Some
devices may need continuous real-time
communication, some others random
wide band connection.
One fundamental requirement set
on the factory f loor network is
flexibility. The structure of the network
should support the dynamics of
production devices so that removing
and adding single devices does not
bring about demand for modifications
or complicated re-configuration for the
network. Possibly, devices might self-
announce their entrance to the network
and the services they offer. Also, they
might be able to actively search for
information about the services that they
need. In addition, the structure of the
network should support the distri-
bution of applications by providing the
communication services to be used by
application modules.
As a whole, in order to implement
efficiently the special features of the
factory f loor environment, new
network architectures are necessitated.
One starting point is the so-called
service architectures, in which the idea
is to place different kinds of available
networking services in the network. A
user may ask for authentication, for
example, or for real-time communi-
cation service. The benefit of the use of
service architectures is rapid network
service development, because advanced
low-level services can be used through
simple application programming
interfaces.
VTT Automation has participated in
a development project in which one
research area is the implementation of
service architecture in the factory floor
network. The research part of VTT
Automation will concentrate on the
implementation and use of application
services in the service architecture,
which has traditionally been directed
towards the presentation and pro-
duction of communication-level ser-
vices.
So far, new communication methods
have been applied to the factory floor
network in our test environment at VTT
Automation in Oulu. The aim of the
research has been the networking
flexibility of production devices. One
interesting experiment is the use of
instant-messaging based systems within
the control of production devices. The
communication system provides both
messaging and presence services. The
service is based on a peer-to-peer
communication model [Oram 2001],
which has proven to be useful to
extend traditional client-server model.
In the research, the XML-based instant
messaging system called Jabber has
been used. Jabber not only provides
platform- and applicationindependent
communication but is also expanding
towards being a platform of middle-
ware services. [Jabber] [Goldfarb
2000].
ATR_2001.p65 5.12.2001, 11:3322
23
AU
TOM
ATIO
N T
EC
HN
OLO
GY
RE
VIE
W 2
001
Timo NäyhäSenior ResearchScientistVTT Automation
XMLXML (Extensible Markup Language) is
a flexible way to create common
information formats and share both the
format and the data on the World Wide
Web, intranets, and elsewhere. XML is
similar to the language of today’s Web
pages, Hypertext Markup Language
(HTML). Both XML and HTML contain
markup symbols to describe the
contents of a page or file. XML is
“extensible” because the markup
symbols are unlimited and self-defining.
MiddlewareIn the computer industry, middleware
is a general term for any programming
that serves to “glue together” or
mediate between two separate and
usually already existing programs. A
common application of middleware is
to allow programs written for access
to a particular database to access other
databases. Messaging is a common
service provided by middleware
programs so that different applications
can communicate.
ConclusionWe have implemented remote control
for AGV transporting components. A
remote user may launch a trans-
portation task using a JAVA applet as a
graphical user interface, GUI. This GUI
will be downloaded from the
transportation task manager standing
behind the firewall, which will guard
the gate of production network.
Remote control for a device is possible
when we forget the hard real-time
feedback and ask for services or tasks
to be completed. Also, continuous live
video image feedback can be provided
over the Internet within an acceptable
delay-time.
After the first “remote AGV” extranet
application we plan to integrate
transportation and production tasks on
the factory floor. Second tier appli-
cation, “Robot and AGV”, has already
been demonstrated at Automaatio-
päivät (September 2001, Helsinki). A
robot launched a transportation task
when a remote user requested a
production task when the robot did not
have correct component in hand.
Although the future of Wireless
Factory Floor seems to be in a light
upwind, we may provide more flexibi-
lity-enabling communication between
production devices, production (or
transportation) tasks and remote users.
Flexibility also means new ways of
organising production:
• to let the production devices com-
municate and be more autonomous
and
• to allow partners to have a customised
view into production processes and
to interact with the interfaces they
need to serve better.
References1. Dolmen 1998 B.C.F. Wind, F. Lucidi, P. Reynolds: Open Service Architecturefor Mobile and Fixed Environments, DOLMEN Consortium, http://www.fub.it/dolmen/delpages/asd4.htm
2. Goldfarb 2000 Charles F. Goldfarb, Paul Prescod:The XML handbook,second edition, ISBN 0-13-014714-1, Prentice-Hall, Inc.,2000
3. Jabber http://www.jabber.com/downloads/whitepapers/jabber_tech_whitepaper.pdf
4. Oram 2001 Ed. Andy Oram: Peer-to-Peer; Harnessing the Power ofDisruptive Technologies, 0-596-00110-X, O’Reilly & Associates, Inc. 2001http://www.oreilly.com/catalog/peertopeer/desc.html
Leila RannanjärviResearch ScientistVTT Automation
Kristiina ValtanenResearch ScientistVTT Automation
Wireless Flexible Factory Floor:Remote Control and Monitoring for Wirelessly Connected Production Devices
ATR_2001.p65 5.12.2001, 11:3323
AU
TOM
ATIO
N T
EC
HN
OLO
GY
RE
VIE
W 2
001
24
Seamless Mobile ServicesPasi Viitanen
A rich variety of mobile information services areavailable to users. A user can order different ring tones,logos, and have a chat with friends as well. A lot ofservices are available for business users: stock marketinformation, remote control of devices and on-lineproduction information.
When Mobile Information Society Services are fullyexploited they cover most of users’ daily needs. However,fully integrated services are not yet available. Forexample, a user gets a ticket to a movie from one place(site), and public transportation information from another.They both are obtained manually by the user.
BackgroundFinland has been one of the
forerunners in the development and
application of the latest wireless and
network technologies. Tampere region
forms a centre of excellence in the
field of new generation Information
Society Technologies (IST) services
and applications. Citizens of Tampere
are being introduced to a new wave
of services; hence, the whole area
forms a large-scale information society
laboratory.
The work, home andfreetime roles of the users
mix during the day.
ATR_2001.p65 5.12.2001, 11:3324
25
AU
TOM
ATIO
N T
EC
HN
OLO
GY
RE
VIE
W 2
001
eTampere programmeThe eTampere programme, a five-year
development project costing 130
million euros, will open the new
millennium. Its general objective is
to make Tampere a global leader in
the research, development and
application of issues related to the
Information Society.
The programme focuses on three
themes:
• The development of public online
services, and making these available
to all residents.
• The strengthening of the
knowledge base of research and
training.
• Generating new business related
to the Information Society.
These three themes consist of seven
modules: the Information Society
Institute, the eBusiness Research
Centre, the Research and Evaluation
Laboratory - RELab, the eTampere
Business Incubator, Technology
Engine programmes, Infocity, and the
eTampere office.
The European model for eTampere
is the eEurope programme launched
Figure 1. eTampere structure.
in December 1999. For eEurope, the
eTampere undertaking will provide an
extensive, and possibly the first, local
application. While eEurope builds the
Information Society for an entire
continent, eTampere builds a future
welfare city for Tampere citizens.
Research & EvaluationLaboratory RELabThe Research & Evaluation
Laboratory (RELab) is one of the
Figure 2. Progress of seamless services.
!� )+)���/*$) )8�66�6669
�66: �66� �66� �66� �66�
!� )+)������ ��") ;/&%/8�6�6669
!� )+)���/*$��8�6669
seven modules of the eTampere
programme. RELab will act as a
“melting pot” for the various efforts
of the different modules of eTampere
and, in varying scales, as a testing and
development environment for
companies large and small.
Objective The objective of RELab is to bring
new information society services
closer to everyday life by creating
ATR_2001.p65 5.12.2001, 11:3325
AU
TOM
ATIO
N T
EC
HN
OLO
GY
RE
VIE
W 2
001
26
new solutions that will ease the daily
life of citizens.
In the future, information
technology will be reachable by us
all whenever it is needed, and will
adapt to us and to our needs. We will
use the equipment and services in a
way that seems most natural to us.
Development ProcessRELab will start the development of
seamless services in three stages:
campus, suburb and the entire city
of Tampere, Fig. 2. The numbers at
each stage refer to the potential
number of users of IT services.
OperationRELab’s operation is based on both
national and international networking,
and will establish co-operation at a
European level through EU-projects.
It will build only a minimum
evaluation and testing environment
for its own use; in addition to this, it
will use other available resources as
agreed with its partners and
infrastructure owners. VTT will equip
RElab’s office building with the latest
network technology for testing and
evaluation purposes. The whole city
of Tampere will become one big
testing and development environment
for domestic and international players.
Test environmentThe campus of Hervanta suburb (in
Fig. 2) consists of the premises of the
Technical University of Tampere,
Tampere Technology Centre and the
Technical Research Centre of Finland.
The total number of workers and
students on campus is about 14000,
of which 2000 are considered to be
potential test users. The network
infrastructure includes WLAN,
Bluetooth, cellular (GSM, GPRS,
UMTS), HiperLAN and fast-wired
connections.
The second testing area is planned
for the downtown area, which offers
an excellent opportunity to reach
ordinary citizens. Public transport
goes through the central square,
where the City of Tampere is
equipping two bus routes as test lines
with real time travel information
systems. The time schedules are
updated on the basis of GPS receiver
information on buses.
The Seamless services require
different network environments,
from very local to global, Fig. 3.
Seamless servicesSeamless services are formed from
various content sources, and operate
in different environments without
interruption. The environments
include fixed Internet, WLAN, and
cellular networks, and move from
one environment to another and,
Figure 3. Network coverage in seamless services environment.
ATR_2001.p65 5.12.2001, 11:3326
27
AU
TOM
ATIO
N T
EC
HN
OLO
GY
RE
VIE
W 2
001
especially, they move from one
operator to another without notice.
Seamless services are not based on
a single technical insight, but rather
on several developments in
technology, and on extensive co-
operation between different actors.
What problems will the seamless
services solve? They will help users
in their daily tasks, at first, in their
daily routines. For example, arranging
a meeting is generally a routine task,
but it can be daunting, due to the
mismatching participants’ calendars,
perhaps. Things get even worse if the
time of the meeting suddenly
changes. Rearrangement takes a lot
of effort, which the user could use
on something more productive if the
seamless service took care of the
rearrangement. In fact, the question
is about time: how and where time
is spent, and on what.
Seamless services will be designed
for ordinary users. However, they will
also emerge in business environments,
which have the money required, and
also the need.
Experience gained from seamless
services and their applications is
limited because of the missing
seamless environments. The final goal
- really seamless environments - is
some years away, but something can
be done today. VTT Automation has
started a research project that tackles
the seamless problem from the user’s
point of view. The environment
References1. eTampere Programme Plan, 2000,19 p.
2. www.eTampere.fi
consists of work, car and home. The
target is to get user feedback from
the current system; the next step will
be to develop the current system on
the basis of the feedback received.
VTT Automation has made its very
first seamless car application, where
a user can order a taxi either to his
address or to geographic co-ordinates
defined by the user. The taxi is
equipped with both WLAN- and
cellular connections. When the car
is beyond the WLAN-connection it
switches automatically to the cellular
network and vice versa.
ConclusionsThe big challenge in seamless
services is co-operation between
different players. This includes the
value-chain and, especially, how the
different players get their revenues.
Technological problems will be
solved within a few years, but the
contents of seamless services are the
question mark. Intellectual property
rights, liability and responsibility
questions are hard to solve in what
is a very complex environment.
Seamless services are being given
an important role for the next EU 6th
framework program. ‘Ambient
intelligence’ is the term the EU is
using, but the content is the same:
technology serving people.
Pasi ViitanenSenior Research ScientistVTT Automation
Seamless Mobile Services
ATR_2001.p65 5.12.2001, 11:3327
AU
TOM
ATIO
N T
EC
HN
OLO
GY
RE
VIE
W 2
001
28
Future Mobility Services in Urban Areas -Selected CasesTapani Mäkinen, Jari Kaikkonen and Henrik Huovila
Safe, fluent and easily accessible mobility services are objectives that are not yetmet anywhere in industrialised countries. Industries, research institutes and publicservice providers are currently working in a great number of joint research projectsaround the world aiming at alleviating mobility problems and producing added valuefor traffic services. New technology areas that are being developed and graduallyimplemented range from active vehicle safety systems to traffic management andservices directed towards bringing added value to the mobility of travellers. eTamperewill also host a number of experiments aimed at improving mobility services usingwireless automation technology in urban areas. This paper presents and assessessome scenarios for potential solutions that could improve the level of traffic servicesand bring added value to the mobility of residents and visitors in urban areas. Thesescenarios include the following services: 1) Emerging smart tyre technology thatcan be used to improve safety services in urban areas; 2) Intelligent Speed Adaptation(ISA), which is taking its first steps currently; 3) For those arriving in the city area,wireless location-based information services, giving information on the location ofboth model-specific and general repair and maintenance services; 4) A centralisedparking management system that will inform drivers of free parking space accordingto the location of the vehicle prior to entering the city area.
ATR_2001.p65 5.12.2001, 11:3328
29
AU
TOM
ATIO
N T
EC
HN
OLO
GY
RE
VIE
W 2
001
Two cases on alreadyfeasible technologies
Smart tyres providinginformation to driversMost people ignore the task of
checking their tyres at regular
intervals. Even though tyre condition
is in fact one of the most crucial
elements in operating a car or a truck,
people tend to neglect this and
assume that all is well unless they
appear totally flat, at which point it
is too late. Modern radial tyres can
be underinf lated and still look
normal. According to the United
States Department of Energy, tyres
lose about 70 mbar per month and
70 mbar for every 5 °C drop in
temperature [1]. The reason for
people not checking their tyres is
probably not just ignorance or not
knowing about the importance of
tyres. The reason is more probably
just the fact that checking tyres is
inconvenient. It’s time consuming,
often dirty and the tyre pressure
gauges at gas stations are often
difficult to use.
Car tyres that provide drivers with
some specific real time information,
or the car information system itself,
are often called smart or intelligent
tyres. Smart tyres include some
sensors and electronics, with which
the information is measured and
transmitted to the driver. At the
moment, there are already some cars
that have pressure-monitoring
systems as standard equipment. The
number of tyre monitoring systems is
increasing because of the need for
better driving safety and economy. For
instance, in USA the tyre monitoring
systems will become obligatory in
new cars during 2003. These systems
will normally have a display that
shows the state of the tyres. However,
if the tyre information interface is
standard, the information could be
read by other display systems also. For
instance when you arrive at the filling
station, the gas terminal could show
your tyre status, or the tyre pressures
could be checked by just driving to
the site where you can adjust your
tyre pressures. The display detects
your cars tyres and reads the pressures
and displays them, so you do not even
have to get out of the car to check
the pressures. The system can be used
in cars, vans and trucks. This means
that it is suitable for vehicles with 4
or more tyres. The intelligent module
can be attached on the rim with a
special bonding method. One module
weighs about 33 gram with battery.
The battery lasts at least 3 years, in
best conditions 5 years, and can be
easily replaced. With a different type
of special housing of the electronics,
the system will later on be suitable
for forestry equipment also, and for
other kind of vehicles with tube type
tyres.
The basic property of BlueTooth is
that the tyre modules within a 10
meter range form a net. All mobile
phones within the range and with
access to the specific tyre network
can inquire about pressure and
receive the warning.
With the mobile phone the user
can set high- and low-pressure alarm
limits. These limit values can be
acquired, for example, from the
Internet. If tyre pressure changes
more than 0,2 bar, the system informs
�+�)%��%1*�0�+)���
��+%/')3�&% �++)
�)&�� �
�/%%) (
�
�
����������������
the driver. If the low- or high- warning
limit is exceeded, an alert will be sent
to the mobile phones within 10 m.
VTT Automation is developing
together with Nokian tyres a safety
system called RoadSnoop. The main
asset of the system is that it is universal
– it works with any BlueTooth-enabled
terminal. The system also gives the
driver an easy way to monitor tyre
pressures and it improves safety with
the alert property. It is also easy to
install. No extra equipment is needed
in the car.
Using the mobile phone as a
receiver for the tyre information is a
suitable solution for the after market.
For the original equipment solution,
the tyre information must be
delivered to the display of the car’s
on-board computer. BlueTooth will
also be introduced in cars as a
standard wireless interface for, for
example, the engine diagnostics and
infotainment systems. It will be
possible to use the same BlueTooth
receiver for the tyre information also,
and to deliver the information to the
CAN-bus of the car for display to the
driver.
RoadSnoop is not only a tyre
pressure monitoring system, but a
safety system as well. Thanks to the
constant Internet connection of the
receiver, RoadSnoop has more to
Figure 1. Basic Layout of the Sensor Module.
ATR_2001.p65 5.12.2001, 11:3329
AU
TOM
ATIO
N T
EC
HN
OLO
GY
RE
VIE
W 2
001
30
offer drivers than just basic tyre
information. It will be possible, for
example, to fetch the correct tyre
pressure recommendations for the
specific car or tyre type from the
Internet and store this information
in the memory of the tyre modules.
The services on the Internet can also
help the driver calculate the correct
tyre pressures according to the
loading of the vehicle, when
preparing a holiday trip.
The tyre modules can also store
information as to whether summer or
winter tyres fitted, and give the driver
information about the suitability of
the tyres to the weather conditions.
For this, the weather information can
be acquired from a weather service
on the Internet, and the location of
the vehicle can be obtained from the
GSM-network or GPS. Also a message
to the driver can be sent when it is
time to change between winter and
summer tyres.
These examples are just a few
solutions that RoadSnoop Safety
System can offer to the consumer. For
a commercial customer, for example
a truck company, the system can be
used as a telemetry service, where the
fleet manager can monitor the tyre
condition of the whole fleet or even
get information as to which trailer is
connected to which truck. It is also
possible to arrange public displays on
roadsides or gas stations where
drivers, upon passing these displays
or arriving at filling stations, can
immediately see the tyre pressure or
other information needed for tyres.
The pressure and temperature are
the parameters measured first by the
present tyre monitoring systems.
Later on, the tyres will also measure
more advanced parameters like tyre
wear; these parameters can also be
shown on both the car and public
displays mentioned above.
Intelligent speed adaptation - ISAThe relationship between speed level
and the number of accidents has
been conclusively shown over the
years in a number of studies. Reduced
speeds resulted in significant
reductions in the number of
accidents [2,3,4,5,6]. Speeding is a
common phenomenon in Europe
and it is especially widespread on
urban roads and on motorways [7].
Traditional measures for speed
management have been proven
rather ineffective. Differentiated
speed limits, theoretically, might be
an effective measure against
inappropriate speeds in different
critical situations. Time-differentiated
speed limits have been found
effective [8,3], but these are still not
flexible enough, since they cannot be
adjusted to the prevailing weather
and road conditions. Visible
enforcement usually results in an
immediate reduction in speeding, but
its extent in time and space is very
small [9,10,11]. The effects of most
of the physical engineering measures
turned out to be lasting in time, but
their effects mostly ceased outside
the vicinity of the measures [12].
Besides, they cannot be used in
critical road and weather conditions.
It is more logical to get to the vehicle
itself, and control its speed directly.
Recent technological advances have
allowed the application of information
technology and modern wireless
communications to transport facilities.
Tools based on wireless data transfer
may offer a much greater flexibility and
give broader possibilities to manage
speed, even in adverse road and
weather conditions, in place and time-
related critical conditions (e.g. in
Figure 2. Speeds of subjects driving through four intersections with the limiter switched on and off in free driving conditions,and when driving in a platoon.
ATR_2001.p65 5.12.2001, 11:3330
31
AU
TOM
ATIO
N T
EC
HN
OLO
GY
RE
VIE
W 2
001
school zones), and in critical interactive
situations with other road-users (e.g.
pedestrians crossing the road).
Field trials by VTT and Lunds
tekniska högskolan in three
European countries, the Netherlands,
Spain and Sweden, were carried out
in order to investigate the effects of
an in-car speed limiter. The trials were
carried out on urban and rural roads
including motorways. A so-called
unobtrusive instrumented car was
used, where all the measuring
equipment was hidden [13]. The
speed limiter used in this experiment
was an active gas pedal which provides
drivers with counter-force whenever
they try to exceed the pre-set speed
limit. The pedal resistance is sufficient
to remind drivers of the speed limit, and
the extra effort required to go faster is
sufficient to deter them from speeding.
When the speed of the vehicle
approaches the pre-set limit, the
counter-force of the accelerator
gradually increases. The speed limiter
also restricts the engine’s fuel injection
when the vehicle reaches the actual
speed limit. The speed limiter of the
test car was automatically triggered by
radio-transmitters attached to speed
limit signs.
The results generally indicated that
the difference between the limiter-off
and the limiter-on conditions was
greatest when the subjects could
choose their speeds freely - as might
be expected. The differences between
the two conditions were momentarily
really great, even up to 40 km/h.
Where traffic was congested, the
effects were considerably smaller.
However, on roads with the posted
limit from 30 km/h up to 70 km/h,
the effects of the limiter were
consistently seen no matter whether
the subjects were driving in platoons
or outside them. The analysis of
variance revealed that the speed
limiter had a statistically significant
effect on the mean travel speeds in
urban areas on 30 km/h stretches
F(1,131)=4.69; p<0.05; on 40 km/h
stretches F(1, 39)=22.84; p<0.001; on
50 km/h stretches F(1,131)=21.07;
p<0.001; and on 60 km/h stretches F(1,
38)=6.71; p<0.05.
The recorded number of the
limiter interferences by driver shows
that practically every subject tried to
exceed the posted limit at some
point, some drivers at all times and
the others only occasionally. The
main conclusion is that automatic
speed limiting via in-car equipment
is promising within built-up areas.
On the other hand, very little data is
yet available on driving in truly rural
conditions where the speeds are
highest. The acceptance of the
system amongst drivers is the highest
in built-up areas.
VTT is currently further
developing the ISA concept. The
recently constructed ISA-car has
been tested with a number of subjects.
VTT’s car has some new options for
intelligent speed adaptation. The
possibilities for the speed adaptation
are:
• Warning function: The posted
speed limit is shown on the display
of the car. A spoken warning
message is conveyed to the driver
when he/she exceeds the posted
limit. The message is repeated at 10
second intervals until the driver
has adjusted his speed according
to the limit.
• Recording function: The posted
speed limit is shown on the display
of the car. Moreover, the display
indicates also the time travelled
over the limit as percentage points.
• Speed limiting function: The posted
speed limit is shown on the display
of the car. When the maximum
allowed speed is reached, a yellow
dot is shown on the display
indicating there is a counter-force
on the gas pedal preventing the
driver from speeding.
The results of the Finnish pilot study
are currently under evaluation.
Two cases oftechnologies forfeasibilityconsideration
Wireless location-basedinformation services to launchthe car repair processAlong with the development of
vehicle technology, the reliability of
motor vehicles has been increased
significantly. Going back to “the good
old days” in the 1950’s to 1970’s it
was customary that, upon leaving for
a trip, you would be well prepared
for a possible technical failure during
your trip. As late as in the 1980’s,
there was an extensive network of
service stations along the roads, and
they could really do some repairs to
your car without regard to the make
and model.
Currently, the situation is much
better in terms of failure frequency
of the cars. Also preventive
maintenance has developed thanks
to the sophisticated diagnostic
systems, and several cars have a “limp
home” feature that makes it possible
to drive home even if something
breaks down. But when your car
really breaks down, you will certainly
have a problem. Due to the fact that
people trust their cars and are not
prepared for breakdowns, the
situation is more inconvenient that
in the early days.
The need for model specific
knowledge and expensive special
repair tools has lead to the situation
in which you will have to find a
specialist who has an understanding
of your make and model of car.
Because there is only a limited
number of authorised repair shops
available, they usually have long
waiting lists, even several weeks long.
The situation gets even more difficult
if you are a tourist: you may have
quite a job even to find a suitable
repair shop. Of course, there are
Future Mobility Services in Urban Areas - Selected Cases
ATR_2001.p65 5.12.2001, 11:3331
AU
TOM
ATIO
N T
EC
HN
OLO
GY
RE
VIE
W 2
001
32
dealer lists, yellow pages and mobile
phones available to get help, but you
have to manage quite a process
before your car is running again. At
this point, modern information
technology could bring added value:
the management of the repair
process could be made automatic so
you would just need to activate the
process with your special
requirements, and things would be
arranged for you.
When we think of the possibilities
of modern information technology
and the schema of value added
services, we can finally harness them
to the service of the car user by
making maintenance arrangements
easier and giving some extra support
in possible breakdown situations.
Three opportunities for improving
customer service are pointed out:
1) Services that will make it easy to find
a suitable repair or maintenance place,
and make the necessary practical
arrangements, even in a strange
country and when the user is without
any special knowledge of the local
language. This could include methods
that will find the service that has, at that
time, the capacity to take the car in
almost immediately. 2) Following the
actual breakdown situation, help that
will offer the driver the means to
continue his trip easily, and to get the
car moved to a safe place. 3) During
the time when the car is in the
possession of service personnel,
information that will help the driver
know where the car is and when it will
be ready for pickup.
The advantages for the driver are
obvious. In addition, there are also
advantages for the car manufacturer
and for the respective car service
stations. The manufacturer will get an
image of a good manufacturer who
really is interested in serving the car
user in every situation where their
car is involved. The car repair shops
will benefit by getting more loyal
customers who will use their
services because, if the customers are
satisfied and can trust the services
of one repair shop, they will be
unwilling to try another.
Centralised parkingmanagement systemAn increasing number of
municipalities want to reduce the
amount of private cars in the central
city area. However, consumers want
to get as close to their destination, e.g.
a mall or a theatre, as possible. Here
we have the traditional conf lict
situation. We also face the well-known
congestion, noise and pollution
problems.
If we approach the problem from
the car user’s point of view, the
reason for driving “to the door step”
is usually to act according to the
principle of least effort. Sometimes
this is justified, such as when the user
has to carry heavy bags or boxes, or
when he just has a reduced capability
to move. The other common reason
is an attempt to save time, which is a
consequence of our busy life style:
you just do not want to wait half an
hour to get a bus, rather you want to
leave right away. If we think about
congestion, for example, we know
that both these attempts will be
confronted by a serious problem.
The above leads us to presume the
following: if we could guarantee the
availability and quality of public
transportation, and could increase its
attractiveness by offering some
appealing features, then we could
give positive impetus to people to
change their travel habits in city
areas. Definitely, not all of those who
could use public transportation
would change their travel patterns;
however, there is a possibility of
changing the behaviour of a
sufficient number of travellers to
improve the situation in city centres.
It is a fact that people entering the
city centre from the commuter belt
will often use their cars because
there is no adequate public transport
network outside the city centre.
Another reason is that they will
probably buy things that are difficult
to carry in buses or metros. But how
could these people be guided easily
to benefit from the public transport
network available in the city centre?
The following aspects should be
taken into account: 1) They should
be easily able to park their car; 2)
they should be made aware of the
possibilities of public transport; 3)
payments relating to the use of public
transport should be easy and
reasonable; 4) and leaving the car in
a parking lot, parking house or even
at home should be made more
attractive than currently.
A power cure for this problem
could be the following: 1) to inform
people of this kind about the new
services, and to make access to these
services extremely easy; 2) to equip
people who come to town with
information about the nearest large
parking area located close to their
main destination and with enough
spaces guaranteed; 3) to give them
clear guidance to the area if they are
not familiar with it; 4) to provide a
central payment system that makes
it possible for users to leave their cars
for as long as they need without
coming back to add time to the meter
or buy a new ticket; 5) to give
sufficient information as to the
possibilities of how to continue from
the parking lot to the destinations
users are going to; 6) to provide
order-delivery help if needed to, for
example, carry heavy bags to users’
cars, or help users when it rains; 7)
to make all these services available
at such a cost that users will not
hesitate to pay when they become
available.
In cases when a person is going
to a certain shop, sports event,
theatre, or restaurant this is quite
easy. However, it must be kept in
mind that many people do not decide
their destinations at home, but visit
shops impulsively; also, this solution
must not make travelling to the city
centre more difficult than when
using the car as usual.
ATR_2001.p65 5.12.2001, 11:3332
33
AU
TOM
ATIO
N T
EC
HN
OLO
GY
RE
VIE
W 2
001
Henrik HuovilaSenior Research ScientistVTT Automation
References1. http://www.schrader-bridgeport.com/clemson.html
2. Salusjärvi, M. (1981) The speed limit experiments on public roads in Finland. Technical Research Centre of Finland.Publication 7/1981. Espoo, Finland.
3. Nilsson, G. (1982) The effect of speed limits on traffic accidents in Sweden. VTI Report 68, Linköping, Sweden.
4. Elvik, R., Vaa, T. and Östvik, E. (1989) Trafikksikkerhetshåndbok, Transportøkonomisk Institutt. Oslo, Norway.
5. Finch, D.J., Kompfner, P., Lockwood, C.R. and Maycock, G. (1994) Speed, speed limits and accidents. Project Report58. Transport Research Laboratory, Crowthorne, UK.
6. O’Cinnéide, D. and Murphy, E. (1994) The Relationship between Geometric Road Design Standards and Driver/Vehicle Behaviour, Level of Service and Safety. Traffic Research Unit, University of Cork.
7. Draskóczy, M. and Mocsári, T. (1997) Present Speeds and Speed Management Methods in Europe. Deliverable R2.1.1 in the MASTER project. VTT, Espoo, Finland.
8. Hansén, L and Hydén, C. (1976) Hastighetsbegränsning vid skolor (Speed limiting at schools, in Swedish). Bulletin18, Lund University, Lund.
9. Hauer, E., Ahlin, F.J. and Bowser, J.S. (1982) Speed enforcement and speed choice. Accident Analysis and Prevention,14(4), pp. 267 - 278.
10. Östvik, E. and Elvik, R. (1990) The effects of speed enforcement on individual road user behaviour and accidents.Proceedings of the International Road Safety Symposium on Enforcement and Rewarding Strategies and Effects.Copenhagen, Denmark.
11. Teed, N., Lund, A.K. and Knoblauch, R. (1993) The duration of speed reductions attributable to radar detectors.Accident Analysis and Prevention 25 (2), pp. 131 - 137.
12. Comte, S.L., Várhelyi, A., Santos, J. (1997) The Effects of ATT and Non-ATT Systems and Treatments on Driver SpeedBehaviour. Working Paper R 3.1.1 in the MASTER project. VTT, Espoo, Finland.
13. Rathmayer, R. and Mäkinen, T. 1995. Measuring driving behaviour without disturbing it. Nordic Road & TransportResearch. Vol. 7 (1995) Nr: 2, pp. 20 - 22. VTT Communities and Infrastructure.
Jari KaikkonenGroup ManagerVTT Automation
Tapani MäkinenSenior Research ScientistVTT Automation
Future Mobility Services in Urban Areas - Selected Cases
ATR_2001.p65 5.12.2001, 11:3333
AU
TOM
ATIO
N T
EC
HN
OLO
GY
RE
VIE
W 2
001
34
Next Generation Industrial Automation -Needs and OpportunitiesTeemu Tommila, Olli Ventä and Kari Koskinen
Despite years of activity, truly open and intelligentcontrol systems seem still to be a promise of the future.Agreement on common architectures and applicationobjects is needed to raise open control systems fromexchanging raw data to the level of real interoperabilityof off-the-shelf components. Future control platformsand programming languages should have new built-inmechanisms that support implementation of intelligentfunctions, such as flexible resource management andexception handling. This article argues that many ofthese challenges can be met by taking full advantageof emerging software engineering technologies. Thisalso means that the modelling techniques and designpractices of software engineering should be combinedwith the traditional ways of thinking in automation.
Challenges inautomationCurrently, industry is striving towards
product quality, safety and
environmental protection. Tight profit
margins and networked manufacturing
emphasise the need for integration and
global optimisation of production
facilities. The role of information
technology in achieving these goals
has become critical. Large and
complex production systems can’t be
efficiently and safely managed
without computers in information
management and process control.
ATR_2001.p65 5.12.2001, 11:3334
35
AU
TOM
ATIO
N T
EC
HN
OLO
GY
RE
VIE
W 2
001
Figure 1. A system model for a distributed control application (from IEC 61499-1).
End users expect to get improved
functionality at reasonable cost.
Management of knowledge and real-
time information, integration with
condition monitoring and plant
maintenance, high availability, and
flexibility of upgrades and life-cycle
support are examples of key
requirements. System integrators
need efficient tools for building
applications. Manufacturers face the
challenge of satisfying customers’
needs while still maintaining a sound
and profitable product structure in a
rapidly changing technical and
business environment.
A control system is a collection of
distributed devices interconnected by
means of a communication network,
Fig. 1. Recent trends are characterised
by geographical distribution and
functional integration. On the
technical level, the goal is to be able
to easily connect devices and software
components from different vendors.
Functionally, there is a need for
interoperability of control functions
on different hierarchical levels ranging
from field equipment to Enterprise
Resource Planning (ERP). Many
customers already use, within certain
security limitations, web-based ‘thin’
clients for on-line knowledge
management, remote monitoring and
maintenance. Future systems will be
based on an even stronger distribution
to the field level, on mobility, and on
co-operation with components in
distant locations.
For many years, integrated,
intelligent and dependable control
systems have been the focus of
standardisation organisations,
industrial consortia and research
groups. The solutions have, however,
been hard to find, partly due to the
complexity of the issue and partly
because of conflicting commercial
interests. Technically, open control
systems still focus on ways of making
bits f low between devices from
different vendors. Only a few efforts
are underway to agree about common
architectures and application objects
that could make systems really
‘understand’ each other. There is also
a gap between research results and
real-life applications. Instead of
elegant control theories, practical
automation projects often struggle
with low-level technical problems.
Important issues are, for instance, how
to find out user requirements, how to
interface different products, and how
to reuse existing (sometimes poorly
structured) application software.
To summarise, there is a need for
an integrated and low-cost system
platform that allows intelligent
features to be easily implemented.
Some starting points can be found
from software engineering. As
illustrated in Fig. 2, earlier generations
of digital control systems have been
combinations of existing automation
practices and advances in electronics
and information technology. With the
emergence of microprocessors in late
70’s, faceplates of pneumatic
controllers were transferred to
computer screens of distributed
control systems (DCS). Since then, PC
technology has, after a long debate,
found its way to industrial
applications. Current control systems
are typically a mixture of many
techniques.
We can expect that a similar
situation will exist in the future also.
Increasing processor power,
consumer electronics, mobile
communication networks and
programming languages provide the
tools to implement smart functions
that have been unrealistic before. The
sections below discuss the functional
features needed and give a short
review of the most important
implementation technologies. As a
conclusion, some suggestions are
made for future development.
Concepts forintelligent functions
Intelligent machinesIn an industrial plant, physical
process systems consist of machines
and process equipment. They are
individual devices or larger
subsystems of their own. This leads
to a wholes-parts hierarchy as
shown in Fig. 3. Process systems can
be in different operational states,
such as ‘maintenance’, ‘starting up’ or
‘operating’. In each state, they
provide a set of capabilities that can
be combined to perform the various
stages of the process. In the course
of control system design, control
tasks identified in co-operation with
users and other engineering
disciplines are allocated to the
control system and human operators.
The automated parts should form a
structured set of control activities
corresponding to the physical
equipment and processing tasks.
ATR_2001.p65 5.12.2001, 11:3435
AU
TOM
ATIO
N T
EC
HN
OLO
GY
RE
VIE
W 2
001
36
Figure 2. New generations of control systems have typically been combinations of new information technology and moretraditional thinking in automation.
Figure 3. Process entities and automation activities are arranged as hierarchies. Capabilities of process equipment are usedto carry out process phases. In the control system, production oriented functions of operations management send commandsto process control functions associated with physical equipment.
For example, process units in a
multipurpose batch plant or
machines in a flexible manufacturing
system have several capabilities and
can be used for different purposes. A
pool of resources can even be re-
configured into ‘virtual production
lines’. In the control domain, various
product recipes are defined on the
basis of services programmed on the
equipment control level of the
control system. Each process control
component takes care of the process
system it represents, including for
instance:
• reading process measurements
and performing control actions
• managing physical resources of
the process system
• controlling operational state and
operating mode
• condition monitoring and
exception handling
• services for other activities.
This arrangement makes the physical
manufacturing resources behave in
an ‘intelligent’ way. The actual
implementation may vary. For
example, controls can be provided by
the equipment manufacturer and
ATR_2001.p65 5.12.2001, 11:3436
37
AU
TOM
ATIO
N T
EC
HN
OLO
GY
RE
VIE
W 2
001
Figure 4. Mechanisms for connecting automation components include terminals,services and event notifications.
Next Generation Industrial Automation - Needs and Opportunities
embedded into the physical device
(e.g. intelligent valves). The functions
can also be included into the ‘main
control system’. In both cases, the
know-how for controlling and
maintaining the equipment is
typically in the interest of the
equipment manufacturer. Open
control systems are one way to make
the combination of tangible products
with related support services, i.e. the
concept of extended products more
practical.
Components in automationControl functions like sequences,
PID-loops, and displays, can be
described as automation
components. They are similar to
objects in object-oriented
programming with the difference
that they, in addition to responding
to external messages, have internal,
periodic or continuous activities. The
external interface of automation
components includes terminals
(ports), services and event
notifications, Fig. 4. Components
read and assign values of terminals
of other activities. This ‘wiring’ is the
common paradigm currently used in
function block programming. In
addition, a component can request a
service from another component by
sending a request message. To
propagate events, components send
notification messages to interested
listeners. Also other distributed
programming models, such as
message queues or shared memory,
might be considered for control
applications. These approaches have,
so far, been more common in
information systems and object-
oriented programming, although
they are becoming more familiar to
control engineers, along with
standards like DCOM, OPC and
CORBA.
Automation components are
organised in a hierarchical manner.
They may consist of lower-level
components or be basic components
implemented in other techniques,
such as function blocks (IEC 61131-
3, IEC 61499-1) or a general purpose
programming language. A product
recipe, for example, may consist of
several unit recipes. A component
logically contains the components
controlling the lower-level process
systems even if they are allocated to
other control devices. The allocation
is usually static, but an activity can
also be re-allocated to another
computer if one platform fails or
becomes overloaded. During
operation, a component can acquire
external resources owned by other
components. Managing shared
resources is essential for
implementing flexibility in control
systems. The hierarchical relations to
owners and clients should be
maintained in the run-time
environment. For example, they can
be used to propagate component
status allowing upper levels and
clients to react to device failures.
‘Plug & play’ featuresCurrent applications are
combinations of control products
from different vendors. Furthermore,
product and process changes are
more frequent than earlier. Therefore,
control devices and automation
components must be able to describe
themselves to designers, human
operators and other automation
components. During system
operation, newly inserted devices
and software components must have
ways of looking up the rest of the
application and to advertise their
own capabilities. Devices and
network segments can be arranged
to ‘system areas’ in a hierarchical
fashion. Root devices in each area can
maintain directories of other nodes.
This results in a distributed directory
service embedded into the control
system itself.
Exception handlingIn addition to changes in products
and production schedules, control
systems should cope with other
types of unexpected situations,
namely disturbances originating
from process fluctuations, failure of
process equipment or faults in
control system hardware and
software. The scope of exception
handling covers issues from fault
avoidance in design to problem
identification and display, diagnosis,
corrective action and recovery
during system operation, and
continuous improvement on the
basis of problem reports. Even if as
much as 50 % of control system
software and design costs is related
to exception handling, only a few
practical approaches and tools are
available. For example, alarm floods
are still a problem. Current
programming languages provide
virtually no support for managing
exceptions. While actual algorithms
ATR_2001.p65 5.12.2001, 11:3437
AU
TOM
ATIO
N T
EC
HN
OLO
GY
RE
VIE
W 2
001
38
Figure 5. Layers for distributed automation.
depend on the situation and
equipment under control, the control
platform should include some built-
in features that make abnormal
situation management easier. For
example, automation components
can have a ‘health’ attribute that
describes, on a qualitative scale, its
performance. Intelligent, situation-
aware, event generation implemented
on the spot would limit the generation
of nuisance alarms and the need for
alarm filtering at the user interface.
The designer should be able to specify
monitors running in parallel with the
normal actions. Depending on the
situation, they could force control
functions, e.g. sequences, to an
appropriate exception routine.
General principles of Quality of
Service (QoS), like request priorities,
performance figures for services,
deadlines for network messages, could
also be included in the basic control
platform.
Enabling technologiesImplementation of the features
outlined above is outside the scope
of this paper. Instead, we list some
relevant developments to show their
potential for industrial automation.
Open control systemsIn the control business, the
requirement for openness has led to
numerous development and
standardisation efforts, both by
industrial groups and standardisation
organisations (see Table 2). Among
the most important are:
• Programming languages for
programmable logic controllers
• Field bus systems
• Reference models for batch
automation and manufacturing
execution systems
• OLE for process control, OPC.
These working groups have done an
excellent job by documenting best
automation practices and defining
new concepts, although sometimes
their work has been slowed down by
commercial interests and the need to
comply with existing products.
Another problem is the difficulty of
co-ordinating the different working
groups and research projects. This has
led to contradictory and overlapping
documents. An example is the
emergence of several standards for
field buses instead of just one, or a few,
for different types of applications.
Another tendency has been
focussed on the lower levels of
communication. Only a few attempts
have been made to agree about the
application level objects for specific
domains in automation. A good
example is given by the guidelines
for batch process control developed
by the SP88 committee of ISA. In
addition, committee SP95 is actively
working on the higher levels of
control, writing a recommendation
for enterprise and control system
integration. As a further example,
draft IEC 61850 defines rules for
exchanging real-time data in
electricity distribution. It includes
application object models of
common functions and components,
such as voltage regulators and
protection relays. While still on a
rather abstract level, these efforts can
be seen as domain-specific reference
architectures paving the way for
more concrete implementation
architectures in the future.
CommunicationA multitude of protocols have
emerged for communication in
process, manufacturing and building
automation. The key standards for
field devices level are IEC 61158 and
EN 50170. The benefits include
reduced wiring, improved diagnostic
and measurement data, the possibility
of remote diagnostics and improved
control at field level. The problem lies
in their limited interoperational
capability. FoundationFieldbus is
ATR_2001.p65 5.12.2001, 11:3438
39
AU
TOM
ATIO
N T
EC
HN
OLO
GY
RE
VIE
W 2
001
EN 50170 1996. General purpose field communication system.
IEC 61131-3 2001. Programmable controllers - Part 3: Programming languages, 2nd edition, finaldraft.
IEC 61158 2000. Digital data communications for measurement and control - Field bus for use inindustrial control systems, parts 1 to 6.
IEC 61850-1 2001. Communication networks and systems in substations – Part 1: Introduction andoverview. Committee draft.
IEC 61499-1 2000. Function blocks for industrial-process measurement and control systems - Part1: Architecture. Publicly available specification, draft.
IEC 61804-1 1999. Function blocks for process control - Part 1: General requirements. Committeedraft.
ANSI/ISA-88.01 1995. Batch Control - Part 1: Models and Terminology.
ANSI/ISA-95.00.01 2000. Enterprise-Control System Integration - Part 1: Models and Terminology.
ISO/DIS 15745-1 2000. Industrial automation systems and integration – Open systems applicationintegration frameworks – Part 1: Generic reference description. Draft international standard.
Table 1. Standards and drafts related to open control systems
Next Generation Industrial Automation - Needs and Opportunities
currently the only one that allows
distribution of control functions to
field devices.
Ethernet with the TCP/IP protocol
has become the de facto standard in
the office and IT world. It has found
its way to the upper levels of
automation networks and is now
invading the field level also. Recent
enhancements to the field bus
architecture have included the
integration of Ethernet with the
existing protocols. For example, High
Speed Ethernet (HSE) uses standard,
low-cost equipment.
Standardised use of the TCP/IP
protocol guarantees a seamless
transition into the Internet. Despite
its non-deterministic features, the
improved network speed has made
it a feasible solution for many real-
time control applications. A Web
server can be integrated into even
small control devices. Familiar
technologies, such as programming
languages and WWW browsers can
be used easily. The eXtensible
Markup Language (XML) is a
promising tool for application
integration in the business field. It is
also used or considered in several
standards and R&D projects in the
control domain.
Support for distributionHowever, these standards don’t
include the application and user
layers needed to achieve device
interoperability. Therefore, a number
of working groups are developing
specifications for the adoption of
Ethernet to industrial automation. To
take one example: IDA, Interface for
Distributed Automation, is an
industrial group that combines
Ethernet and Web technologies to
develop distributed control
architectures. The function blocks of
IEC 61499 are used as a reference
model . The communicat ion
mechanisms include data distribution,
event notifications and remote
method invocation. XML based object
descriptions and SOAP Remote
Procedure Call (RPC) are used for
configuration and diagnostics. Each
device contains an embedded Web
server. Web browsers and Java Applets
are used for operator interfaces. Real-
time communication is based on the
Real-Time Publish/Subscribe (RTPS)
model that is built on top of the UDP/
IP protocol, a solution that saves
network and processor load. The RTPS
middleware layer takes care of
discovering and publishing objects in
the network and sending data from
publishers to subscribers. Reliability
features include priorities, time-out,
guaranteed data delivery and
redundant service/data providers.
Mobile communicationThe markets for wireless
communication have been growing
rapidly. It has many benefits when
compared to traditional wiring,
including, for example, flexibility and
savings in cabling costs. On the other
hand, radio communication is more
susceptible to electromagnetic
interference and often has limitations
in data transfer capacity and
maximum distances between
devices. Moreover, radio waves
spreading outside the factory area
may be a threat to system security.
In industrial automation, radio
links have been used for selected data
transfer needs for a long time.
Recently, text messages delivered via
Short Message Service (SMS) have
been applied to supplying alarm and
ATR_2001.p65 5.12.2001, 11:3439
AU
TOM
ATIO
N T
EC
HN
OLO
GY
RE
VIE
W 2
001
40
Table 2. Selected WWW links
Industrial Automation Open Networking Alliance, IAONA http://www.iaona-eu.com
Interface for Distributed Automation, IDA http://www.ida-group.org
Open Device Net Vendor Association, ODVA http://www.odva.org
Open Modular Architecture Controls, OMAC http://www.arcweb.com/omac
Association Connecting Electronics Industries, IPC http://www.ipc.org
Factory Information Systems, NEMI http://www.fis.nemi.org
Open System Architecture for Controls within http://www.osaca.org/osacaAutomation Systems, OSACA
OPC Foundation http://www.opcfoundation.org
Standardization in Industrial Control Programming, PLCopen http://www.plcopen.org
Function Block-Based, Holonic Systems Technology http://www.holobloc.com
Object Management Group, OMG http://www.omg.org
World Batch Forum, WBF http://www.wbf.org
Instrumentation, Systems and Automation Society, ISA http://www.isa.org
diagnostic information to remote
operators and maintenance
technicians. Wireless versions of field
buses are being developed and
efforts are underway to explore the
potential of the newest wireless
techniques such as BlueTooth. In
spite of growing interest, the use of
wireless solutions for industrial
control systems is currently limited
to less critical applications such as
portable user interfaces, remote
monitoring and messaging services.
Software componentsThe powerful principles of object-
oriented programming provide new
possibilities for the control sector, too.
In spite of the wide interest, object
orientation and related design
approaches, like UML, are not
commonly used in the design of
control applications. In general, object-
oriented design and programming has
not fulfilled all its promises for better
reusability, quality and efficiency of
software development. In order to
enable easier runtime reuse, alleviate
problems of implementation
inheritance and the difficulties in
implementing distributed objects,
software component technologies
have emerged. Slightly different from
every-day language, software
components are typically defined as
reusable, self-contained pieces of
software that can be used in binary
format by third parties through well-
defined interfaces. They may be results
of in-house development or
commercial components off-the-shelf
(COTS). Applications can, at best, be
easily made up by connecting the
interfaces provided and expected by
components.
The need for distributed
components typically leads to the use
of a middleware product providing
the necessary object distribution
models. Currently, the most common
are the Distributed Component
Model (DCOM) from Microsoft and
CORBA (Common Object Request
Broker Architecture) developed by
the Object Management Group
(OMG). A third one is the Enterprise
JavaBeans (EJB) model from Sun
Microsystems. So far, the software
component business in automation
is still at its beginning. Large vendors
are using some principles of
component software in their product
development. Also, some references
to components can be found in
advertising. However, while common
standards and frameworks are
missing, no real component markets
can emerge.
ConclusionsIn current discussions, intelligence
often refers to the use of advanced
techniques, such as neural networks,
or to a new way of implementing
existing functions. We suggest that
new functionality, e.g. management
of hierarchical structures and
exception handling, should be
included in the basic control
platform and engineering tools. The
current ‘flat’ collection of application
modules like loops and sequences
should be organised in a more
hierarchical fashion based on process
structure. Each process system is
seen as an intelligent resource
capable of performing different
processing tasks. The interaction
ATR_2001.p65 5.12.2001, 11:3440
41
AU
TOM
ATIO
N T
EC
HN
OLO
GY
RE
VIE
W 2
001
Olli VentäGroup ManagerVTT Automation
Teemu TommilaSenior Research ScientistVTT Automation
Kari KoskinenD.Tech., ProfessorHelsinki University ofTechnology
Next Generation Industrial Automation - Needs and Opportunities
mechanisms between different
automation activities are defined on
the basis of object-oriented analysis
and design, and emerging
international standards. A
standardised distribution middleware
takes care of the needs specific to the
control domain. Above that, a higher-
level working environment for the
other system components of the
control platform is needed, Fig. 5.
Their role is to provide the container
for application components. The
principles of component-based
software development can be
applied to the development of
control products and applications. To
be useful in a multi-vendor situation,
an agreement about application
environments and data presentation
is needed.
AcknowledgementFinancial support for this work was
provided by the Finnish National
Technology Agency (TEKES),
industrial participants and VTT
Automation. The current project on
new component-based automation
architectures is a part of the ongoing
TEKES technology programme on
‘Intelligent Automation Systems’.
ATR_2001.p65 5.12.2001, 11:3441
AU
TOM
ATIO
N T
EC
HN
OLO
GY
RE
VIE
W 2
001
42
Customized Dynamic Simulator SupportsProcess & Control Engineering at Mill SiteSami Tuuri, Jari Lappalainen and Kaj Juslin
Dynamic simulation is being adapted as a commontechnology for the design and analysis of dynamicbehaviour of processes. In a wider perspective, adetailed dynamic simulator is an ideal tool for studyingand developing the co-operative performance of processoperators, automation systems, and the productionprocesses. The potential of dynamic simulators is notrestricted to the design and operator training phase,but extends to the whole mill life cycle. However, therehas been a threshold to taking dynamic simulation toolsinto use by mill personnel, although there would bepotential benefits in doing so. This article focuses onpresenting experiences in using dynamic simulation ina paperboard mill for the purpose of improving millperformance during transient operating conditions.
IntroductionDynamic simulation as a method is
making its breakthrough into process
industries. The potential of the
dynamic simulation has been
regarded as very high throughout the
process life cycle: pre-design, detailed
design, commissioning, operator
training, and operation support. Since
1994, VTT Automation has been
systematically developing a modelling
and simulation platform for pulp and
paper processes. The platform, called
APMS, has been designed to meet
most of the dynamic simulation needs
ATR_2001.p65 5.12.2001, 11:3442
43
AU
TOM
ATIO
N T
EC
HN
OLO
GY
RE
VIE
W 2
001
Figure 1. An example of a production cycle of a board machine.
of the pulp and paper industry.
The high-quality and validated
model library ensures the accuracy of
the simulation results of even
processes that are in the design phase
/1/. The simulator provides data that
helps in developing solutions where
process and automation work better
together, leading to higher product
quality, better runnability and more
economical process solutions.
Dynamic simulation as a method can
integrate process and automation
design procedures: it offers a common
platform for process and control
engineers to discuss, analyse and
further elaborate the solutions under
design. When the model building and
utilisation starts in this early phase,
dynamic simulation offers a cost and
time effective way of aiding
engineering during the whole life
cycle of the process.
At some point in near future, mills
will start to demand the simulator-
aided automation system checkout
from the automation system providers.
Tools based on OPC (OLE for Process
Control) specification facilitate an easy
and flexible connection between the
simulator and control system. The
quality of the automation application
improves when the control and logic
systems can be checked out early
enough, ensuring time for proper re-
design.
In addition, operator training
assisted by customized training
simulators is becoming a standard part
of a delivery. The same simulator that
has been used in design and check-out
can be further utilised in the training
of operation personnel, both prior to
start-up and as on-going training after
plant commissioning /2/. With
authentic machine-like behaviour, the
operators can get hands-on experience
of the process. The trainee interface can
be either an original automation
system display, or an emulated display.
A process simulator could be used
for interactive problem solving and
process optimisation at the mill site.
This is especially lucrative if the
process simulator has already been
constructed in the design phase.
Alternative operational procedures and
control strategies could be studied and
tested parallel to the real process.
Satisfactory results in simulation
studies could be used in justifying
minor automation parameter changes,
or even investment decisions.
However, dynamic simulation has not
yet gained ground at mill sites. Mill
personnel set higher requirements for
dynamic simulation tools than process
or automation design engineers. The
tool should be very easy to use, so that
the threshold of using it is low enough
even for casual users. Above all,
however, is the requirement for very
high modelling accuracy. Mill
personnel know the behaviour of their
process too well to be satisfied with
generic simulator behaviour.
Confidence is lost if simulator
behaviour is not identical to real mill
behaviour.
This article describes some
experiences of using the dynamic
simulator tool in a paperboard mill
environment as an aid for process
engineering. First, the challenges of
multi-product paperboard mill
operation and its automatic control are
described. Next, a simulator user from
the mill site describes his experiences
in assessing how dynamic simulation
has been useful, and expresses his
views on the future. Finally, some
representative simulation results of
that case are presented.
Automation In Multi-Product Paper MillsToday, paper and paperboard mills
produce a variety (dozens) of product
types or grades to meet different
customer specific quality
requirements. Customer order sizes
are decreasing, resulting in smaller
batch sizes. Demands for cost-
effectiveness urge shorter delivery
times. In addition, the variety of raw
materials, of which some are recycled,
ATR_2001.p65 5.12.2001, 11:3443
AU
TOM
ATIO
N T
EC
HN
OLO
GY
RE
VIE
W 2
001
44
Figure 2. The automatic grade change program calculates new target values for low-level controls and coordinates theirmutual ramping during the grade change.
has increased. Instead of a few product
types, paper and board machines are
more often dedicated to multi-grade
production. The advantages of high-
volume or bulk production, which
prevailed in the paper industry in the
past, have been removed to some
extent.
The goal at a paper or board mill is
to produce the correct amount of the
“old” grade, and then start producing
the “new” grade as soon as possible in
order to minimize losses, i.e. of paper
or board, fulfilling the quality
requirements of neither the “old” nor
the “new” grade. This change from the
“old” quality to the “new” one is called
a grade change. The consecutively
produced grades should be as close
to each other in quality specifications
as possible. To achieve this, production
planning has an essential role. Since
most of the grade changes are usually
basis weight transitions, the
production cycle looks typically like
the one presented in Fig. 1. Despite
careful planning, it is usually not
possible to completely avoid losses
due to production of off-specification
qualities during grade changes.
Another key matter from the grade
change point of view is whether the
machine has a limited or large range
of products. In order to minimize
product inventories, and to ensure
production is exactly on time, the
mills having a large repertoire of
grades have to make grade changes
more frequently. In such multi-grade
machines, grade change performance
plays an important role when
maximizing the overall production
efficiency of the mill.
Grade changes are among the most
distinctive and critical types of tasks
performed in modern paper or board
mills. Although skilled and
experienced operators are able to
carry out grade changes manually,
their performance varies considerably
from case to case. Automatic grade
change is potentially faster and more
reliable than manual operation.
Typically, an automatic grade
change program is based on open-
loop control, i.e. the quality controls
are switched off during the grade
change. The manipulated variables
are ramped with pre-planned targets,
ramping rates and mutual timing.
Also, more advanced methods for
grade change automation have been
proposed /3/.
Firstly, the grade change program
has to calculate the new target values
for the manipulated variables. The
target values should be such that all
quality variables reach and remain
the new grade’s specifications as
soon as possible. Secondly, the
ramping of different variables must
be coordinated by using suitable,
variable-specific waiting times and
ramping rates in order to compensate
the effects of different time constants
ATR_2001.p65 5.12.2001, 11:3444
45
AU
TOM
ATIO
N T
EC
HN
OLO
GY
RE
VIE
W 2
001
Tommi Myller, Process Engineer from
Stora Enso’s Imatra Mills:
“The simulator helps justifyingchanges by trying variousalternatives”
During the past two years, we have taken the first steps in familiarizing ourselves with dynamic simulation. I wouldsay, that for any typical pulp and paper mill, dynamic simulation is a totally new method. Accordingly, we are veryinterested to see what its possibilities are, and how easy simulation tools are to use for a mill engineer.
We are quite satisfied with the first experiences we have gained using a customized board machine simulator asan engineering tool at our site. We wanted to use the simulator as an aid in our efforts to speed up grade changesand stabilize paper moisture variations. However, we also got some other benefits we couldn’t have thought of whenwe started the simulation project.
Already the modelling phase - the configuration of the process and automation model into the simulation software- has turned out to be useful. Many different kinds of engineering data were accumulated in the model database.Often the process behaviour had to be thought over from a new point of view. This kind of re-engineering clearlyrefreshed and deepened our understanding of our process and its automation.
We carried out quite extensive model validation against process measurement data from actual grade changes.In the beginning, it often happened that the simulated grade changes didn’t fit well with the measurements. Wefigured out the reason was that some operators used the automatics in a different way from others, and fromourselves, in the simulations. Of course, this is quite natural when such a long time has passed since the systemwas taken into use and training given. Anyway, we could pinpoint some undesirable practices in using the gradechange automatics. Thus, the simulations helped us to unify operation procedures closer to the best practice.
We tried out different parameters for the grade change program on the simulator. After validations, we had quitea high level of confidence in the simulation results. When the parameters gave satisfactory results in the simulator,we had the courage to put them into the real automation system. So far, we have managed to shorten grade changetime considerably in a number of cases. Many of the ideas on how we could improve the performance alreadyexisted before simulations. However, the simulations gave us the needed evidence and justification to put theseideas into practice. The simulator simply encouraged us to step ahead from ideas to implementation.
We continue to use the simulator in helping us in further improving the grade change performance. We havestarted using the simulator as a tool in other engineering projects, as well. It is not a single-purpose tool, but can beused as a continuous process development aid.
and delays of the process.
It is difficult to estimate the target
steam pressures accurately enough to
reach the desired paper moisture,
without the need to correct the
moisture with feedback control after
the grade change ramps. In addition, it
may be difficult to prevent the paper
moisture from fluctuating during
ramping of the basis weight and the
machine speed. The paper moisture is
usually the difficult part, because
changes in raw materials, dirtiness of
wires and cy l inder sur faces ,
condensate layer thickness in cylinders,
felts etc. should be taken into account
in the estimation method. In practice,
it is a very challenging task to tune up
the grade change control to give good
performance for multi-grade machines.
Challenges Of APaperboard MachineA board machine at Imatra Mills of
Stora-Enso Ltd produces a large
variety of three-ply liquids packaging
boards. The basis weight range is
large: 170 - 350 g/m2. Additionally, the
products vary in the mixture of
furnish materials used, how the three
plies have been proportioned to each
other, coatings, etc. The machine
speed varies between 200 and 450
m/min. The yearly production
capacity of the machine is 200000
tons paperboard. Fig. 2 shows a
simplified process flow diagram.
Due to the large number of board
grades and the size of customer-
orders, a production “batch” is
typically quite small. So, it is
necessary to carry out a grade change
relatively often - averaging once a day.
Practically all grade changes for
several years have been carried out
using an automatic grade change
program. Fig. 2 shows the variables
the program controls and the inputs
it uses. After the operator has initiated
the grade change, the program
coordinates the mutual delays and
ramping of the controlled variables.
Customized Dynamic Simulator Supports Process & Control Engineering at Mill Site
ATR_2001.p65 5.12.2001, 11:3445
AU
TOM
ATIO
N T
EC
HN
OLO
GY
RE
VIE
W 2
001
46
This coordination is pre-planned by
giving a start delay, maximum
stepping rate, and a stop delay for
each control variable. After a grade
change, when a certain waiting time
has passed, the quality controls are
switched on.
The grade change program has
been operating reasonably well.
However, considerable variations on
paper moisture content have taken
place in many cases. It has been noted
that clearly the moisture bounds up
or down when the machine speed
starts to change, and again just after
reaching the target speed. Additionally,
it has been generally thought among
the personnel that the time of grade
changes could still be reduced.
An intuitive reaction was to think
that by faster ramping of the variables,
and better mutual timing, the situation
could be improved. But, how? For
example, it is very challenging to try
to figure out the right actions needed
to fix the moisture fluctuations using
just a mind model. The number of
tuning parameters of the grade
change program in a three-ply
machine is remarkable. Different ideas
compete and conflict. This kind of
“opinion engineering” is a quite fragile
basis on which to start experiments
with the real machine.
Dynamic Simulator As AProblem Solving Tool AtMill SiteA customized dynamic simulator was
seen as an attractive tool to test
different ideas before doing anything
on the actual board machine.
Extensive test runs and trials on real
machine are expensive due to the
potential of production losses, and
sometimes impossible to carry out.
The paperboard machine model
was configured using the APMS
software /4/. The simulation model
consists of major process units and
piping from stock preparation to the
end of the baseboard drying. In
addition, major automatic control
loops were included. A special
attention was paid in the modelling
of drying section and grade change
automation. The grade change
automation was modelled in detail:
the model included all the tuning
parameters that are available in the
actual program. It was important to
make the simulator easy to play with
various parameter combinations.
For the time being, the simulator has
helped the mill people to shorten the
grade change times considerably. The
simulator is used at mill sit as a tool
aiding in other tasks related to process
and automation development.
Model validaty was extensively
tested against actual process
measurement data, and the simulator
showed very good agreement with the
measurements. Fig. 3 shows some of
the major quality and manipulated
variables during one representative
grade change.
After validation phase, the
simulator was moved into the mill
site as a process engineer’s tool. The
Figure 3. Development of major quality variables and manipulated variables duringpaperboard grade change.
Figure 4. Graphical user interface of APMS modeling and simulation tool.
ATR_2001.p65 5.12.2001, 11:3446
47
AU
TOM
ATIO
N T
EC
HN
OLO
GY
RE
VIE
W 2
001
Jari LappalainenSenior Research ScientistVTT Automation
Sami TuuriSenior Research ScientistVTT Automation
Kaj JuslinGroup ManagerVTT Automation
References1. Kokko, T., M. Hietaniemi, J. Ahola, T. Huhtelin, P. Lautala, Development of Paper Machine Wet End UsingSimulation, Proceedings of 3rd ECOPAPERTECH Conference, June 4 - 8, 2001, Helsinki, Finland, pp. 135 - 146.
2. Tervola, P., Lappalainen, J., Rinne, J., Leinonen, T., Peltonen, S., Karhela, T., Juslin, K., Bleach Plant Training SimulatorFeaturing Enhanced Linkage Between Simulator and DCS, Proceedings of 1999 TAPPI Pulping Conference, October 31.– December 4., 1999, Orlando, Florida, USA, TAPPI Press, pp. 1031 - 1045.
3. Välisuo, H., Niemenmaa, A., Lappalainen, J., Laukkanen, I. and Juslin, K., “Dynamic simulation of paper and board mills: acase study of an advanced grade change method”, Proceedings of TAPPI Engineering Conference 1996, September 16 - 19,Chicago, USA, pp. 491 - 498.
4. Lappalainen, J., Myller, T., Vehviläinen, O., Tuuri, S., Juslin, K., Enhancing grade changes using dynamic simulation,Proceedings of TAPPI 2001 Engineering/Finishing & Converting Conference, December 2-6, 2001, San Antonio,Texas, USA, TAPPI press, 11 p.
http:\www.vtt.fi\aut\tau\ala\apms.htm
Customized Dynamic Simulator Supports Process & Control Engineering at Mill Site
simulator was installed on a standard
desktop PC, where it was operated
much like a real mill, though using
the simulator’s design graphics as a
user interface (Fig. 4).
ConclusionsDynamic simulation is making a break-
through as a multi-purpose tool in pulp
and paper technology. VTT Automation
is committed to bringing the large
potential of dynamic simulation into
reality in close co-operation with
process and automation system
providers, engineering companies, and
the process industry. The on-going
projects at VTT Automation address
• increase user-friendliness of the
simulation system,
• developing tools and working
methods for simulator aided
engineering,
• integrating various simulation
software tools into a common
platform,
• improving compatibility of
automation systems with
simulators, and
• developing tools and model
libraries for easy delivery over the
Internet.
ATR_2001.p65 5.12.2001, 11:3447
AU
TOM
ATIO
N T
EC
HN
OLO
GY
RE
VIE
W 2
001
48
SCM is about integrationIlkka Seilonen, Juha Nurmilaakso, Jari Kettunen, Stefan Jakobsson, Petri Kalliokoski, Markku Mikkolaand Veli-Pekka Mattila
Today, many companies are either developing orconsidering the introduction of SCM (Supply ChainManagement) systems. This kind of development effortrequires a many-sided knowledge of both the businessprocesses and the IT systems of interacting parties.Based on our experience gained in two projects, Gnosis-VF and Dedemas, we present results from two differentcase studies. The case studies indicate how an adequateintegration of IT systems enables an enhancement ofbusiness processes in distributed and networkedorganizations. Both projects also contained an evaluationphase in order to estimate how and to what extent SCMapplications produce actual business benefits.
IntroductionWhen developing SCM systems,
companies have to carefully match
the requirements of their business
processes and the capabilities of
current technology. As researchers,
we have studied this task by means
of case studies. We hope that the
descriptions of these case studies,
including their requirements,
prototype solutions and evaluations,
provide useful insights both into the
manufacturing companies considering
ATR_2001.p65 5.12.2001, 11:3448
49
AU
TOM
ATIO
N T
EC
HN
OLO
GY
RE
VIE
W 2
001
adoption of SCM systems and into the
IT vendors developing them.
The case studies described in this
article were carried out within two
projects supported by the European
Union Esprit program (EP 24986,
Dedemas: Decentralized Decision
Making and Scheduling [2] and EP
28448, GNOSIS-VF: Towards the Virtual
Factory: Delivering Configuration,
Scheduling and Monitoring Services
Through a Web-Based Client-Server
Architecture [3]).
Integration in supplychainsThe starting point of the integration
of IT systems in company networks
is the business processes of the
companies and their development
needs.
In our first test case, ABB Control,
the development target is the
customer order delivery process. The
production is usually carried-out with
company’s own resources. Only
during overload situations does it use
external resources belonging either to
other ABB companies or to local
subcontractors. This combined
utilization of their own and external
resources is a business process new
to ABB Control. The basic idea of the
IT system integration in this test case
is to assure the transparency of
internal and external resources. We
mean that with this transparency
production managers at ABB Control
are able to plan and manage external
resources as easily as their internal
ones. Fig. 1 i l lustrates the
communication relations in this test
case.
Our second test case involves a
multi-site steel tube manufacturer,
Rautaruukki Metform, with many
alternative production sites that are
accessed from several sales offices
simultaneously. The production of a
single customer order of any sales
office might be divided to several
production sites. The aim of the IT
systems integration in this test case Figure 1. Subcontracting relations of ABB Control.
is to support sales offices in planning
suitable production sites for their
orders. The production sites need to
increase their co-operation with sales
offices by providing more information
about their production situation and
inventory. IT tools are needed to run
this planning process. The underlying
business goal of this test case is
enhanced customer service and
better utilization of the company’s
resources.
Communication withXMLThe need for more IT support for
communication is evident in both
test cases. In the context of the ABB
Control test case we developed a
prototype of an XML/XSLT-based
integration server [5]. The purpose
of this server was to evaluate the
extent to which the new technology
enables the development of more
and cheaper communication tools to
support data transfer between ABB
Control and it’s suppliers. We also
experimented with some novel
software design ideas.
The prototype integration server
conforms to layered software
architecture that could be described
as engine/processors architecture
(see Fig. 2). We expect that this type
of architecture is more configurable
and maintainable than the traditional
type, thus resulting in reduced IT
costs.
The top layer of our architecture is
formed by an engine that processes
the interaction requests, and executes
them according to the configured
interaction definitions. These define
the interaction-handling logic in terms
of parameters and operations. The
bottom layer of the architecture
contains a set of processors that are
able to perform the operations. The
integration server can load these
processors “on demand” from the local
file system or the Internet. This makes
it possible to add or update the
functionality of the integration server
without code changes. The server
configuration information has to be
located on the same machine as the
server itself, whereas all other
configuration data can be retrieved
from the Internet.
A large par t of the actual
functionality of the prototype
integration server is defined with XML-
based configuration languages instead
of lower level programming languages
like Java. Also, the motivation of the
XML-based configuration languages is
easier maintainability.
The configuration data of the
prototype server is divided into three
levels. The system level configuration
ATR_2001.p65 5.12.2001, 11:3449
AU
TOM
ATIO
N T
EC
HN
OLO
GY
RE
VIE
W 2
001
50
Figure 2. Software architecture of the XML/XSLT-based integration server.
data specifies which interactions are
defined for the server. The interaction
level configuration data defines which
parameters and operations are
required for execution of an
interaction. At the operational level,
the configuration language is specific
to the processor that executes the
operation. The hierarchical structure
of the configuration data provides an
advantage by providing independence
between the different levels.
The example in Fig. 3 is from an
“order-query-cbl” interaction definition.
This interaction checks the password
of an user, retrieves a given order from
a database, transforms the order into a
document in xCBL format, and sends
this document to a given target.
The Translator is an essential
processor from the enterprise
integration point of view. This
processor is configured with XSLT, an
XML-based language for transforming
an XML document into another XML
document (and also into other
formats). The pattern-matching
mechanisms of XLST provide an
effective tool for the definition of
such translations. Companies usually
<interaction-definition>
<parameter name=”user” type=”String”/>
<parameter name=”password” type=”String”/>
<parameter name=”order-id” type=”String”/>
<parameter name=”receiver-id” type=”String”/>
<operation name=”access” processor=”file:AccessProcessor.class”
configuration=”file:permissions.xml”>
<input name=”user”/>
<input name=”password”/>
</operation>
<operation name=”query” processor=”file:QueryProcessor.class”
configuration=”file:order-query-cbl-database.xml”>
<input name=”order-id”/>
</operation>
<operation name=”translation” processor=”file:Translator.class”
configuration=”file:order-query-cbl-translation.xsl”>
<input name=”query”/>
</operation>
<operation name=”send” processor=”file:Messenger.class”
configuration=”file:partners.xml”>
<input name=”receiver-id”/>
<input name=”translation”/>
</operation>
<result name=”send”/>
</interaction-definition>
Figure 3. An example of an XML-based integration logic definition.
ATR_2001.p65 5.12.2001, 11:3450
51
AU
TOM
ATIO
N T
EC
HN
OLO
GY
RE
VIE
W 2
001
have internal data models that are
specific for their own IT systems. It
is possible to transform business
documents by means of XSLT from
or into external formats that are
shared by the companies in a VE.
Although XSLT is relatively simple,
a lot of manual work may be needed
to generate configuration information
for the XSLT translation. A rough
estimate is that about 80 % of the XSLT
code is easy to write while the
remaining 20 % could need
considerably more effort at least in the
learning phase. The translation task is
often considerably alleviated by
splitting the task into a set of
consecutive translations. In the
prototype integration server this is
easily accomplished by adding
operat ions to an in teract ion
specification. However, understanding
the semantics of the company specific
or other data structures (such as xCBL
[1]) is often rather difficult.
Coordination with theMediatorIn the Rautaruukki Metform test case,
IT support for communication is not
enough. There is also a clear need for
better co-ordination between decision-
making in sales offices and production
sites. During this test case a prototype
system based on the concept of the
Mediator [see e.g. 4] was developed to
fulfil this co-ordination requirement.
The Mediator in our test case is a
shared server whose role is to provide
mechanisms to support collaborative
decision-making between sales offices
and production sites [6]. It is
essentially a co-ordination broker with
some decision-support functionality.
In order to fulfil its role, the Mediator
provides a selection of decision-
support mechanisms. The mechanisms
are targeted for the various steps in the
order-planning process. In this test
case, the Mediator provides negotiation
and rule-based decision support, and
monitoring mechanisms. These
decision-support mechanisms are
based on an underlying XML-based
communication solution.
In order to understand how the
Mediator works one needs to know
that the Mediator is implemented
with a decentralized software
architecture as illustrated in Fig. 4.
This architecture is motivated by
scalability and modifiability needs.
The architecture is also reflected in
the decision-support mechanisms.
Although in this study there is only
one Mediator, it might be possible to
have many of them co-operating with
each other in a more large-scale
system. However, this would entail a
more complex functionality for the
Mediator.
In order to support the order-
planning process, the Mediator
provides a Contract-Net-based
negotiation mechanism [7]. This
mechanism is based on the
metaphor of an auction. The sales
offices announce their tasks, and the
production sites reply with bids.
Finally, the sales offices make their
choices between bids.
The negotiation mechanism of the
Contract Net had to be extended for
usage in this kind of real applications.
While in the basic Contract Net the
bidders have to accept the tasks as
such, in a real situation they can make
counter proposals. In general, the
bidders can adjust the time, cost, and
content of the announced tasks.
Figure 4. Decentralized software architecture of the Mediator.
However, one should note that this
approach leads to more complicated
iterat ive negot iat ions i f used
deliberatively.
The Mediator’s role as a negotiation
hub offers the possibility of making
the negotiation a service for the sales
offices and production sites. The
Mediator can hold the data structures
and run the processes of negotiations.
The Mediator can also select suitable
production sites for the negotiations.
In order to do this, the Mediator has
access to data that characterizes the
production sites and to rules that
describe the logic to select them. As a
consequence, the Mediator can make
the negotiations transparent to the
sales offices.
The negotiation service of the
Mediator becomes particularly useful
when combined with other decision-
support mechanisms. Rule-based
decision-making mechanisms can be
used to make some of the decisions
during the negotiations. Furthermore,
the Mediator may also act as a
monitoring broker.
Evaluation of the ITimpactIn both test cases, we performed an
evaluation of the possible costs and
benefits of the integration tools if
developed into operational systems.
SCM is about integration
ATR_2001.p65 5.12.2001, 11:3451
AU
TOM
ATIO
N T
EC
HN
OLO
GY
RE
VIE
W 2
001
52
In the ABB Control test case, a natural
reference to the developed prototype
was the exist ing EDI-based
communication system. In the
Rautaruukki Metform case, there
were not any clear comparison
targets because the prototype
contained new functionality.
The evaluation of the XML/XSLT-
based integration server in the ABB
Control test case server was carried out
by a set of test users with help from
researchers. The prototype contained
the configuration of four purchasing
related business documents and their
interactions. Three of the configured
business documents were also
supported by EDI at the time of this
study (purchase order, purchase order
response and invoice) while the fourth
one was not (demand forecast). The
implementation of send, receive and
query operations for these business
documents was quite feasible with the
presented integration server
architecture and its configuration
mechanisms. Thus, the test case would
suggest the feasibility of the chosen
approach as an interaction modelling
and implementation method.
The implementation costs of EDI
in the test case appeared to be clearly
higher than those of an XML/XSLT-
based integration server. With regard
to the amount of necessary work only,
the introduction of a new message
type (corresponds to the definition of
a new interaction) requires a new EDI
module and about 200 hours’ work
of related specification and testing.
Opening a new connection (a
particular message type is taken in use
between two partners) requires a
couple of days of testing. Provided that
the testing times are about the same
in both cases, it can be assumed that
establishing an EDI connection for a
new message type is perhaps three to
four times more expensive than
defining and implementing a new
interaction in the XML/XSLT
integration server. And when use
charges are taken into account, the
XML-based approach is certainly less
expensive.
The prototype of the Mediator in
the Rautaruukki Metform test case
was assessed by a number of sales
people and some managers of the
company. Many sales people saw
good things in the Mediator.
However, some of the sales people
did not feel that it would benefit
them that much in their own
operations in all cases. Most of the
managers thought that the Mediator
was a very useful and helpful tool for
many operations and could be even
more helpful if developed further.
However, the company had many
other IT development projects at the
time, and the question of the further
development of the Mediator was left
open.
ConclusionsIn this article we have presented two
case studies involving new XML/XSLT-
based communication technology and
novel co-ordination architectures, i.e.
the Mediator. Both of these test cases
emphasize the potential business
benefits of Internet-based information
systems. While our evaluation of the
test cases suggests that basic XML/
XSLT-based communication systems
seem to enable significant cost savings,
the situation is more uncertain with
more advanced co-ordination oriented
systems like the Mediator. However, we
hope that these experiments might
g i ve m o t i va t i o n fo r f u r t h e r
development of this kind of system.
They seem to have potential.
ATR_2001.p65 5.12.2001, 11:3452
53
AU
TOM
ATIO
N T
EC
HN
OLO
GY
RE
VIE
W 2
001
References1. Commerce One, Inc. xCBL Specification. WWW page. http://www.xcbl.com.
2. Dedemas Consortium. WWW page. http://dedemas.ifw.uni-hannover.de.
3. GNOSIS-VF Consortium. WWW page. http://www.vtt.fi/aut/tau/gnosis
4. Maturana, F. & Norrie D.H. 1996. Multi-Agent Mediator Architecture for Distributed Manufacturing, Journal ofIntelligent Manufacturing, Vol. 7, pp. 257 - 270.
5. Seilonen, I., Nurmilaakso, J., Jakobsson, S., Kettunen, J. & Kuhakoski, K. 2001. Experiences from the Development ofan XML/XSLT-based Integration Server for a Virtual Enterprise Type Co-Operation. To be presented at the 7thInternational Conference on Concurrent Enterprising (ICE 2001), Bremen, Germany.
6. Seilonen, I., Teunis, G. & Leitao, P. 2000. Mediator-based communication, negotiation and scheduling fordecentralised production management. MCPL 2000 Conference. Grenoble, France.
7. Smith, R.G. 1980. The Contract Net protocol: high-level communication and control in a distributed problem solver.IEEE Transactions on Computers, Vol. c-29, No. 12, pp. 1104 - 1113.
Ilkka SeilonenResearch ScientistVTT Automation
Jari KettunenResearch ScientistVTT Automation
StefanJakobssonResearch ScientistVTT Automation
Petri KalliokoskiResearch ScientistVTT Automation
Markku MikkolaResearch ScientistVTT Automation
Veli-Pekka MattilaDevelopment ManagerVTT Automation
SCM is about integration
ATR_2001.p65 5.12.2001, 11:3453
AU
TOM
ATIO
N T
EC
HN
OLO
GY
RE
VIE
W 2
001
54
Implementing ERP Systems in SME EnterprisesMagnus Simons and Raimo Hyötyläinen
Small and medium sized enterprises have difficulties inachieving the full potential of integrated informationsystems. A central task is to develop the informationculture of the enterprise. Also, the vendor of the systemneeds to rethink implementation and service strategies.VTT Automation is providing a service, tools andmethods to support both users and vendors in theimplementation of these systems and in the process ofcontinuous learning during their use.
IntroductionEnterprise Resource Planning (ERP)
systems that integrate functions like
sales, production, purchasing,
materials management with each
other, and with f inancial
management, are used in small
companies as well as larger ones.
Integration of systems between
companies i s a lso becoming
increasingly common. This has
brought new requirements for
people and act iv i t ies in the
organisation. The use of an integrated
system al lows for systemat ic
planning and rationalisation of
business processes often covering
work performed in a multitude of
organisational functions (Davenport,
1993). This integration of information
systems requires a change of culture
in the user organisat ions. By
connectedness we mean that each
person using the system has to
under stand that he or she is
participating in an integrated
information processing activity
involving a wide range of people over
a certain period of time. The “how,
when and what” of feeding
information into the system is
ATR_2001.p65 5.12.2001, 11:3454
55
AU
TOM
ATIO
N T
EC
HN
OLO
GY
RE
VIE
W 2
001
relevant for many people in the
organization, and to feed information
properly one has to understand the
expectations of the “information
customers”. As a user of information
you also have to understand the
context in which the information
was created and fed into the system.
To understand why the system
expects the user to act in a certain
way, he has to know the business
requirements of the enterprise in
order to understand his own role in
the act iv i ty system, and to
understand what the technical
constraints set by the designers of
the system are.
The use of large integrated
information systems is wide spread
among large companies around the
world but small and medium sized
companies (SME) are a lso
increasingly using complex systems.
Figure 1. The integrated, cooperative learning process.
This creates a need for a re-evaluation
of ERP systems and their use and
implementation. Small companies are
moving and changing fast, they rely
more on personal contacts between
managers and workers, and less on
the structures, hierarchies and formal
methods that are the basis of ERP
systems to day. Also, small companies
have fewer resources than large ones,
and fewer possibilities are open to
them to acquire knowledge of the
new technology. Implementing a
system in this environment requires
proper methods and tools. This is a
challenge not only for a small
company implementing an ERP
system, but also for the vendors of
such systems. If the implementation
of the system fails, the consequences
can be fatal for both parties. Research
has shown that the involvement of
users is crucial to the results of
implementation (Lyytinen 1986,
Checkland & Holwel l 1998,
Hyötyläinen 1998). Our practical
experience from working with many
Finnish SME-companies has shown
that this is an area left open by
vendors of ERP-systems. Product
quality and service has not yet
achieved a central status in the
business of software production.
In the HanskaHanskaHanskaHanskaHanska project, VTT
Automation, together with the
Universities of Turku and Jyväskylä,
are developing means for managing
di f ferent phases of the
implementation process of ERP
systems. The goal is to create an
integrated, cooperative learning
process linking strategic thinking,
system requirements planning and
choice of system, system
configuration and user training, with
continuous development during the
ATR_2001.p65 5.12.2001, 11:3455
AU
TOM
ATIO
N T
EC
HN
OLO
GY
RE
VIE
W 2
001
56
Figure 2. A business process modelling technique for system requirement specification.
use of the system. The goal of the
project is to create tools and methods
that support a cooperative learning
process involving all parties, both on
the user side and on the vendor side.
The IntegratedLearning ProcessIn Fig. 1, the user’s implementation
process and the vendor’s delivery
process are combined into one
integrated, cooperative learning
process. The whole process can last
many years and involve tens of people
even in rather small enterprises.
Cooperation between the user
organisation and the vendor can and
should take place in all phases of the
learning process. There is always a
potential for learning more about
user’s needs or about existing systems.
To support this kind of learning, the
two organisations need a common
concept and language, including
common tools and methods for
defining needs and for describing the
potential of an ERP-system.
The learning process should also
integrate the different phases of the
process with each other. As Fig. 1
shows, the key actors in each phase
vary from phase to phase. In the
beginning, the management in the
user organisation is at the centre of
activities, later, middle management
and key users become involved, and
in the two last phases, other users
also participate. On the vendor’s side,
the organisation of the process has
similar features. By ‘integration of the
learning process’ we mean that the
tools and methods used in the
learning process have to also support
the diffusion of strategies, goals,
target process models, etc. from one
group of people to the next. The
integrative process is one of using
simple tools to combine dialogue and
documentation into a continuum
between the various phases of the
learning process.
A challenge for the learning
process is to create true feedback for
all parties involved during the
di f ferent phases of the
implementation process. The process
can last several years between initial
strategic design and feed-back from
actual use of the system. During this
time many things change: strategy,
goals, system concept, for example.
Double loop learning (Argyris &
Schön, 1978) in this environment
stresses the need for continuous
involvement of al l par t ies,
documentation of goals and concepts
in each phase and systematic feed
back techniques.
Business Processbased Tools andMethods for ERP-implementation in SMEenterpriseIn the Hanska project, we strive to
develop simple means and methods
for communication within and
between the actors of the
implementation process. The goal is
to find solutions suitable for SME-
enterprises. Business-process thinking
is central to ERP-implementation
(Scheer, 1989). It has also been a
ATR_2001.p65 5.12.2001, 11:3456
57
AU
TOM
ATIO
N T
EC
HN
OLO
GY
RE
VIE
W 2
001
Table 1. Tool for use-analysis information collection.
Table 2. Tool for information analysis in use-analysis.
Implementing ERP Systems in SME Enterprises
central method in the Hanska project.
Based on this we strive to create tools
to support the different phases of the
implementation process.
Fig. 2 describes a business process
modell ing technique used to
describe the activity of a company.
During the modelling process, target
models for defined business
processes were described. This was
used as a basis for defining systems
requirements and for choosing the
system and system vendor. The
technique connected strategic goals
with business processes, defined
processes at three different levels and
showed the connection between
tasks in processes and the supporting
IT-tools. The tasks, where the use of
the IT-systems was especially intense,
were analysed in detail to find out
different scenarios in which the
system could be used in the future.
A method developed in the Hanska
project is the business process based
method for analysing the level of use
of an ERP-system in an organisation.
The use-analysis method is based on
the same kind of business process
modelling technique as illustrated in
Fig. 2. The difference is that while the
previous phases process modelling is
used to create an image of a process
to-be, the use-analysis method starts
from a process ‘as-is’. In Tables 1 and
2 some additional tools for gathering
information about the use of the
system in different phases and tasks
of the business processes are listed.
The use of process-based methods
and tools for implementing ERP
systems has so far taught us that
process modelling is a useful but not
sufficient means for supporting
communicat ion dur ing the
implementation process. Process
models are useful tools to create a
picture of the ERP-user organisation’s
activities.
Building process models of their
own activities is, in general, seen as a
fruitful learning process that gives the
actors in the enterprise an improved
picture of activities as a whole in their
organisation. For creating a picture of
the requirements the user imposes on
the system, business models are not
sufficient. Firstly, the process models
do not give a picture of the needs at
the level of the data and information
needed in the system. Secondly, the
process model does not describe
what is going to be produced and for
whom. Thirdly, not all activities can be
described as a business process
(Nurminen et al., 2001).
Modelling of data and information
needs are centra l features of
information system modell ing
(Scheer, 1989). In the process of
implementing a standard ERP-system
in a small or medium sized enterprise,
though, it is not reasonable to think
that the modelling of thousands of
data and pieces of information should
be performed “from scratch” by the
user organisation. A solution offered
during implementation in one
Hanska case is to look for special
features in the business processes of
the company, and to model these
features in more detail. Here the
problem is to define what is special
in relation to what. For a small
company with restricted possibilities
as far as benchmarking activities with
other companies is concerned, and
with few resources to study different
vendors’ ERP-software, defining what
is special can be overwhelming.
Key Process:Sub-process:Workflow:Process ChartNumber:
Task Current state, problems, Target state, specialspecial requirements requirements
Materials Management Process
Phase Task Responsibility System, Problems in Problems andmodule implementation comments regarding
and use of mode of operation orthe system way of action
ATR_2001.p65 5.12.2001, 11:3457
AU
TOM
ATIO
N T
EC
HN
OLO
GY
RE
VIE
W 2
001
58
The process model does not
describe the products. Nevertheless,
one of the most difficult tasks during
implementation is often to define the
product structure. Since an SME-
company often mixes subcontracting,
its own products and services in a
varying set, deciding and defining
what is a product and how it be
should structured is often tedious. To
help SME-companies do this a tool is
needed. The basic structure used is the
bill-of-material.
A study of activities in a hospital
has showed that service related
activities could not always be
described as a business process. For
instance, connecting the goals of
rationalising business processes to
the life cycle of a patient at the
hospital can give bizarre results
(Nurminen et al., 2001). These
findings can probably also be
connected with other activities
where the goal is not to end a process
as soon as possible, but rather, for
instance, to prolong the life cycle of
a product.
ConclusionsIn the Hanska project researchers
from VTT Automation, the University
of Turku and the University of
Jyväskylä are studying the process of
implementing ERP-systems in SME-
enterpr i ses . Exper imenta l
development research has been
carried out in twelve case companies
and organisations during the project.
The work in each case supports
method and tool development for
one or more phases of the user’s
implementation process or vendor’s
delivery process. The use of business
process modelling is a central tool,
but is not alone sufficient to describe
user needs or system potential. The
work with SME-enterprises shows
that simple tools are necessary to
involve groups of people not used to
either information technology or to
the tools used for designing these
tools. On the other hand, these small
organisations need to be supported
in finding out what is special in their
own activities compared to the line
of business they are in. Here some
kind of pre-made reference models
could be used. The work in the
Hanska project will continue until
the end of year 2002. Until then
further research will be conducted
to gain more knowledge about the
methods and tools useful to the
integrated, cooperative learning
process of implementing ERP-
systems.
The use of process modelling in
the different phases of the
implementation process in the
Hanska cases may provide a basis for
an integrated and cooperative
learning process. Business process
models can be used to describe both
the user’s activity and the mode of
activity supported by the vendor’s
ERP-system. Both user and vendor
can use the business process models
in all phases. Further research is
needed to evaluate how these
process-models work and to discover
any other needs relating to the
learning process.
ATR_2001.p65 5.12.2001, 11:3458
59
AU
TOM
ATIO
N T
EC
HN
OLO
GY
RE
VIE
W 2
001
ReferencesArgyris, C. & Schön, D. A. 1978. Organizational Learning: A Theory of Action Perspective. Reading, Massachusetts -Menlo Park, California - London - Amsterdam - Don Mills, Ontario - Sydney: Addison-Wesley.
Checkland, P. & Holwell, S. 1998 Information, Systems and Information Systems. Making Sense of the Field.Chichester: John Wiley and Sons.
Davenport, T. H. 1993. Process Innovation. Reengineering Work through Information Technology. Boston: HarvardUniversity Press.
Hyötyläinen, R. 1998. Implementation of Technical Change as Organizational Problem-Solving Process. Espoo: VTTPublications 337.
Lyytinen, K. 1986. Information Systems Development as Social Action: Framework and Critical Implications. JyväskyläStudies In Computer Science, Economics and Statistics, No 8.
Nurminen, M., Järvinen, O. 2001. The strength and borders of process thinking. In: Kettunen, J., Simons, M.: ERPimplementation in small and medium-sized enterprises –From technology push to the management of knowledgeand expertise. Espoo: VTT Publications. (in Finnish)
Scheer, A.-W. 1989. Business Process Engineering. Reference Models for Industrial Enterprises. Second, CompletelyRevised and Enlarged Edition. Berlin: Springer-Verlag.
Implementing ERP Systems in SME Enterprises
Magnus SimonsSenior Research ScientistVTT Automation
RaimoHyötyläinenGroup ManagerVTT Automation
ATR_2001.p65 5.12.2001, 11:3459
AU
TOM
ATIO
N T
EC
HN
OLO
GY
RE
VIE
W 2
001
60
Managing Electrostatic Discharges byProtective ClothingSalme Nurmi, Terttu Peltoniemi, Markku Soini, Mika Tukiainen, Tuija Luoma, Inga Mattilaand Raija Ilmén
Managing personal static electricity is an important issue in the community at large,as well as in the electronics industry. Electrostatic discharge (ESD) protectivegarments are generally used in areas where the relative humidity may fall below 20 %RH. Garments should maintain their ESD protective performance, be ESD reliable inuse, and still be reliable after 20 – 50 washes. The surface resistance and chargedecay time of different kinds of ESD protective garments have been studied as afunction of washings. Measurements have been made in humidities of 5 - 10 % RH.The objective of both surface resistance and charge decay measurements was toobtain reliable information on the performance of the clothing in ESD control.According to the results, careful attention should be paid to the technical developmentand manufacturing, as well as to the washability properties of the ESD garments.Several washings may change the electrostatic properties of some ESD garments,leading to a highly restricted capability of dissipating charge.
ATR_2001.p65 5.12.2001, 11:3460
61
AU
TOM
ATIO
N T
EC
HN
OLO
GY
RE
VIE
W 2
001
IntroductionIn the recent years there has been a
growing interest in the
comprehensive management of
electrostatic discharge (ESD). Static
electricity occurs commonly both in
industry and daily life. Many of the
effects are harmless, and either pass
completely unnoticed or are simply
a nuisance, but static electricity can
also give rise to hazardous situations.
It can be generated by the contact
and separat ion of sol ids , the
movement of a person, the flow of
liquids or powders or by induction
phenomena. The accumulation of
electrostatic charge can give rise to
hazards and problems in a wide range
of industries; in particular it can give
rise to ignition and explosion hazards
in the chemical, pharmaceutical,
petroleum, food processing and
electronics industries especially. ESD
is an event that usually affects the
device in question during its non-
operating condition, and causes
electrical and physical subsurface
damage in sensitive electr ical
devices. Garments on which high
levels of static electricity can be
generated are one of the causes of
ESD damage. The ordinary materials
used in protective textile clothing are
insulators exhibiting rather low
electrical conductivity and thus the
charge dissipates with difficulty.
Electrically conductive fibres have
been blended with normal textile
fibres, and yarns have been used to
enhance the electrical properties of
traditional fabrics. In practical use,
several washings may change the
electrical properties of materials and
garments. Managing static electricity
at low relative humidity is a problem
as well. In the winter-time in
northern countries the relative
humidity can be less than even 5 %
for several weeks.
In this paper we have studied the
surface resistance and charge decay
times of modern ESD protective
clothing used in Nordic countries at
low relative humidity. The same
garments were new at the beginning
of the test series and by the end of
the series had been washed several
times (0,20,50). The electrical
measurements were taken at 5 - 10 %
RH using IEC methods.
ExperimentalSeveral measurements were
performed to examine the surface
resistance, point-to-point resistance
and discharge time of ESD textile
materials and garments. All tests were
done in an ambient relative humidity
of 5 - 10 % and a temperature of 20-
21 ºC. Before testing, the samples
were conditioned for three days in
these conditions. All garments were
washed in water in a laundry at 60 ºC
for 6 minutes us ing washing
chemicals, and a non-chlor ine
bleaching agent (peracetic acid); a
scavenging agent was also used in the
final flushing. Drum drying was the
used at 150 ºC for 15 minutes (the
temperature of the garment was
about 90 ºC); finally, the garments
were cooled to a temperature of 40 ºC.
The ESD protective garments chosen
for the study were examples of
materials that had been used in
electrical industries after several
washings.
PA-carbon fibres were the surface
conductive type. Textile materials A,
B and C were staple fibre fabrics and
D was a filament fabric. PES =
Polyester, PA = Polyamide and CO =
Cotton.
A 65 % PES / 34 % CO / 1 % PA-carbon, plain weave, grid 3 mm x 4 mm
B 65 % PES / 34 % CO / 1 % PA-carbon, soft finish, plain weave, grid 3 mm x 4 mm
C 70 % PES / 22 % CO / 8 % PA-carbon, plain weave, grid 3 mm x 4 mm
D 95 % PES / 5 % PA-carbon, warp knit, diagonal grid 3 mm x 3 mm
Table 1. ESD-garments used in the study.
Figure 1. Experimental arrangement of the surface resistant measurement.
ATR_2001.p65 5.12.2001, 11:3461
AU
TOM
ATIO
N T
EC
HN
OLO
GY
RE
VIE
W 2
001
62
Figure 2. Experimental arrangement of the point-to-point measurement.
Surface resistanceSurface resistance was determined
pursuant to Chapter 4 in Appendix A
of Technical Report IEC/EN 61340-5-
1 [1]. The garment was opened and
placed on a piece of 10 mm acrylic
board on a table. A resistance
measur ing instrument and a
cylindrical electrode was used for the
measurement. The total weight of the
cylindrical electrode was 2.5 kg ±
0,25 kg. The ring electrode was
supplied by a voltage of 100 V
collected by the centre electrode. The
measuring voltage was 100 V. The
current between the electrodes in
the sample surface was measured by
a sensitive ammeter. Five parallel
determinations were made on the
right side of the garment; the result
was an average of these. See Fig. 1.
Point-to-pointresistancePoint-to-point resistance was
determined pursuant to IEC/EN
61340-5-1 [1] Technical Report,
Annex A, Chapter 3. Measuring was
conducted using a resistance-
measur ing instrument. Two
cylindrical electrodes of weights 2,5
kg ± 0,5 kg, respectively were used.
The diameter of the electrodes was
75 mm. The point-to-point
measurements on an insulating sub-
plate were taken from sleeve to
sleeve and from sleeve to hem. The
other weight sensor was supplied by
a voltage of 100 V collected by
another weight sensor. See Fig. 2.
Charge decay timeCharge decay in a garment was
determined as the duration of the
voltage decrease from 1000 V to 100
V [1]. The point-to-point discharge
time was measured for a hanging
garment from sleeve to hem and for
a garment on an insulating sub-plate
from sleeve to sleeve. The point-to-
point discharge time for a level
surface was measured with 2,5 kg ±
0,5 kg on the sensors [2]. The sample
was placed on an insulating sub-plate.
Measurements were taken from
sleeve to sleeve. The weight sensors
were placed on both sleeves. One
sleeve was charged to a
predetermined voltage (about 1500
V) using a charged plate monitor
Simco EA-3 charge analyser. Then the
other sleeve was grounded and the
voltage of the charged garment was
monitored as a function of time by
the field-meter of the device.
Measuring was interrupted if the
discharging had not taken place
within 60 seconds, and the charge
value at that moment was noted. A
diagram was drawn of the data
measured at the discharge time. The
decay times presented in this work
are averages of both positive and
negative charging/discharging
procedures with one garment.
Discharge time for the hanging
material was measured using a
garment hanging on a wooden
clothes hanger. The width of the
clamps were 5,0 cm. Using a Simco
EA-3 charge analyser, a voltage
exceeding 1000 V was connected to
one of the clamps (the analyser gives
an alternating 1000 V – 1800 V
charge). The garment was earthed
through the other clamp, and the
discharge time from 1000 V to 100 V
was measured. Measuring was
interrupted if discharging had not
taken place within 60 seconds. A
diagram was drawn of the measuring
data over the discharge time. The
decay times presented in this work
are averages of both positive and
negative charging/discharging
procedures with one garment.
Measurements were taken from
sleeve to hem. See Fig. 3.
ResultsThe measured surface resistances of
the materials are presented in Fig. 4.
An IEC standard [1] recommends
that the surface resistance should be
less than 1 x 1012 Ù, in order to
minimize ESD risks. According to the
results, garments A, C and D of the
materials studied fulfil the criterion
at 5 -10 % RH after 50 washings. The
textile properties of garment B were
changed during the washings, so that
after 50 washings the fabric was
stiffer than before the washings.
Garment B was compensated with
garment A.
The measured surface point-to-
ATR_2001.p65 5.12.2001, 11:3462
63
AU
TOM
ATIO
N T
EC
HN
OLO
GY
RE
VIE
W 2
001
point resistances from sleeve to
sleeve are presented in Fig. 5. All
garments, A, C and D, fulfil the IEC
criterion after 50 washings.
The measured charge decay times
from 1000 V to 100 V of the garments
after the grounding are presented in
Figs. 6 and 7. A good ESD
management level would require that
the charge decay occured within two
seconds [1]. From Figs. 6 And 7, it can
be seen that at 5 – 10 % RH, and after
50 washings, this happens only in the
case of garment D. The discharge
times of garments A and C after 20
washings had increased from a
couple of seconds to nearly 100
seconds measured from the hanging
garments. The discharge time of
garment A, measured on an insulating
support, was also nearly 100 seconds
after 50 washings.
The results show that garments A,
C and D fulfil the IEC [1] criterion at
low relative humidity both as new
and after 20 washings. Between 20
and 50 washings, the charge dissipate
properties changed so much that,
after 50 washings, the charge decay
time for hanging garments A and C
had increased to a level of 100
seconds.
Shrinkage of the materials studied
during several washings can change
the structure, the surface, and other
textile properties of the fabrics.
Garment D kept the properties on
the same level best, despite several
washings. Its washing shrinkage was
only about 1 %. Also, the surface of
material D was bald, with no staple
fibre ends seen penetrating from the
surface. After 50 washings, the
washing shrinkage of garments A, B
and C was about 7 – 11 %, with a large
number of staple f ibre ends
penetrating from the surfaces.
(Fabrics A and B had the most hairy
surfaces.) Hair y surfaces may
decrease the surface contact
between the mater ial and the
electrode used in the measurement.
Figure 4. Surface resistance of the materials as new and after 50 washes.
Figure 5. Point-to-point resistance of the garments as new and after 20 and 50 washes.
Figure 3. Experimental arrangement of the charge decay time measurement forthe hanging garment.
Managing Electrostatic Discharges by Protective Clothing
ATR_2001.p65 5.12.2001, 11:3463
AU
TOM
ATIO
N T
EC
HN
OLO
GY
RE
VIE
W 2
001
64
Figure 7. Charge decay times of the garments hanging as new and after 20 and 50washes.
ConclusionsThe results of the surface resistance
and charge decay measurements for
different kinds of textile materials
after several washes show that
special attention should be paid to
ESD management and personal ESD
protection at low relative humidity.
Several washes may change the
electrical properties of the garments
and expose them to ESD risks.
Figure 6. Charge decay times of the garments on an insulating support as new andafter 20 and 50 washes.
AcknowledgementsThe authors would like to thank to
Mrs. T. Peltoniemi, Mr. M. Soini and
Mr. M. Tukiainen for their helpful
discussions, and Dr. K. Tappura, Dr.
J. Paasi and Dr. T. Varpula for their
careful reading of the manuscript.
This work was part of the Finnish
STAHA (Managing Static Electricity)
technology programme supported
by the National Technology Agency,
TEKES.
ATR_2001.p65 5.12.2001, 11:3464
65
AU
TOM
ATIO
N T
EC
HN
OLO
GY
RE
VIE
W 2
001
References[1] EN/IEC 61340-5-1 Electrostatics Part 5-1: Protection of electronic devices from electrostatic phenomena –General requirements (IEC 61340-5-1:1998 + corrigendum 1999). Brussels: European Committee for ElectrotechnicalStandardization, 2001. 157 pages. (IEC = International Electrotechnical Commission, EN = European Standard)
[2] EN 100 015-1 Basic Specification: Protection of electrostatic sensitive devices. Part I: General requirements.Frankfurt am Main: CENELEC Electronic Components Committee, 1992. 57 pages.
www.vtt.fi/rm/projects/staha
Managing Electrostatic Discharges by Protective Clothing
Salme NurmiSenior Research ScientistVTT Automation
Tuija LuomaResearch EngineerVTT Automation
Inga MattilaTechnicianVTT Automation
Raija IlmenTechnical AssistantVTT Automation
Markku SoiniManaging DirectorMix-Vaate Oy
Mika TukiainenMarketing ManagerLaitosto Oy
Terttu PeltoniemiProject ManagerNokia Oyj/Nokia Networks
ATR_2001.p65 5.12.2001, 11:3565
AU
TOM
ATIO
N T
EC
HN
OLO
GY
RE
VIE
W 2
001
66
Integrated Safety Management in TeamworkOrganisationJarmo Karlund
Group and teamwork in Finland expanded during the 1990´s. Changes in the role ofmanagement in the work organisation have had a marked impact on safety measures.Teamwork safety issues were studied in nine Finnish companies, and the effect ofteam organisation on safety culture and occupational accidents was evaluated. As aresult, rules and responsibilities were defined, and measures to integrate safetywith teamwork, as well as criteria for successful team activities, were specified.
IntroductionThe current trend in organising the
managerial and supervisory functions
in industrial plants involves changing
the locus of control by establishing
self-regulating production teams.
These teams are comprised of
production workers. The degree of
autonomy of their daily production
work is increased within the teams.
Supervisors are taken away from
their daily routines of controlling and
monitoring work and workers. The
role of supervisors is changing
towards coaching and supporting the
teams. The new role of supervisors
gives them more responsibility in the
general development of work and
working conditions.
The idea behind teamwork has been
to improve workers’ commitment, to
give workers more freedom in their
daily work, to encourage them to
develop and use their knowledge and
skills, and to improve the social-
psychological climate at work sites.
There are, however, examples of serious
and even fatal accidents in teamwork
organisations.
Figure 1. Outlook from production lineof the mineral wool.
ATR_2001.p65 5.12.2001, 11:3566
67
AU
TOM
ATIO
N T
EC
HN
OLO
GY
RE
VIE
W 2
001
material plants will be presented here.
The total work force of the plant was
84. The number of workers was 61.
The safety culture was evaluated by
the Safra questionnaire (Seppälä, 1992).
This is composed of the following
scales: background data, evaluation of
safety hazards, rating of safety values
and attitudes, appraisal of daily safety
activity and appraisal of company-level
safety measures and functions. Items
that concern teamwork, co-operation
and stress were added to the basic
questionnaire.
The inquiry was conducted as a
pilot project at the end of 1998, and
again at the end of 1999. The number
of completed questionnaires was 53
the first time, and 27 the second time.
New Organisation ofTeams and Supervision
Management and the role ofsupervision The plant moved into a new mode of
team organisation in May 1998. The first-
Figure 2. Team organisation creates new challenges and variable functions for the safety personnel. Implementation of theinformation is one important sector.
The main aim of the study was to
create for large and small size
enterprises models that ensure safety
in team organisations.
Method and MaterialsA research project was targeted to
clarify safety issues in teamwork
organisations. The project started in
September 1988 and ended in
September 2000. The aims of the
study were
1) to clarify the ways and means to
organise teamwork, and to define
the role of supervisors in general,
2) to define the safety responsibilities,
3) to develop safety rules,
4) to help teams make risk analyses,
and to integrate safety training
into job instruction, and
5) to clarify safety culture and
accidents in team organisations.
The study was carried out in nine
industrial enterprises; four plants
producing building materials, one
packaging-material factory, three metal
factories and one rubber production
plant. The results of one of the building-
line supervisors, who had earlier worked
in shifts, were transferred to a day shift
only. These supervisors formed a
production management team together
with production engineers, maintenance
managers, quality control and production
planning personnel. Thus the organisation
had three levels of personnel instead of
the earlier five. The production
management team is responsible for the
daily production of the plant. The former
first-line supervisors are now called
process technicians. Process technicians
were trained in co-operation and
communication skills at the beginning of
the change period.
The worker teamsThe worker teams continued to work
in three-shift terms. The principles and
rules of the worker teams were defined
before the change by the teams and the
process technician of the team in co-
operation with the VTT researchers.
The rules were specified in the
following subject areas: occupational
safety, production and quality, pro-
duction schedule, maintenance and
repair, worker substitutes (vacation,
ATR_2001.p65 5.12.2001, 11:3567
AU
TOM
ATIO
N T
EC
HN
OLO
GY
RE
VIE
W 2
001
68
sick-leaves etc.), overtime work, train-
ing and job instruction, booking of
work hours, reporting and information.
Safety responsibilitiesThe safety responsibilities of the
production management and workers
were defined. Attention was paid
especially to risk identification, safety
rounds, monitoring and controlling of
safety work habits, and informing of
technical faults and deficiencies.
Safety rulesSafety rules were specified in two
meetings with the process tech-
nicians, the representatives of the
teams, and the company safety
representative of the workers. The
major point in the definition was to
specify the role and responsibilities of
the worker teams in matters relating
to safety. Duties were defined in the
following safety related areas: risk
identification, informing of perceived
defaults, near-miss reporting, accident
reporting, production disturbances
and deviances, preventive maintenance
and repair, working alone, takeover and
proficiency, catastrophes, social
problems of work groups, problems
in the division of labour, improvement
of work capacity and information flow.
Risk analyses and jobinstructionThe teams were trained to identify and
assess the safety hazards in their work
in a one-day session. Process technicians
also participated in the session. The
hazards were divided into four main
groups: general hazards in the work
environment, hazards related to work
methods, hazards related to specific
process disturbances and break-downs,
and catastrophic hazards, such as fires,
gas explosions etc. During the training
session the teams did a preliminary
general analysis and one specific case
analysis concerning their own work
methods and work process. The
analyses were discussed together, and
plans were made to improve the work.
Safety cultureThe safety culture questionnaire
revealed that there were many problems
in the areas of co-operation and
communication within the teams, and
between the team and the process
technician. The ease and fluency of
handling safety problems deteriorated
from 28 % to 15 % during the change. In
the traditional organisation, the daily
safety responsibility lies with the
supervisors. 56 % of the workers
evaluated the supervisors fairly well in
terms of the care they took in fulfilling
these duties before the change into the
team organisation. After two years of
team organisation, the ratings decreased
to 41 %. Support from the supervisors
in problem situations deteriorated from
44 % to 22 %.
The supervisors’ negligence of daily
safety was seen in the ratings of
preventive safety as well. The ratings of
a good level of preventive safety work
deteriorated from 55 % to 37 %. Some
increase towards passivity is also seen
in a question about one’s own attitude
towards fixing things. 94 % of the pre-
team personnel reported that they try
first to fix faults and deficiencies by
themselves, compared with 85 % of the
teamwork personnel.
Risk training had a positive effect
on the results of the evaluations.
Safety analysis and hazard identi-
fication improved from 42 % to 54 %,
informing of technical faults and
deficiencies improved from 77 % to
93 %. Teamwork emphasises everyone’s
Figure 3. The figure shows production rules are based on experience of the team, information and ability of the team, the wayof acting and the company’s targets.
ATR_2001.p65 5.12.2001, 11:3568
69
AU
TOM
ATIO
N T
EC
HN
OLO
GY
RE
VIE
W 2
001
Figure 4. Rules and values are control action. In situations where the rules are missing, the values of thecompany define the actions.
personal responsibility for safety. 85 %
of the team workers said that they
consider safety when working
compared with 83 % before the team
change. The use of safety devices was,
however, reported to have decreased,
from 71 % to 58 %, and the use of safety
switches from 67 % to 56 %. On the
other hand, team workers reported that
they wore personal safety equipment
more often than before (52 % vs. 44 %).
Team workers also told their co-
workers about good solutions more
often than before (56 % after, 46 %
before). They also received more
support from their co-workers (41 %
after, 34 % before).
Teamwork itselfAs to the teamwork itself, 67 % of the
workers were quite satisfied with the
independence of their team, 78 %
considered that their team functioned
flexibly, and 82 % effectively. About half
of the workers (48 %) were satisfied
with the division of work within the team.
But only 44 % considered the climate of
the team to be rewarding and
encouraging. 67 % considered that the
commitment within the team was good,
18 % average and 15 % poor.
Stress and accidents
The change of the organisation into a
team model affected the stress reactions
of the workers. 58 % of the workers
reported that before the change that
they had little if no stress reactions, while
after the change only 44 % reported
little or no stress reactions. Moderate
stress reactions were reported by 33 %
before the change, and 44 % after the
change.
The accident frequency (acci-
dents/1 million work hours) was 0,658
in 1996 and 0,405 in 1997. During the
year of the team change the accident
frequency was 0,380. The severity of
accidents decreased from 100 days of
absence in 1997 to 46 in 1998. The year
1999 was more problematic. There
were 13 accidents reported in 1999,
compared to 5 in 1998. Also, the
severity of the accidents increased;
accidents in 1999 led to 855 hours of
absence from work, compared to only
245 hours of absence in 1998.
Two Models to EnsureSafety in TeamsIn the research study two models that
ensure safety in teams were developed.
The first safety management model
is targeted towards large and medium
size enterprises. This safety mana-
gement model consists of five main
items:
1) definition of line organisations and
support personnel roles and duties
2) identification and deletion of safety
risks by team workers
3) the role of safety personnel in team
organisation
4) definition of safety policy and
safety value
5) occupational instruction and
guidance model with responsi-
bilities and follow-up descriptions.
Team organisation creates new
challenges and variable functions for
the safety personnel. Communication
of information and implementation
of safety suggestions is one important
sector (Fig. 2). For instance, infor-
mation about the identification and
deletion of safety risks by team workers
must reach the safety personnel.
The production rules (the second
model) are targeted at small size
Integrated Safety Management in Teamwork Organisation
ATR_2001.p65 5.12.2001, 11:3569
AU
TOM
ATIO
N T
EC
HN
OLO
GY
RE
VIE
W 2
001
70
ReferencesHätinen M. & Ruoppila I. 2000.Attitudes towards safety in theFinnish metal industry. Universityof Jyväskylä.
Nurmi, M. 2000. Occupationalsafety in new kinds oforganisations. In the 1st
International Conference onOccupational Risk Preventionproceedings. February 23rd -25th,Tenerife, Spain.
Tammi, M. 2000. Safety practicesassociated with new system oforganisation within the metalindustry. In the 1st InternationalConference on Occupational RiskPrevention proceedings. February23rd -25th, Tenerife, Spain.
www.occuphealth.fi/e/dept/t/safety_management/managem.html
Figure 5. The figure shows how the rules can be created in different steps. Importantprinciples must be agreed on by rule groups in order to start creating the rules.Later other rules are added.
enterprises and include safety rules
also. The production rules are based on
the experience of the team, information
and ability of the team, the way of acting
and the company’s targets (Fig. 3). The
aim of production rules is to get all
personnel to put even their passive
knowledge into action.
In team organisations, the definition
of values can be seen as a significant
part of the management system. The
rules of the teams are created according
to the values of the company. In
situations where there are no rules
available, the values of the company
define the actions (Fig. 4).
Safety issues must involved when the
rules of the teams are chosen. The rules
are needed by the teams in their
everyday work. The rules can be
created in different stages. Some
important subjects must be agreed on
by rule groups in order to start
creating the rules. Later other subjects
are added (Fig. 5).
ConclusionsThe change into a team organisation
means many changes in the daily
work life. These changes are quite
well reflected in the perception of
the workers. Deter ioration in
communication, in safety values, and
in overall care of safety is also reflected
in the accident numbers. During the
change, when a lot of interest was aimed
at safety subjects, the accident numbers
were low. But as soon as the teams
started on their own, the amount of
accidents increased. There were no
other major changes in the organisation
or in the production process during the
study.
The results point to the importance
of continuous monitoring and revision
of safety practices in team organisations.
On the basis of the results reported here,
the plant has revised its team meetings
and decided to have more thorough
follow-up meetings of production and
safety twice a year. They also plan to have
a trained safety agent in each of the
worker teams.
In the research project, two models
were developed to ensure safety in the
teams. The safety management model
that consists of five main items is
targeted at large and medium size
enterprises, while the production rules
are targeted at small size enterprises.
The models were tested in two large
enterprises and two small enterprises.
The creation of these models was also
a major result of the study.
Jarmo KarlundResearch ScientistVTT Automation
Integrated Safety Management in Teamwork Organisation
ATR_2001.p65 5.12.2001, 11:3570
71
AU
TOM
ATIO
N T
EC
HN
OLO
GY
RE
VIE
W 2
001
VTT AUTOMATIONThe aim of VTT Automation is to promote the competitiveness of our partners bydeveloping new products, production systems, and operating principles. Our aimis to be a technology forerunner, a reliable partner and an attractive employer. Wehave a high quality staff consisting of 340 people. Our turn over is 28 millioneuros. The institute has five research fields.
Industrial AutomationThe mission of Industrial Automation is to promote theability of industrial companies to operate efficiently andprofitably in changing business environments. We developtools, methods, and practices that help companies tomanage and control their production processes andengineering activities. Our customers often need integratedsolutions, so a multitechnical and multidisciplinaryapproach is required. Our staff includes not only graduateengineers, but also researchers with academic degrees insubjects such as psychology, sociology and mathematics.Industrial Automation performs research in areas such asthe following: new generation of automation systemarchitectures, advanced control of industrial processes,system reliability, modelling and simulation of industrialprocesses and systems, organisational, economic andmanagerial issues of technological systems, manufacturingnetworks, and man-machine systems.
Machine AutomationThe mission of Machine Automation is to increase theautomation level in work machines, vehicles and devicesby integrating different mechatronic technologies. MachineAutomation covers instrumentation and the systems usedto control and automate machine operations. The mainfields of interest are navigation, location and guidancesystems for vehicles, manipulator control methods andsystems, environment perception and signal processingtechnologies, including connections to modern networks.Other important areas are production automation andwireless factory solutions, telematic applications andservices for mobile users, and robotics for spaceapplications.
Safety EngineeringThe research field of Safety Engineering of VTT Automationoffers to enterprises and communities a safety-conscioustotal service to develop businesses, enterprise networks,production systems, and machines, as well as occupational,environmental, and working methods. It also studies anddevelops new technologies, products and methods tomanage indoor atmospheric contaminants and noise. Thework is done in several industrial sectors, the mostimportant of which are the metal, machine manufacturing,basic metal and electronics industries. The focus of theSafety Engineering research field activity is the integrationof safety and productivity requirements to the optimumlevel. The research field is split into five research groups:Production Development, Industrial Ventilation, Noise
Control, Machine Development and Electronics ProductTechnology.
Measurement TechnologyMeasurement Technology develops sensors andmeasurement instruments for industry and researchinstitutes. The sensor research is focused onmicromechanics, low temperature sensors, and detectorsbased on quantum effects. Not only sensors, but alsoreadout electronics usually based on a capacitive methodor the use of the electrical or mechanical resonances, aredeveloped. Recently the research has been focused on radiofrequency components and antennas mainly intended forwireless communication and for remote sensing. In additionto sensors, measurement instruments for the electrical andprocess industries are developed. Optical instruments aredesigned, manufactured and verified for space, scientificand industrial applications. One of the groups inMeasurement Technology is developing advanced imageinterpretation methods using remotely sensed data. Themethods are transferred to the user community. Specialareas of expertise include multi-source data applicationsfor environmental monitoring and forestry, as well asapplications for sea ice monitoring. The MeasurementTechnology research field will be merged with VTTInformation Technology in 2002.
Risk ManagementRisk management includes risk analyses and operationalreliability studies for the industry as well as expert servicespromoting the product-related activities of the company.The starting point of activities is close cooperation withprogressive Finnish and foreign companies. The aim of riskanalyses is to reduce the exposure of humans, plant-lifeand the environment to potential risks. Studies ofoperational reliability aim to reduce the occurrence ofdisturbances, thereby increasing efficiency and shorteningthrough-times. Product-related activities of the companiesare promoted by analysing and developing the productionmethods of these companies in order to enable them toachieve economically feasible manufacturing of highquality products that are safe and easy to use, and thatmeet terms agreed with customers. In the field of medical-device technology, we also perform conformity studies ofthe products.
ATR_2001.p65 5.12.2001, 11:3571
AUTOMATION
AUTOMATIONP.O. Box 13001
33101 Tampere, Finlandwww.vtt.fi/aut
forename.surname@vtt.fi
AU
TO
MA
TIO
N T
EC
HN
OLO
GY
RE
VIE
W 2
00
1V
TT
AU
TO
MA
TIO
N
AAAAAUTUTUTUTUTOMAOMAOMAOMAOMATIONTIONTIONTIONTIONTECHNOLOGYTECHNOLOGYTECHNOLOGYTECHNOLOGYTECHNOLOGYREVIEW 2001REVIEW 2001REVIEW 2001REVIEW 2001REVIEW 2001
kansi aukeamana.p65 5.12.2001, 11:311
Recommended