UDC 621.3 REAL-TIME REGULATION SYSTEMS BASED ON ...

UDC 621.3
ІНФОРМАЦІЙНІ СИСТЕМИ І МОДЕЛЮВАННЯ
REAL-TIME REGULATION SYSTEMS BASED ON INTERNET – OPTIMIZATION
ALGORITHM
Vince T., research, Ing., assist. Prof., Molnár J., research, Ing., assist. Prof., Bučko R., research, Ing., assist. Prof.
Technical University of Košice
Park Komenského 3, 04200 Košice, Slovak Republic
E-mail: tibor.vince@tuke.sk , jan.molnar@tuke.sk , radoslav.bucko@tuke.sk
The article presents one of the ways for implementing Internet as a communication bus between real-time system
elements. It handles the advantages and disadvantages of Internet as a control and communication bus in different level
of the information hierarchy. The paper describes algorithm, which helps to build network for a real-time system. This
algorithm is suitable for creating new network, or own domain of system elements. For implementing system in existing
network, there would be few modifications in algorithm.
Key words: Internet, real-time system, algorithm.
Introduction. The Internet begins to playing a very
important role in industrial processes manipulation, not
only in information retrieving. New concept of
controlling, which has been paid much attention in these
years is distance remote via Internet, or other words,
Internet-based control. Such a type of control bus allows
remote monitoring or regulation of plants or single
devices over the Internet. With the progress of the
Internet it is possible to control and regulate from
anywhere around the world at any time. The design
process for the Internet-based control systems includes
requirement specification, architecture design, control
algorithm, interface design and possibly safety analysis.
Due to the low price and robustness resulting from its
wide acceptance and deployment, Ethernet has become
an attractive candidate for real-time control networks.
Requirements to put Internet depend on the system.
Since each system has its own requirements, there is no
prototype of such Internet bus. Therefore there should
exist an algorithm for building suitable network.
Specific algorithm depends on the number of system
elements (if system is complex or not), if system should
be implemented in existing network, or it is necessary
build a new one etc. This contribution solves the cases,
when system needs to build a new network, or all
system elements could be included in system own
network domain.
Network Performance. Network performance is
referred in more parameters in mutual relationship. One
of performance parameters is Latency. Latency means a
time required to transfer an empty message between
relevant computers. Another parameter is Data transfer
rate. Data transfer rate is the speed at which data can be
transferred between sender and receiver in a network.
The unit of this parameter is Bits/sec. For message
transfer time calculating is equation 1. A third parameter
of network performance is Bandwidth. Bandwidth is a
total volume of traffic that can be transferred across the
network. Maximal data rate formula is shown in
equation 2. This maximum is only theoretical, not
reachable in practice [5].
LM
MTT = LT + (1)
DTR
where MTT is Message transfer time, LT is latency, LM
is Length of message and DTR is Data transfer time.
sg
MDR = CB . log 2 (1 + )
ns
(2)
where MDR is maximal data rate (bps) and CB is carrier
Bandwidth, sg is signal magnitude and ns is noise
magnitude.
The all parameters are pointing on the main
disadvantage of controlling via the Internet – packets
delivery delay. When packets are concurrently
transported over an ordinary Ethernet, packets may
experience a large delay due to contention with other
packets in the local node where they originate and
collision with other packets from the other nodes. By
data transmission, four sources of delay spring up at
each hop: nodal processing, queuing, transmission delay
and propagation delay. The most significant part of total
delay belongs to queuing. By queuing is considered the
following equation 3:
A
TI = L . (3)
W
where TI is traffic intensity, L is packet length (bits), A
is average packet arrival rate, and W is link bandwidth
(bps).
If ratio L*A/W will be very small almost 0, average
queuing delay is small. If ratio L*A/W rise up to 1,
delays become large (exponentially) and if ratio L*A/W
is bigger than 1 average delay is infinite, more “work”
arriving than can be serviced. Today, there is a term
Quality of Service (QoS). Quality of Service is the
ability to provide different priority to different
applications, users, or data flows, or to guarantee a
certain level of performance to a data flow.
Information Architecture. It is becoming
increasingly necessary to think in terms of integration of
information and control, across the entire plant site. The
term of integration, especially integration of information
and control across the entire plant site, becomes more
and more significant. In the manufacturing industries
this is often referred to as "Computer Integrated
Manufacturing" (CIM). On the surface, it would seem
that the increasing use of microprocessor-based plant
level devices such as programmable controllers,
distributed digital control systems, smart analyzers,
personal computers, etc., would make this easy. After
all, most of these devices have "RS232" connectors,
which enable connection to computers. Unfortunately,
the real world situation is somewhat more challenging.
Вісник КДУ імені Михайла Остроградського. Випуск 4/2010 (63). Частина 3
150

If we began to hook all these RS232 ports together,
there would soon be an unmanageable mess of wiring,
custom software and little or no communication. This
problem solution results in integration these devices into
a meaningful "Information Architecture". This
Information Architecture can be separated into 4 levels
with the sensor/actuator level as shown in Fig. 1, which
are distinguished from each other by“4Rs” principle
criteria.
Local
Computer
fast
more
slow
less
Response time Resolution Reliability Reparability
ІНФОРМАЦІЙНІ СИСТЕМИ І МОДЕЛЮВАННЯ
Scheduling,
Management
easy
low
Optimisation,
Maintenance
Supervisory
Control
high
Interlocking
& Regulatory
hard
Figure 1- Information Architecture
Sensor,
Actuator
The 4Rs criteria are: Response time, Resolution,
Reliability and Reparability.
Response time: as one moves higher in the
information architecture, the time delay, which can be
tolerated in receiving the data, increases. Conversely,
information used at the management & scheduling level
can be several days old without impacting its usefulness.
Resolution: an Abstraction level for data varies
among all the levels in the architecture. The higher the
level is, the more abstract the data is.
Reliability: Just as communication response time
must decrease as one descends through the levels of the
information architecture, the required level of reliability
increases. For instance, host computers at the
management & scheduling level can safely be shut
down for hours or even days, with relatively minor
consequences. If the network, which connects
controllers at the supervisory control level and/or the
regulatory control level, fails for a few minutes, a plant
shutdown may be necessary.
Reparability: The reparability considers the ease
with which control and computing devices can be
maintained.
Local computer on supervisory control level is able
communicate with higher levels of information
architecture via Internet, but there is also possibility to
use the Internet also in lower levels of the Information
architecture. The Internet can be linked with the local
computer system at any level in the information
architecture, or even at the sensor/actuator level. These
links result in a range of 4Rs (response time, resolution,
reliability, and reparability). For example, if a fast
response time is required a link to the control loop level
should be made. If only abstracted information is needed
the Internet should be linked with a higher level in the
information architecture such as the management level
or the optimization level.
Industry’s Buses. There are various types of buses
using in Industry to connect of different devices, which
are supported on the type of interface/bus [6]:
AS-Interface – the Actuator Sensor Interface offers
many of the benefits of more powerful and expensive
fieldbuses, but at much lower cost and as a much
simpler installation. Transmission of analog signals via
time multiplex procedure. Data and power via the same
line. No termination necessary. Address setting
automatically from the master or via service tool. ASI
conforment power supply required. It supports 62 nodes,
max 300m with 3 repeaters. Date rate is 167kbit/s.
Addressing is Master/Slave.
CANopen – is a CAN-based higher layer protocol. It
was developed as a standardized embedded network
with highly flexible configuration capabilities. Node
removal without severing the network is possible.
Provisions for the typical request/response orientated
network communications. Provisions for the effiecient
movement of data framentation for moving larger bodies
of information It supports 127 nodes, 25-5000m
(depending on boudrate). Data rate is from 10kbit/s to
1Mbit/s, addressing is Master/Slave, Peer-to-Peer,
Multi-cast and Multi-master.
ProfiBus – is a Multi-Master System and makes
possible the mutual operation of several automation,
engineering or visualizing systems at a Bus. The
Masters, also designated as active devices, define the
data traffic on the Bus. When in possession of the access
permission (Token), they can send data without external
requests. The Slaves, designated as passive devices,
have no Bus access permission. They can only confirm
received messages or send messages when requested by
a Master. Baud rates from 9.6 kBaud up to 12 MBaud
are supported. A maximum of 126 devices can be
operated at the Bus. Profibus also supports Broadcast
and Multicast communication. Network length is
possible 100-1200m, addressing is DP: Master/Slave,
Cyclic, Polling, DPV1: Cyclic, Polling + acyclic data
transfer.
EthernNet/IP – The Industrial Ethernet Protocol
(Ethernet/IP) has been developed by ODVA with strong
support from Rockwell Automation. It uses the Control
& Information Protocol (CIP) which is already well
known from ControlNet and DeviceNet. Network size
(number of nodes) is scalable and nearly unlimited.
Network size can be 100m (10/100 Base-T) or 35-
2000m (fiber optic). Standard layers 1-4 providing
Ethernet data transmisson, bus access, internet protocol
(IP) and TCP & UDP protocols. CIP "implicit" and
"explicit" messaging with encapsuslation technology.
Message routing between EtherNet/IP, DeviceNet &
ControlNet.
Network Improve Methods. If the network doesn’t
meet system requirements, it should be improved.
Among the ordinary methods of improvement we can
include: improvement of network architecture,
improvement of network bandwidth and special
protocols implementation. The sequence of these
methods in algorithm can vary depending for example
on the cost of implementing these methods.
Improvement of network architecture can lead to the
increase of the number of nodes, branches and changing
architecture. The goal is to decrease network traffic jam
and that way decrease packets time delay in.
Вісник КДУ імені Михайла Остроградського. Випуск 4/2010 (63). Частина 3
151

Classic Ethernet has 10MBps bandwidth and has
reserve in 100MBps and also 1GBps bandwidth.
Widening Ethernet bandwidth is another way to
decrease packet delivery time. It is possible to decrease
time delay by using special protocols, for example
RSVP- Resource Reservation Protocol, IPv6-support
packet priority.
Applied special protocols must be supported by all
network elements. Special protocols are meaningful
only in complex systems.
Time deterministic Ethernet is a special way for
decreasing time delay. For example: Industrial Ethernet,
Ethernet Powerlink etc. Deterministic timing is achieved
by applying a cyclic timing schedule for all connected
nodes to access the physical layer. To apply
deterministic type of Ethernet, all system elements in a
real-time domain must support such type of Ethernet.
The last chance to use Internet is to modify system
requirements. Sometimes, with small system changes,
can modified system requirements meets the network
parameters.
Algorithm.
The whole diagram is shown in figure Fig. 2
The first step is dataflow analysis. The most
important question that needs to be answered is the
following: What amount of data will be transported over
the net and what is the greatest acceptable time delay?
The next step is an architecture proposal. The
architecture should follow dataflow analysis. The first
architecture proposal should be as simple as possible.
Next there is network type proposal. Network type
determines network bandwidth, type of wiring etc. The
first proposal of network should be with low bandwidth.
After first proposals, there is a testing, if the network (as
simple as possible) meets the system requirements. This
is the most difficult part of the algorithm – to answer
this question.
The answer should include the most inopportune
system state with the biggest possible network traffic
jam. To answer the question for simple systems it is
possible to use calculation, but it is better to use some
simulation tools (OMNeT++, ns2, OPNET …) for more
complex systems.
In case of special (not standard) situations the time
delay exceeds the system requirements, the system
needs to be maintained to handle such situations.
Algorithm is shown in figure Fig. 3. It is possible to
ignore such situations when they appear rarely and time
delay exceeding does not cause system to malfunction.
Conclusion. It is become to be a standard, that
many control elements have been embedded with
Internet-enabled functions, for example, PLC with
TCP/IP stack, smart control valves with a built-in
wireless communication based on TCP/IP protocol, and
process control computer (DCS) with an Internet
gateway. There possibility that some mechatronic
system could be connected directly to the Internet
(without a necessity of a server computer).
These days, there is effort to utilize Internet also in
real-time systems. However there is no guide for
building such a network yet. This algorithm describes
ІНФОРМАЦІЙНІ СИСТЕМИ І МОДЕЛЮВАННЯ
one of the possible ways. The whole algorithm aims to
build a network that is as simple as possible and
gradually improve it until the network parameters meet
the system requirements.
Requirements of feedback less than tens of
miliseconds can be achieved only with deterministic
type of Ethernet, but all devices in real-time domain
must support such type of Ethernet. Time delay
exceeding can be ignored, when it does not cause
system malfunction.
Is is possible to apply
special protocols to
speed up the
communication ?
are the
system
requirements
satisfied ?
Вісник КДУ імені Михайла Остроградського. Випуск 4/2010 (63). Частина 3
152
yes
apply suggest
network
no
no
no
Is it possible to
apply time
deterministic type
of Ethernet ?
no
Is it possible to
change system
requirements ?
no
new network
creation
algorithm
data flow
analysis
architecture
proposal
network type
proposal
Is is possible to
apply higher
bandwidth net ?
no
Is it possible to
change the net
architecture to
decrease network
delay ?
yes
apply the special
protocols
change the system
requirements
yes
apply the higher
bandwidth net
apply time deterministic
type of Ethernet
yes
change the
architecture
yes
maintain
are the
requirements
satisfied in
standard state of
system and net ?
no
in these circumstances is
not possible to build system
bus based on the Ethernet
yes
yes
Figure - 2 The algorithm
Acknowledgement. The paper has been prepared by
the support of Slovak grant projects VEGA 1/0660/08,
KEGA 3/6386/08, KEGA 3/6388/08, KEGA 003-
003TUKE-4/2010

yes
maintain the system for
time-delay exceeding
apply suggest
network
no
maintain
Is it possible to
maintain the system in
handling with
situations, when time
delay exceeds the
system requirements ?
Exceeding a
time delay
limit would
cause system
malfunction ?
Figure - 3 The system maintain algorithm
ІНФОРМАЦІЙНІ СИСТЕМИ І МОДЕЛЮВАННЯ
no
yes
in these circumstances is
not possible to build
system bus based on
the Ethernet
REFERENCES
1. Yang, S.H., Tan, L.S., Chen X: Requirements
Specification and Architecture Design for Internet-based
Control Systems, proceedings of the 26th Annual
International Computer Software and Applications
Conference (COMPSAC’02), 2002.
2. Kweon, S. K, Cho M., Shin K. G.: Soft Real-
Time Communication over Ethernet with Adaptive
Traffic Smoothing, IEEE Transactions on parallel and
distributed systems, VOL. 15, NO. 10, October 2004.
3. Hall E.: Internet Core Protocols: The Definitive
Guide, O'Reilly & Associates (February, 2000) USA,
ISBN: 1-56592-572-6.
4. Yang, X., Chen Q.: Virtual Reality Tools for
Internet-Based Robotic Teleoperation, proceedings of
the 8th IEEE International Symposium on Distributed
Simulation and Real-Time Applications (DS-RT’04).
5. Fonda C., Postogna F.: Computer networking
basics, ICTP workshop on telecommunications: science,
technology and applications. Trieste, 15th September -
3rd October 1997
6. http://www.anybus.com/technologies/technolog
ies.shtml, 10.2.2008
7. Molnár, J., Kováčová, I.: Distance remote
measurement of magnetic field. In: Acta Electrotechnica
et Informatica, 2007, Vol.7, No.4, pp. 52-55, ISSN
1335-8243.
8. Molnár, J.: Telemetric system based on
internet. In: OWD 2009 : 11 International PhD
Workshop : Wisa, 17-20 October 2009, p. 38-41. ISBN
83-922242-5-6.
9. Kováčová, I., Kováč, D.: Non-harmonic power
measuring. In: Acta Electrotechnica et Informatica. Vol.
8, No. 3 (2008), pp. 3-6. ISSN 1335-8243.
10. Tomčík, J., Tomčíková, I.: Safety politics in the
enviroment of the automatized and SCADA systems ( In
Slovak). In: EE Journal. Vol. 14, No. 1 (2008), pp. 46-
47. ISSN 1335-2547
11. Vince, T., Molnár, J., Tomčíková, I.: Remote
DC motor speed regulation via Internet. In: OWD 2008:
10th international PhD workshop : Wisla, 18-21 October
2008. p. 293-296. ISBN 83-922242-4-8.
Вісник КДУ імені Михайла Остроградського. Випуск 4/2010 (63). Частина 3
153
Стаття надійшла 20.04.10 р.
Рекомендовано до друку к.т.н., доц.
Гладирем А.І.
СИСТЕМЫ РЕАЛЬНОГО ВРЕМЕНИ НА ОСНОВЕ АЛГОРИТМА ИНТЕРНЕТ-
ОПТИМИЗАЦИИ
Винс Т., научный сотрудник, инж., ассист., Мольнар Дж., научный сотрудник, инж., ассист., Бучко Р.,
научный сотрудник, инж., ассист.
Технический университет в Кошице
Park Komenského 3, 04200, Košice, Slovak Republic
E-mail: tibor.vince @ tuke.sk, jan.molnar @ tuke.sk, radoslav.bucko @ tuke.sk
Данная статья представляет один из способов применения Интернета в качестве коммуникационной шины
между элементами системы реального времени. Рассмотрены преимущества и недостатки Интернета как шины
управления и коммуникационной шины на различных уровнях информационной иерархии. В статье
описывается алгоритм, который облегчает создание сети для систем реального времени и пригоден для
создания, как новой сети, так и собственного домена элементов системы. При внедрении системы в
существующую сеть потребуются некоторые изменения в алгоритме.
Ключевые слова: Интернет, система реального времени, алгоритм
СИСТЕМИ РЕАЛЬНОГО ЧАСУ НА ОСНОВІ АЛГОРИТМУ ІНТЕРНЕТ-
ОПТИМІЗАЦІЇ
Вінс Т., науковий співробітник, інж., асист., Мольнар Дж., науковий співробітник, інж., асист., Бучко Р.,
науковий співробітник, інж., асист.
Технічний університет в Кошице
Park Komenského 3, 04200, Košice, Slovak Republic
E-mail: tibor.vince @ tuke.sk, jan.molnar @ tuke.sk, radoslav.bucko @ tuke.sk
Дана стаття представляє один зі способів застосування Інтернету в якості комунікаційної шини між
елементами системи реального часу. Розглянуто переваги і недоліки Інтернету як шини управління та
комунікаційної шини на різних рівнях інформаційної ієрархії. У статті описується алгоритм, який полегшує
створення мережі для систем реального часу і придатний для створення, як нової мережі, так і власного домену
елементів системи. При впровадженні системи в існуючу мережу будуть потрібні деякі зміни в алгоритмі.
Ключові слова: Інтернет, система реального часу, алгоритм.