Table Of ContentD.P. Buse and Q.H. Wu
IP Network-based
Multi-agent Systems
for Industrial
Automation
InformationManagement, ConditionMonitoring
and Controlof Power Systems
123
D.P.Buse,PhD
Q.H.Wu,PhD,CEng,FIEE
DepartmentofElectricalEngineeringandElectronics
TheUniversityofLiverpool
BrownlowHill
LiverpoolL693GJ
UK
BritishLibraryCataloguinginPublicationData
Buse,DavidP.
IPnetwork-basedmulti-agentsystemsforindustrial
automation:informationmanagement,conditionmonitoring
andcontrolofpowersystems.-(Powersystems)
1.Electricsubstations-Automaticcontrol-Computer
networkresources2.Intelligentagents(Computersoftware)
3.Electricpowertransmission-Dataprocessing
4.Distributedartificialintelligence
I.TitleII.Wu,Qing-Hua
621.3’126’02854678
ISBN-13:9781846286469
LibraryofCongressControlNumber:2006939313
PowerSystemsSeriesISSN1612-1287
ISBN 978-1-84628-646-9 e-ISBN 978-1-84628-647-6 Printedonacid-freepaper
©Springer-VerlagLondonLimited2007
ActiveX®andVisualBasic®areregisteredtrademarksoftheMicrosoftCorporation,OneMicrosoftWay,Redmond,
WA98052-6399,USAhttp://www.microsoft.com/
ARCHON™isatrademarkofallpartnersoftheARCHONproject,http://www.archon.org/
CORBA®isaregisteredtrademarkoftheObjectManagementGroup,Inc.,140KendrickStreet,BuildingA,Suite300,
Needham,MA02494,USAhttp://www.omg.org/
Fieldbus™isatrademarkofFieldbusFoundation,Austin,Texas,USAhttp://www.fieldbus.org/
JACK™isatrademarkoftheAgentOrientedSoftwareGroup,153OwensvilleRoad,WestRiver,MD20778,USA
http://www.agent-software.com/
Java™andJini™aretrademarks ofSunMicrosystems, Inc.,4150NetworkCircle, SantaClara, CA95054,USA
http://www.sun.com/
LabVIEW™isatrademarkofNationalInstrumentsCorporation,11500NorthMoPacExpressway,Austin,TX78759-
3504,USAhttp://www.ni.com/
Profibus®isaregistered trademarkofProfibusInternational, Haid-und-Neu-Str.7,76131Karlsruhe,Germany,
http://www.profibus.com/
Apartfromanyfairdealingforthepurposesofresearchorprivatestudy,orcriticismorreview,aspermittedunderthe
Copyright,DesignsandPatentsAct1988,thispublicationmayonlybereproduced,storedortransmitted,inanyform
orbyanymeans,withthepriorpermissioninwritingofthepublishers,orinthecaseofreprographicreproductionin
accordancewiththetermsoflicencesissuedbytheCopyrightLicensingAgency.Enquiriesconcerningreproduction
outsidethosetermsshouldbesenttothepublishers.
Theuseofregisterednames,trademarks,etc.inthispublicationdoesnotimply,evenintheabsenceofaspecific
statement,thatsuchnamesareexemptfromtherelevantlawsandregulationsandthereforefreeforgeneraluse.
Thepublishermakesnorepresentation,expressorimplied,withregardtotheaccuracyoftheinformationcontained
inthisbookandcannotacceptanylegalresponsibilityorliabilityforanyerrorsoromissionsthatmaybemade.
987654321
SpringerScience+BusinessMedia
springer.com
To our parents and families
Preface
Duetothecomplexityofdistributedsystems,suchasrailways,aerospacesys-
tems, navigation systems, gas transmission systems, power utility and power
plants, the conventional automation system is not capable of providing in-
formation management and high-level intelligent approaches.This is because
achieving these functionalities requires comprehensive information manage-
ment support and coordination between system devices, and the control of
many different types of task, such as data transportation, data display, data
retrieval,informationinterpretation,controlsignalsandcommands,documen-
tationsortinganddatabasesearchingetc.Theseoperateatdifferenttimescales
andarewidelydistributedovertheglobalsystemanditssubsystems.Without
reasonably designed system software architectures and hardware structures,
it is impossible to handle these tasks efficiently, safely and reliably, with the
possibility of online reconfiguration and flexibly embedding applications.
Withadvancesincommunicationtechnologies,inparticularInternettech-
nology,in recentyears,Internet Protocol(IP)networkshave been considered
for use in conventional automation systems. This book is substantively con-
cerned with developing a novel concept−e-Automation−that capitalises on
the advantages of IP networks and agent technology for system integration
and leads to a new generation of industrial automation systems. In contrast
to conventional automation systems, an e-Automation system can provide
integrated functionalities for distributed information management condition
monitoringandcontrolwithanopenarchitectureofsystemsoftwareandhard-
wareforimplementation ofvarioustaskswithin Wide Area Networks(WAN)
andLocalAreaNetworks(LAN).Thee-Automationsystemisabletoprovide
greatgridabilityandcommunicationcapabilitytoresolvetheproblemsoftask
implementation and information management for a wide range of distributed
complex industrial systems.
The basic idea of the multi-agent-based e-Automation system was first
consideredin 1998,basedonworkundertakenin theIntelligenceEngineering
and Automation research group, The University of Liverpool, over the pre-
vious ten years, in the areas of distributed control and automation systems,
viii Preface
computationalintelligence,intelligentsystems,informationmanagement,and
power system control and operation. Since then, e-Automation architecture
has evolved with initial support from National Grid Transco (NGT). As the
e-Automation system is in its early stage of development, some fundamen-
tal issues need to be studied. What is the optimal hardware structure of the
system? What is the most suitable software platform that can accommodate
the intelligent agents and give room for future development? How can one
develop a stable, reliable, and robust system? Is there any theory to support
the system design? Many unsolved issues need to be investigated.
This book describes a substation automation architecture based on the
concept of e-Automation, using the multi-agent systems methodology. The
bookbeginsbypresentingthehistoricalbackgroundofsubstationautomation
systems, along with the newer network-based approaches and architectures.
Agents,multi-agentsystemsandmobileagentsarealsointroduced.Themain
contributions of the book are concerned with the development of an agent-
based architecture and its components for power system automation. The
implementation of a substation information management system with multi-
agent-based architecture is also presented.
We would like to thank John Fitch and Zac Richardson of NGT, for sup-
porting this work andprovidingassistancewith the substationsimulator and
InformationManagementUnit (IMU), and BrianBaker,also of NGT, for his
support of the work during its initial stages. We would also like to thank
Pu Sun, who worked on the hardware architecture and the data acquisition
system for the prototype, and with whom we collaborated on the writing
of several papers based on the work, and Jun Qiu Feng, who worked on
the human−machine interface and personal agents for the prototype system,
which are included as part of Chapter 6 in this book. Finally, thanks go to
Chen Ma for a large amount of time spent on the preparation of diagrams
and typesetting as well as the provision of the standards of agent develop-
ment platforms.
This work was supported by NGT. The facilities of the Department of
Electrical Engineering and Electronics and the e-Automation Laboratory at
the University of Liverpool were essential to the completion of the work.
Special thanks go to Anthony Doyle (the Senior Editor), Sorina Moos-
dorf (the Production Editor) and Simon Rees (Editorial Assistant) for their
professional and efficient editorial work on this book. Our thanks are also
extended to all colleagues in the Intelligence Engineering and Automation
research group, The University of Liverpool, for all assistance provided, and
which have not been specifically mentioned above.
University of Liverpool, UK, David P. Buse
Qing-Hua Wu
August, 2006
Contents
List of Figures................................................. xiii
List of Tables ..................................................xvii
1 Introduction............................................... 1
1.1 Industrial Automation ................................... 2
1.2 Automation Systems in Electricity Transmission............. 4
1.3 Network-basedPower System Automation .................. 5
1.4 e-Automation ........................................... 8
1.5 Book Outline ........................................... 9
2 Agents, Multi-agent Systems and Mobile Code ............ 13
2.1 Overview of Agent Technology ............................ 13
2.2 Intelligent/Autonomous Agents ........................... 13
2.2.1 Definitions ....................................... 13
2.2.2 Intelligent Agents in Information Processing and
Problem Solving .................................. 14
2.3 Agent Architectures ..................................... 16
2.3.1 Deliberative Architectures .......................... 16
2.3.2 Reactive Architectures ............................. 19
2.3.3 Learning-basedArchitectures ....................... 20
2.3.4 Layered Architectures.............................. 20
2.4 Standards for Agent Development ......................... 20
2.4.1 Foundation for Intelligent Physical Agents Standards .. 20
2.4.2 Mobile Agent Standards............................ 25
2.5 Mobile Agent Technology................................. 27
2.6 Multi-agent System...................................... 27
2.6.1 Architectures ..................................... 27
2.6.2 Multi-agent Programming .......................... 28
2.6.3 Middle Agents: Brokers and Facilitators .............. 31
x Contents
2.7 Agent Application Architectures........................... 32
3 An Agent-based Architecture forPower SystemAutomation 35
3.1 Agents in Power Systems ................................. 36
3.1.1 Control .......................................... 36
3.1.2 Negotiation and Pricing ............................ 37
3.1.3 Information Management........................... 37
3.1.4 Condition Monitoring.............................. 37
3.2 Tasks Performed ........................................ 38
3.3 A Multi-agent System for Power System Automation ........ 41
3.3.1 Agent Platform ................................... 42
3.3.2 Data Acquisition and Control System ................ 44
3.3.3 Information Management System and User Interface ... 48
3.3.4 Combined Multi-agent Architecture.................. 50
3.4 Agents, Tasks and Interaction Protocols.................... 52
3.5 Data and Knowledge..................................... 60
3.5.1 Available Data and Knowledge...................... 60
3.5.2 Knowledge Representation.......................... 62
3.5.3 Ontologies........................................ 66
3.6 Agent Platform Implementation ........................... 67
3.6.1 Standard FIPA Platform ........................... 68
3.6.2 JiniTM-based Platform ............................. 69
3.6.3 UDP-based Platform............................... 71
3.6.4 Combined FIPA and UDP-based Platform............ 71
3.7 Summary............................................... 72
4 Static Components of Architecture ........................ 75
4.1 Database Agents ........................................ 76
4.1.1 Description ....................................... 76
4.1.2 Agent Specification ................................ 78
4.1.3 Agent Implementation ............................. 78
4.2 Document Agents ....................................... 80
4.2.1 Description ....................................... 80
4.2.2 Agent Specification ................................ 81
4.2.3 Document Agent Issues ............................ 82
4.3 Ontology Agents ........................................ 83
4.4 Device Agents and Node Agents........................... 83
4.4.1 Description ....................................... 83
4.4.2 Agent Specification ................................ 84
4.5 Plant Agents............................................ 85
4.5.1 Description ....................................... 85
4.5.2 Agent Specifications ............................... 86
4.5.3 Data Acquisition System/Plant Mappings ........... 87
4.6 User Interface Agents .................................... 91
4.6.1 Description ....................................... 91
Contents xi
4.6.2 Agent Specification ................................ 91
4.7 Summary .............................................. 92
5 Applications of Mobile Agents............................. 93
5.1 Mobile Agent Performance................................ 93
5.2 Mobile Agent for Data Analysis ........................... 95
5.2.1 Agent Algorithms and Implementation ............... 97
5.2.2 Benchmarks ......................................100
5.2.3 Discussion .......................................105
5.2.4 Related Work.....................................106
5.2.5 Conclusions.......................................107
5.3 Mobile Agent for Remote Control of Power Systems .........108
5.3.1 Agent Algorithms and Implementation ...............110
5.3.2 Experiment.......................................112
5.3.3 Conclusions and Related Work......................123
5.4 Summary...............................................124
6 Multi-agent-based Substation Information Management
System ...................................................127
6.1 Introduction ............................................127
6.2 System Architecture and Agents...........................128
6.2.1 Information Management System Agents .............129
6.2.2 Personal Assistant Agents in Substation Information
Systems..........................................132
6.3 System Ontology ........................................133
6.4 Examples of Usage ......................................135
6.4.1 Querying IMU for a Data Set .......................136
6.4.2 Mobile Agent-based Analysis of Data ................141
6.4.3 Searching for Documents ...........................142
6.4.4 Performing an Action Using Data Acquisition Agents ..145
6.4.5 Reading a Plant Property Using Data Acquisition
Agents ...........................................146
6.4.6 Human−MachineInteractionUsingPersonalAssistant
Agents ...........................................148
6.5 Implementation Issues ...................................153
6.6 Summary...............................................154
7 Evaluation and Analysis ...................................157
7.1 Functionality ...........................................158
7.1.1 National Grid Transco Requirements for Substation
Control Systems...................................158
7.1.2 Haacke “Opportunity Matrix” ......................158
7.1.3 Summary of Functionality Results ...................160
7.2 Performance ............................................160
7.2.1 Data Acquisition Performance ......................160
xii Contents
7.2.2 Responding to User Queries ........................163
7.2.3 Data Display .....................................164
7.3 Modifiability............................................165
7.3.1 Modifying the Substation...........................165
7.3.2 Modifying the Data Sources ........................166
7.3.3 Modifying the User Interface........................167
7.3.4 Summary.........................................168
7.4 Security, Reliability and Availability .......................168
7.5 Integration into Existing Substations.......................169
7.6 Possible Applicability to Other Industries...................170
7.7 Discussion..............................................171
7.7.1 Advantages.......................................171
7.7.2 Disadvantages ....................................172
7.8 Summary...............................................173
References.....................................................175
Index..........................................................185
List of Figures
1.1 “CIM pyramid” model of an automation system.............. 3
1.2 Supervisory control and data acquisition (SCADA) system .... 5
1.3 Client−server automation system........................... 7
1.4 e-Automation for large-scale distributed complex systems...... 9
2.1 Structure of a BDI agent .................................. 17
2.2 Overview of the FIPA standards ........................... 21
2.3 FIPA abstract architecture mapped to different concrete
realisations .............................................. 22
2.4 FIPA agent management reference model constitution......... 23
2.5 FIPA message transport reference model .................... 24
3.1 Outline view of architecture ............................... 36
3.2 Functional decomposition of power system automation system . 39
3.3 User interaction.......................................... 40
3.4 Allocation of systems to substations and wide area network.... 41
3.5 Generic object model for data acquisition system (using UML
class diagram notation) ................................... 45
3.6 Multi-agent system for data acquisition and control........... 47
3.7 Information management multi-agent system................. 49
3.8 Combined multi-agent system.............................. 51
3.9 Agent collaboration....................................... 52
3.10 User interaction.......................................... 54
3.11 Intervention (including output data interpretation) ........... 55
3.12 Data acquisition.......................................... 55
3.13 Input data interpretation.................................. 56
3.14 Automatic control........................................ 57
3.15 Data storage............................................. 57
3.16 Querying................................................ 58
3.17 Document retrieval ....................................... 59
3.18 Document storage ........................................ 59
