Towards the Next Generation of Industrial Cyber-Physical Systems

23 Pages·2015·1.72 MB·English

Chapter 1
Towards the Next Generation of Industrial
Cyber-Physical Systems
ArmandoW.Colombo,StamatisKarnouskos,ThomasBangemann
Abstract IntelligentNetworkedembeddedsystemsandtechnologies,rangingfrom
components and software to Cyber-Physical Systems (CPS) [1], are increasingly
gainingimportancefortheICTsupplyindustry,thesystemintegratorsandallma-
jormainstreamsectorsfortheeconomy[9].Developmentofnewtechnologiesfor
provisioninginnovativeservicesandproductscanleadtonewbusinessopportuni-
tiesfortheindustry.MonitoringandControlareseenaskeyforachievingvisions
inseveralCPSdominatedareassuchasindustrialautomationsystems,automotive
electronics, telecommunication equipment, smart-grid, building controls, digitally
drivensmartcities,homeautomation,greenertransport,waterandwastewaterman-
agement,medicalandhealthinfrastructures,onlinepublicservices,andmanyothers
[10].Thischapterintroducesthecloud-basedindustrialCPSanddescribesthefirst
resultsofmakingitarealitytowardstheNextGenerationSOA-basedSCADA/DCS
systems.Thereaderistaughtabouttheresearch,developmentandinnovationwork
carriedoutbyasetofexpertscollaboratingundertheumbrellaoftheIMC-AESOP
project,forspecifying,developing,implementinganddemonstratingmajorfeatures
ofIntelligentMonitoringandControlSystemsandtheadvantagesofimplementing
themindifferentindustrialprocesscontrolenvironments.
ArmandoW.Colombo
Schneider Electric & University of Applied Sciences Emden/Leer, Germany e-mail: armando.
[email protected],e-mail:[email protected]
StamatisKarnouskos
SAP,Germanye-mail:[email protected]
ThomasBangemann
ifak,Germanye-mail:[email protected]
1
Colomboetal.
1.1.CurrentParadigmsandTechnologiesassociatedtoCPS
1.1 CurrentParadigmsandTechnologiesassociatedtoCPS
The advances in computation and communication resources have given rise to a
new generation of high-performance, low-power electronic components that have
advanced communication capabilities and processing power. This has led to new
possibilitiesthatenableimprovedintegrationofheterogeneousdevicesandsystems,
withparticularemphasisonplatformindependence,real-timerequirements,robust-
ness,security,stabilityofsolutions,amongothermajorrequirements.
Industrialists,researchersandpractitionersareassociatingtheseadvanceswitha
4thIndustrialRevolution(referredtoasIndustrie4.0inGermany[15])happening
today,wherephysical“Things”getconnectedtoInternet[12]allowingthatthereal
touchableworldintegratespartofthecyber-space. WiththesefoundationsinCPS
[1] and IoT, a number of different system concepts and architectures (e.g. www.
iot-a.eu) have become apparent in the broader context of Cyber-Physical Systems
[20,1,4]overthepastcoupleofyearssuchasCollaborativeSystems[11],Service-
Oriented Architectures (SOA) [3], networked cooperating embedded devices and
systems[22],cloudcomputing[2]etc.
Fig.1.1 CollaborativeManufacturingModel
Theumbrellaparadigmunderpinningnovelcollaborativesystemsistoconsider
the set of intelligent system units as a conglomerate of distributed, autonomous,
intelligent,pro-active,fault-tolerantandreusableunits,whichoperateasasetofco-
operatingentities[6].Theseentitiesarecapableofworkinginapro-activemanner,
initiatingcollaborativeactionsanddynamicallyinteractingwitheachotherinorder
toachievebothlocalandglobalobjectives,andthisalongthethreebasiccollabora-
2
Thisisapreprintversion,whichmaydeviatefromthefinalversionwhichcanbeacquiredfromhttps://www.springer.com/gp/book/9783319056234
Colomboetal.
1.1.CurrentParadigmsandTechnologiesassociatedtoCPS
tionaxes(asdepictedinFig.1.1)associatedtoanyapplicationdomainandrelated
infrastructure,i.e.enterprise,supply-chainandlifecycleaxes[11].
Fromthephysicaldevicecontrolleveluptothehigherlevelsofthebusinesspro-
cessmanagementsystem,asdefinedinISA-95(www.isa-95.com):fromsuppliers
through the enterprise to the customer [5], and from design through operation to
recyclingphasesofanengineeringsystemlifecycle,collaborationwillbeenabled
if, on one hand, the involved systems act and react on their environment, sharing
someprincipalcommonalitiesand,ontheotherhand,havesomedifferentaspects
thatcomplementeachothertoformacoherentgroupofobjectsthatcooperatewith
eachothertointeractwiththeirenvironment[22].
As we are moving towards Smart Cyber-Physical Systems and the “Internet of
Things”,millionsofdevices,notalltimesmart,areinterconnected,providingand
consuminginformationavailableonthenetworkandareabletoexchangecapabili-
tiescollaboratinginreachingcommongoals.Asthesedevicesneedstointeroperate,
bothatcyberandalsoatphysicallevel,theservice-orientedapproachseemstobe
apromisingsolution,i.e.eachdeviceshouldofferitsfunctionalityasstandardser-
vices,whileinparallelitispossibletodiscoverandinvokenewfunctionalityfrom
other services on-demand. These technologies can be leveraged to build advanced
functionalityintosmartCyber-PhysicalSystems,thusenablingbuildingad-hocnew
distributedapplicationparadigmsbasedoninterconnected“smartcomponents”with
ahighlevelofautonomy.
Thisevolutiontowardsglobalservice-basedinfrastructures[6]indicatesthatnew
functionality will be introduced by combining services in a cross-layer form, i.e.
servicesrelyingontheenterprisesystem,onthenetworkitselfandatdevicelevel
will be combined. New integration scenarios can be applied by orchestrating the
servicesinscenario-specificways.Inaddition,sophisticatedservicescanbecreated
atanylayer(evenatdevicelayer)takingintoaccountandbasedonlyontheprovided
functionalityofotherentitiesthatcanbeprovidedasaservice[18,7].Inparallel,
dynamicdiscoveryandpeer-to-peercommunicationwillallowtooptimallyexploit
thefunctionalityofagivendevice.Itisclearthatwemoveawayfromisolatedstand-
alonehardwareandsoftwaresolutionstowardsmorecooperativemodels.However,
inordertoachievethat,severalchallengesneedtobesufficientlytackled.
TheconvergenceofsolutionsandproductstowardstheSOAparadigmadopted
for smart Cyber-Physical Systems contributes to the improvement of the reactiv-
ity and performance of industrial processes, such as manufacturing, logistics, and
others. This is leading to a situation where information is being available in near
real-timebasedonasynchronousevents,andtobusiness-levelapplicationsthatare
abletousehigh-levelinformationforvariouspurposes,suchasdiagnostics,perfor-
manceindicators,traceability,etc.SOA-basedverticalintegrationwillalsohelpto
reducethecostandeffortrequiredtorealizeagivenbusinessscenarioasitwillnot
requireanytraditionalhigh-costsolutionssuchascustom-developeddevicedrivers
orthird-partyintegrationsolutions.
3
Thisisapreprintversion,whichmaydeviatefromthefinalversionwhichcanbeacquiredfromhttps://www.springer.com/gp/book/9783319056234
Colomboetal.
1.2.AService-orientedcross-layerAutomationandManagementInfrastructure
1.2 AService-orientedcross-layerAutomationandManagement
Infrastructure
AService-orientedcross-layerAutomationandManagementInfrastructureadopts
the “collaborative automation” paradigm, combining cloud computing and Web
services technologies [16], among others. The aim is to effectively develop tools
and methods, to achieve flexible, reconfigurable, scalable, interoperable network-
enabled collaboration between decentralized and distributed Cyber-Physical Sys-
tems.Afirststeptowardsthisinfrastructureistocreateaservice-orientedecosys-
tem.Thatis,networkedsystemsthatarecomposedbysmartembeddeddevicesthat
are Web service compliant (as depicted in Fig. 1.2 and Fig. 1.3), interacting with
bothphysicalandorganisationalenvironment,abletoexpose,consumeandsome-
timesprocess(compose,orchestrate)services,pursuingwell-definedsystemgoals.
Fig. 1.2 An industrial component virtualised by a Web service interface embedded into smart
automationdevice(adaptedfrom[23])
Taking the granularity of intelligence to the automation device level allows in-
telligent system behaviour to be obtained by composing configuration of devices
that introduce incremental fractions of the required intelligence. From a run-time
infrastructureviewpoint,theresultisanewbreedofveryflexiblereal-timeembed-
deddevices(wired/wireless)thatarefault-tolerant,reconfigurable,safeandsecure.
Amongothercharacteristicsofsuchsystems,automaticconfigurationmanagement
isanewchallengethatisaddressedthroughbasicplug-and-playandplug-and-run
mechanisms. The approach favours adaptability and rapid reconfigurability, as re-
programming of large monolithic systems is replaced by configuring loosely cou-
pledembeddedunits.
Theuseofdevice-levelService-OrientedArchitecturecontributestothecreation
of an open, flexible and agile environment, by extending the scope of the collab-
orative architecture approach through the application of a unique communication
infrastructure[26],downfromthelowestlevelsofthedevicehierarchyupintothe
manufacturingenterprise’shigher-levelbusinessprocessmanagementsystems[18].
Theresultofhavingasingleunifyingapplication-levelcommunicationtechnology
4
Thisisapreprintversion,whichmaydeviatefromthefinalversionwhichcanbeacquiredfromhttps://www.springer.com/gp/book/9783319056234
Colomboetal.
1.2.AService-orientedcross-layerAutomationandManagementInfrastructure
Fig.1.3 Anindustrialsystemviewedasadistributedsetofsmartservice-compliantdevicesand
systems
acrosstheenterprise,labelledas“servicebus(network)”inFig.1.4,transformsthe
traditional hierarchical view of the industrial environments into a flat automation,
controlandmanagementinfrastructure.Thatis,devicesandsystemslocatedindif-
ferentlevelsareallhavingthesameWebserviceinterfaceandareabletointeract.
Thisfunctionalinteractioniscompletelyindependentfromthephysicallocationin
thetraditionallyimplementedenterprisehierarchy.
Fromapurefunctionalperspective,oneofthemajorchallengesisfocussed,on
onesideonmanagingthevastlyincreasednumberofintelligentdevicesandsystems
populatingthecollaborativeSOA-basedsystemandmasteringtheassociatedcom-
plexity.Ontheotherside,followingemergingrequirementsofcontrol,automation,
managementandbusinessapplications,otherchallengesaretheengineering,devel-
opmentandimplementationoftherightinfrastructuretomakeusabletheexplosion
ofavailableinformationexposedasservicesinthe“servicecloud”originated,e.g.
ontheSOA-basedshopfloor[24].Industrialapplicationscannowberapidlycom-
posed/orchestrated,byselectingandcombiningthenewservicesandcapabilities
offeredasserviceinanautomationcloud,whichrepresentsthepartialortotalvir-
tualisationoftheautomationpyramid,asdepictedinFig.1.5andexplainedinthis
book’sChapter3.
5
Thisisapreprintversion,whichmaydeviatefromthefinalversionwhichcanbeacquiredfromhttps://www.springer.com/gp/book/9783319056234
Colomboetal.
1.2.AService-orientedcross-layerAutomationandManagementInfrastructure
Fig.1.4 Aservice-orientedviewofanindustrialsystem(adaptedfrom[23])
Fig.1.5 Buildingsupervisorycontrolandmanagementfunctionsasapplicationsusingservices
exposedbydevicesandsystemsinthephysicalworldandbytheIMC-AESOPcloudinthecyber
world
6
Thisisapreprintversion,whichmaydeviatefromthefinalversionwhichcanbeacquiredfromhttps://www.springer.com/gp/book/9783319056234
Colomboetal.
1.3.IntelligentService-orientedMonitoringandControl–TheIMC-AESOPApproach
1.3 IntelligentService-orientedMonitoringandControl–The
IMC-AESOPApproach
Theworldmarketfortechnologies,products,andapplicationsalonethatarerelated
towhattheInternetofThingsenables,i.e.monitoringandcontrol(M&C),willin-
creasesignificantlyinthenextyears.WorldM&Cmarketisexpectedtogrowreach-
ing500eBillionin2020.TheM&CEuropeanmarketfollowsthesametrendsas
the M&C world one in terms of product repartition and also market product evo-
lution.TheEuropeanmonitoringandcontrolmarketwillbereaching143eBnin
2020[25].Whenanalysingthemajorapplicationdomainsforreal-timemonitoring
andcontrol,fromthelargeprocessindustryviewpoint,theseindexesandtherelated
expectationsoutlinethetremendouspotentialandvalue.
Largeprocessindustrysystemsareacomplex(potentiallyverylarge)setof(fre-
quently) multi-disciplinary, connected, heterogeneous systems that function as a
complex distributed system whose overall properties are greater than the sum of
its parts, i.e. very large scale integrated devices (not all time smart) and systems
ofwhichthecomponentsarethemselvessystems.Multidisciplinaryinnature,they
link many component systems of a wide variety of scales, from individual groups
ofsensorstoe.g.wholecontrol,monitoring,supervisorycontrolsystems,perform-
ingSCADAandDCSfunctions.Theresultingcombinedsystemsareabletoaddress
problemswhichtheindividualcomponentsalonewouldbeunabletodoandtoyield
controlandautomationfunctionalitythatisonlypresentasaresultofthecreation
ofnew,“emergent”,informationsources,andresultsofcomposition,aggregationof
existingandemergentfeature-andmodel-basedmonitoringindexes.
TheseverylargescaledistributedprocessautomationsystemsthatIMC-AESOP
isaddressing,constitutesystemsofsystems[14],andarerequiredtomeetabasic
setofcriteriaknownasMaier’scriteria[21],i.e.:
1. Operationalindependenceoftheconstituentsystems
2. Managerialindependenceoftheconstituentsystems
3. Geographicaldistributionoftheconstituentsystems
4. Evolutionarydevelopment
5. Emergentbehaviour
Suchsystemsshouldbebasedonprocesscontrolalgorithms,architecturesandplat-
forms that are scalable and modular (plug & play) and applicable across several
sectors, going far beyond what current Supervisory Data Acquisition and Control
(SCADA),andDistributedControlSystems(DCS)anddevicescandelivertoday.
AfirstfastanalysisofcurrentimplementedSCADAandDCSsystemsdetectsa
setofmajorhindersfornotcompletelyfulfillingsomeofallthosecriteria:thebig
numberofincompatibilitiesamong thesystems,“hardcoded”data, differentview
onhowsystemsshouldbeconfiguredandused,co-existenceoftechnologiesfroma
verylongperiodsoftime(oftenmorethan20years),useofreactiveprocessautoma-
tioncomponentsandsystemsinsteadofhavingthemworkinginaproactivemanner.
Ifwebeganhookingallthesehinders,wewouldsoonhaveanunmanageablemess
ofwiring,andcustomsoftware,andlittleornooptimalcommunication.Todaythis
7
Thisisapreprintversion,whichmaydeviatefromthefinalversionwhichcanbeacquiredfromhttps://www.springer.com/gp/book/9783319056234
Colomboetal.
1.3.IntelligentService-orientedMonitoringandControl–TheIMC-AESOPApproach
hasbeentheusualresult,where“pointsolutions”havebeenimplementedwithout
anoverallplantointegratethesedevicesintoameaningful“InformationArchitec-
ture”.
Looking at latest reported R&D solutions for Control and Automation of very
largedistributedsystems,itispossibletoidentifytodaythattherearealreadymany
known possibilities for covering some and if possible many or all the criteria ad-
dressed above. The IMC-AESOP concept is pointing to optimisation at architec-
turalandfunctionallevelsofthelogicalandphysicalnetworkarchitecturesbehind
the process automation systems, mainly towards a potential optimal configuration
andoperation,e.g.ofenergyconsumption[17]incurrentcomplexandpowerhun-
gry process industries, based on service-oriented process control algorithms, scal-
ableandmodularSOA-basedSupervisoryDataAcquisitionandControl(SCADA)
and Distributed Control Systems (DCS) platforms, going far beyond what current
mainlycentralizedSCADAandDCScandelivertoday[16].
To address integration of very large numbers of subsystems and devices, the
IMC-AESOPapproachtakesitsrootsinpreviousworkinseveralresearchandde-
velopmentprojects[13,7,18],whichdemonstratedthatembeddingWebservicesat
thedevicelevelandintegratingthesedeviceswithMESandERPsystemsatupper
levelsofanenterprisearchitecturewasfeasiblenotonlyatconceptualbutalsoatin-
dustrialapplicationlevel.Thefirstresultsshowninpilotapplicationsrunninginthe
car manufacturing, electromechanical assembly and continuous process scenarios
havebeenverysuccessful,confirmingthattheuseofcross-layerService-Oriented
Architectures in the Industrial automation domain is a very promising approach,
abletobeextendedtothedomainofcontrolandmonitoringofbatchandcontinu-
ousprocesses.
Thisapplicationdomainoflargeprocesssystemscomposedofverylargenum-
bersofsystemsisverychallengingintermsof:
• Distributedmonitoringandcontrolofverylargescalesystems(tensofthousands
ofinterconnecteddevicesareencounteredinasingleplant)enablingplanteffi-
ciencycontrol,productandproductionqualitycontrol.
• A multitude of plant functions requesting information and functionality due to
continuouslychangingandincreasingbusinessrequirements.
• Integrationofexistingdeviceswhichgeneratesthedataandinformationneces-
saryforthemultitudeofplantfunctionalitieslikeplantoperation,maintenance,
engineering,businessandtechnology,i.e.systemofsystemsintegration,opera-
tionandevolution.
• Theverylargespreadindeviceandsystemperformancerequirementsregarding
e.g.responsetime,powerconsumption,communicationbandwidth,security.
• Legacy compatibility (20 years old systems have to interoperate with modern
ones).
WhenusingService-OrientedArchitecturesinProcessControlapplications,sev-
eraladvantagesareexpected.Foropenbatchand/orprocessautomationmonitoring
andcontrolsystemstheseinclude:
8
Thisisapreprintversion,whichmaydeviatefromthefinalversionwhichcanbeacquiredfromhttps://www.springer.com/gp/book/9783319056234
Colomboetal.
1.4.PositioningtheIMC-AESOPApproachwithintheIndustrialAutomationLandscape
• theabilitytobeaccessedbyanyothersystemoftheenterprisearchitectureable
tocallotherservices
• improved ease-of-use and simplified operation and maintenance of SOA-based
SCADAandDCSsystemembeddeddevicesduetotheuniversalintegrationca-
pabilitiesthattheservicesareoffering
• anextgenerationofSOA-basedprocessautomationcomponentsofferingplug-
and-playcapabilities,providingself-discoveryofalldevicesandservicesofthe
completeplant-widesystem.
Forproactivebatchand/orprocessautomationmonitoringandcontrolsystemsthese
include:
• theabilitytoexposetheirfunctionalitiesasservices
• the ability to compose, aggregate and/or orchestrate services exposed by them-
selvesandfromotherdevicesinordertogeneratenewdistributedSCADAand
DCSfunctions(alsoexposedas“services”attheshopfloor)
• attheshopfloorthesesystemsareinteroperablewithSOA-basedsystemsofthe
upper levels of the enterprise architecture (e.g. integrating ERP and MES with
theSCADAandDCS).
• anextgenerationofSOA-baseddevicesandsystemexposingSCADAandDCS
self-adaptable(emergent)functionalities(asaresultofautomaticservicecompo-
sitionororchestration),takingcareofreal-timechangesinthedynamicsystem
• thegenerationofnewmonitoringindexesandcontrolfunctionsatdifferentlevels
of the plant-wide system, as a result of using event propagation, aggregation /
orchestration/compositionofservicesandmanagementpropertiesoftheSOA-
baseddistributedSCADAandDCS.
Allofthesystemscanbenefitfromcost-effectiveness,thankstooptimizedSCADA
andDCSdistributionatthedevicelevelontheshopfloorandatupperITsystemlev-
els.Anadditionalbenefitstemsfromtheeasiernetworkmanagementoflarge-scale
networkedsystems.Basedontheseadvantagesaclearpossibilityistogeneratesys-
temenergyusageoptimisation.WiththeSOA-approachintegrationofsubsystems
havingtheappropriateinformation,itcanbedonebothattheoperatorlevelandat
thebusinesslevel,wheredifferentapproachestoenergyoptimisationcanbeapplied.
1.4 PositioningtheIMC-AESOPApproachwithintheIndustrial
AutomationLandscape
The degree of reliability and efficiency of energy consumption / utilisation in the
operationofindustrialenvironmentsdependsnotonlyontheoperationoftheindi-
vidualmechatronic/hardwarecomponentsbutalsoonthestructureandbehaviourof
theembeddedsupervisorycontrolsystem.Supervisorytaskshavetobeperformed
attwodifferentandseparatebutnetworkedlevels,i.e.theshopfloorandtheupper
levels of the enterprise architecture. At each of those levels is possible to identify
a set of functional and logical components that are responsible for performing the
9
Thisisapreprintversion,whichmaydeviatefromthefinalversionwhichcanbeacquiredfromhttps://www.springer.com/gp/book/9783319056234
Colomboetal.
1.4.PositioningtheIMC-AESOPApproachwithintheIndustrialAutomationLandscape
following functions: sensing, information collection, signal and information pro-
cessing,decision-makinganddiagnosis,anddiscrete-eventcontrol.
Each level (enumerated as 1–6 in Fig. 1.6), is having its own time-constraints
(frommicro-secondstodaysandweeks)anditsowndomainofdataandinforma-
tionprocessing.Monitoringofoperations,ofthebehaviourofthemechatronic/hard-
ware components, and ofthe system as a whole, is anessential function of such a
supervisory control system. Consider the definition of “monitoring” as the act of
identifying the characteristic changes in a process and in the behaviour of mecha-
tronic/hardwareresourcesbyevaluatingprocessandcomponentsignatureswithout
interruptingnormaloperations[8].
Fig.1.6 SchneiderElectricEnterpriseSystemArchitecture“TransparentReadyTM”
Intheplant,thereareasetofprocesscontrolstationsthatcontroldifferentpro-
cesssectionsintheplant,numberedas2and3inFig.1.6).Theyareconnectedto
various devices, distributed I/O stations, PLCs etc. that are themselves connected
totheprocessequipmentlabelledwiththenumber1.Formuchlargerandprocess-
specific equipment, the supplier also includes dedicated and unique devices, sys-
temsorcompletecontrol.Intheoverallplantmonitoringandcontrol,othersystems
andsectionsarealsointegratedlikelubricationsystems,transformers,switchgears,
valves,ventilation,heatingetc.Foroperators,engineers,maintenancepersonneland
management, there are one or several control and engineering rooms available, as
wellasmobiledevicesforlocalmonitoringandcontrol,depictedunderthenumber
4inFig.1.6).Attheenterpriselevel,thereareinformationaccess,controlandanal-
ysis through various management and enterprise information and control systems,
identifiedbythenumbers5and6inFig.1.6).
10
Thisisapreprintversion,whichmaydeviatefromthefinalversionwhichcanbeacquiredfromhttps://www.springer.com/gp/book/9783319056234