
AtuningstrategyformultivariablePIandPIDcontrollersusingdifferentialevolutioncombinedwithchaoticZaslavskiimap
LeandrodosSantosCoelho ,MarceloWicthoffPessôa
IndustrialandSystemsEngineeringGraduateProgram,PPGEPS,Ponti calCatholicUniversityofParaná,ImaculadaConceição,1155,Zipcode80215-901,Curitiba,Paraná,Brazil
articleinfoabstract
Atechniquefortuningofdecoupledproportional-integral(PI)andproportional-integral-derivative(PID)multivariablecontrollersbasedonachaoticdifferentialevolution(DE)approachispresentedinthispaper.Duetothesimpleconcept,easyimplementationandquickconvergence,nowadaysDEhasgainedmuchattentionandwideapplicationinsolvingcontinuousnon-linearoptimizationproblems.However,theperformanceofDEgreatlydependsonitscontrolparametersanditoftensuffersfrombeingtrappedinlocaloptimum.TheapplicationofchaoticsequencesbasedonchaoticZaslavskiimapinsteadofran-domsequencesinDEisapowerfulstrategytodiversifythepopulationandimprovetheDE’sperformanceinpreventingprematureconvergencetolocaloptima.TheoptimizedPDandPIDcontrollersshowsgoodclosed-loopresponsesincontrolofthebinaryWood–Berrydistillationcolumn,amultivariableprocesswithstronginteractionsbetweeninputandoutputpairs.SomecomparisonresultsofPDandPIDtuningusingchaoticDE,classicalDEandgeneticalgorithmarepresentedanddiscussed.
Ó2011ElsevierLtd.Allrightsreserved.
Keywords:
DifferentialevolutionChaoticsequencesPIDtuningOptimization
Evolutionaryalgorithms
1.Introduction
Theproportional-integral-derivative(PID)controllerhasre-mained,byfar,asthemostcommonlyusedcontrollerinpracticallyallindustrialcontrolapplications.Morethan90%ofallcontrolloopsarePID,withawiderangeofapplications:processcontrol,motordrives,magneticmemories,automotive, ightcontrol,amongothers(Åström&Hägglund,2001).Thereasonisthatithasasimplestructurewhichiseasytobeunderstoodbytheengi-neers.Inthepastdecades,differenttuningmethodologiesofPIandPIDcontrollershavebeenproposedintheliteraturesuchasauto-tuning,self-tuning,patternrecognition,andcomputationalintelli-gence(Åström&Hägglund,1995,2001;Cominos&Munro,2002).Unfortunately,ithasbeenquitedif culttotuneproperlythegainsofPIDcontrollersbecausemanyindustrialplantsareoftenburdenedwithproblemssuchashighorder,timedelays,poorlydamped,nonlinearities,andtime-varyingdynamics.
Overtheyears,severalauthorshaveproposedthetuningofPIDtocontrolmonovariableprocessesbyoptimizationmethods,such
lu,Erdog an,Hapog lu,&asgeneticalgorithms(Alt nten,Ketevanliog
Alpbaz,2008;Huang&Chen,1997;Hwang&Thompson,1993;Jan,Tseng,&Liu,2008;Krohling&Rey,2001;Li,Jan,&Shieh,2003;Takahashi,Peres,&Ferreira,1997;Wang&Kwok,1993;Zhang,Zhuang,Du,&Wang,2009),particleswarmoptimization(Chang,Correspondingauthor.Tel.:+554132711333;fax:+554132711345.
E-mailaddresses:leandro.coelho@pucpr.br(L.S.Coelho),celowp@gmail.com(M.W.Pessôa).

0957-4174/$-seefrontmatterÓ2011ElsevierLtd.Allrightsreserved.doi:10.1016/j.eswa.2011.04.156
2009;Kim,Maruta,&Sugie,2008),tribesalgorithm(Coelho&Bernert,2009a),harmonysearch(Coelho&Bernert,2009b),evolu-tionstrategy(Coelho&Coelho,1999),antcolony(Duan,Wang,&Yu,2006),amongothers(Bianchi,Mantz,&Christiansen,2008;Toscano,2005;Toscano&Lyonnet,2009;Xu,Li,Qi,&Cai,2005).Moreover,thereisliteratureaboutthetuningofmultivariablePIandPIDcontrollersusinggeneticalgorithms(seedetailsinZuo(2005),Vlachos,Williams,andGomm(2002),Chang(2007),Herre-ros,Baeyens,andPerán(2002)).
Recently,anewevolutionarycomputationtechnique,calleddif-ferentialevolution(DE)algorithm,hasbeenproposedinStornandPrice(2005,1997).TheDEhasthreemainadvantages:itcan ndnearoptimalsolutionregardlesstheinitialparametervalues,itsconvergenceisfastanditusesfewnumberofcontrolparameters.Inaddition,DEissimpleincoding,easytouseanditcanhandleintegeranddiscreteoptimization(Storn&Price,2005).DuetothegoodfeaturesofDEalgorithm,nowadaysithasbeenemergedasanewandattractiveoptimizerandappliedinvarietyofresearch elds.
ThispaperpresentsahybridoptimizationapproachgivenbycombinationofDE(Storn&Price,1997,2005)withchaoticZaslav-skiimap(Zaslavskii,1978)(DECZ)todeterminePIDcontrolgainsinamultiloopcontrolscheme.Themainbene tsofDECZovertra-ditionalDEapproachesareallowingthediversitymaintenanceandaidsinslowingprematureconvergenceandexploringthesearchspace.Alackofdiversityinapopulationcorrespondstosamplesolutionsbeingverysimilarwithrespecttothedistancemetric.Conversely,whensamplesarenotverysimilarthenthedegreeof
Multivariable control ... 暂无评价 15页 免费 ...CHOI AND LEE: ANTIWINDUP STRATEGY FOR PI-TYPE ..., PID Controllers: Theory, Design, and Tuning. ...
(GA) for tuning proportional–integral–derivative (PID) controllers to ...6. Conclusions A multivariable control strategy for simultaneously controlling ...
(GA) for tuning proportional–integral–derivative (PID) controllers to ...6. Conclusions A multivariable control strategy for simultaneously controlling ...
热门文档