山东大学博士学位论义
hasgoodnumericalstabilityandlowtimecomplexity,whichprovidesbasisfor‘‘control’’in“controlwhileidentification”thought.
SimulationresultsoftheChinaEPRI8-machine36-bussystemdemonstratethattheproposedon -linerecursiveclosed- loopidentificationalgorithmCaneffectivelyidentifyand
lowfrequencytrackreduced.orderstatespacemodelwhichcontainssystem
oscillationmodes,anddominantCanbeusedtoadjustparametersofsupplementarydampingcontrolleron-line.Additionally,supplementarydampingcontrollerCaneffectivelydampsystemintearealowfrequencyoscillations.
3)Witll
intotheregardstoshortcomingofoptimalcontrolwhichcannottakecontrolconstraintsadaptiveoptimization,wide-areadampingcontrolstrategybasedonmodelpredictivecontrol(MPC)isproposed,whichattempts
identification'’thought.Basedontosolving‘‘control’’in‘‘controlwhileidentifiedmodel,systemoutputdisturbancemodelis
topreventoutputsobtainedbyaugmentingsystemstateswithoutputdisturbances
steady.stateoffsets.Kalmanfilterisused
Accordingtotofromestimatestatestheaugmentedmodel.theidentified
anmodel,predictiveequationsforclosed loopsystemarefunctionrepresentingformulated.Thenobjectivethecostofdeviationofsystem
responsesfromthereferencetrajectoryandthecostofimposingdampingcontrolsin
oninfinitehorizonisestablished.Consideringconstraintscontrolinput,optimalcontrol
usesCanbeobtainedbysolvingthisoptimalcontrolproblemwhichcurrentstateofpower
aresystemastheinitialstate.Onlinemodelidentificationandcontroloptimizationrepeated
ineachtimeinterval.Theproposedmethodisanorganicintegrationofonlinemodelidentificationandonlineupdatingcontrol
powersystemlowfrequency
、析thfixed
controlparameters.Itareaccomplishesadaptivecontrolofoscillation.Thus,theofflineidentificationtovariationinherentshortcomingsofcontrollersparametersbasedonovercomeandtheproblemoftheperformance
aredegradationdueofthecomplexoperationconditionsandtime-varyinganduncertaincharacteristicofsystemofcontrolparametersaresolved.Negativeeffectsavoidedbyadoptingmodelprediction,recedingoptimizationandfeedback
.compensationstrategyinthismethod.
SimulationresultsoftheChinaEPRI8-machine36-bussystemdemonstratethattimeconsumptionforupdatingofsystem
lessthanstatespacemodelandmodelpredictivecontrollawissamplinginterval.Thus,modelpredictivedampingcontrollerssatisfyrequirementsforonlineapplication.It
oscillationmodes.ItalsoCaneffectivelydampinter-arealowfrequencytohavetheabilitycoordinatewithPSSsandothermodelⅥI