简介:BasedonthemethodofseparatingvariablesandthetechniqueofMuller’scalculatingroots,thepropagationcharacteristicsofhybridmodesinhomogeneousandinhomogeneouscircularFaradaychirowaveguidesareexaminedindetail.Numericalresultsarepresentedtodemonstratethecombinedeffectsofdifferentconstitutiveparametersonthemodebifurcation,symmetryandnonreciprocity,backward-waveandattenuationofhybridmodes,whicharedifferentfromthecasesofordinarygyrotropicandreciprocalchiralwaveguides.
简介:Anewalgorithmofnonuniformitycorrectionforinfraredfocalplanearray(IRFPA)isreported,whichisacombinedalgorithmbasedonboththetwo-pointcorrectionandartificialneuralnetworkscorrection.Thecombinedalgorithmiscalibratedbytwo-pointcorrection,andthecalibratedcorrectioncoefficientsareautomaticallymodifiedbyBPalgorithm.Soitisnotonlycalibrated,butalsoreal-timeprocessed.Inadaptivenonuniformitycorrectionalgorithm,thephenomenaghostartifactandtargetfade-outareavoidedbyedgeextraction.Inordertogetintensifiedimage,themodifiedmedianfiltersareadopted.Thesimulateddataindicatestheproposedschemeisaneffectivealgorithm.
简介:Anactiveandreactivecombinedspotpricemodelandthecorrespondingalgorithmareintroducedinthispaperbasedonthetheoryofoptimalpowerflow.Differentfromthetraditionaleconomicdispatchinthefieldofspotprice,theobjectivefunctionistominimizegenerationcostsbasedonpricingreactivepower.Thenconsideringthecharacteristicsofwindturbines,processingmethodsofwindfarmsinoptimalpowerflowisdiscussed.Finally,thefeasibilityofthemodelandthealgorithmisverifiedthroughthesimulationresultsofIEEE30system.
简介:Adefinitionofcombinedphasecenterforhornfeedsisgiven.FormulasofE-planeandH-planecombinedphasecenterforconicalhornsandthecorrespondingOptimalmodelarepresented,andafastoptimizationmethodforsolvingthismodelisdescribed.Byusingthismethod,thephasecenterofcorrugatedhornisdiscussedandcalculated,andthevariationofthephasecenterwithdistanceandoperatingfrequencyisgiven.
简介:Aclassofmodifiedparallelcombinedmethodsofreal-timenumericalsimulationarepresentedforastiffdynamicsystem.Bycombiningtheparallelismacrossthesystemwiththeparallelismacrossthemethod,andrelaxingthedependenceofstagevaluecomputationonsamplingtimeofinputfunction,aclassofmodifiedreal-timeparallelcombinedmethodsareconstructed.StiffandnonstiffsubsystemsaresolvedinparallelonaparallelcomputerbyaparallelRosenbrockmethodandaparallelRKmethod,respectively.Theirorderconditionsandconvergencesarediscussed.Thenumericalsimulationexperimentsshowthatthisclassofmodifiedalgorithmscangethighspeedandefficiency.
简介:Thispaperpresentstheanalysisofopenmicrostripstructuresbyusingdiakopticmethodoflines(ML)combinedwithperiodicboundaryconditions(PBC).Theparametersofmicrostrippatchareobtainedfrompatchcurrentexcitedbyplanewave.ImpedancematrixelementsarecomputedbyusingfastFouriertransform(FFT),andreducedequationissolvedbyusingdiakoptictechnique.Consequently,thecomputingtimeisreducedsignificantly.TheconvergencepropertyofsimulatingopenstructurebyusingPBCandthecomparisonofthecomputertimebetweenusingPBCandusualabsorbingboundarycondition(ABC)showthevalidityofthemethodproposedinthispaper.Finally,theresonantfrequencyofamicrostrippatchiscomputed.Thenumericalresultsobtainedareingoodagreementwiththosepublished.
简介:
简介:ThedeterminationofLuzhou-flavorliquoragesiscarriedoutbythree-dimensionalfluorescencespectroscopycombinedwithnon-negativematrixfactorization(NMF).37samplesofagedliquorswithweightedagesof15,20and25yearswerepreparedbyblendingthreeLuzhou-flavororiginalbaseliquorswithstorageagesof10,20and30yearsindifferentproportions.Thefluorescencespectraofthesamplesweremeasured,andthenfactorizedintobasismatrixandcoefficientsmatrixbymultiplicativeiterativeNMF.Thefluorescencespectra,reconstructedfromthebasismatrix,aresimilartotheoriginalspectra.Thecoefficientsmatrixistakenastheinputofsupportvectormachine(SVM)toestablishapredictionmodelforthedeterminationofliquorages.Comparedwiththeprincipalcomponentanalysis,thepredictionmodelbasedonSVMhasapredictedaccuracybetterthan91.7%.Thismethodcanprovidehelpsforthemarketsupervisionontheagedliquors.
简介:EstimationprecisionofDisplacedPhaseCenterAlgorithm(DPCA)isaffectedbythenumberofdisplacedphasecenterpairs,thebandwidthoftransmittingsignalandmanyotherfactors.DetailedanalysisismadeonDPCA'sestimationprecision.AnalysisresultsshowthatthedirectionalvectorestimationprecisionofDPCAislow,whichwillproduceaccumulatingerrorswhenphasecen-ters'trackisestimated.Becauseofthisreason,DPCAsuffersfromaccumulatingerrorsseriously.Toovercomethisproblem,amethodcombiningDPCAwithSubApertureImageCorrelation(SAIC)ispresented.Largesyntheticapertureisdividedintosub-apertures.Microerrorsinsub-apertureareestimatedbyDPCAandcompensatedtorawechodata.Bulkerrorsbetweensub-aperturesareesti-matedbySAICandcompensateddirectlytosub-apertureimages.Afterthat,sub-apertureimagesaredirectlyusedtogenerateultimateSASimage.Themethodisappliedtothelake-trialdatasetofa20kHzSASprototypesystem.ResultsshowthemethodcansuccessfullyremovetheaccumulatingerrorandproduceabetterSASimage.