基于混合小波包的电能质量数据压缩算法

29320107AdvancedTechnologyofElectricalEngineeringandEnergyVol.29No.3July20102009-11-151981-1944-/、。310027MATLAB。。TM769A1003-3076201003-0008-04。。1。。。2。34。5。6CombinedWaveletPack-etsCWP。。。11.1/。2Nhngnwntn=013…n=0w0=tw1=ψtWP139w2nt=槡2Σ2N－1n=0hnwn2t－nw2n+1t=槡2Σ2N－1n=0gnwn2t－n11Fig.1Constructionofcombinedwaveletpackets1.27。Ψ1t…ψKtVjj∈zφ1t…Vjj∈zφKtk10φkt∈V0V1ψkt∈V1k=12…Kφ12t－nV1hkn∈Rgkn∈R1.1WP1wk2nt=Σ2N－1n=0hknwkn2t－nwk2n+1t=Σ2N－1n=0gknwkn2t－n2wk2ntWP1w2ntwk2n+1w2n+1t。56。1WP1k。db1db427W20W21。W2116W2129。1db4。1db1db4Tab.1ComparisonbetweenCWPandtraditionalwaveletWP1db1WP2db4CombineWP11.54770.13720.084720.74660.05010.0427db422W21db4db4W20。2Fig.2Coefficientsofcombinedwaveletpackets102922.19。M。M。7λx=－Σk|xk|2lg|xk|23MxMx=‖x‖2λx+lg‖x‖24Mx。7。MxMx。MxMx。2.2ψ1t…ψKt。11。。dbsymbiorrbiocdf1336w0w1w2w3w4w5w6w74811。50010025050。2.1MxaMXbc。3Fig.3Resultsofgeneticalgorithmforcombinedwaveletpacketsdeb。340。Mx。3121.CRCR=Sraw/Scmp5SrawScmp2.MSESNRdbMSE=1NΣN－1i=0Xi－Xci槡2×100%6SNR=10logΣN－1i=0Xi2ΣN－1i=0Xi－Xci()27XiXciN。IEC210。12.8k45db。3112Tab.2Disturbancesignalmodelsxt=A1－αut－t1－ut－t2sinωtt1t20.1≤α≤0.9T≤t2－t1≤8Txt=A1+αut－t1－ut－t2sinωt0.9≤α≤0.99T≤t2－t1≤8Txt=A1+αut－t1－ut－t2sinωt－0.5≤α≤－0.1or0.1≤α≤0.5t2－t1=0.01Txt=Asinωt+α3sin3ωt+α5sin5ωt0.1≤α3≤0.20.05≤α5≤0.14db4db4315%。db4db4。200。3200。。3Tab.3ExperimentalresultsofdifferentalgorithmsforPQeventdatacompressionsDb4Db4SNRdbMSE%SNRdbMSE%SNRdbMSE%51.05070.280256.72320.145860.57720.093661.15320.087661.47540.084461.50600.084156.02410.158057.55100.132660.66550.092646.39720.478853.04190.222860.66600.092651.07430.279451.00960.281559.48560.106146.87020.453446.11690.494547.35800.4286。8meanpp-mean/meanηη10p。。4CR。★db4▲db4。db4db4。54Fig.4RelationbetweenCRanddecompositionlevel/。21cont.onp.213FRP21smallerthanthatofsteeltower.AlternatingthedimensionofFPRtowerthemechanicalpropertiesandelectricpropertiesareanalyzedthereforetheminimumsectionwidthoftoweris37.6cmtheminimumpipethicknessis3.06cmanditsuggestedthatthenominalheightis11mthewidthofbottomcrossarmis4m.KeywordspowertransmissionlineFRPmechanicalstrengthelectricalstrengthsimulationstructureparame-櫆櫆櫆櫆櫆櫆櫆櫆櫆櫆櫆櫆櫆櫆櫆櫆櫆櫆櫆櫆櫆櫆櫆櫆櫆櫆櫆櫆櫆櫆櫆櫆櫆櫆櫆櫆櫆櫆櫆櫆櫆櫆櫆櫆櫆櫆ters11cont.fromp.11References1RobertsonDCCampsOI.WaveletsandelectromagneticpowersystemtransientsJ.IEEETrans.onPowerDe-livery19961121050-1058.2StéphaneGMallat.AwavelettourofsignalprocessingMNewYorkAcademicPress1999.3MariantoniaCotroneiLauraBMontefuscoLuigiaPuc-cio.MultiwaveletanalysisandsignalprocessingJ.IEEETrans.onCircuitsandSystems—IIAnaloganddigitalsignalprocessing1998458970-987.4LiuZhigangQianQingquan.Compres-sionoffaulttransientdatainelectricpowersystembasedonmultiwaveletJ.Proc.CSEE2003231022-26.5ChenYuyuZhangBo.Band-limitedcombinedorthogonalwaveletbasesAAustraliaIEEETENCON’1996C.NewYorkAcademicPress1996.593-597.6ChenYuyuZhangBo.CombinedwaveletpacketsandbestbaseJ.JSoftware199893161-168.7TheMathWorksInc.Matlab2007bHelpDocumentZ.2007.8LiuYingmeiBaiXiaominYiSuetal..CompressionofpowersystemdisturbancesignalsbasedonleastmaximummethodJ.PowerSys-temTech.200428333-37.9CoifmanRRWickerhauserMV.Entropybasedalgo-rithmsforbestbasisselectionJ.IEEETrans.onInfor-mationTheory19923827l3-718.10DaubechiesWSweldens.FactoringwavelettransformsintoliftingstepsJ.TheJournalofFourierAnalysisandApplications199643247-269.11PanZhengjunKangLishanChenYuping.EvolutionarycomputationM.BeijingTUP1998.12SalomonDavid.DatacompressionMPublishingHouseofElectronicsIndustry2003.13PascalGetreuer.FiltercoefficientstopopularwaveletsOL.http//scholar.google.cnMay2006.14TaoShunXiaoXiangning.Syntheticgradingeval-uationofpowerqualitybasedonincentivemechanismJ.Adv.Tech.ofElec.Eng.＆Energy200827216-20.ApowerqualityeventdatacompressionbasedoncombinedwaveletpacketsZHENGWei-yanWUWei-linCollegeofElectricalEngineeringZhejiangUniversityHangzhou310027ChinaAbstractAsthetime-frequencycharacteristicofsinglewaveletbasecannotmatchwithcomplexsignalfeaturethecompressioneffectcouldhardlybeimprovedfurtherinwaveletcompressionmethodsothispaperproposedcom-binedwaveletpacketsbasedonwaveletpacketsalgorithmforpowerqualitysignalcompression.Fordifferentkindsofpowerqualitydataweusinggeneticalgorithmsearchfortheoptimalcombinedwaveletpacketsbaseassemblagebyusingwaveletentropyfunctionascostfunction.Thesimulationresultsshowthatnewalgorithmgetsbetterresultswhencomparetothetraditionalwaveletandwaveletpacketmethodatthesameconditions.Keywordspowerqualitycombinedwaveletpacketsentropyfunctiongeneticalgorithm