資源描述:
《基于改進的ARMA遞推算法的低頻振蕩模式在線辨識.pdf》由會員上傳分享,免費在線閱讀,更多相關內容在行業(yè)資料-天天文庫。
1、第40卷第l期電力系統(tǒng)保護與控制Vo1.40NO.12012年1月1日PowerSystemProtectionandControlJan.I,2012基于改進fl~JARMA遞推算法的低頻振蕩模式在線辨識陳剛2,龔嘯。,李軍,古志明(1.南方電網(wǎng)科學研究院,廣東廣州510080;2.輸配電裝備及系統(tǒng)安全與新技術國家重點實驗室(重慶大學)重慶400044;3.重慶市電力公司,重慶400015)摘要:為提高電力系統(tǒng)低頻振蕩現(xiàn)象的實時監(jiān)測水平,提出采用一種基于自回歸滑動平均模型的兩段加權遞推最小二乘算法進行低頻振蕩模式辨識,并通過
2、估計ARMA譜的方法以提取低頻振蕩的主導模式。該改進算法先采用加權遞推最小二乘算法擬合高階AR模型單獨得到白噪聲估值,再將該估值用于常規(guī)加權遞推最小二乘算法中,提高了算法參數(shù)辨識的精度和收斂速度。New—England39節(jié)點系統(tǒng)的時域仿真測試驗證了該改進算法對低頻振蕩模式辨識的有效性,并通過與常規(guī)加權遞推最小二乘算法辨識效果的比較驗證了該改進算法對低頻振蕩模式的辨識具有更好的精確性且提高了收斂速度。最后通過對某電網(wǎng)PMU實測數(shù)據(jù)的辨識分析,驗證了該改進算法能夠準確地辨識系統(tǒng)的低頻振蕩主導模式頻率和阻尼比,具有實際的2;-程意
3、義。關鍵詞:自回歸滑動平均模型;加權遞推最小二乘算法;ARMA譜;低頻振蕩在線辨識;主導模式ImprovedARMArecursivealgorithmforonlineidentificationoflowfrequencyoscillationmodesCHENGang’,GONGXiao,LIJun3GUZhi—ming(1.ChinaSouthernPowerGridScienceResearchInstitute,Guangzhou510080,China;2.StateKeyLaboratoryofPowerTra
4、nsmissionEquipment&SystemSecurityandNewTechnology,ChongqingUniversity,Chongqing400044,China;3.ChongqingMunicipalElectricPowerCompany,Chongqing400015,China)Abstract:Inordertoimprovethelevelofreal—timemonitoringoflowfrequencyoscillation,thispaperproposestouseatwoparts
5、weightedrecursiveleastsquare(wm~s)algorithmbasedonauto—regressivemoving·average(ARMA)modeltoestimatethelowfrequencyoscillationmodes,andextractsthedomainmodesoflowfrequencyoscillationbythemethodofARMAspectrumestimation.TheimprovedalgorithmUSeStheobtainedwhitenoiseest
6、imatesbyfittingthehigherorderautoregressive(AR)modelbytheWRLSmethodintheconventionalWRLSmethod,andthenithasthepreferableaccuracyandthefastconvergencerateofparameteridentification.ThevalidityofproposedalgorithmisdemonstratedwithsimulationdatafromNew—England39一bussyst
7、em.Comparisonwithproposedalgorithmandconventionalwei【ghtedrecursiveleastsquarealgorithmshowstheadvantagesofthealgorithminthispaper.Finally,identificationanalysisofpracticalsignalmeasuredofPMUinsomegriddemonstratesthattheproposedalgorithmcanaccuratelyestimatethefrequ
8、encyanddampingratioofpowersystemlowfrequencyoscillationdomainmodes,SOtheproposedalgorithmhasthepracticalprojectsignificance.Keywords:ARMAm