1?GOLDWINDSCIENCE&TECHNOLOGYCO.,LTD.測(cè)風(fēng)數(shù)據(jù)長(zhǎng)期訂正對(duì)風(fēng)資源評(píng)估的影響Theimpactoflong-termcorrectionofwindmeasurementdataonwindresourceassessment北京金風(fēng)科創(chuàng)風(fēng)電設(shè)備有限公司BEIJINGGOLDWINDSCIENCE&CREATIONWINDPOWEREQUIPMENTCO.,LTD風(fēng)資源技術(shù)部WindResourceAssessmentDepartment高級(jí)風(fēng)資源工程師：楊長(zhǎng)鋒SeniorWindResourceEngineer：ChangfengYang長(zhǎng)期訂正的作用及意義Thefunction

andsignificanceoflong-termrevision01算法匯總Algorithmsummary02結(jié)果驗(yàn)證Results03結(jié)論及相關(guān)進(jìn)展Conclusion&Relatedprogress04CONTENTS目錄3長(zhǎng)期訂正的作用及意義Thefunctionandsignificanceoflong-termrevision數(shù)據(jù)清理參數(shù)計(jì)算風(fēng)速風(fēng)頻分布極端風(fēng)速入流角湍流切變水平外推垂直外推降低LCOE計(jì)算的不確定度，為項(xiàng)目評(píng)估提供支持提高載荷輸入?yún)?shù)準(zhǔn)確性算法匯總AlgorithmsummaryAlgorithmTypeAbbrev.DescriptionLinearLeastSquaresLLSTheclassicleastsquaresfittothescatterplotoftargetandreferencespeedsTotalLeastSquaresTLSAslightmodificationofLLSthatminimizesorthogonaldistancetothebestfitVerticalSliceVSApiecewiselinearfittotthescatterplotoftargetandreferencespeedsVarianceRatioVRAsimplelinearmappingthataimstopreservethevarianceofthetargetdataBulkSpeedRatioBSRThesimplestpossiblealgorithm,basedontheratioofmeanwindspeedsWeibullFitWBLApowerlawfitwhoseparametersderivefromtheWeibullparametersofthetargetandreferencedataSpeedSortSSAlinearfittotherelationshipbetweentargetandreferencecumulativefrequencydistributionsMatrixTimeSeriesMTSAnimplementationoftheclassicmatrixmethodthatwemodifiedtoproducerealistictimeseriesdata整體最小二乘TotalLeastSquaresAlgorithm線(xiàn)性最小二乘LinearLeastSquaresAlgorithm分風(fēng)速段線(xiàn)性回歸VerticalSliceAlgorithm方差比值法VarianceRatioAlgorithm風(fēng)頻擬合法WeibullFitAlgorithm速度追蹤法SpeedSortAlgorithm風(fēng)速序列聯(lián)合概率分布MartixTimeSeriesAlgorithmCASE1CASE2CASE3CASE4Duration6years6years3years3yearsR^20.6440.5750.6210.598RecoveryRate99.29%99.93%99.69%99.54%Timestepsofcomparison60min60min60min60mintimes212199評(píng)價(jià)指標(biāo)WSKWPDAEPSD√？☆★RMSE√？☆★計(jì)算結(jié)果及驗(yàn)證思路Ideasforverifyingcalculationresults1month3month6month9month12month15monthBSR

0.2124

0.1319

0.0672

0.0442

0.0319

0.0292MTS

0.2067

0.1362

0.1105

0.0686

0.0243

0.0404SS

0.1965

0.1214

0.0590

0.0461

0.0330

0.0323TLS

0.2047

0.1239

0.0602

0.0496

0.0350

0.0332VR

0.2128

0.1311

0.0685

0.0502

0.0347

0.0340WBL

0.2265

0.1386

0.0716

0.0493

0.0342

0.0349LLS

0.1833

0.1073

0.0701

0.0463

0.0377

0.0317時(shí)間長(zhǎng)度對(duì)計(jì)算結(jié)果的影響Theeffectoflengthofdataonthecalculationresults結(jié)果驗(yàn)證-風(fēng)速Result–WindSpeedMast1Mast2Mast3Mast4BSR0.05720.06050.03250.0406LLS0.04850.05610.02580.0361MTS0.05470.06510.03320.0471SS0.05460.06950.03340.0424TLS0.05960.06790.02910.0426VR0.05870.06960.02870.0430VS0.04450.05150.02500.0367WBL0.06110.06710.02750.0429New0.04590.05250.02390.0327結(jié)果驗(yàn)證-K值Result–KvalueMast1Mast2Mast3Mast4BSR0.06900.21550.08990.0783LLS0.53200.58290.45980.4321MTS0.22140.20340.20350.1821SS0.11250.08470.11370.0681TLS0.12130.10240.11200.0980VR0.11350.07710.10590.0817VS0.62020.60130.49020.4373WBL0.10580.07830.10150.0741New0.06100.05790.06780.0383結(jié)果驗(yàn)證-風(fēng)功率密度Result–WindPowerDensityMast1Mast2Mast3Mast4BSR7.934317.397314.76118.3066LLS38.670835.319033.557631.0287MTS18.912116.255417.735715.2753SS8.50776.90768.38656.9202TLS8.48539.23768.98768.4063VR8.09767.36147.76037.1570VS42.350635.116835.485530.6848WBL8.15046.21838.55056.7247New7.56876.12577.04754.2817結(jié)果驗(yàn)證-發(fā)電量Result–AnnualEnergyProductionMast1Mast2Mast3Mast4BSR152467.2145961.380031.43154478.9LLS105321.1372920.395588.4894103.6MTS126250.5245128.956274.3136752.6SS139641107285.556274.3136752.6TLS140613.8100627.778606.19143524.6VR139840.9108498.678719.25150406.5VS214058492175.366243.53188772.1WBL150526111711.556960.36150968.8New92524.47107033.546819.97104563.3季節(jié)因素對(duì)長(zhǎng)期訂正的影響十分重要，隨測(cè)風(fēng)時(shí)間增加，風(fēng)速訂正不確定度會(huì)明顯降低。Theinfluenceofseasonalfactorsonlong-termcorrectionis

important.Asthewindmeasurementtimeincreases,thewindspeedcorrectionuncertaintywillbesignificantlyreduced.在進(jìn)行長(zhǎng)期訂正時(shí)，不僅要注意風(fēng)速的變化值，同時(shí)要注意K值的變化幅度，避免造成風(fēng)功率密度產(chǎn)生較大差異。Inthelong-termcorrection,toavoidalargedifferenceinthecalculationofthewindpowerdensity.PayattentiontothevariationoftheKvaluewhilepayingattentiontothechangevalueofthewindspeed.結(jié)論Conclusion線(xiàn)性回歸方法在風(fēng)速訂正上與其他方法差異不大，但會(huì)較大幅度改變K值，對(duì)發(fā)電量計(jì)算產(chǎn)生較大影響。尤其是實(shí)測(cè)風(fēng)速K值較小時(shí)，誤差會(huì)更加明顯。Thelinearregressionmethodhaslittledifferencewithothermethodsinwindspeedcorrection,butitwillchangetheKvaluegreatly,whichhasagreatimpactonthecalculationofpowergeneration.EspeciallywhenthemeasuredwindspeedKvalueissmall,theerrorwillbemoreobvious.優(yōu)化后的算法與其他MCP算法相比，在風(fēng)速、K值和發(fā)電量計(jì)算上誤差較小且表現(xiàn)較為穩(wěn)定。ComparedwithotherMCPalgorithms,theoptimizedalgorithmhaslesserrorinwindspeed,Kvalueandpowergenerationcalculati