基于小波特征提取的可调节负荷辨识方法

RenewableEnergyResources

可再生能源

Vol.37No.6Jun.2019

基于小波特征提取的可调节负荷辨识方法

许春蕾1袁刘文颖1袁郭

渊1援袁袁袁曰袁袁曰虎1袁刘福潮2袁郑

102206曰2援遥曰FCMC尧伟2尧袁袁遥遥袁730050冤袁袁摘要院遥曰DBDB渊FCM冤关键词院中图分类号院TK81文献标志码院A文章编号院1671-5292渊2019冤06-0845-07

0前言

遥袁尧[1]袁袁袁-[2]袁[3]遥袁[4]遥袁袁袁遥[6]尧袁

遥袁遥遥

C尧尧袁尧渊FCM冤袁袁袁遥遥[5]袁袁袁遥袁DB袁袁袁[7]遥遥袁1典型可调节负荷的特征参量提取1.1可连续调节负荷的特征

袁袁袁袁遥

PL袁kt院PL袁kbasek袁袁遥袁t袁尧遥[10]袁[11][8]袁[9]PL袁kt=PL袁kbase+Iadj袁kt滋ktPL袁kadjk渊1冤收稿日期院2018-10-28遥

基金项目院渊2015BAA01B04冤曰渊52272216002D冤曰渊SGGSKY00JNJS1700149冤遥

通讯作者院渊1955-冤袁袁袁袁尧遥E-mail院liuwenyingls@sina.com

窑845窑

可再生能源

滋ktkt曰PL袁kadjk曰

,曰Iadj袁kt滋mtm曰PL袁madjm袁2019袁37渊6冤

曰

tkt袁袁Iadj袁kt=1袁袁Iadj袁kt=0遥1遥

mt袁袁Iadj袁kt=1袁Iadj袁mt=1袁袁Iadj袁mt=0遥PL袁.m1t袁PL袁m2t扇设设设设设设设设设设设缮设设设设设设设设设设设墒扇设设设设设设设设设设设缮设设设设设设设设设设设墒

曰Iadj袁mtPL袁.m1t=

P0+渊t-ton冤驻up袁t1约t臆t1+渊P+P0冤/驻upP袁

P0袁t臆t1P-渊t-tr冤驻down袁t2约t臆t2+渊P-P0冤/驻downP0袁P0袁P袁

t跃t2+渊P-P0冤/驻downt臆t1t1+渊P-P0冤/驻up约t臆t2渊3冤

048

12

t

162024

1.2可离散调节负荷的特征

袁袁遥m袁tPL袁mt院PL袁mbasePmaxMP0驻up

驻down

图1可连续调节负荷示意图

Fig.1Schematicdiagramofcontinuouslyadjustableload

PL袁.m2t=

P0-渊t-t1冤驻down袁t1约t臆t1+渊P0-P冤/驻downP+渊t-t2冤驻up袁t2约t臆t2+渊P0-P冤/驻upP0袁

t跃t2+渊P0-P冤/驻upt1+渊P0-P冤/驻up约t臆t2渊4冤

PL袁mt=PL袁mbase+Iadj袁mt滋mtPL袁madjm渊2冤院P0曰t2曰驻up曰驻down曰t1遥

2遥

P0M

Pmin驻down

驻up

渊a冤t1t2图2可离散调节负荷示意图

Fig.2Schematicdiagramofdiscretelyadjustableload

渊b冤t1t21.3可时移负荷的特征

袁袁遥h袁tPL袁ht院PL袁ht袁trPL袁ht=Iadj袁ht滋htPL袁ht袁trh曰滋htht渊5冤

袁1袁袁Iadj袁ht=0遥

0袁

PL袁ht=PL袁ht袁

0袁

扇设设设设设缮设设设设设墒

袁Iadj袁ht=

tL袁h+TL袁h约ttL袁h臆t约tL袁h垣TL袁htL袁h跃t

渊6冤袁t曰Iadj袁hth窑846窑

院PL袁hh遥

h曰tL袁h曰TL袁hh许春蕾袁等基于小波特征提取的可调节负荷辨识方法

3遥0

4院00

8院00

12院00t

16院0020院00

24院00

1.4Fig.3可调节负荷特征参量提取Schematic图diagram3可时移负荷示意图

oftransferableadjustableload

袁尧尧袁尧尧尧袁遥k1驻袁max袁袁扇设设N设设设nsgn[渊Pmax-P驻max=缮设设n设设移=1驻

设移N2rated冤/驻n-t]+1袁姿=1

设渊7冤设墒设设n=1驻nsgn[渊Prated-P2min冤/驻n-t]+1袁姿=0院N曰驻n曰姿=1袁

姿=0袁P曰P曰Pratedmaxmin尧遥kM袁2M=max渊VP冤渊8冤

院VP遥

k3tadj袁max袁袁袁tadj袁max=tret-ton=max渊tret袁n-ton袁n冤

渊9冤

院ton曰tret曰tret袁n尧袁ton袁n遥

k4渍遥渍袁袁袁袁曰渍袁袁袁遥96袁渍96渍=

蓘移i=1yi渊k冤讦k沂peakn蓡-蓘移96i=1yi渊k冤讦k沂valleyn蓡渊10冤院yi渊k冤讦k冤讦k沂peak曰yi渊k沂valley22.1基于小波特征提取的负荷辨识方法

遥

参量的提取基于DB小波变换的可调节负荷能量值特征2.1.1DBDaubechies遥袁袁袁袁[12]~[16]袁袁遥2.1.2淤遥袁渊xDBij冤k袁袁Yin=

移n袁YXkn遥

=k=1Xk-n滋

姨n滓i渊11冤

院滋i袁滓ixij尧遥

xFnn寅肄袁Yn抑N渊0袁1冤遥渊x冤=xijP{YijY袁Yn渊11冤臆x}袁=渊Yij冤遥

于P遥袁袁窑847窑

可再生能源

2019袁37渊6冤

袁k袁ui袁5袁ui=

院淄i遥姨移i淄

2渊12冤

淄沂uYY袁遥k袁1袁k2袁k3袁k4k52.2识方法

基于可调节负荷特征参量和遥

FCM的负荷辨DB袁FCM[17]袁袁袁袁2.2.1袁遥FCMFCM遥[0袁1]c袁袁遥FCM遥2.2.2FCM遥FCM淤袁遥

袁袁袁遥nPK杉山山山k11k煽衫K渊k山山ij冤=

山山kk11噎山山M21噎kk1p衫衫2p衫衫山山山山kn1M22O衫衫衫删

kn2噎Mk衫衫衫渊13冤

np衫衫闪

窑848窑

院袁袁kijij遥噎袁n冤袁Kki=渊ki1袁ki2d渊ki袁kj冤=

-k袁噎袁d渊kkipi袁冤kTj渊i=1袁2袁k=1渊k冤ikjk冤2袁蓘移

p蓡1/2渊14冤

袁Simcos渊ki袁kj冤=cos渊ki袁kj冤=kkikjikj渊15冤袁袁袁遥袁s袁r遥C=[CC1袁C2袁噎袁Cs]T袁遥

s=[k1袁k2袁噎袁kr]s于=[khgg袁1袁kg袁2袁噎袁kg袁r噎袁冤袁]袁渊g=1袁2袁噎袁n袁gH袁l祝=[l1s]袁H=[袁l2袁h1袁h2袁噎袁hn]T遥

hiLiC沂R1伊s袁L袁I沂Rn伊s袁LI=[L1袁L2袁噎袁Ln]T遥

Di=[di袁g袁1袁di袁g袁2i袁g袁s袁噎袁di袁g袁sd]遥

di袁g袁s=

蓘kg袁1k-袁2i袁1ki袁12+

kgk-kg袁i袁2ki袁22+噎+ks-i袁ki袁ss2蓡1/2Di袁渊16冤

Li袁k袁aL=[Li1袁Li2袁噎袁Lin]渊17冤

许春蕾袁等基于小波特征提取的可调节负荷辨识方法

di袁k袁a-1Li袁k袁a=渊a=1袁2袁噎袁s冤

di袁g袁1-1+di袁g袁2-1+噎+di袁g袁s-1盂袁院njss渊18冤

Li袁k袁Pi袁j遥L袁njs]袁Nj=[nj1袁nj2袁

Nj忆遥4Pi袁j=渊Li袁k冤-1Nj袁袁hi袁渊19冤Li袁k遥Y钥

NPi袁j袁NjNj忆遥

遥

FCM3仿真验证

5图4负荷辨识方法流程图

Fig.4Floatchartofloadidentificationmethod

遥遥

袁3.1特征参量的计算

图5甘肃电网系统结构图

Fig.5SystemstructurediagramofGansupowergrid

袁30遥2017袁6袁尧尧尧尧袁遥袁1遥

窑849窑

可再生能源

表1各类负荷特征参量的取值范围

Table1Valuerangeofvariousloadcharacteristicparameters

2019袁37渊6冤

驻M驻臆0.50.2约M臆0.8驻臆0.5M跃0.8驻臆0.5

0.2约M臆0.8驻臆0.5

M臆0.2

3.2负荷辨识仿真结果分析

淤袁C120袁杉山山煽山衫C=

山0.60820.13760.4292衫山衫山衫山衫山0.87950.99770.3258衫山衫山衫山0.00230.54550.4125-0.09830.67390.05470.34120.0230

-0.10210.27040.48120.07290.54610.8218

衫山衫山衫删

衫衫闪

4袁袁0.200L1=[0.2312袁0.49袁0.1034袁5]遥

201785袁N0.5391=[0.2151袁[0.2192袁0.0250袁0.2207]袁表3袁0.54926袁0.0225袁0.0250]袁2N2=L2P遥

1=Table2Various桃树村变试验负荷点各类型负荷构成比例typesoftestoftypicalpointTaoshucun

loadcompositionproportion

N1N0.2152

0.21913

0.5390.5492

60.0250

0.02550.2200.2257

0

2袁袁于遥

袁1[7]2袁k袁渊k12袁k3袁k4渊冤k1袁kk2袁3袁k4袁k5冤袁遥袁FCM袁61~6袁窑850窑

tadj袁max渍utadj袁maxt跃0.1adj袁maxt跃0.1渍渍跃0.1

臆00.61约0.46约uadj袁maxt跃0.1

0.32约u臆0.臆0.78adj袁max臆0.10约渍渍臆0

臆0.1

0.21约uu臆0.57

臆0.88

袁遥遥3遥

表Table33采用不同特征参量的负荷辨识方法的正确识别率methodThecorrectwithdifferentidentificationcharacteristicrateofloadparametersidentification

/%

123456

173.6281.2176.2379.8272.5977.2

86.31

.47

87.16

85.17

82.74

84.46

3袁袁.47%袁4结论

遥袁袁尧尧袁DB袁袁FCM遥袁袁遥袁袁袁

遥

参考文献院

[1]袁袁.[2]

[J].袁袁袁袁.袁2006袁30渊16冤院6-9.[J].袁[3]2016袁34渊11冤院1617-1625.66-75.袁袁袁.[J].袁2017袁41渊7冤院

许春蕾袁等基于小波特征提取的可调节负荷辨识方法

[4][5][6][7][8]

袁袁袁.1504-1508.

[J].袁2018袁36渊10冤院

[11][12][D].院袁.袁袁2016.袁.measurementapproach袁袁袁HeRM袁MaJ袁HillDJ.Compositeloadmodelingvia

..Systems袁2006袁21渊2冤院663-672.1495-1501.57.[J].[J].IEEETransonPower

86-90.[J].C院-袁2016渊16冤院

袁2016袁40渊5冤院[13]袁2013.袁[M].袁袁.袁袁[J].袁2010袁34渊7冤院52-

[14]PanigrahiBK袁PandiVR.Optimalfeatureselection

forclassificationofpowerqualitydisturbancesusingalgorithm[J].GenerationTransmission&Distribution[15][16][17]Iet袁2009袁3渊3冤院296-306.30.袁袁袁袁...waveletpacket-basedfuzzyk-nearestneighbour

2016袁34渊9冤院1356-1362.

[J].袁

clustering

BidokiSM袁Mahmoudi-KohanN.Evaluatingdifferent

techniques[A].Transmission

for

electricityand

Distributioncustomer

classification[9]

2010.1-5.

ConferenceandExposition[C].NewOrleans院IEEE袁

[J].袁2017袁30渊4冤院25-CclassificationofloaddynamiccharacteristicsbasedonliftingwaveletpackettransforminpowersystemloadEnergySystems袁2014袁62院353-363.

.WangZ袁BianS袁LeiM袁etal.Featureextractionand袁袁袁袁modeling[J].InternationalJournalofElectricalPower&[10]2017袁41渊9冤院90-95.[J].袁

[J].C袁2014袁32渊2冤院179-184.

Adjustableloadidentificationmethodbasedon

waveletfeatureextraction

渊1.StateKeyLaboratoryofNewEnergyPowerSystem袁NorthChinaElectricalPowerUniversity袁Beijing102206袁China曰2.StateGridGansuElectricPowerCompany袁Lanzhou730050袁China冤

XuChunlei1袁LiuWenying1袁GuoHu1袁LiuFuchao2袁ZhengWei2Abstract院Loadidentificationmethodcanidentifythedifferenttypesofadjustableloadsinpowersystemwithhighproportionofrenewableenergy.Thispapermakesanin-depthanalysisofvarioustypicaladjustableloadcharacteristics袁thecharacteristicparameterssuchasmaximumadjustablespeed袁maximumadjustabletimeanddepthofadjustmentareproposed.Inordertoimprovetheidentificationprecisionofadjustableload袁theDaubechiesfastwavelettransformisusedtoextractenergyvalueasanewfeatureparameter.Onthisbasis袁anadjustableloadidentificationmethodbasedonwaveletfeatureextractionisproposed袁andfuzzyCmeansclusteringmethod渊FCM冤isusedtoidentifytheload.Finally袁typicalintegratedloadpointsinHexiareaaresimulatedtoverifythattheproposedidentificationmethodcaneffectivelyimprovetheidentificationaccuracyofadjustableload.Hence袁itisofgreatpracticalsignificancetousetheadjustableloadcapacitytomakeuseofthecurtailedwindpower.

Keywords院adjustableloadidentification曰DBwavelettransform曰featureextraction曰waveletenergyvalue曰FCM

窑851窑