您好,欢迎访问三七文档
当前位置:首页 > 行业资料 > 国内外标准规范 > 一种结合DS证据信息融合的变压器故障诊断方法
DS,,,(400050):,。:;;;:TM406:B:1001-8425(2010)08-0059-05TransformerFaultDiagnosisMethodofCombiningDSEvidenceInformationFusionCHENXin-gang,TIANXiao-xiao,ZHANGChao-feng,ZHAOYang-yang(ChongqingUniversityofTechnology,Chongqing400050,China)Abstract:BasedonDGAoftransformeroil,thefaulttypesandtheirlocationinoperatingtransformerarediagnosed.Keywords:Transformer;Neuralnetwork;Informationfusion;Faultdiagnosis1,。、,。。[1];[2]DS,;[3]。,,。2,,。:。、、。DS、。,,。BP,。1BP。、,,。,jikY1Y2YLLwkiwijMXmX3X2X1q1BPFig.1TopologicalstructureofBPneuralnetworkTRANSFORMER47820108Vol.47AugustNo.8201047123123m1m2mn3Fig.3Faultdiagnosisbasedonevidentialtheory。,,,。BP2。BP,,。(1)。(2),。(3),,。(4),。,,,,,,。3DSDS,。DSU。UA,U。m:2U→[0,1]:(1)m(准)=0(2)A哿UΣm(A)=1mU,m(A)A。Bel(A)A,A。Bel:(1)Bel:2U→[0,1](2)Bel(A)=B哿AΣm(B),(坌A哿U)Ap1,:(1)p1(A)=1-Bel(A-)(2)p1(A)=A∩B≠0Σm(B)坌A哿Up1,[Bel(A),pl(A)]A,A。,。Bel1,Bel2…BelnUn,m1,m2…mn,:A1,A2,…Ak,B1,B2,…Br,C1,C2,…Ct,:K1=Ai∩Bji,j…p∩…Cp=ΦΣm(Ai)m(Bj)…m(Cp)(1)K1<1,A,B,…C:m(X)=Ai∩Bji,j…p∩…Cp=ΦΣm(Ai)m(Bj)…m(Cp)1-K1(2)K1=1,A,B,…C,。DS,,,,。3。“”“”,、、2BPFig.2FaultfusionmodelofBPnetwork60、、DS8-0.0008-0.00161.0062-0.0008-0.0001-0.0019-0.0017-0.00021.0027-0.0017-0.0010-0.00220.00030.00050.99660.00030.0010-0.0001。,,,,;,,D-S,;,“”,,。Dempster,Dempster,,Dempster,DS。,BPDS。4DS,BPDS,。4BPDS,4。BPDSDSDS。:BP,,(0,1),DS,。5。,BPDS,,。5SFSZ7-20000/110,1。5.1、、、、。matlab,A,A,,traincgf,10e-5,:TRAINCGF-srchcha,Epoch0/300,MSE1.23101/1e-005,Gradient137.309/1e-006TRAINCGF-srchcha,Epoch50/300,MSE0.0399393/1e-005,Gradient2.9845/1e-006TRAINCGF-srchcha,Epoch81/300,MSE3.55253e-006/1e-005,Gradient0.00203845/1e-006TRAINCGF,Performancegoalmet.A=BP6。6,10e-5,BPDS4DSFig.4FusionmodelofneuralnetworkandDSevidentialtheoryBPBPDSDS5Fig.5Fusionstructureoftrnasformerfaultdiagnosis1Table1ChromatographicdataoftransformerH2CH4C2H6C2H4C2H2COCO32.5122.7217.217.30417491.2357.29.089.3213.624.40298.4378.4385.740.9104.3253.926.10.41552.2344.3388.7μL/L6147。。022,,DS。5.2,、、、,F1、F2、F3、F4、F5F6。BP2。2BP,,(0,1),DS。:ijOi(j),j:mi(j)=Oi(j)qj=1ΣOi(j)(3)DS3。23F2、F3、F4、F6F1F5,F1,F5,。F1F5:M1(F1,F5)=0.1,M2(F1,F5)=0.2,M3(F1,F5)=0.1。(2)(3)M1M2,4。M1M2M3,5。5,DS,F10.0998,F50.9900,F1F50.10.0079,,F5,,,。6,BP,,DS,;,BP,,DS,,DSBP10-610110010-110-210-310-410-5010302040506070806Fig.6Networktraining2BPTable2OutputvaluesofBPneuralnetworkF1F2F3F4F5F60.17890.00090.1005-0.00020.69210.09780.32210.00360.10030.00770.6397-0.00330.2318-0.0023-0.00020.00150.72210.15113DSTable3BasicprobabilitydistributionofDSevidencesF1F2F3F4F5F6M10.16720.00080.09390.00020.64680.0914M20.30100.00340.09370.00720.59780.0031M30.21000.00210.00020.00140.65410.13694M1M2Table4DataafterfusionofM1andM2F1F5F1F5M10.16140.81630.02845M1、M2M3Table5DataafterfusionofM1,M2andM3F1F5F1F5M10.09980.99000.0079(70)62477,,,,。:[1],.[J].,2006,30:423-425.[2].DGA[J].,1999,36(1):30-34.[3],,.[J].,2005,42(9):43-46.[4],,,.[J].,2007,33(8):52-55.[5],,,.[J].,2008,28(7):141-147.[6],,,.[J].,2002,22(10):121-124.[7],,.BP[J].,2008,23(2):72-75.[8],.[J].,2005,29(18):73-75.[9]DL/T722-2000,[S].[10].[M].:,1990.[11],.[J].,2002,26(7):24-29.1601401201008060402001601401201008060402005Fig.5Historycurves:2009-08-11:(1982-),,,,;(1962-),,,,、;(1974-),,,,、、。,,,。:[1],.[J].,2006,26(3):119-123.[2],.PSO-BPD-S[J].,2006,30(7):46-50.[3],.[J].,2003,39(1):49-52.[4],.[J].,2002,22(7):115-118.[5],.[J].,2008,27(2):79-82.[6].matlab[M].:,2005.[7],.[J].,2006,(1):174-179.[8],.[M].:,2006.:2010-01-27:(1968-),,,,、、、;(1986-),,,,、。(62)!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!70
本文标题:一种结合DS证据信息融合的变压器故障诊断方法
链接地址:https://www.777doc.com/doc-11103030 .html