基于圖像特征的GIS局部放電嚴重程度評估的研究.pdf

1、華北電力大學(xué)(北京)碩士學(xué)位論文基于圖像特征的GIS局部放電嚴重程度評估的研究姓名：于樂申請學(xué)位級別：碩士專業(yè)：電氣工程；高電壓與絕緣技術(shù)指導(dǎo)教師：詹花茂201103華北電力大學(xué)碩十學(xué)位論文AbstractGasinsulatedsubstation(GIS)hasbeenwidelyappliedinthefieldofhi曲voltagepowertransmissionHoweverGISequipmentsarewitnessi

2、ngmoreandmoreinternalfaultsandfailuresasthevoltageclassofpowernetworkandcapacityofpowersystemkeepupgradingAgreatneediSthereforeinexistencetosearchforaneffectivewayinassessingtheinternalinsulatinconditionsofGISequipmentsS

3、ervingasthemostimportantmeasuremeansinthisregard，theseverityofpartialdischargeismainlyjudgedbyvisualmethodofcomparingspectraandcombiningwithexperienceNeverthless，thismethodiSsubjectivylackofscientificandaccuracy，and10Wef

4、ficiencyTherefore，amethodtoassesstheseverityofpartialdischargeinGISbasedonthefeaturesofspectraiSputforwardbyanalyzingtherelationshipbetweenfeaturesofspectraandseverityofpartialdischargewithdigitalimageprocessingtechnique

5、s，themainresearchworkandachievementsinthisthesisareasfollow：1Theimagepreproeessmethodsrelatedtospectraarediscussed，includingimagescale，crop，grayscale，binaryandedgedetectionTheseareconvenientfortheunifiedtreatmentofspectr

6、aandef!ficientextractionoffeatures2Thefeaturesofspectrainsuchaspectsascolortextureshapeandspaceareanalyzedconcretelycolormoments，graylevelgradientCOoccurrencematrixandPseudoZernikemomentsandSOonareextractedbyMATLABprogra

7、mming3UsingthemethodofIterativeSelfOrganizingDataAnalysisTechniquesAlgorithmclusteranalysistoachievethefurtheroptimalselectionoffeatures4Theimagecomparabilityisindicatedbycalculatingthecorrelationcoemcientofengenvectorwh

8、ichcanrealizeimageretrievalbythis5AsoftwareoftheassessmentoftheseverityofpartialdischargeinGISbasedonthefeaturesofspectraisdevelopedfinallyandtheaccuracyofthesystemisverifiedbyphysicalexperimentsofbadcontactofhigllvoltag