lif技术与elm算法的电力变压器故障诊断研究 闫鹏程

TheOriginal-ELMfittingexcellence is0.99695,themeansquareerror MSEisР0.399%,andthecomparisoncanbefoundthat MSCisbetterthanSNV afterprocessing.The MSC-KPCA-ELM modelРperformsbest,thepredictionvalueisclosertotherealvalue.Themeansquarerooterroristhesmallest.TheresultsshowthatРtheMSC-KPCA-ELM model,combinedwithlaser-inducedfluorescencespectroscopytechnology,ismoresuitablefortherapidРdiagnosisofwhetherornotthepowertransformerhasfailed,whichtypeoffaultisaccuratelydetermined,andtheoperationРsafetyofpowerequipmentisguaranteed.РРKeywordsР Laserinducedfluorescencespectroscopy;Extremelearningmachine;Transformeroil;KPCA;PCAРР (ReceivedApr.21,2021;acceptedJun.8,2021) РР*Correspondingauthor