第60卷第6期2023年6月15日电测与仪表ElectricalMeasurement&InstrumentationVol.60No.6Jun.15,2023基于CEEMDAN和HT的谐波检测新方法张乐乐，王海云，王维庆（新疆大学可再生能源发电与并网技术教育部工程研究中心，乌鲁木齐830047）摘要：经验模态分解（EMD）作为希尔伯特-黄变换（HHT）的重要组成部分，为了克服其在谐波检测中出现的模态混叠、端点效应问题，提出采用自适应噪声完备集合经验模态分解（CEEMDAN）和希尔伯特变换（HT）相结合的谐波检测新方法。文章首先在理论上对比分析了EMD、EEMD以及CEEMDAN算法，研究CEEMDAN算法的特性。再用CEEMDAN算法对原始信号进行分解，得到固有模态函数（IMF）。最后用HT算法对每阶IMF分量进行分析，检测到谐波中包含的瞬时幅频信息。算例仿真结果表明,相对于HHT算法对信号的处理能力,文中提出的方法在谐波检测中有效地克服了EMD算法的弊端，提高了信号分解精度。关键词：经验模态分解；希尔伯特-黄变换；自适应噪声完备集合经验模态分解；希尔伯特变换；谐波D0I:10.19753/j.issn1001-1390.2023.06.021中图分类号：TM935AnovelharmonicdetectionmethodbasedonCEEMDANandHT(EngineeringResearchCenterofMinistryofEducationforRenewableEnergyPowerGenerationandGridTechnology,Abstract:Empiricalmodedecomposition(EMD)isanimportantpartofHilbert-Huangtransform(HHT),inordertoo-vercomeitsmodalaliasingandendpointeffectinharmonicdetection,anovelharmonicdetectionmethodcombingadaptivenoisecompleteintegrationofempiricalmodedecomposition(CEEMDAN)andHilberttransform(HT)isproposedinthispaper.ThismethodfirstlycomparesandanalyzesEMD,EEMDandCEEMDANalgorithmintheory,andstudiesthechar-acteristicsofCEEMDANalgorithm,Then,CEEMDANalgorithmisadoptedtodecomposetheoriginalsignalstoobtaintheintrinsicmodalfunction(IMF).Finally,theHTalgorithmisusedtoanalyzetheIMFcomponentofeachorder,andtheinstantaneousamplitude-frequencyinformationcontainedintheharmonicisdetected.ThesimulationresultsshowthatcomparedwiththesignalprocessingcapabilityofHHTalgorithm,theproposedmethodovercomesthedisadvantagesofEMDalgorithminharmonicdetectionandimprovestheaccuracyofsignaldecomposition.Keywords:empiricalmodedecomposition,Hilbert-Huangtransform,completeensembleempiri...
