基于核密度估计的谐波分布特征表述方法  

Harmonic Distribution Characteristic Representation Method Based on Kernel Density Estimation

作　　者：冯跃 钱永亮 唐明淑 鲁化喆 Feng Yue;Qian Yongliang;Tang Mingshu;Lu Huazhe(Wenshan Power Supply Bureau,Yunnan Power Grid Co.,Ltd.,Wenshan 663000,China;School of Electric Power Engineering,Kunming University of Science and Technology,Kunming 650500,China)

机构地区：[1]云南电网有限责任公司文山供电局,云南文山663000 [2]昆明理工大学电力工程学院,云南昆明650500

出　　处：《云南电力技术》2023年第6期28-30,共3页Yunnan Electric Power

摘　　要：当前对单次谐波电流数据的统计、评估主要依据95%概率值,将谐波统计数据归一化,并不能很好反应出谐波的分布特征。基于此提出了一种针对谐波数据统计分析的新方法,首先,以3 s或3 min为基本记录间隔,记录一个完整周期内的谐波监测数据,然后采用核密度方法刻画针对单次谐波的分布特征,最后建立针对谐波分析特征的概率区间分布。通过实测数据验证了该方法能够更好地反应谐波的概率分布。At present,the statistics and evaluation of single current data are mainly based on 95%probability values,and the normalization of harmonic statistics does not reflect the distribution characteristics of harmonics.Based on this,a new method for statistical analysis of harmonic data is proposed,firstly,the harmonic monitoring data in a complete cycle is recorded with 3 s or 3 min as the basic recording interval,and then the kernel density method is used to characterize the distribution characteristics of single harmonics,andfinally the probability interval distribution for harmonic analysis features is established.The measured data verify that the proposed method can better express the probability distribution of harmonics.

关 键 词：核密度估计 谐波分布 概率密度 

分 类 号：TM74[电气工程—电力系统及自动化]