摘要
配电网重合闸失败导致用户短时间遭受连续多次电压暂降,给电压暂降持续时间评估带来挑战。此外,配电网故障率高,故障分布差异大,对电压暂降幅值评估影响显著。因此,本文提出一种基于过程免疫时间(PIT)和优化K近邻(KNN)估计的配电网电压暂降频次估计方法;提出了基于PIT曲线的连续电压暂降聚合方法,根据过程参数变化曲线分析工业过程在连续电压暂降下的后果状态,基于线性化PIT曲线求解连续电压暂降等效持续时间;提出了基于优化KNN算法的线路故障分布估计法,基于历史故障样本数量,采用交叉验证法自适应寻找最优KNN参数,估计线路故障分布概率密度函数,进而提出配电网电压暂降频次估计方法。最后,应用IEEE RBTS-6测试系统母线5配电网验证了所提方法。
The failure of reclosing in distribution network leads to continuous multiple voltage sags in a short time,which brings challenges to the evaluation of voltage sag duration.In addition,the distribution network has high fault rate and large difference in fault distribution,which affects significantly on voltage sag evaluation.This paper proposes a method of voltage sag frequency estimation based on process immunity time(PIT)and optimized K-nearest neighbor(KNN)estimation algorithm.An aggregation method for continuous voltage sags based on PIT curve is proposed.The consequence state of industrial process under continuous voltage sags is analyzed according to process parameter curve,and the equivalent duration of continuous voltage sag is calculated based on linearized PIT curve.An estimation method for line fault distribution based on optimized KNN algorithm is proposed.The cross-validation method is used to search the optimal parameter of the algorithm adaptively based on the number of historical fault samples,and the probability density function of line fault distribution is calculated.Then,the method of voltage sag frequency estimation is proposed.Finally,IEEE RBTS-6 test system is used to verify the proposed method.
作者
罗珊珊
陈兵
汪颖
陈韵竹
LUO Shan-shan;CHEN Bing;WANG Ying;CHEN Yun-zhu(Power Research Institute of State Grid Jiangsu Electric Power Co.,Ltd.,Nanjing 211103,China;College of Electrical Engineering,Sichuan University,Chengdu 610065,China)
出处
《电工电能新技术》
CSCD
北大核心
2022年第7期25-37,共13页
Advanced Technology of Electrical Engineering and Energy
基金
国家电网公司科技项目(202024211A)。
关键词
电压暂降频次
配电网
连续电压暂降
过程免疫时间
K近邻算法
交叉验证法
voltage sag frequency
distribution network
continuous voltage sag
process immunity time
K-nearest neighbor algorithm
cross validation method
