摘要
针对分布式光伏接入后,异常线损率的判定阈值不合理导致漏检率较高的问题,提出考虑光伏波动性的10 kV配电网线损率波动阈值评估方法。首先,分析分布式光伏电量的并网形式以及可能存在的电量异常情况;其次,采用长短时记忆网络(LSTM)拟合输入特征与线损率之间的非线性关系预测线损率的方法,取代基于量测数据的潮流计算线损率的方法;然后,基于天气类型SCF指数,采用Monte Carol模拟体现光伏发电量的波动性,求得考虑光伏发电随机性的配电网线损率概率密度函数,并以此划定线损率的异常判定阈值。最后以IEEE 37节点配电系统为例,实际评估光伏接入后在考虑气象因素条件下配电线路日线损率的概率密度分布,并与电网的传统判定阈值进行对比分析。
An evaluation method for the fluctuation threshold of 10 kV distribution network line loss rate is proposed in this article,which takes into account the fluctuation of photovoltaics,to address the issue of high leakage rate caused by the unreasonable judgment threshold of abnormal line loss rate after distributed photovoltaic access.Firstly,the grid connection form of distributed photovoltaic electricity and possible abnormal electricity situation are analyzed.Secondly,the long short term memory(LSTM)network is used to fit the nonlinear relationship between input features and line loss rate to predict line loss rate,replacing the method of calculating line loss rate based on measured data in power flow.Then,based on the weather type SCF index,Monte Carlo simulation is used to reflect the volatility of photovoltaic power generation,to obtain the probability density function of distribution network line loss rate considering the randomness of photovoltaic power generation,and use this to determine the abnormal judgment threshold of line loss rate.Finally,taking IEEE 37 bus distribution system as an example,the probability density distribution of daily line loss rate of distribution lines afterphotovoltaic access is actually evaluated under the condition of considering meteorological factors,and is compared with the traditional determination threshold of the grid.
作者
周群
陈灿玉
青倚帆
佃钰林
冷敏瑞
刘雪山
Zhou Qun;Chen Canyu;Qing Yifan;Dian Yulin;Leng Minrui;Liu Xueshan(School of Electrical Engineering,Sichuan University,Chengdu 610044,China)
出处
《太阳能学报》
EI
CAS
CSCD
北大核心
2024年第10期354-362,共9页
Acta Energiae Solaris Sinica
关键词
分布式光伏
线损
配电网
长短时记忆网络
异常光伏计量点
SCF指数
distributed photovoltaic
line loss
distributed network
long short-term memory network
abnormal photovoltaic metering points
SCF index
