基于超阈值模型的中国电网停电事故极值风险分析  

Extreme value risk analysis of blackout in China power grid based on over threshold model

作　　者：于群 肖松青 曹娜 何剑 张建新 YU Qun;XIAO Songqing;CAO Na;HE Jian;ZHANG Jianxin(College of Electrical Engineering and Automation,Shandong University of Science and Technology,Qingdao 266590,Shandong,China;China Electric Power Research Institute Co.,Ltd.,Beijing 100192,China;State Grid Shanghai Electric Power Company,Shanghai 200122,China)

机构地区：[1]山东科技大学电气与自动化工程学院,山东青岛266590 [2]中国电力科学研究院有限公司,北京100192 [3]国网上海市电力公司,上海200122

出　　处：《山东大学学报（工学版）》2022年第6期157-166,175,共11页Journal of Shandong University（Engineering Science）

基　　金：国家电网有限公司总部科技项目(5100-202155482A-0-5-ZN)。

摘　　要：在已有的极值模型的基础上,利用超阈值(peaks over threshold,POT)提出基于相对值法的POT-CVaR_(Bootstrap)模型,并对中国电网停电损失风险进行评估。结合电网装机容量的发展趋势,对比利用绝对值法与相对值法统计的停电损失负荷数据,表明相对值预测更符合未来电网损失;基于相对值数据建立POT模型,提出W^(2)&A^(2)-Hill图的阈值选取方法。为了解决电网停电损失负荷样本数量稀少的问题,结合对有限样本重复抽样的Bootstrap法,建立CVaR_(Bootstrap)停电损失风险评估模型,利用中国电网停电损失负荷数据验证了模型的有效性,得出CVaR_(Bootstrap)模型的评估结果更优,在此基础上给出我国未来10 a的停电事故风险评估,未来很长一段时间我国的停电损失负荷将仍呈现增长趋势,且华北、华东和南方地区是停电事故发生较严重的地区。On the basis of the existing extreme value model,a POT-CVaR_(Bootstrap) model based on the relative value method was proposed by using peaks over threshold(POT),and the outage loss risk of China power grid was evaluated.Combined with the development trend of the installed capacity of the power grid,the comparison of the outage loss load data calculated by the absolute value method and the relative value method showed that the relative value prediction was more in line with the future power grid loss.Based on the relative value data,the threshold model was established,and the threshold selection method of W^(2)& A^(2) Hill graph was proposed.In order to solve the problem that the number of power grid outage loss load samples was scarce.A CVaR_(Bootstrap) outage loss risk assessment model was established by combining the bootstrap method with repeated sampling of limited samples.The effectiveness of the model was verified by the data of power grid outage loss load in China,and the evaluation result of the CVaR_(Bootstrap) model was better.On this basis,the risk assessment of power outage accidents in China in the next ten years was given.For a long time in the future,China′s power outage loss load would still show an increasing trend,and North China,East China and South China were the areas where power outage accidents were more serious.

关 键 词：POT模型 阈值 相对值法 大停电事故 CVaR_(Bootstrap)风险评估 

分 类 号：TM714[电气工程—电力系统及自动化]