广义Pareto分布在南方电网大停电事故分析中的应用

Application of Generalized Pareto Distribution in Analyses of Blackouts in China Southern Power Grid

为研究南方电网停电规模的数理分布特征,通过历史数据拟合,发现我国南方电网大停电事故损失负荷序列具有“厚尾”特性.基于广义Pareto分布模型分析南方电网停电事故损失负荷序列,进而实现对电网停电损失风险的评估和预测.通过综合运用经验平均超出函数图和Hill图,实现了对门限值u的精准定位.利用BMM模型和GPD模型下的VaR指标对风险进行度量,评估结果结合真实数据比较,验证了此方法的可行性与精确性.仿真结果可为电网调度人员评估电网风险、划分事故等级提供参考.

In order to study the mathematical distribution characteristics of the power failure scale of the southern power grid,it is found that the loss load sequence of the blackouts of the southern power grid in China has the characteristic of“fat tail”through the fitting of historical data.Based on the generalized Pareto distribution model,the loss load sequence of power failure in southern power grid is analyzed,and then the risk of power failure is evaluated and predicted.The de-termination of threshold value u is realized through the comprehensive application of empirical mean excess function graph and hill graph.Risks are measured by using VaR.indicators under the BMM model and GPD model,and the re sults are compared with real data to verify the feasibility and accuracy of the method in this paper.The simulation re-sults can provide a reference for grid dispatchers to assess grid risks and classify accident levels.