基于Pareto分布的磁暴感应地电场极值测度模型

磁暴感应地电场为电力系统安全运行带来风险，研究极端磁暴地电场水平能够为量化并防御电力系统磁暴灾害风险提供理论支持.提出了基于极值理论的磁暴感应地电场极值测度模型.首先利用地磁观测数据构造年度最大地电场样本数据，建立地电场广义Pareto分布模型，然后利用剩余寿命图和Q-Q图检验模型，并通过Bayes方法估测模型参数，最后计算模型的分位数，即地电场极值测度.利用中国地磁台观测的52年磁暴数据，结合中国喀什地区大地典型结构，估算出喀什地区五十年一遇的感应地电场可达3.288V·km-1，百年一遇地的感应地电场可达3.332V·km-1.利用相同方法还估测了中国部分地磁台站所在区域的地电场极值.研究结果可为估计极值地磁暴提供新思路，也可为分析电力系统磁暴灾害风险提供数据支撑. 

Extreme Value Model of Mid- and Low-latitude Geoelectric Field Due to Magnetic Storms

Estimation for the level of extreme geoelectric field can provide a basis to quantify the risk of the grounding technology systems due to magnetic storms. The extreme value model of geoelectric field is presented on the basis of the EVT. Firstly, the years maximum sample data of geoelectric field are constructed by using the geomagnetic data, and the general Pareto distribution model is established. Secondly, the model is examined by the mean residual life plot and the Q-Q plot, and the parameters are estimated by Bayes method. Finally, the return period i.e., geoelectric field extreme value, was calculated. The 52 years' data of geomagnetic storms and the Earth resistivity model of Shache, Xinjiang are used to estimate the geoelectric field. The geoelectric field of 50-year is 3.288V·km-1 and that of once-in-a-century is up to 3.332V·km-1 in mid- and low-latitude area. By contrasting with the existing literature analysis, the credibility of model is validated.