冰风暴灾害下输电线路故障概率预测

针对冰风暴灾害下输电线路运行故障问题,提出了基于极端学习机和Copula函数的断线倒塔概率预测模型。该模型运用广义极值分布刻画冰风暴灾害下冻雨量、风速、输电线和铁塔冰、风荷载的随机特性,并通过ELM网络预测出实时变化的GEV分布的形状参数、尺度参数和位置参数。随后考虑冰、风荷载之间的概率相关性,借助Clayton-Copula建立冰、风荷载的联合概率分布,从而实现输电线和铁塔的实时故障概率预测。结合湖南郴州电网的历史数据展开算例分析,验证了该预测方法的有效性和准确性。

Probabilistic Prediction of Transmission Line Fault Resulted from Disaster of Ice Storm

In allusion to transmission line fault resulted from the disaster of ice storm,based on extreme learning machine(ELM) and Copula function a probabilistic prediction model for transmission line breakage and tower topple over is proposed.By use of generalized extreme value(GEV) distribution,the proposed model depicts the probabilistic characteristics of excessive freezing rainfall,wind speed,wind load and ice load of transmission line and tower under the disaster of ice storm,and by means of FLM network the varying parameters of GEV distribution’s shape scale and location are predicted,and then considering probability correlation between wind load and ice load and by means of Clayton-Copula function the joint probability distribution of ice load and wind load is established,thus the realtime fault probability prediction of transmission line and towers is implemented.Based on historical data of Chenzhou power network in Hunan province,both effectiveness and accuracy of the proposed prediction method are verified by the results of calculation example.