【摘要】 不同覆冰类型对架空输电线路的危害程度各异，准确预测覆冰类型有助于了解未来覆冰风险，保障电网安全运行。为此提出一种基于多元线性回归和随机森林算法的架空输电线路未来覆冰类型预测方法。融合终端微气象数据与天气预报格点数据，建立多元线性回归方程实现了微气象预测。融合覆冰图像拍摄时间前72 h间隔6 h微气象数据与覆冰类型数据，建立了随机森林覆冰类型分类方法，模型的精确率P、召回率R、宏平均F₁分数为94%、95%和86%。两种模型结合可实现输电线路未来覆冰类型预测。结果表明，所提方法可为电网防冰除冰提供参考，为电网覆冰期安全运行提供理论支持。更多还原

【Abstract】 Different types of icing pose varying degrees of damage to overhead transmission lines. Accurately predicting the type of icing can help understand future icing risks and ensure the safe operation of the power grid. A prediction method for future icing types on overhead transmission lines based on multiple linear regression and random forest algorithm was proposed. By integrating terminal micro meteorological data with weather forecast grid data, a multiple linear regression equation was established to achieve micro meteorological prediction. A classification method for random forest ice covering types was built by integrating micro meteorological data and ice covering type data 72 h before the shooting time of ice covering images at intervals of 6 h. The accuracy P, recall R, and macro average F₁ scores of the prediction models were 94%, 95%, and 86%, respectively. The combination of two models can achieve the prediction of future icing types on transmission lines. The results indicate that the proposed method can provide reference for anti-icing and de-icing of power grids, and provide theoretical support for the safe operation of power grids during ice covering periods.更多还原