为充分挖掘微气象与线路跳闸的关联规律，文中提出一种输电线路微气象跳闸预警及成因量化分析方法。以微气象历史数据为基础，在跳闸前基于轻梯度提升机（LightGBM）构建微气象与输电线路跳闸概率预测模型，从线路层面为LightGBM损失函数中的失稳样本赋予更高的权重以应对数据集不平衡的问题，并添加线路抵御能力差异惩罚项以削减对薄弱线路的学习|在跳闸后基于Sigmoid云模型量化微气象对跳闸事件的置信度，并从区段层面构造微气象耦合系数实现跳闸成因的量化分析。文中方法可在给定微气象条件下对跳闸事件进行实时预警，并进一步量化各个微气象对引发跳闸事件的贡献程度，为检修维护工作提供决策支持。

In order to fully explore the correlation between microclimate and transmission line tripping, this paper proposed a method of early warning for microclimate-based tripping and cause quantitative analysis for transmission line. On the basis of meteorological history data, the probability prediction model of microclimate and transmission line tripping was constructed before the tripping based on Light Gradient Boosting Machine (LightGBM). From the circuit level, higher weights were given to the instable samples in LightGBM loss function to cope with the unba-lanced data sets, and the line resistance difference punishment item was added to cut thin lines for learning. After the tripping, the confidence degree of the micrometeorology to the tripping event was quantified based on the Sigmoid cloud model, and the micrometeorology coupling coefficient was constructed at the sectional level to achieve the quantitative analysis of the cause of the tripping. The proposed method can provide real-time early warning for tripping events under given micrometeorology conditions and further quantify the contribution of each micrometeoro-logy to trigger tripping events, thus providing decision support for maintenance work.