融合注意力机制的输电线路故障概率预测模型

Prediction Model of Transmission Line Fault Probability Applying Attention Mechanism

输电线路运行环境恶劣,发生故障的概率和气象条件直接相关。由于目前极端气象条件在全世界范围内表现活跃,研究不同气象条件下输电线路发生故障的概率对提高电网运行稳定水平有重要意义。该文从分析气象相关输电线路典型故障的作用机理和统计特征出发,提出使用融合注意力机制的深度神经网络进行输电线路故障概率预测,使用停电数据对模型进行检验并与采用反向传播(back propagation,BP)算法的多层感知机模型进行对比。通过算例仿真验证了模型在输电线路故障概率预测方面的有效性,为电力事业部门更好地进行防护措施建设和修复计划制定提供了可能性,有利于电力系统的安全稳定可靠运行。

The fault probability of transmission lines is directly related to meteorological conditions.Due to the extreme weather conditions showed activity in the world scope,for the research of the fault of the transmission line under different meteorological conditions and probability of stable level has important significance to improve the operation.Starting from the analysis of typical weather related transmission line fault mechanism and statistical characteristics,this paper proposed a deep neural network based on attention mechanism to predict the fault probability of transmission lines.We use the power outage data to test the model and compares it with the multi-layer perceptron model using back propagation algorithm,which is referred as the BP network in the paper.The result validates the effectiveness of the model on transmission line fault probability prediction,which provides the possibility for the electric utilities to better carry out the construction of protective measures and the repair plan and shows that the method is conducive to the safe,stable and reliable operation of the power system.