微型电脑应用

快速准确地识别输电线路雷击故障有利于减少故障存续时间和经济损失，为此，提出一种基于快速傅里叶变换-格拉姆角场(FFT-GAF)和卷积神经网络(CNN)的输电线路雷击故障判别方法。将线路雷击过电压数据经过FFT,得到过电压频域特征数据；利用GAF处理频域特征数据，得到不同过电压类型的特征图像；提出结合迁移学习的CNN对特征图像进行分类，实现输电线路的雷击过电压识别；利用PSCAD软件获得雷击故障波形进行验证。实验结果可知，所提方法的识别准确率高达98.16%,优于其他模型。

Rapid and accurate identification of transmission line lightning faults is conducive to reducing the duration of faults and economic losses. Therefore, this paper proposes a lightning faults recognition method for transmission lines based on fast Fourier transform-Gramian angular field(FFT-GAF) and convolutional neural network(CNN). The lightning overvoltage data of the lines are subjected to FFT to obtain the overvoltage frequency domain characteristic data. The GAF is used to process the frequency domain feature datas, and the feature images of different overvoltage types are obtained. A CNN combined with transfer learning is proposed to classify the feature images to realize the lightning overvoltage recognition of transmission lines. The lightning fault waveform is obtained by PSCAD software for verification. The experimental results show that the recognition accuracy of the proposed method is as high as 98.16%, which is superior to other models.