Research on Fault Identification of the Transmission Line Based on Improved Convolutional Neural Network

In view of the high false alarm rate of the power grid transmission line warning system and the reliance on the post-analysis of operation and maintenance personnel, a fault identification model of the power grid transmission line based on an improved convolutional neural network (CNN) is proposed. Firstly, preprocess the current time series data of power grid transmission lines. Secondly, improve the convolutional neural network through dual-channel fusion, multi-layer convolution and pooling, and combine the batch normalization method in the convolutional layer to separately detect short circuit faults in the line feature extraction is performed on the false alarm data, and then classified and identified through the soft-max classifier to construct an intelligent and efficient fault identification model, which effectively reduces the false alarm rate. Finally, the actual data of the State Grid Dispatching Center is used to verify the effectiveness of the proposed method.