Fault Prediction in Electric Power Communication Network Based on Improved DenseNet
Peng Zhao Gao, Li Guan, Jiakai Hao, Qian Chen, Yang Yang, Zhenying Qu, Ming Jin · 2023
Power communication network plays a crucial role in the power system infrastructure. However, its rapid development has led to a significant increase in real-time alarm volume, posing great challenges to network management personnel. In the event of a fault, the impact on normal work and daily life of users can be substantial. Thus, fault prediction for the power communication network is essential for effective equipment management and maintenance, as well as for improving system reliability. In this paper, we propose a multichannel fault prediction model for power communication network based on an improved DenseNet and Transformer. This method encodes the features into multiple channels, and proposes the Improved Channel Attention Module(ICAM) to the DenseNet architecture to suppress non-significant features. Transformer is then utilized for prediction, thereby improving the accuracy of fault prediction. Simulation experiments conducted on the Telstra dataset demonstrate that the algorithm proposed in this paper outperforms the compared algorithms, and can effectively predict faults in the power communication network, providing early warnings for network operators.