基于卷积神经网络的智能变电站通信网络故障诊断研究

Fault Diagnosis of Communication Networks in Smart Substations Based on Convolutional Neural Networks

摘要: 对智能变电站通信网络的故障进行快速诊断和定位，可有效提高变电站的检修效率和运行可靠性。文章以智能变电站二次系统通信网络为研究对象，利用卷积神经网络（convolutional neural network，CNN）在复杂数据处理与信息融合方面的优势，提出了一种针对智能变电站通信网络故障诊断的CNN模型，给出了该模型的构建及训练方法。通过比对模型预测与实际故障类型的相似度评估模型效果，结果显示该模型具有较高的精确度。

Abstract: Rapid fault diagnosis and localization in smart substation communication networks can significantly enhance maintenance efficiency and operational reliability. Focusing on the secondary-system communication network of smart substations, this study leverages the advantages of convolutional neural networks (CNNs) in complex data processing and information fusion to propose a CNN-based fault diagnosis model. The model's performance is evaluated by comparing the similarity between its predicted fault types and the actual ones. The results demonstrate that the model achieves high accuracy.