基于深度学习的智能变电站通信网络故障诊断与定位方法

Fault Diagnosis and Positioning for Communication Network in Intelligent Substation Based on Deep Learning

为提高智能变电站通信网络运维效率,提出了基于深度学习的智能变电站通信网络故障诊断与定位方法。从通信网络故障状态的冗余监测出发,分析基于不同监测节点的故障特征信息,提出了通信网络故障特征信息表征方式。基于涌现原理,根据通信网络物理连接、逻辑连接及报文订阅关系实现故障样本的自动生成,并结合深度学习理论中训练规则,建立基于深度置信网络的通信网络故障诊断模型,基于此给出实时故障分析处理流程。以典型110kV智能变电站过程层网络为例进行验证,仿真结果验证了所提故障诊断方法的有效性和精确性,且在部分信息不可信时仍能得到准确诊断结果,容错性能较好。

In order to improve efficiency of operation and maintenance of intelligent substation communication network, this paper proposes a fault diagnosis and positioning method based on deep learning for the network. Based on redundant monitoring of fault state of communication network, the fault feature information based on different monitoring nodes is analyzed and a fault feature information representation mode of communication network is proposed. Furthermore, on the basis of emergence principle, fault samples are generated automatically according to physical connection, logical connection and massage subscription relationship of the communication network. Combined with the training rules in deep learning theory, a fault diagnosis model is established and a real-time fault analysis processing flow is proposed. An 110 kV intelligent substation process layer network is taken as an example to verify accuracy and effectiveness of the proposed method. It is proved that the fault tolerance performance is better even when dealing with fault feature information mixed with some incredible information.