基于深度学习的GIS设备故障识别技术

GIS equipment fault identification technology based on deep learning

传统GIS设备故障识别方法无法处理负载突变问题,且负载调节不及时,导致故障识别精度偏低。提出基于深度学习的GIS设备故障识别技术,构建动态响应特征检测模型、GIS设备故障特征融合模型以及故障统计特征分析模型。计算故障状态下的电压外环和负载突变参数,提取GIS设备故障属性特征;采用深度学习判断故障识别过程中的参数寻优控制和收敛性,实现对变电站GIS设备故障优化识别和检测,提高GIS设备故障检测识别能力。实验结果表明,所提变电站GIS设备故障识别方法的精度较高,对故障属性类别判断的准确性较好。

The traditional fault identification methods of GIS equipment can not deal with the problem of load change,and the load adjustment is not in time as well,resulting in the low accuracy of fault identification.In this paper,the technology of GIS equipment fault recognition based on deep learning is proposed,which constructs a dynamic response feature detection model,GIS equipment fault feature fusion model and fault statistical feature analysis.The parameters of outer voltage loop and load change under fault state are calculated,and the fault attribute characteristics of GIS equipment are extracted.Then,using deep learning to judge the parameter optimization control and convergence in the process of fault identification,which realizes the optimal identification and detection of substation GIS equipment fault,and improves the ability of GIS equipment fault detection and identification.The experiment results show that the accuracy of the proposed method is higher,and the accuracy of fault attribute classification is better.