基于同步挤压变换和深度迁移学习的GIS隔离开关故障诊断

为实现气体绝缘组合电器(GIS)隔离开关机械故障的智能诊断，基于GIS隔离开关分合闸过程中的振动信号，提出了基于深度迁移学习的GIS隔离开关机械故障诊断方法。首先应用二阶同步挤压傅里叶变换(FSST2)获取GIS隔离开关振动信号的时频分布，然后基于深度迁移学习的思想构建预训练模型并进行优化，建立了GIS隔离开关机械故障智能诊断模型。对某GR角型GIS隔离开关正常和典型机械故障状态下的振动信号的分析结果表明，基于FSST2得到的GIS隔离开关振动信号时频表示具有较好的能量聚集性，所建立的GIS隔离开关机械故障智能诊断模型识别准确率高且模型复杂度低，可用于GIS隔离开关机械故障的高效诊断。

Fault diagnosis of GIS disconnector based on synchrosqueezing transform and deep transfer learning

To recognize the typical mechanical faults of gas insulation switchgear(GIS) disconnector with high efficiency, a deep transfer learning-based mechanical fault diagnosis model on the basis of the vibration signals of GIS disconnector in the opening and closing operation is proposed. The second-order short-time Fourier transform-based synchrosqueezing transform(FSST2) is applied to obtain the time-frequency distribution of the vibration signals of GIS disconnector. Then with the thought of deep transfer learning, the pre-trained model is built and optimized. A deep transfer learning-based mechanical fault intelligent diagnosis model of GIS disconnector is proposed. The analysis results of vibration signals of a GR angle-type GIS disconnector under normal and typical mechanical fault conditions show that the time-frequency representation of vibration signals of GIS disconnector has better energy aggregation. The proposed mechanical fault intelligence diagnosis model is capable of recognizing the different mechanical faults with a high recognition rate and has a low model construction complexity, and can be used to efficiently diagnose the mechanical fault of GIS disconnector.