Isolation switch fault diagnosis method based on class perception depth time sequence clustering

The invention discloses an isolating switch fault diagnosis method based on class perception depth time sequence clustering. The method specifically comprises a depth time clustering algorithm and a CA-DTCR-based semi-supervised training model. According to the method, a rapid and accurate isolation switch method is established, early discovery and early warning of faults of the isolation switch are facilitated, operation scheduling of an electric power system is met, maintenance cost and management cost of the electric power system are reduced, and economic benefits are increased. 本发明公开了基于类感知深度时间序列聚类的隔离开关故障诊断方法，具体包括深度时间聚类算法、基于CA-DTCR的的半监督训练模型两部分；建立了较为快速及精度的隔离开关方法，有利于隔离开关故障提前发现与预警，满足电力系统的运行调度，同时减少了电力系统的维护成本与管理成本，经济效益。