Data-Driven Fault Detection Technology for Distribution Automation Master Station System

Aiming at the hidden faults such as communication avalanche, dual-machine synchronization suspension and feeder automation (FA) service suspended animation in distribution automation master station system under large-scale distributed source-load access and complex topology changes, this paper collects multi-dimensional time-series operation performance data of the master station operating system, network message traffic and system bus text logs, and constructs a unified operation state feature matrix through text template extraction and spatiotemporal alignment. Verified in simulation environment and actual distribution network dispatching, the results show that the proposed technology can effectively shorten the location time of hidden faults.