Data-Driven Islanding Detection Using a Principal Subspace of Voltage Angle Differences

The likelihood of an unintentional power system islanding is increased in systems with significant penetration distributed generation. To mitigate the adverse effects islanding, a quick and reliable detection method needed. This paper first analyzes covariance matrices linearized model, relates them to principal component analysis experimentally obtained matrices. Additionally, new model-independent proposed that uses measurements voltage angle differences between multiple locations system. are preprocessed remove nonstationarity. Thereafter, probabilistic model trained using acquired measurements. residual spaces extracted from used discriminate other events applicability demonstrated by real gathered several transmission grid.