On-line dynamic voltage instability prediction based on decision tree supported by a wide-area measurement system

In this study, a new decision tree-based method is proposed to assess the small disturbance (limited to load disturbance) voltage stability of power systems by only using synchronised phasors (synchrophasors) measured by phasor measurement units in a wide-area measurement system. To properly predict the voltage instability considering various load increase directions as well as load increase step sizes, a mathematical morphology algorithm is used to extract appropriate features from synchronised measured voltage phasors during the few seconds just after the disturbance. These features along with the pre-disturbance ones are then applied to a trained decision tree to predict the voltage instability of the power system. The proposed method is applied and tested on the Nordic32 test system. To examine the efficiency and accuracy of the proposed method, the dynamic behaviours of on-load tap changing transformers as well as the stator current limiters of synchronous generators are considered. The simulation results demonstrate that this method effectively predicts the voltage instability of the power system during the first two seconds just after the disturbance, for the test system.