ResNet-GRU Based Transient Voltage Stability Margin Evaluation of Power System

The safe and stable operation of power system faces serious challenges such as the variability of operation mode and the lack of disturbance attenuation ability. To achieve the accurate voltage stability evaluation and provide reliable reference for emergency control, this paper proposes a transient voltage stability margin evaluation method based on deep residual network (ResNet) and gated recurrent unit (GRU). Firstly, a practical transient voltage stability margin index based on two-element table is presented to construct sample label in the case of large disturbances. Secondly, a ResNet-GRU hybrid model is established to quantitatively evaluate voltage stability margin. The complementary advantages of ResNet in spatial coupling feature extraction and GRU in temporal compliance relationship learning are fully exploited. Finally, case studies are performed in Northwest China local region power grid. A series of experiment and simulation results verify the effectiveness and accuracy of the proposed method for voltage stability margin evaluation.